Main Purpose of Study: The main purpose of the community surveillance component of the ARIC Study is to continuously monitor and independently validate coronary heart disease (CHD) events (acute myocardial infarction [MI] and death due to coronary heart disease) and acute decompensated heart failure occurring among residents of four geographically defined communities in the United States in order to evaluate trends in mortality, incidence, case fatality rates, and medical care by age, gender, race, community, and time.
The ARIC Study also includes monitoring and validating events among cohort participants (random sample of 15,792 men and women from four communities enrolled in 1987–1989). In addition to the CHD endpoints above, clinically recognized strokes among cohort participants are also identified and validated through surveillance procedures. All hospitalizations among the cohort are identified and recorded, but only those mentioned above are independently validated. Investigations using endpoints captured in cohort surveillance has led to over 800 published reports regarding the prevalence of cardiovascular disease and risk factors among the cohort as well as the role of established and novel risk factors in predicting disease. The information relates to the ARIC Study’s community surveillance component unless noted otherwise.
Sample: ARIC community surveillance is a continuous retrospective, hospital-based surveillance study. Population denominators are estimated using interpolation and extrapolation for 1990 and 2000 U.S. Census population estimates.
Through community surveillance, the ARIC Study enumerates and validates cases (events) of hospitalized MI and CHD deaths occurring after January 1, 1987, in 35- through-74-year-old male and female residents of the four ARIC Study communities: Forsyth County, North Carolina; Jackson, Mississippi; suburbs of Minneapolis, Minnesota; and Washington County, Maryland.
The population of men and women age 35–74 in these four communities was approximately 406,000 in 2007. The eligible age range of ARIC surveillance was expanded to age 84 beginning in 2005, bringing the surveillance population to a total of approximately 447,000 men and women age 35–84.
Community surveillance was also expanded to include surveillance of hospitalized heart failure beginning in 2005 for community residents age 55 years and older. The ARIC Study is currently funded to conduct surveil-
lance for events occurring through December 31, 2010. A renewal is pending that will extend surveillance through December 31, 2014.
All residents of the four geographically defined communities are included in ARIC community surveillance regardless of race or ethnicity. The white minority in Jackson and the black minority in Forsyth County are oversampled. The number of persons reported to be neither black nor white has been small in these communities and currently produce unstable event rates. Both men and women of all race or ethnic groups are eligible for selection in ARIC community surveillance.
Frequency of Collection and Sources of Data: ARIC community surveillance identifies, samples, and investigates hospitalizations on a continuous basis and produces annual event rates in the four communities. Sources include hospitalized events for acute myocardial infarction are identified from electronic lists of discharges obtained for catchment area hospitals on an ongoing basis. Hospitalizations selected (sampled) for investigation are identified and medical records for those events are obtained by trained ARIC medical record abstractors. Medical records are abstracted using a standardized, web-based data entry system.
Fatal events for investigation are identified through electronic lists of deaths obtained from local or state health departments. Deaths are sampled based on underlying cause of death codes. Sampled death certificates are obtained and abstracted by trained ARIC staff. Death with an underlying cause of death code that is related to CHD and occurring out-of-hospital are targeted for further investigation through telephone interview with next of kin or witness. The decedent’s physician is also identified from the death certificate and sent a questionnaire requesting information relevant to classification of the death. If a coroner or medical examiner, information is also obtained if appropriate. For ARIC cohort participants, hospitalizations for investigation are identified through an annual follow-up telephone interview conducted by ARIC staff. In addition to the events noted above, stroke related hospitalizations are also indentified among cohort members. Any hospitalization reported by a cohort member through annual follow-up telephone interviews or found through routine community surveillance is identified and obtained for abstraction. Deaths among cohort members are identified either through annual follow-up contact, monitoring of electronic death files from local and state health departments, or by on-going monitoring of obituaries in the study communities by ARIC staff.
Mode of Data Collection: Data on hospitalizations and deaths are collected manually through detailed abstraction of medical record or death records. Data from next of kin for out-of-hospital deaths are collected through a telephone interview. Mailed surveys are also used to solicit information from physicians identified on death certificates for selected cases.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: In ARIC community surveillance, information collected on sampled hospitalized myocardial infarction and heart failure events include specific items on medical history and comorbid conditions such as chronic pulmonary disease and diabetes. Furthermore, all discharge codes for sampled cases are also recorded. However, hospitalizations for chronic pulmonary disease, asthma, and/or diabetes are not specifically investigated in ARIC community surveillance.
Among ARIC cohort participants, all hospitalizations for any reason are indentified and ICD-9-CM discharge diagnoses and procedure codes recorded. Those related to myocardial infarction, heart failure, and stroke are investigated further as noted above. During the annual follow-up interview of cohort participants, questions are asked that relate to pulmonary signs and symptoms as well as self-reported physician diagnosis of diabetes.
Information Obtained: Measures available through ARIC community surveillance are summarized as follows:
- Incidence: Annual incidence rates of hospitalized acute myocardial infarction (1987–2014)
- Annual incidence rates of hospitalized acute decompensated heart failure (2005–2014)
- Annual mortality rates due to coronary heart disease (1987–2014)
- Annual mortality due to sudden cardiac death (1987–2014)
- Case fatality: Annual case fatality rates (through one year) after hospitalized acute MI (1987–2014)
- Clinical care information: Procedures during hospital stay for hospitalized acute MI (* includes data on time since event onset): Cardiac catheterization, coronary angiography, coronary angioplasty,* coronary atherectomy,* Swan-Ganz catheterization, echocardiography, coronary bypass surgery,* intracoronary thrombolytic therapy,* intravenous thrombolytic therapy,* aortic balloon pump, MRI scan of heart, exercise
stress test, Holter monitoring, coronary stent placement,* implanted defibrillator,* coronary CT, closed chest massage or cardioversion, and other procedures recorded as open text
- Procedures during hospital stay for hospitalized acute decompensated heart failure: Chest X-ray, echocardiography, coronary angiography, cardiac radionuclide ventriculogram, magnetic resonance imaging, cardiac CT, stress tests
- Medical therapy during hospital stay or discharge for hospitalized acute MI: Nitrates, calcium channel blockers, beta-blockers, digitalis, lidocaine, Coumadin, aspirin, ACE or angiotensin II inhibitors, heparin infusion, antiplatelet agents, lipid lowering medication
- Medical therapy during hospital stay or discharge for hospitalized acute decompensated heart failure: CE inhibitors, angiotensin II receptor blockers, beta blockers, digitalis, diuretics, aldosterone blocker, lipid lowering agents, nitrates, hydralzaine, IV inotropes
- Diagnostic information for hospitalized acute MI: Biomarkers (total CK, CK-MB, total LDH, troponin I, tropoinin T, BNP, Pro-BNP, serum creatinine)
- Electrocardiographic information: Copies of up to three electrocardiograms with Minnesota coding for each
- Diagnostic information for hospitalized acute decompensated heart failure: Diagnostic findings including measures of ejection fraction from imaging, biomarkers including BNP, proBNP, troponin I, troponin T, serum creatinine, BUN
Information Not Obtained: Measures available through ARIC community surveillance are summarized as follows:
- Prevalence: ARIC community surveillance designed to capture incident and recurrent hospitalized MI and heart failure. Prevalence measures of these conditions in the population are not obtained.
- Functional Health Outcomes: Not captured in ARIC community surveillance
- Risk Factor (including stressor): Not captured in ARIC community surveillance
Demographic Characteristics Collected: Age, sex, race or ethnic group, Hispanic or Latino origin, type of health insurance, patient address (geocoded to provide latitude and longitude and linked to census data on neighborhood socioeconomic status).
Who Pays for Data Collection?: ARIC surveillance is funded by a contract with the National Institutes of Health, National Health, Lung, and Blood Institute.
Dissemination of Data: Public use data files are created and updated annually and are available through the NHLBI project office.
Main Purpose of Study: The Cardiovascular Health Study is an NHLBI-funded observational study of risk factors for cardiovascular disease and stroke in adults 65 years or older.
Sample: In June 1990, four field centers completed the recruitment of 5,201 participants. Between November 1992 and June 1993, an additional 687 African Americans were recruited using similar methods. The sample was drawn from four field centers. They are located in Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania. The only geographical level of collection of data was by field center. Minorities are oversampled. Men and women are included. A stipulation of the study was that participants be over 65 or older.
Frequency of Collection and Sources of Data: Until 1999, semiannual contacts alternated between clinic examinations and telephone contacts, during which information about hospitalizations and potential cardiovascular events was collected. Since 1999, participants have been contacted twice a year by telephone to collect limited data, including medication data, and to identify all hospitalizations and potential cardiovascular events. Participants were also invited to participate in a clinic or home visit as part of an ancillary study called CHS All Stars Study in year 18; semiannual phone calls continue. Presently collecting data for year 21. The CHS All Stars Study is an ancillary study that focuses on reexamining the long-term survivors of CHS to determine the likelihood of maintaining
function later in life. The primary source of data is participant examinations and interviews. Hospital records were obtained to confirm self-reported events, and CMS (Medicare) records were searched to capture any missed events.
Mode of Data Collection: The baseline examination consisted of a home interview and a clinic examination that assessed not only traditional risk factors for cardiovascular disease but also measures of subclinical disease, including carotid ultrasound, echocardiography, electrocardiography, MRIs, and pulmonary function. Major exam components were repeated during annual follow-up examinations through 1999. Cranial MRI scans, retinal photography, and tests of endothelial function were added as new components.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Cardiovascular disease was measured and adjudicated by a panel of doctors based on hospital records. Chronic pulmonary disease was measured by self-report and periodic spirometry measurement. Asthma was measured through self-report. Diabetes was measured through self-report, periodic measurement of blood glucose level, and annual review of medications.
Information Obtained: Incidence, prevalence, functional health outcomes, risk factors, clinical care information, demographic characteristics.
Who Pays for Data Collection?: CHS was originally funded through a contract from the National Heart, Lung, and Blood Institute (NHLBI). Additional funding has been obtained through a renewal of the NHLBI grant to continue cohort follow-up, from NINDS, the NIA, and a wide number of ancillary studies.
Dissemination of Data: There are two means of obtaining data: (1) Researchers can contact the Collaborative Health Studies Coordinating Center to be sponsored by a CHS investigator. They would then submit a paper proposal to be reviewed by the CHS Publications and Presentations Committee and the Steering Committee. Proposed papers would be reviewed for consistency with the goals of CHS, lack of overlap with other work, scientific integrity, and evidence of collaborative authorship, including junior investigators. (2) Data are also available as a limited access data set with NHLBI. Applicants can make a direct enquiry to NHLBI to obtain that data.
Additional Comments: Pittsburgh population entirely urban; other three field centers recruited mixed urban and rural populations; participating academic institutions include the University of Washington, the University of California–Davis, the Johns Hopkins University, Wake Forest University School of Medicine, University of Pittsburgh, University of Vermont, University of Maryland, Baltimore, University of Arizona, University of Wisconsin, and Tufts New England Medical Center; standard protocols for the identification and adjudication of cardiovascular events were implemented during follow-up. The adjudicated events are CHD, angina, heart failure (HF), stroke, transient ischemic attack (TIA), claudication, and mortality. Other patient information was collected through a wide array of measures.
Main Purpose of Study: Scientific researchers are conducting an investigation to find the genes that cause a susceptibility to developing COPD. This groundbreaking study, a $37 million grant awarded by the National Heart, Lung, and Blood Institute, has the potential of changing what is known about COPD and COPD treatment. Dr. James Crapo of the National Jewish Health in Denver, Colorado, along with Dr. Edwin Silverman of the Brigham and Women’s Hospital in Boston, Massachusetts, are conducting a genetic epidemiology study of COPD—a study to characterize COPD in the U.S. population and to find the genes that create a risk for developing this disease. The study brings together a large cohort of individuals who have COPD or are at risk for developing COPD. The study will analyze genetic variations across the entire human genome to identify the primary genes that determine why some individuals are more susceptible to developing COPD than other individuals. The COPD Foundation Registry serves as a valuable supplemental source of patients for the COPDGene® Study cohort. The COPDGene® Study has a highly characterized cohort of patients, including African Americans.
The COPDGene® Study is in the process of creating the largest, well-characterized set of COPD and control subjects ever assembled for pulmonary disease research. In addition to identifying COPD susceptibility genetic determinants, important advances in characterization of the natural history of COPD and its phenotypes can be anticipated, along with identification of new, well-characterized COPD subtypes. Improved understanding of COPD subtypes and genes controlling susceptibility to COPD could lead to novel pathophysiological insights, refined diagnostic criteria, and new approaches for pharmacological treatments for COPD.
An interesting aspect of the COPDGene® Study is that participating patients have agreed to be contacted for other studies. So in essence, this study acts as a cohort of patients who are ready to participate in other studies. Thus, the COPDGene® Study cohort is also a registry for participation in other studies.
The COPDGene® Study includes the following: Creation of a large cohort of subjects at risk for or expressing one of the various stages of COPD (GOLD grades 1–4). This cohort will be clinically phenotyped and all subjects will undergo HRCT, which will be quantitatively analyzed to divide the cohort into unique subtypes of COPD. A genome-wide association study is being conducted on the cohort resulting in fine mapping of the genetic determinants for susceptibility to develop COPD. The impact of these identified COPD susceptibility genes on each of the CT-defined COPD subtypes will be determined. The cohort of patients is being followed longitudinally to identify the clinical phenotypes, CT subtypes, and genotypes that determine risk for COPD progression, morbidity, and mortality. While the initial emphasis in the design of the cohort was on COPD, all subjects in the cohort will be smokers, and the cohort analysis will be expanded to include other smoking-related diseases such as cardiovascular disease and cancer. Finally, the ultimate goal of this project is to carry out translational studies to validate new therapies that are personalized for subjects with specific subtypes of COPD.
The study, which was originally scheduled to be completed in 5 years, has surpassed enrollment projections and will be concluding enrollment in the next month, 1 year ahead of schedule. In order to identify the genetic basis of COPD, a study group of 10,000 individuals will be formed. Nineteen medical centers across the U.S. host the clinical evaluations of eligible individuals that want to participate in the study. The clinical evaluations are being conducted physiologically and radiographically, meaning that the individual takes a spirometry test, tests on a 6-minute walk, and scores on a BODE scale (body mass index, degree of airflow obstruction, degree of dyspnea, and exercise capacity). The individual is also given a chest CT scan, completes a set of questionnaires, and donates about 30cc (6 teaspoons) of blood for genetic analysis.
The aims of the COPDGene® Study are as follows:
- Specific Aim 1: Build Cohort of 10,000 Smokers (10+ pack years). Clinically phenotype COPD cases and control subjects for genetic, epidemiologic, natural history, and pharmacologic intervention studies.
— Sample Size by GOLD Criteria
Non-Smokers without COPD (100 subjects)
Smokers without COPD (4,000 subjects)
Smokers with possible COPD, GOLD grade 1 (800 subjects)
Smokers with possible COPD, GOLD grade U (1,100 subjects)
Smokers with COPD, GOLD grades 2–4 (4,000 subjects)
— Sample Size by Race
COPD case/control cohort—Non-Hispanic white (n = 6,700 total)
COPD case/control cohort—African American (n = 3,300 total)
- Specific Aim 2: CT Phenotype Cohort—Characterization of Subtypes of COPD.
— Use HRCT (inspiration and expiration scans) to subdivide the cohort into groups expressing unique airway disease (inflammation) and lung parenchymal subtypes (emphysema) of COPD.
— Validate quantitative indexes (by HRCT) of the emphysema and airway inflammatory disease subtypes.
Percent of specific lung regions with low attenuation
Airway wall thickness of 3rd–6th generation airways
Percent gas trapping identified by expiratory CT scan
- Specific Aim 3: Genotype Cohort. Genome-wide association study using a phased approach.
— A genome-wide panel of SNPs will be tested for association with COPD.
— Confirmation of SNPs to identify genomic regions for intensive investigation.
— Mapping of 50 genomic regions to identify susceptibility genes for COPD.
— Fine mapping of candidate genes to identify susceptibility alleles and/or high risk haplotypes.
— Assess the association of genetic variants in the identified COPD susceptibility genes with CT-defined COPD subtypes.
- Specific Aim 4: Longitudinal Follow-up of This Cohort to
— Assess risk factors for COPD progression, morbidity, and mortality
— Determine which CT phenotypes and genotypes identify asymptomatic subjects who are at high risk for progression to overt COPD
— Develop biomarkers for COPD and COPD subtypes that relate to disease progression and prevention
- Specific Aim 5: Expand Analysis of Cohort to Evaluate Other Smoking-Related Diseases and/or Comorbidities.
— Cardiovascular disease
— Musculoskeletal disease
— Depression/cognitive dysfunction
- Specific Aim 6: Use Phenotyped and Genotyped Subcomponents of Cohort for Translational Studies to Validate New Therapies.
— Pharmacologic interventions in specific subtypes of COPD
— Identification of asymptomatic smokers at high risk for disease progression and development of preventive therapies
— Use identified biomarkers that relate to disease progression and prevention to assess efficacy of new therapies
Sample: The goal of the COPDGene® Study is to enroll 10,000 patients. The participating centers enrolling patients are located across the United States. Geographic levels included are state and national (United States only). Goal is to frequency match for gender distribution within each racial group for the case-control component of this project. Approximately 33% of the final study sample will be African American.
Subjects must be at least 45 years old, and both men and women may participate.
Frequency of Data Collection and Sources of Data: Data are collected at the initial study visit, and longitudinal data are collected at 6-month intervals following the visit. Sources of data include questionnaires collected at the visit, blood and CT scan analysis, longitudinal follow-up questionnaires, and subject medical records if available.
Mode of Data Collection: In-person examination or biomarker data collection, administrative data, mailed surveys, and automated telephony surveys
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Questions are asked related to comorbidities that include CVD, chronic pulmonary disease, asthma, and diabetes.
Information Obtained: Incidence, prevalence, functional health outcomes, and risk factors, including stressors, clinical care information, and demographic characteristics
Cost of Survey/Registry: Almost 40 million dollars
Who Pays for the Data Collection?: The National Institutes of Health (NIH) and the COPPD Foundation
Dissemination of Data: Data files are available online to authorized study investigators either through the COPDGene® Study website or through the National Institutes of Health dbGaP resource. Authorized COPDGene® investigators have access through the COPDGene® website. All other data users must access data through dbGaP. All data to be provided externally shall be de-identified. This study is novel and unique since it will allow transparent and public access the study’s data. Investigators, government, and industry shall have access to the data.
Main Purpose of Study: The Coronary Artery Risk Development in Young Adults (CARDIA) Study is a study examining how heart disease develops in adults.
Sample: 5,115 black and white men and women across four cities: Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California. The study was designed so that there are approximately 50% blacks, 50% whites; 50% males, 50% females; 50% aged 18 to 24, 50% aged 25–30; 50% who have not completed high school, and 50% who have completed high school.
From the official sample size for the study n = 5,115: Birmingham, Alabama: 3,252 eligible participants contacted, 1,811 (55.7%) made appointments; of those 1,178 (65.0%) examined; Chicago, Illinois: 2,205 eli-
gible participants contacted, 2,149 (97.5%) made appointments; of those 1,109 (51.6%) examined; Minneapolis, Minnesota: 2,473 eligible participants contacted, 1,777 (71.9%) made appointments, of those 1,402 (78.9%) examined; Oakland, California: 2,203 eligible participants contacted; 2,047 (92.5%) made appointments; of those 1,426 (69.7%) examined. A majority of the initial group (1986) has participated in follow-up examinations: 1987–1988 (90%); 1990–1991 (86%); 1992–1993 (81%); 1995–1996 (79%); 2000–2001 (74%); 2005–2006 (72%)
Frequency of Collection and Sources of Data: Study initiated in 1986 (Year 0), with the same participants asked to participate in follow-up examinations in 1987–1988 (Year 2); 1990–1991 (Year 5); 1992–1993 (Year 7); 1995–1996 (Year 10); 2000–2001 (Year 15); 2005–2006 (Year 20). Core study exam results; Medical history; Obesity questionnaires; Psychosocial; Physical measurement; Pulmonary function testing and questionnaire; Electrocardiogram; Echocardiography; Genetic studies
Mode of Data Collection: In-person interview on medical and family history; medical exams; anthropometry; blood test
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Core Study: blood pressure, lipids, lipoproteins, apoproteins, cotinine, SMAC 12, fasting insulin, fasting glucose, oral glucose tolerance test, CBC, Lp(a), fibrinogen, stored plasma, stored serum, serum creatinine, urinary creatinine, uric acid, c-reactive protein, albuminuria, interleukin-6; Pulmonary function; Electrocardiogram; Echocardiography, HbA1C
Information Obtained: Incidence, prevalence, quality of life, discrimination, height, weight, skinfold fat, blood pressure, cholesterol, other lipids, other chemistries (insulin and glucose), physical activity/fitness, diet history, food frequency, obesity questionnaire, personal history, substance use (tobacco and alcohol), behavioral and psychological variables, stress, anxiety. Subclinical atherosclerosis measured via echocardiography during Year 5 and 10, computed tomography during Year 15 and 20, carotid ultrasound during Year 20. Demographic information includes age, race/ethnicity, geographic region, education.
Who Pays for Data Collection?: Group of contracts funded by the National Heart, Lung, and Blood Institute (NHLBI); brain MRI component is included in the Year 25 exam that NIA is funding.
Dissemination of Data: CARDIA uses a distributed data system with current data. In this system the Field Center PIs, the Steering Committee Chair, and the Project Office (NHLBI) get a full copy of the data. Proposals for manuscripts are approved by the study’s Publications and Presentations Committee. All approved manuscripts require a CARDIA-approved investigator to be associated with the manuscript. Investigators with approved manuscripts can request manuscript specific data sets from the Coordinating Center. Investigators can also request a data repository data set (formerly known as Limited Access Dataset) directly from NHLBI. These data, however, have a 5-year lag (soon to be changed to a 3-year lag) and have been winsorized to protect participant confidentiality.
Purpose: The Eight Americas explore the causes of disparities in race-counties that inform specific public health intervention policies and programs.
Sample: The sample includes the entire U.S. population (all ages) divided into eight distinct subgroups. Building blocks for these groups were a combination of race and county of residence. Race-county units were combined based on socioeconomic and geographical indicators, including location of county of residence, population density, race-specific county-level per capita income, and cumulative homicide rate. Smaller counties were merged with adjacent counties to form units with total population of at least 10,000 males and 10,000 females. Merged county units also formed to account for changes in county status and county lines. There are 2,072 counties.
Souces of Data: For 1982–1989, interpolated age-, sex-, race-, and county-specific population figures using 1980 and 1990 Census figures. Used bridged-race population estimates released by National Center for Health Statistics (NCHS). NCHS mortality statistics. Behavioral Risk Factor Surveillance System for health plan coverage and health care utilization. Global Burden of Disease database for life expectancy and probabilities of death for international comparisons.
Information Obtained: Race, race-specific county income, average income per capita, percent completing high school, sex, and age.
Who Pays for Data Collection?: Centers for Disease Control and Prevention, the Association of Schools of Public Health, the National Institute on Aging.
Main Purpose of Study: The Framingham Heart Study is a prospective, longitudinal observation of three generations of living population for CVD endpoints and other common chronic diseases.
Sample: Size at entry: 15,447 men and women in five cohorts residing in or near Framingham. Local level data only. Minorities were not oversampled. Response rate: 70–80% over the 60+ years of the study. Age groups included at entry:
- Original Cohort: 29–62 years
- Offspring Cohort: less than 10 to 70 years
- Third Generation and New Offspring Spouses Cohort: 19 to 79 years
- Omni Group 1 Cohort: 20 to 79 years
- Omni Group 2 Cohort: 20 to 70+ years
Frequency of Data Collection: Health history updates every 2 years, clinic exam every 2 to 6 years
Sources of Data: Hospital discharge, medical records, population interviews, disease registries
Mode of Data Collection: Interviews and measurements from in-person examinations and biomarker collection; health history updates by mailed questionnaire or telephone interview; follow-up medical records from health care providers.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Extensive (variables posted on SHARe website, http://www.ncbi.nlm.nih.gov/sites/entrez?db=gap)
Information Obtained: Incidence, prevalence, functional health outcomes, and risk factors, including stressors, clinical care information
Who Pays for Data Collection?: NHLBI Contract and NIH grants
Dissemination of Data: Data request applications are available online. Data are available through research proposals submitted online and approved by relevant review committees. FHS variables are posted on the Framingham Heart Study website, www.framinghamheartstudy.org. Queries may be sent to the FHS research committee. Principal investigators with proposals approved by the review committees and with IRB approval and data distribution agreements can obtain access.
Purpose: In the late 1990s, the U.S. Department of Defense and Congress identified the need for coordinated epidemiological research to determine how military occupational exposures, including deployment-related exposures, affect long-term health. The Institute of Medicine more specifically defined the importance of a large, prospective study for evaluating exposures and a broad spectrum of important health outcomes. The Millennium Cohort Study was designed, in collaboration with all military services and the Department of Veterans Affairs, to meet these research challenges.
The objectives of the Millennium Cohort were (1) to evaluate chronic diagnosed health problems, including hypertension, diabetes, and heart disease, among military members, in relationship to exposures of military concern and (2) to evaluate long-term subjective health, including chronic multi-symptom illnesses, among military members, especially in relationship to exposures of military concern.
Sample: Launched in the summer of 2001, the Millennium Cohort Study began enrolling a representative sample of U.S. military personnel, both active duty and Reserve and National Guard members, who agreed to participate in follow-up well past their time in service, for up to 21 years. There are more than 150,000 consenting Millennium Cohort Study members. Currently, participants that enrolled in the 2001 survey cycle, Panel 1, include over 77,000; participants that enrolled in the 2004 survey cycle, Panel 2, include an additional 31,100; and participants that enrolled in the 2007 survey cycle, Panel 3, include approximately 43,400 participants. The total cohort is over 152,000 members and will be adding approximately 60,000 U.S. military members, both active
duty and Reserve and National Guard, as well as a Millennium Cohort Family Study component of approximately 10,000 spouses later this year.
The cohort is based on a random sample of U.S. service members residing in all 50 states and territories; no selection or stratification is based on geographic location. Service members are often assigned and deployed to many geographic locations outside of the United States. Team members are working with professionals at Defense Manpower Data Center (DMDC) and U.S. Army Center for Health Promotion and Preventive Medicine (USACHPPM) to differentiate location of deployment and will include whatever data are available in these analyses. Basic geographic information, including duty and home address for all participants, is included.
Women were oversampled in Panels 1, 2, and 3. In Panel 2, to achieve a higher proportion of Marines, the population was comprised of 20% Marines and 80% other service branches (Army, Navy, Air Force, and Coast Guard). Different races/ethnicities have not been oversampled in any panel.
The mean age for Panel 1 at baseline is 35 years, Panel 2 is 24 years, and Panel 3 is 24 years, with an average age of 27.6 years for all 3 panels at baseline.
More than 70% of cohort members who submitted baseline data have submitted at least one follow-up questionnaire.
Approximately 88% of participants completed the questionnaire via the online questionnaire in the 2007 survey cycle.
Frequency of Data Collection: All panels will be followed with repeat surveys at 3-year intervals through 2022.
Source of Data: The Millennium Cohort Study is a research platform that prospectively combines self-reported data with inpatient, outpatient, pharmacy, vaccination, personnel, deployment, and occupational data.
Multiple standardized instruments (to compare to other civilian and veteran populations) are included in the questionnaire, including the posttraumatic stress disorder (PTSD) Checklist Civilian Version; the Patient Health Questionnaire to assess depression, panic, anxiety, eating disorders, and alcohol-related problems; the Medical Outcome Study Short Form 36-Item Health Survey for Veterans to assess functional health; and the CAGE questionnaire to assess potential problem drinking.
Mode of Data Collection: Participants are given the option to complete a web or paper survey.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The Millennium Cohort Study baseline survey asks the following related questions:
- “Has your doctor or other health professional ever told you that you have any of the following conditions?”
— Hypertension, high cholesterol requiring medication, coronary heart disease, heart attack, angina (chest pain), any other heart condition: please specify, chronic bronchitis, emphysema, asthma, diabetes or sugar diabetes
- If marking yes to any of these conditions the participant is asked what year they were first diagnosed and if they were ever hospitalized for the condition.
- Other questions on the survey instrument relating to cardiovascular disease, chronic pulmonary disease, asthma, and/or diabetes include:
— “In the last 12 months have you had persistent or recurring problems with the following?” Shortness of breath, chest pain
— “During the last 4 weeks, how much have you been bothered by any of the following problems?” Shortness of breath, chest pain
— “Have you ever been diagnosed with gestational diabetes by a glucose tolerance test during pregnancy?”
Information Obtained: Research published in the November 2009 issue of Hypertension investigating the association between deployment and newly reported hypertension found that deployers who reported multiple combat exposures, especially those who personally witnessed death due to war or disaster, were at higher risk for newly reported hypertension, possibly indicating a stress-induced hypertensive effect.
Research findings published in the December 2009 issue of the American Journal of Epidemiology highlighted the first prospective population-based study to include changes in smoking as a covariate for investigation of increased risk for respiratory symptoms, asthma, bronchitis, and emphysema. This study suggested an elevated risk for respiratory symptoms, including persistent and recurring cough and shortness of breath, among Army and Marine Corps personnel deployed in support of the operations in Iraq and Afghanistan.
Follow-up of at-risk populations will allow for better understanding of the potential episodic nature of potentially acute and transient or early stages of chronic respiratory and cardiovascular illnesses.
Demographic data such as age, geographic home of record, and race/ethnicity are collected through the Defense Manpower Data Center. Income, education level, marital status, and military occupation are all asked on the 2010 Millennium Cohort Study survey.
Who Pays for Data Collection?: The Millennium Cohort Study is funded by the DOD through the U.S. Army Medical Research Materiel Command (USAMRMC) Military Operational Medicine Research Program (MOMRP) and conducted at the Naval Health Research Center (NHRC).
Dissemination of Data: The Millennium Cohort Study team has published over 30 manuscripts in notable peer-reviewed journals, produced greater than 150 conference presentations and posters since 2001, and has received multiple research awards at notable public health conferences.
Findings from the Millennium Cohort Study have been regularly presented to various organizations within the Department of Defense, as well as leading scientific institutions that advise military leaders and policy makers.
Additionally, several press outlets, including Reuters Online, the Atlanta Journal Constitution, Forbes.com, and the Defense Military Health System, featured articles on the study’s research.
The Millennium Cohort Study surveys are available online to participants during survey cycles. Participants can gain access to the survey by logging-in with their assigned subject ID number and the last four digits of their Social Security number.
In order to protect participants’ privacy and within rules governing human subjects research, identifiable data are not shared outside of the NHRC. Collaborations are encouraged and current collaborations exist with many external DOD, VA, academic, and other civilian institutions. Establishment of a memorandum of understanding or other data use agreement and support from one or more Millennium Cohort Study co-investigators must be obtained before data transfer of de-identified data can take place, though collaboration on various projects where data are not transferred is easily accomplished by outside researchers.
Main Purpose of Study: The Multi-Ethnic Study of Atherosclerosis was initiated in July 2000 to investigate the prevalence, correlates, and progression of subclinical cardiovascular disease (CVD) in a population-based sample (Bild et al., 2002). The objectives of MESA are (1) to determine characteristics related to progression of subclinical CVD to clinical CVD; (2) to determine characteristics related to progression of subclinical CVD itself; (3) to assess ethnic, age, and sex differences in subclinical disease prevalence, risk of progression, and rates of clinical CVD; (4) to determine relations of newly identified factors with subclinical disease and to determine their incremental predictive value over established risk factors; and (5) to develop methods, suitable for application in future screening and intervention studies, for characterizing risk among asymptomatic persons.
Sample: 6,500 men and women from six regions in the United States: Baltimore City and Baltimore County, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles County, California; New York, New York; St. Paul, Minnesota; Approximately 1,083 eligible participants per site (Bild et al., 2002). Minorities are not oversampled. Age range is 45–84, and includes men and women.
Response Rate: MESA Exam 1: Among those households contacted, recruiters explained the study to 29.0%, and among those, the participation rate was 39.8%. Among those screened and deemed eligible, the participation rate was 59.8% (from MESA study website).
Frequency of Collection and Sources of Data: Examination 1 (July 2000–July 2002); Examination 2 (July 2002–January 2004); Examination 3 (January 2004–July 2005); Examination 4 (July 2005–July 2007).
Sources of data include blood samples, medical exam records, and interviews
Mode of Data Collection: Medical exam, physical exam, laboratory exam, personal interview.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Baseline measurements will include measurement of coronary calcium using computed tomography; measurement of ventricular mass and function using cardiac magnetic resonance imaging; measurement of flow-mediated brachial artery endothelial vasodilation, carotid intimal-medial wall thickness, and distensibility of the carotid arteries using ultrasonography; measurement
of peripheral vascular disease using ankle and brachial blood pressures; electrocardiography; and standard CVD risk factors; blood samples; identification and characterization of CVD events, including acute myocardial infarction and other coronary heart disease, stroke, peripheral vascular disease, and congestive heart failure; therapeutic interventions for CVD; and mortality (Bild et al., 2002).
Information Obtained: Prevalence, anthropometry, blood pressure, personal history, diet assessment medical history. Whether it includes clinical care age, race/ethnicity, geographic region, sex.
Who Pays for Data Collection?: National Heart, Lung, and Blood Institute (NHLBI) and the National Institutes of Health (NIH).
Main Purpose of Study: The National Longitudinal Study of Adolescent Health (Add Health) is a study of a nationally representative sample of more than 20,000 individuals that began with in-school questionnaires administered to adolescents in grades 7–12 in the United States during the 1994–1995 school year. Since then, four additional waves of in-home interviews in 1995, 1996, 2001–2002, and 2008 have been conducted, tracking the initial cohort as they transitioned out of adolescence into young adulthood and adulthood.
Add Health combines longitudinal survey data on respondents’ social, economic, psychological, and physical well-being with contextual data on the family, neighborhood, community, school, friendships, peer groups, and romantic relationships, providing unique opportunities to study how social environments and behaviors in adolescence are linked to health and achievement outcomes in young adulthood. The later waves of interviews expanded the collection of biological data in Add Health to understand the social, behavioral, and biological linkages in health trajectories as the Add Health cohort ages through adulthood.
Add Health was developed in response to a mandate from the U.S. Congress to fund a study of adolescent health. The original purpose of the study was to examine adolescent health and health behavior with special emphasis on the effects of multiple contexts of adolescent health. As participants aged into adulthood, however, the scientific goals of the study expanded and evolved.
Waves I and II, conducted when respondents were between 12 and 18 years old, focus on the forces that may influence adolescents’ health and risk behaviors, including personal traits, families, friendships, romantic relationships, peer groups, schools, neighborhoods, and communities.
Wave III, conducted when respondents were between 18 and 26 years old, focuses on how adolescent experiences and behaviors are related to decisions, behavior, and health outcomes in the transition to adulthood. Specific aims of Wave III included obtaining relationship, marital, childbearing, and educational histories, and dating key labor force events.
Wave IV, the most recent follow-up, was conducted when respondents were between 24 and 32 years old and assuming adult roles and responsibilities. The fourth wave of interviews focused on obesity, stress, and health risk behaviors and expanded the collection of biological data in Add Health to better understand genetic, stress, and pre-disease pathways.
Sample Size and Sample for Each Wave
- Wave I, Stage 1. 90,118 Adolescent In-School Questionnaires, 164 School Administrator Questionnaires, 20,745 Adolescent In-Home Interviews, 17,669 Parent Questionnaires (parent specific component), 17,713 Parent Questionnaires (child specific component)
— A stratified, random sample of all high schools in the United States was undertaken. A school was eligible for the sample if it included an 11th grade and had a minimum enrollment of 30 students. A feeder school—a school that sent graduates to the high school and that included a 7th grade—was also recruited from the community.
— A sample of 80 high schools and 52 middle schools from the United States was selected with unequal probability of selection. Incorporating systematic sampling methods and implicit stratification into the Add Health study design ensured that this sample is representative of U.S. schools with respect to region of country, urbanicity, school size, school type, and ethnicity.
- Wave I, Stage 2
— An in-home sample of 27,000 adolescents was drawn consisting of a core sample from each community plus selected special oversamples. Eligibility for oversamples was determined by an adolescent’s responses on the In-School Questionnaire. Adolescents could qualify for more than one sample. In addition, parents were asked to complete a questionnaire about family and relationships.
- Wave II: 128 School Administrator Questionnaires, 14,738 Adolescent In-Home Interviews
— The Wave II in-home interview sample is the same as the Wave I in-home interview sample, with a few exceptions: the majority of 12th-grade respondents were removed from the Wave II sample, as they exceeded the grade eligibility requirement; the Wave I disabled sample was not reinterviewed at Wave II; the Wave II sample contains a small number of adolescents who did not participate in the first wave; and no parent interview was conducted at Wave II.
— In addition, school administrators were contacted by telephone to update school information. Information about neighborhoods/communities was gathered from a variety of previously published databases.
- Wave III: 15,197 Young Adult In-Home Interviews and biomarker collection
— The in-home Wave III sample consists of Wave I respondents who could be located and reinterviewed six years later. A sample of 1,507 partners of original respondents was also interviewed. Wave III also collected High School Transcript Release Forms as well as samples of urine (for sexually transmitted infections) and saliva (for HIV testing and, for full siblings and twins, DNA extraction).
- Wave IV: 15,701 Adult In-Home Interviews and biomarker collection
— All original Wave I in-home respondents were eligible for in-home interviews at Wave IV. Wave IV also included collection of blood pressure readings, anthropometric measures (height, weight, and waist circumference), saliva for DNA, and blood spots from a fingerstick from all consenting respondents. To estimate the reliability of biological measures in the Add Health population, anthropometric measures and collection of biospecimens on 100 Wave IV respondents were repeated.
Minorities Are Oversampled: Adolescents of Chinese, Cuban, Puerto Rican descent, and Blacks from well-educated families were oversampled (see below). Eligibility for oversamples was determined by an adolescent’s responses on the In-School Questionnaire. Adolescents could qualify for more than one sample. Blacks from well-educated families—1,038 black adolescents with at least one parent with a college degree. Chinese—334 adolescents. Cuban—450 adolescents. Puerto Rican—437 adolescents
Age Groups Included: Wave I—Adolescents in grades 7–12 during the 1994–1995 school year. Wave II—Adolescents in grades 8–12 during the 1995–1996 school year. Wave III—Young adults age 18 to 26 years old. Wave IV—Adults age 24 to 32 years old.
Men and women are represented in the study in approximately equal proportions.
Response Rate: Wave I—78.9%. Wave II—88.2%. Wave III—77.4%. Wave IV—80.3%
Source of Data: The primary sources of data collection for Add Health are personal interviews, physical examinations, administrative records (e.g., high school transcripts), and publicly accessible databases (e.g., for information about neighborhoods and communities).
Mode of Data Collection: The primary modes of data collection for Add Health are in-school questionnaires (study participants), mailed questionnaires (school administrators), in-home personal interviews (study participants and their parents, siblings, and romantic partners), in-home biospecimen collection (study participants), and records abstraction.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes:
- Has a doctor ever told you that you have high cholesterol?
- Have you ever been diagnosed with high blood pressure or hypertension?
- Have you ever been diagnosed with diabetes?
- Have you ever been diagnosed with asthma?
- Has a doctor, nurse, or other health care provider ever told you that you have or had: high blood cholesterol or triglycerides or lipids?
- How old were you when the doctor, nurse, or other health practitioner diagnosed you with high blood cholesterol or triglycerides or lipids?
- Has a doctor, nurse, or other health care provider ever told you that you have or had: high blood pressure or hypertension (if female add, when you were not pregnant)?
- How old were you when the doctor, nurse, or other health practitioner diagnosed you with blood pressure or hypertension?
- Has a doctor, nurse, or other health care provider ever told you that you have or had: high blood sugar or diabetes (if female add, when you were not pregnant)?
- How old were you when the doctor, nurse, or other health practitioner diagnosed you with high blood sugar or diabetes?
- Has a doctor, nurse, or other health care provider ever told you that you have or had: heart disease?
- How old were you when the doctor, nurse, or other health practitioner diagnosed you with heart disease?
- Has a doctor, nurse, or other health care provider ever told you that you have or had: asthma, chronic bronchitis, or emphysema?
- How old were you when the doctor, nurse, or other health practitioner diagnosed you with asthma, chronic bronchitis, or emphysema?
Height and weight were also measured at Waves II and III, and DNA was collected from a subset of respondents at Wave III. In addition, Wave IV respondents were asked to participate in an inventory of all prescription medications they had used in the four weeks prior to the interview, anthropometric measures, blood pressure and pulse readings, collection of dried blood spots obtained from a finger prick, and collection of saliva for buccal cell DNA.
- Self-reported health history (Waves III and IV)
- Prescription medication inventory (Wave IV)
- Measured height (Waves II, III, and IV)
- Measured weight (Waves II, III, and IV)
- Calculated BMI (Waves II, III, and IV)
- Waist circumference (Wave IV)
- Arm circumference (Wave IV)
- SBP (Wave IV)
- DBP (Wave IV)
- Pulse (Wave IV)
- Pulse pressure (Wave IV)
- Mean arterial pressure (Wave IV)
- Metabolic measures (Wave IV)
- Inflammatory measures/high sensitivity C-reactive protein (Wave IV)
- Immune measures/Epstein-Barr virus antibodies (Wave IV)
- Candidate-gene and genome-wide genetic measures (Waves III and IV)
- Incidence data for high cholesterol, hypertension, diabetes, and asthma are available for all respondents who provided self-reported health history information at Waves III and IV.
- Prevalence data were captured through self-reported health histories at Wave III, and through self-reported health histories, a prescription drug inventory, and biological measures (anthropometric, cardiovascular, metabolic, inflammatory, immune, and genetic) at Wave IV.
- Functional health outcomes from the Add Health dataset include:
— Respondent’s participation in various types of physical activity
— How much the respondent’s health limits him/her in a range of activities
— Whether the limitation in activities is caused by a condition that has lasted more than a year, or a condition that has developed recently
— Whether the respondent uses a brace, cane, wheelchair, or other device because of a physical condition
— How often a health problem has caused the respondent to miss a day of school or work
Risk Factors, Including Stressors: By combining longitudinal social, behavioral, and environmental data with new biological data, Wave IV greatly expanded the breadth of research questions that can be addressed in Add Health regarding pre-disease pathways, gene-environment interactions, the relationship between personal ties and health, factors that contribute to resilience and wellness, and environmental sources of health disparities.
The following are examples of risk factors and stressors available for analysis in Add Health: Hypertension, Diabetes, Hyperlipidemia, Genetic risk factors, Tobacco use, Alcohol and drug abuse, Obesity, Frequent intake of fast food and sweetened drinks, Low level of physical activity, Depression, Perceived stress, Personality factors, Poor mastery, Weak friendship ties, Religiosity factors, Poor sleep quality, Mental and physical job stress, Poor job satisfaction, Unemployment, Military combat experience, PTSD, Low respondent and/or parent education, Low respondent and/or parent income, Receipt of public assistance by respondent and/or parent, Lack of insurance, Lack of access to health care, Poor school quality, Homelessness, Disability, Poor relationship with parents, Foster care, Parental incarceration, Marriage dissatisfaction, separation and divorce, Unwanted pregnancies, Fertility problems, Parenting stress and single parenting, Child diagnosed with severe disease or chronic condition, Abuse and/or neglect by parent or partner, Victimization (i.e., witnessed violence, victim of physical or sexual assault), Criminal offending, involvement with the criminal justice system, and incarceration, Friend or family suicide, Death of parent, spouse, sibling, or child, Unsafe school, Unsafe living environment, Neighborhood/community crime, Neighborhood/community unemployment, Poor access to public parks and physical activity resources, Unfavorable community climate and weather, Unfavorable social policies and programs.
Clinical Care Information: Health insurance coverage, Access to health services, Length of time since last routine checkup, Length of time since last dental exam, Psychological or emotional counseling in the past 12 months
Demographic Characteristics Collected: Age, Sex, Race/ethnicity, Geographic region, Citizenship, Primary language, Languages spoken/written, Income, Education level, Occupation, Military status, Health insurance status, Family size/structure, Religious affiliation, Sexual orientation.
Who Pays for Data Collection?: The Add Health Study is funded by 24 Federal agencies and foundations:
- Eunice Kennedy Shriver National Institute of Child Health and Human Development
- MacArthur Foundation
- National Cancer Institute
- National Center for Health Statistics, Centers for Disease Control and Prevention, HHS
- National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, HHS
- National Center for Minority Health and Health Disparities
- National Institute of Allergy and Infectious Diseases
- National Institute of Deafness and Other Communication Disorders
- National Institute of General Medical Sciences
- National Institute of Mental Health
- National Institute of Nursing Research
- National Institute on Aging
- National Institute on Alcohol Abuse and Alcoholism
- National Institute on Drug Abuse
- National Science Foundation
- Office of AIDS Research, NIH
- Office of Behavioral and Social Sciences Research, NIH
- Office of Minority Health, Centers for Disease Control and Prevention, HHS
- Office of Minority Health, Office of Public Health and Science, HHS
- Office of Population Affairs, HHS
- Office of Research on Women’s Health, NIH
- Office of the Assistant Secretary for Planning and Evaluation, HHS
- Office of the Director, NIH
- Robert Wood Johnson Foundation
Dissemination of Data: Data from all waves of the Add Health study are disseminated by the Inter-University Consortium for Political and Social Research (ICPSR) as a part of their Data Sharing for Demographic Research
(DSDR) project. The DSDR Add Health webpage contains the Add Health study description, publications list, documentation files, and data sets for analysis: http://www.icpsr.umich.edu/cocoon/DSDR/STUDY/21600.xml.
The Add Health data are available in two forms—public-use data sets and restricted-use contractual data sets. The public-use data sets contain data on only a subset of respondents (to protect the confidentiality of study respondents). The public-use data sets are available to any researcher and may be obtained from either DSDR (at no charge) or Sociometrics (for a fee). The restricted-use contractual data sets contain more extensive data and are distributed by DSDR to certified researchers who commit themselves to maintaining limited access. To be eligible to enter into a contract, researchers must have an IRB-approved security plan for handling and storing sensitive data and sign a data-use contract agreeing to keep the data confidential.
The Add Health public-use data are currently available for download from the DSDR website. To download the data, users must create an online personal account. The system uses an email address as a login ID and requires a password.
The Add Health restricted-use contractual data will be available for download from the DSDR website in Summer 2010. Restricted-use data will be available to certified researchers who sign an online restricted-use data agreement.
Through its Data Sharing for Demographic Research project, ICPSR supports a system for analyzing the Add Health public-use data online. The online analysis system allows users to run both simple and complex analyses, recode and compute new variables, and download subset variables or cases. To access the online analysis system, users must create an online personal account.
Main Purpose of Study: Olmstead County Study includes two cohorts on coronary heart disease (CHD) and heart failure (HF) respectively. The CHD cohort examines the incidence of and survival after ACS (MI and UA) in a geographically defined population; prospectively characterizes the case mix and outcome of ACS in the population; and prospectively examines the value of novel biomarkers for risk prediction in ACS in the population. The HF cohort examines the incidence of all-cause and HF-specific hospitalizations after HF diagnosis (both in- and outpatient cases) in a geographically defined incidence cohort of persons with validated HF identified between 1979 and present; prospectively examines the incidence and total number of hospitalizations according to the type of HF (EF and diastolic function); and prospectively measure health status using a generic (SF-12) and disease specific (KCCQ) instrument to examine the association between health status and hospitalizations.
Sample: CHD cohort: 4,059 incident cases of MI, including both prospective and retrospective cohorts; HF cohort: 3,386 cases, including both prospective and retrospective cohorts.
Men and women age 18 and above residing in Olmstead County, Minnesota, were eligible for the study. Minorities were not oversampled. The response rate for the CHD cohort, prospective component was 82%, retrospective component 97%; and for the HF cohort, prospective component 67%, retrospective component 97%.
Sources of Data: Medical records—the Rochester Epidemiology Project (REP) indexes and links medical records from all sources of care for Olmsted County residents; Informed consent—Questionnaires and biobanking
Mode of Data Collection: Medical record data abstraction by a nurse; questionnaires and biobanking
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: cardiovascular risk factors and comorbidities
Information Obtained: Incidence; some functional health outcomes for Heart Failure; risk factors that include diabetes, hypertension, hyperlipidemia, and body mass index; comorbidities, including malignancies, cerebrovascular disease, and liver disease; detailed clinical care information on medications, labs, and procedures; demographic information including age, sex, race, ethnicity, education, marital status, income level.
Cost: CHD: $4,047,519 (1/1998-5/2010); HF: $3,495,062 (1/2003-6/2010)
Who Pays for Data Collection?: CHD: NIH grant—RO1HL59025; HF: NIH grant—RO1HL72435
Dissemination of Data: Data sharing is contingent upon appropriate compliance with HIPAA regulations. Specifically, on a case-by-case basis, “limited data sets” as defined by the U.S. Department of Health and Human Services can be released to qualified investigators who have obtained approval from all appropriate IRBs. Comment: Case ascertainment: CHD: ARIC criteria; HF: Framingham criteria; Study design: CHD: retrospective—medical
records abstraction; HF: prospective—informed consent with questionnaires and biobanking; Data disseminated through publications.
Main Purpose of Study: The Rancho Bernardo study examines the prevalence of heart disease risk factors and follow-up for common chronic diseases.
Sample: Includes approximately 6,000 men and women aged 30 years and above in a suburb of San Diego. Minorities are not oversampled. The Racho-Bernardo Study is a census-based geographically defined “new town” study. Baseline evaluation took place between 1972–1974, and the most recent 2007–2010. The majority of evaluations were limited to those aged 50 and above at the time of visit. The response rate was 82% of baseline.
Frequency of Collection and Sources of Data: Annual mailer for vital status; Clinical evaluations every 2–4 years depending on funding. Population interviews; physical examination; blood tests; hospital validation of selected diseases
Mode of Data Collection: Annual mailer and periodic in-clinic evaluation
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Has a doctor ever told you that you have had a heart attack, heart failure, angina, stroke, TIA, coronary artery revascularization, diabetes, emphysema, asthma? Rose angina and claudication questionnaire, family history of cardiovascular disease, diabetes, and premature mortality; lifestyle including diet, smoking, alcohol, physical activity, and current medication use.
Information Obtained: Incidence data on CVD, diabetes, death, death certificates coded by nosologist. Prevalence of all common diseases. Functional health outcomes: Instrumental Activities of Daily Living (IADL), Activities of Daily Living (ADL), SF-36 or SF-12 on 3 occasions, ability to walk unaided, balance, chair stand, grip strength, pulse rate, spirometry, cognitive function tests. Risk factors: blood pressure, height, weight, waist girth, pulse rate; Stress, happiness, QWB.
Clinical care information includes blood tested for cholesterol, triglycerides, HDL, CRP-IL-6, fasting and post-challenge glucose, liver and kidney function tests, other biomarkers, ECGs; Selected diseases information validated from hospital records. Demographic information: age, race/ethnicity, geographic region, income, education level, occupation, health insurance status, frequency of health care visits, hospitalizations, surgeries
Who Pays for Data Collection?: NIH grants from various institutes; extramural study
Dissemination of Data: Any legitimate investigator can collaborate; Sometimes data are sent to a small group.
This is the longest running extramural study of aging with good diabetes data and OGTT. Diabetes (as a disease and as a CHD risk factor) has been primary focus since inception. More than 400 publications, and countless junior faculty, students, and visiting scholars have worked with these data. The study has a biobank for biomarkers, including hormones, adipocytokines, etc., and has measured another 40 markers from this cohort awaiting analysis. There are also small studies of Filipinos and African Americans—about 400 each, using the same protocol for race/ethnicity studies focused on diabetes and CVD risk factors. The study have measures of subclinical CVD including ankle-brachial index for lower extremity arterial disease, carotid ultrasound, and coronary artery calcium. Share data for meta-analyses regularly (participated in about 10), mostly from collaborative study groups at Oxford and Cambridge, UK; Unable to afford Medicare data.
Main Purpose of Study: The main purpose is to determine the prevalence and incidence of cardiovascular disease (CVD) and its risk factors in American Indians, and to identify important CVD risk factors in this population.
Sample: Initial cohort: 4,549 (since 1989); Family cohort (Strong Heart Family Study or SHFS): 3,838 from 94 families (since 2001). The sample is approximately evenly distributed among the three sites: Arizona (AZ), North/South Dakota (DK), and Oklahoma (OK). All participants are American Indians. Age groups included: Initial cohort: 45–74 years, Family cohort: ≥ 15 years. Both men and women are included. The response rate for the initial cohort was 62% of the total population aged 45–74 years.
Frequency of Data Collection: Approximately every four years since 1989.
Source of Data: Personal interviews, physical exams, medical records, and hospital discharges.
Mode of Data Collection: In-person exams and interviews and medical record abstraction.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Many questions and procedures (including ECG, echocardiogram, Carotid ultrasound, pulmonary function testing, fasting glucose, and insulin measurements).
Information Obtained: Incidence, prevalence, functional health outcomes, risk factors (including stressors), limited clinical care information, and demographic characteristics (participants are covered by Indian Health Service)
Who Pays for Data Collection?: NHLBI/NIH
Dissemination of Data: At this time, data are available only to investigators who have a paper that has been approved by the SHS Publications and Presentations Committee or who have an ancillary study that has been approved by the SHS Steering Committee.
Main Purpose of Study: CARES is a quality improvement registry to help local EMS administrators and medical directors identify when and where cardiac arrest occurs, which elements of their EMS system are functioning properly in dealing with these cases, and what changes can be made to improve outcomes. The ultimate goal of CARES is to help local EMS administrators and medical directors identify the populations affected by out-of-hospital cardiac arrests as well as when and where these arrests occur, to identify the elements of the system that are functioning properly and those that are not, and to provide information on how to affect systems changes to improve cardiac arrest outcomes. Geographic region information may be obtained by the CARES program managers through data reporting.
Sample: As of March 2010, there are CARES sites in approximately 40 communities in 23 states and the District of Columbia. Communities are located in each region of the United States (East Coast, West Coast, Central, Hawaii, and Alaska). Sites include communities at the county level, city level, state level, and work with two non-U.S. countries. CARES uses a Utstein style of statistics to identify and track cases of out-of-hospital cardiac arrest in defined geographical areas. CARES includes both men and women 18 years or older. Minorities are not oversampled.
Frequency of Data Collection: CARES uses a secure web-based system in which event information is entered directly by the participating sites as the events occur.
Sources of Data: EMS agencies, hospitals, computer-aided dispatch systems
Mode of Data Collection: Completion of the CARES form, direct entry into the website database, or exporting electronic data form field software programs
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The registry is designed specifically for out-of-hospital cardiac arrest and asks questions that deal with the cardiac event such as location of the arrest, whether the arrest was witnessed or not, presumed etiology of the arrest and associated resuscitation attempt information, and arrest rhythms.
Information Obtained: The clinical care information covers the care the patient received from the EMS providers prior to arrival at the hospital. The demographic characteristics collected are age, race/ethnicity, and gender by the individual sites.
Who Pays for Data Collection?: The Centers for Disease Control and Prevention provides funding to Emory University to develop and implement the CARES program. Historically, the CDC has provided between $300,000 and $500,000 annually to the program. The participating sites are not charged a fee to join the registry. The CDC provides the funding annually to the program for operational assistance. None of the sites pays into the program for participation.
Dissemination of Data: All the data are online on a protected web server. Data are made available to the CDC for reporting as part of CDC’s National CVD Surveillance System. The CARES website has both a password protected area and a public area where information may be found about the program.
The participating sites have access to reporting functions that provide performance benchmarking information but not access to any identifiable information. Access is granted through the program managers and is usually provided only to sites who participate in the registry.
Main Purpose of Study: To evaluate the epidemiology, quality of care, and outcomes of cardiovascular disease and to conduct future clinical trials using a community-based mode (Go et al., 2008). CVRN’s overall goals are to provide more robust CVD and related healthcare surveillance data than are currently available; promote research on clinical practice and quality of care; enable assessment of new diagnostic and therapeutic technologies and clinical guidelines on CVD and risk factor incidence, prevalence, clinical management, and patient outcomes over time; facilitate research on determinants of disease for uncommon disease phenotypes; and create opportunities for interested non-network researchers to collaborate with CVRN investigators on high-priority cardiovascular issues. Participants include the HMO Research Network (HMORN) is a consortium of 15 U.S. healthcare delivery systems. HMORN’s primary goal is to transform healthcare practice through population-based research to improve the health and health care of broad populations. HMORN accomplishes this by fostering research collaborations; enhancing awareness of research interests, resources, and capabilities of the member research centers; sharing methodologies, best practices, and consultative expertise; and leveraging the HMORN’s strengths (http://www.hmoresearchnetwork.org/resources/tools/HMORN_Brochure.pdf).
Sample: More than 11 million U.S. citizens; 15 research centers, including the Center for Health Services Research, Detroit, Michigan, Henry Ford Health System, Health Alliance Plan; Department of Research and Evaluation, Pasadena, California, Kaiser Permanente Southern California; Geisinger Center for Health Research, Danville, Pennsylvania, Geisinger Health System; Group Health Research Institute, Seattle, Washington, Group Health Cooperative; Harvard Medical School Department of Ambulatory Care & Prevention, Boston, Massachusetts, Harvard Pilgrim Health Care; HealthPartners Research Foundation, Minneapolis, Minnesota, HealthPartners; Kaiser Division of Research, Oakland, California, Kaiser Permanente Northern California; Kaiser Institute for Health Research, Denver, Colorado, Kaiser Permanente Colorado; Lovelace Clinic Foundation, Albuquerque, New Mexico, Lovelace Health Systems; Marshfield Clinic Research Foundation, Marshfield, Wisconsin, Marshfield Clinic, Security Health Plan of Wisconsin; Meyers Primary Care Institute, Worcester, Massachusetts, Fallon Community Health Plan, Fallon Foundation, University of Massachusetts Medical School; Scott & White Health Care Division of Research, Temple, Texas, Scott & White Health Plan; The Center for Health Research—Northwest, Portland, Oregon, Kaiser Permanente Northwest; The Center for Health Research—Hawaii, Honolulu, Hawaii, Kaiser Permanente Hawaii; The Center for Health Research—Southeast, Atlanta, Georgia, Kaiser Permanente Georgia.
The sample includes men and women of all ages. Minorities are not oversampled. Retention rates of cohort from 2002 vary by CVRN site: 70–90% (1-year); 50–70% (3-year); 40–60% (5-year) (http://www.cvrn.org/info/retention.html).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: CVRN research studies: (1) Hypertension Recognition, Treatment, and Control in Community Practice evaluates three sites involving > 850,000 subjects with recognized and unrecognized hypertension; (2) Community-based Control and Persistence of Warfarin Therapy and Associated Rates and Predictors of Adverse Clinical Events in Atrial Fibrillation and Venous Thromboembolism evaluate quality of care and therapy risk across 5 sites; (3) Implantable Cardioverter Defibrillators for Primary Prevention in Community Practice—Clinical Characteristics, Outcomes, Resource utilization, and Cost; Additional funding to study longitudinal outcomes of patients receiving drug eluting stents, to assess risks and benefits of varying durations of clopidogrel therapy in patients receiving intracoronary stents and to assess the comparative effectiveness of second-line anti-hypertensive agents in patients whose blood pressure is not controlled on a thiazide diuretic alone.
Information Obtained: Administrative database—deaths; Clinical database—hospitalizations, emergency room, laboratory, long-term care, home health care, pharmacy, members with Rx benefit. Age, race, gender.
Who Pays for Data Collection?: Cardiovascular Research Network (CVRN) is a 5-year grant funded in 2007 by the National Heart, Lung, and Blood Institute (NHLBI).
Dissemination of Data: Data are not available online. The CVRN data network is guided by the creation and maintenance of a distributed architecture for data storage based on standardized data definitions; governance based on data holders’ ability to opt in or out of individual activities; and transfer of the minimum amount of required data in order to maximize data security (Magid et al., 2008).
Main Purpose of Study: The Bronchiectasis Research Registry is a consolidated database of non-cystic fibrosis (non-CF) bronchiectasis patients enrolled at multiple U.S.-based clinical institutions. The goal of the Bronchiectasis Research Registry is to support collaborative research and assist in the planning of multicenter clinical trials for the treatment of non-CF bronchiectasis, a progressive, non-curable disease of the lungs that afflicts thousands of patients. The registry will also be used to provide better insight into the diagnosis of the etiology of the different types of bronchiectasis, as well as the pathophysiology of the disorder.
Sample: There are currently 800 patients enrolled in this registry.
Patients are enrolled at 11 sites located throughout the United States. In process of completing and executing regulatory documents to allow three additional U.S.-based sites to join the consortium. The consortium will be expanding to include international sites in 2011. The participating centers enrolling patients are located across the United States. Minorities are not oversampled. All patients aged 18 years or older with a diagnosis of bronchiectasis who complete appropriate informed consent procedures are included. Confirmed diagnosis is defined as cough and/or daily or frequent mucopurulent phlegm plus chest imaging (X-ray or HRCT) showing dilated and thickened airways. Both men and women are encouraged to participate.
Frequency of Data Collection and Sources of Data: Data are collected at baseline visit and at annual follow-up visits. Data acquisition is carried out in one of two modes: (1) Medical records are reviewed and data are abstracted in order to complete the forms on patients seen in the past 3 years by a participating investigator and (2) Forms are completed during a clinical examination scheduled as part of patients’ usual care during the study period, and/or through patient contact.
Mode of Data Collection: The DMS can be used either for electronic capture (EDC, with data recorded directly onscreen and validated during collection) or for distributed data management (with data recorded first on paper forms and then keyed and validated at the participating centers). Sites have the flexibility to select a data collection method that suits their needs. Up-to-date paper versions of each data collection instrument are available for situations in which the computer systems are not appropriate or are inaccessible for any reason.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Data collection includes personal contact information (provided consent has been provided); demographic characteristics; medical history; and clinical procedures relative to the treatment of bronchiectasis, namely, respiratory symptoms, pulmonary function testing, HRCT parameters, diagnostic tests, and therapies.
Information Obtained: Incidence, prevalence, functional health outcomes, risk factors, including stressors, clinical care information, and demographic characteristics.
Cost of Registry: It costs approximately $200,000 a year to maintain the registry and pay sites for enrolling and following up patients.
Who pays for Data Collection: The COPD Foundation and the Richard H. Scarborough Fund as well as industry partners provide funding for the registry including data collection.
Dissemination of Data: A web-based data management system (DMS) is used for the Bronchiectasis Research Registry. The data management system provides all of the capabilities required for research data management, including: data entry, data transfer, data validation, database updating, database closure, data retrieval, data entry, security and confidentiality, and archiving. The server and registry database reside at the coordinating center.
As an alternative to keying data, the system can accept data files that are transferred into the database via a
batch process. The system can accept Excel files, Access tables, delimited ASCII files, or other standard formats. Reading center and laboratory data are often loaded with this process.
Data files are available online. As part of the registry program, each consortium member has the ability to generate a variety of reports and data queries (for reports) through the Interactive Reporting System. This system is web-based and provides an attractive and powerful interface that supports several types of information requests. Features of the interface are based on usage scenarios provided by the consortium membership and include standard reports and freeform queries. Examples of standard reports include frequency distributions of registry participants by demographic variables, cross-tabulations of two or more clinical indexes, and descriptive summaries of lung function parameters. Statistical analysis is supported by the interface to a level of sophistication determined by data appropriateness and usage scenarios. For categorical variables, the focus is on cross-tabulations and measures of association. Means, medians, standard deviations, and minimum and maximum values are generated for continuous variables. Distributions of both categorical and continuous variables can be requested for all patients in the registry database, for only those patients consenting to be contacted about future studies, and for various demographic or clinical subgroups of patients. Consortium members can generate custom analyses through the use of pull-down menus and check-boxes for variable specification and coding. Statistical modeling is also available for a limited set of applications including linear and logistic regression modeling.
Each consortium member and staff at the University of North Carolina who work on the registry have access to the data. Data confidentiality and security are applied at all levels of data acquisition, transfer, and storage. The DMS developed by the University of North Carolina meets exacting data management standards of confidentiality, as well as HIPAA requirements. Access to the DMS requires two levels of identification, one to gain access to the CSCC server and one to gain access to the registry data management system. Users of the system are assigned job-specific permissions such as data entry, reporting, or query resolution. Confidential data collected from the sites are encrypted by the system and only decrypted for display on-screen by authorized users of the system. It is a requirement for all CSCC staff to complete a confidentiality certification procedure upon employment. Policies regarding the confidential nature of the data collected, processed, and stored at the UNC CSCC are explained to all personnel, who must then sign a “confidentiality certification” to be allowed access to confidential information.
Main Purpose of Study: The COPD Foundation Research Registry is a confidential database of individuals diagnosed with COPD or at risk of developing COPD. The registry was established in 2007 by the COPD Foundation to help researchers learn more about COPD and to help people interested in COPD research find opportunities to participate. The COPD Foundation Research Registry serves as an avenue to collect, analyze, and disseminate data on the phenotypes, progressions, and possible treatments of COPD. The information contained in the registry, such as the clinical characteristics, will inform possible research hypotheses. The registry will also assist with the creation of therapeutic clinical trials by providing the information on the number of subjects available with specific inclusion/exclusion criteria as well as information on the symptoms and medication use by those subjects. The registry is a source of identifying subjects who may qualify for specific trials and who may be willing to be a participant in such a trial.
The specific aims of the COPD Foundation Research Registry are as follows: (1) Create and maintain a registry of up to 50,000 people with COPD or who are at risk for developing COPD that are willing to be contacted to ascertain their interest in participating in clinical research; (2) Use the registry as a source of possible subjects for the COPDGene® Study; (3) Use the registry as a source of possible subjects for future clinical research studies that require people with COPD or people who are at risk for developing COPD; and (4) Use the registry to determine demographic data and clinical characteristics of a broad cross section of people with COPD or at risk for developing COPD.
As stated in Aim 2 above, the COPD Foundation Research Registry is used as a means of helping to recruit patients for participation in the COPDGene® Study. Further information about the COPDGene® Study is contained earlier in this document.
Sample: Size of the COPD Foundation Research Registry: Currently, there are 3,000+ patients who have agreed to participate in the COPD Foundation Registry and be contacted for future participation in research studies. The foundation is continuing to advertise and market the registry to expand the cohort. It is currently looking at Pulmonary Rehabs as an additional avenue to recruit more patients for participation.
Patients enrolled in the registry are located throughout the country. Geographic levels include local, state, and national through subject ZIP codes. The majority of registry participants live in the United States. Anyone who is at least 18 years of age and diagnosed with COPD or who may be considered at risk of the development of COPD may participate. Both men and women are encouraged to participate in the registry. African Americans are a group for which minimal COPD information is available. The registry is helping to ensure that the African American population enrollment goal is met.
The COPD Foundation has many mechanisms to enable individuals to participate in the registry; for example, online submission, fax, and mail submission. It is difficult to assess the response rate given the various methods the foundation has implemented to enable individuals to participate. There are increases in enrollment after the registry is advertised and promoted in publications, when administrative or scientific leadership present and give talks/lectures and when the COPD Foundation exhibits at scientific or patient-related conferences.
Frequency of Data Collection and Sources of Data: Data are collected one time for each participant using the registry survey.
Source of Data: The COPD Foundation Research Registry is a patient reported database. As a result, subjects self-report their data using the registry survey.
Mode of Data Collection: Data are collected via a secure, online web-based survey or paper questionnaire/informed consent document. Patients may download the form and mail it to the Data Coordinating Center or they may request that a paper survey be mailed to their home. Data from web-based surveys are directly submitted into the database. For responses that are mailed, a data entry person inputs the data into the Registry.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Specific questions regarding cardiovascular disease and diabetes are not asked. Specific questions related to chronic pulmonary disease are asked. Patients are asked what tests they have had performed (PFT or spirometry, Chest X-ray, Chest CT, or genetic blood test). Patients are asked about their lung symptoms, breathlessness, smoking history, medications, respiratory history, family history, exacerbations, and general demographic information.
Information Obtained: Incidence, prevalence, functional health outcomes, risk factors, including stressors, clinical care information, and demographic characteristics.
Cost of Registry: The COPD Foundation expends about $10,000–$15,000 annually for the maintenance and operation of the registry.
Who Pays for Data Collection?: The COPD Foundation pays for the data collection and the other costs associated with the creation and management of the registry.
Dissemination of Data: Data files are not available online, but patients may enroll and complete the form online. Per the registry’s informed consent, only authorized staff at National Jewish Health may access the registry’s data. A strict proposal review process, including required IRB approval/exemption documentation, is followed when an individual wishes to access the cohort. The process includes review by a committee that approves/disapproves proposals for access to the cohort. After the proposal has been vetted and approved, patients who are eligible for participation are contacted by the registry’s Data Coordinating Center and the patients are informed that they qualify for participation in a specific research study. Patients are referred to the investigator for participation in the proposed study. It is the responsibility of the patient to contact the investigator. Investigators are never given patients’ contact information
Main Purpose of Study: Aortic dissection is the most common acute aortic condition requiring urgent surgical therapy (Hagan et al., 2000). The International Registry of Acute Aortic Dissection (IRAD), a consortium of research centers, enrolls patients at large referral centers to assess current presentation, management, and outcomes of acute aortic dissection (IRAD website, http://www.iradonline.org/irad.html). More specifically, the research
centers study everything from dates and times of symptom onset, presentation, diagnosis, hemodynamic signs of aortic dissection, initial and chronic medical therapy, to diagnostic imaging chosen, and surgical and medical management.
Sample: IRAD includes 24 large referral centers in 12 countries; data available on 1,600 acute aortic dissection cases (IRAD website, http://www.iradonline.org/irad.html). IRAD sites include University of Michigan Hospital, Washington University Hospital, Mayo Clinic, Minneapolis Heart Institute, University of Pennsylvania Hospital, Brigham & Women’s Hospital, Massachusetts General Hospital, Duke University, St. Michael’s Hospital (Canada), University of Tokyo Hospital (Japan), Tromsø University Hospital (Norway), Hospital Bichat (France), Robert-Bosch Krankenhaus (Germany), University of Rostock Hospital (Germany), Medical School Graz (Austria), University of Vienna Hospital (Austria), Hospital General Universitari Vall d’Hebron (Spain), Hospital General Universitario “12 de Octubre” (Spain), Cardiocentro Ticino (Switzerland), University Hospital S. Orsola (Italy), San Giovanni (Italy), and Hadassah University Hospital (Israel).
Sources of Data: Patient history, physical findings, imaging studies and outcomes; Physician review of hospital records
Mode of Data Collection: A questionnaire of 290 variables developed by IRAD investigators
Information Obtained: Clinical care: prior cardiac surgery: aortic valve placement, aortic aneurysm and/or dissection, coronary artery bypass graft surgery, mitral valve surgery; latrogenic: catheterization/PTCA, cardiac surgery; pain symptoms: any pain reported; syncope. Demographic information: age, gender, ethnicity.
Dissemination of Data: Data are not available online.
Main Purpose of Study: The NCDR provides evidence-based quality improvement solutions for cardiologists and other healthcare providers who are committed to measurement, improvement, and excellence in cardiovascular care. NCDR is comprised of over 2,200 hospitals involved in 5 hospital-based registries and more than 600 offices in our practice-based registry.
Sample: NCDR has participating hospitals from every state and participating practices from every state except Alaska. Benchmarking reports are compiled and presented at the national level. Current reports are at the national level. Beginning to look at the state level, especially in the ACTION Registry-GWTG to support STEMI care initiatives. Participating hospitals and practices are required to submit consecutive patients that meet the inclusion criteria for our registries. Oversampling is not necessary since a 100% sample of the patient population is captured.
Current population records include patients 18 years of age and older, both men and women. With the release of ICD V.2 (April 2010) and IMPACT (Winter 2010), NCDR will be gathering data on all age groups.
While response rates vary between registries, the NCDR enjoys a greater than 95% data submission success by its participants. Data submissions to NCDR are scheduled on a quarterly basis. Data collection is an ongoing process in our member facilities.
Sources of Data: The primary source of data is the hospital or practice medical record. Data collection varies among sites. In many cases, sites capture data concurrently via electronic interfaces with ADT and EMR systems. However, depending on the site’s available technology, there are still many who collect data retrospectively via paper forms. The NCDR provides a web-based data collection tool as a part of registry participation. Irrespective of data collection processes, all data are sent via a secure, electronic submission process.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: All of these patient conditions are captured across our registries, with the exception of asthma. Cardiovascular disease, chronic pulmonary disease, and diabetes are collected across the registries; CathPCI Registry®: procedure-based registry collecting data on diagnostic and interventional caths; ICD Registry™: procedure-based registry collecting data on ICD implants and lead information for adult and pediatric patients; ACTION Registry®-GWTG™: process-based registry collecting data on STEMI and non-STEMI patients; CARE Registry®: procedure-based registry collecting data on carotid revascularization procedures (stenting and endarterectomy); IMPACT Registry™: procedure-based registry collecting data on patients with congenital heart disease (adult and pediatric); in a pilot phase, expected launch late
2010; PINNACLE Registry™: practice-based registry collecting data on patients and the care provided to them in the office setting. Particular focus is on coronary artery disease, hypertension, heart failure and atrial fibrillation.
Information Obtained: Incidence and prevalence data are collected. Functional health outcomes: NCDR collects 30-day follow-up in the CARE Registry and are involved with an ICD longitudinal study. Additional studies are being designed to capture functional status. Risk factors: NCDR collects medical risk factors but does not include stress-related data points. Clinical care information: metrics derived from the submitted data help to drive care improvements in member facilities. Demographic characteristics: the registry collects date of birth, race/ethnicity, health insurance status and sex.
Who Pays for Data Collection?: Annual membership fees are $4,000/year/registry with the exception of PINNACLE Registry and ACTION Registry-GWTG, which are no cost. Data collection staff are supplied by and paid for by the member hospitals or practice.
Dissemination of Data: Data files are available for member facilities through a secure, web-based portal. Data extract tools are available for participating members. A series of online query tools will be launched in 2011. Each participating site creates and maintains appropriate access levels for their designated staff/physicians. Access to the data is managed via a formal request process that is available to any interested party. Each registry has a Research and Publications Committee that assesses research requests for their applicability and feasibility.
NCDR supports robust research on the aggregated limited data it houses. Numerous health plans and state regulators require participation in one or more of its registries as part of their preferred provider program or certificate of need program. Additionally, NCDR is the only data collection option for meeting the CMS Primary Prevention coverage for ICDs. Data Quality: Sites submit data on a quarterly basis to the NCDR. The data are processed and assessed for data quality. The data quality report provides the participant with detailed summaries of the completeness for each data field. Participating sites access their data quality report via the web within minutes of their data submission.
Main Purpose of Study: Survey and improve the quality of acute stroke care from onset through hospital discharge
Sample: Hospital data in six states: Georgia, Massachusetts, Michigan, Minnesota, North Carolina, Ohio. Age group is ≥ 18 years of age. Men and women included. Response rate: Request all hospital admissions for acute ischemic stroke, hemorrhagic stroke, and TIA. Response rate varies from 45-100%.
Frequency of Data Collection and Sources of Data: Continuous; concurrent with care and retrospective. The source of data is hospital medical records.
Mode of Data Collection: Health department staff abstract data from hospital medical records.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Medical history of stroke, TIA, myocardial infarction, diabetes, dyslipidemia, hypertension, atrial fibrillation, cardiac valve prosthesis, congestive heart failure, peripheral vascular disease, sickle cell disease, carotid stenosis, tobacco use; taking antihypertensive medication, taking lipid lowering medication; labs: LDL, HDL, total cholesterol, triglycerides, HbA1c; discharge destination.
Information Obtained: Ambulatory status pre- and post-event; in-hospital mortality; discharge destination; comfort measures status; treatment provided, complications (UTI, DVT/PE, pneumonia), dysphagia screening, smoking cessation counseling, 8 NQF endorsed stroke performance measures; arrival mode; place of occurrence; stroke unit care; neurology consult or care; CT results; principle ICD9 code; primary stroke ICD9 code; clinical diagnosis; stroke severity (Glasgow coma score or NIH Stroke Scale Score); clinical exam characteristics; age, race, Hispanic ethnicity, state, insurance status.
Cost of Registry and Who Pays for Data Collection: $600,000/year per state; includes quality improvement program, 5% data reabstraction; program evaluation. Centers for Disease Control and Prevention pays for the data collection.
Dissemination of Data: Key indicators are reported as part of CDC’s National CVD Surveillance System.
Main Purpose of Study: The Behavioral Risk Factor Surveillance System (BRFSS), a state-based system of health surveys, collects information on health risk behaviors, preventive health practices, and healthcare access primarily related to chronic disease and injury. Across many states, BRFSS may be the only available source of timely, accurate data on health-related behaviors.
Sample: More than 350,000 adults are interviewed each year in the 50 states, District of Columbia, Puerto Rico, U.S. Virgin Islands, and Guam. Data are collected at the state, county, and city levels, with selected metropolitan and micropolitan statistical areas (MMSAs). BRFSS includes noninstitutionalized men and women aged 18 years and older.
Frequency of Collection and Sources of Data: The survey is done annually, and there is monthly state telephone surveillance.
Data are collected using a questionnaire organized by core and optional modules.
Mode of Data Collection: Random digit dialing (RDD) landline, RDD cell phones, computer-assisted telephone interviewing (CATI), and web questionnaire (Mokdad, 2009).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Ever been told you have cardiovascular disease/heart attack and stroke (myocardial infarction/heart attack, coronary heart disease, angina, stroke)?; Daily aspirin intake; stomach problem after aspirin intake; signs and symptoms of a heart attack (weak, lightheaded, faint, chest pain, discomfort, trouble seeing in one or both eyes, pain or discomfort in arms or shoulder, shortness of breath, sudden numbness or weakness of face, arm, leg especially on one side, sudden chest pain, trouble seeing in one or both eyes, severe headache with no known cause); outpatient rehab after stroke; Ever been told you had asthma?; routine check up because of asthma; days unable to work or carry out usual activities because of asthma; symptoms of asthma make it difficult to stay asleep; symptoms of asthma (coughing, wheezing, shortness of breath, chest tightness, and phlegm production without a cold or respiratory infection); child asthma status; Ever been told you have diabetes?; age when were told you have diabetes; insulin/diabetes pills intake; Ever been told you have high blood pressure?; taking medicine for your high blood pressure.
Information Obtained: The BRFSS survey reports information on prevalence of asthma/adult asthma history, cardiovascular disease (heart attack/stroke), diabetes, and health risk factors that include cholesterol and hypertension awareness. BRFSS does not have data on chronic bronchitis or emphysema. Risk factor information about physical activity, fruit and vegetable consumption, smoking, weight, height, BMI, alcohol consumption, diabetes and blood pressure medication, and demographic information (race, gender, age, income, education, children in household, employment status, household income, ethnicity, county of residence, marital status) is collected.
Dissemination of Data: Data files are available online.
Main Purpose of Study: To provide population-based data on a broad set of health and health-related indicators for California. The data are actively disseminated to reach a broad constituency, including policy makers, researchers, state and county health agencies, community-based agencies and organizations, health advocates, and others to support evidence-based policy and decision making.
Sample: CHIS interviews approximately 50,000 households every two years.
The CHIS sample design includes 41 individual county strata and 3 multicounty strata that include the remaining 17 counties with small populations. Koreans and Vietnamese have been oversampled in every CHIS cycle and other groups in various cycles. The large CHIS sample and racial/ethnic diversity of California permits estimates for all major racial/ethnic groups, plus many sub-ethnic groups as well. CHIS provides distinct samples for Chinese, Filipinos, Japanese, South Asians, Koreans, and Vietnamese, and for Latinos/Hispanics from Mexico, Central America, South America, and European-origin.
CHIS conducts separate interviews for adults (aged 18 and older), adolescents (aged 12 to 17, interviewed
directly), and about children (age 12, with information provided by parental proxy). Both genders are included in the CHIS sample.
Response rate—Please see the CHIS methodology report 4 on response rate at http://www.chis.ucla.edu/methodology.html and also studies of CHIS data quality at http://www.chis.ucla.edu/dataquality.html.
Frequency of Data Collection: Biennial from CHIS 2001 through CHIS 2009; data collection will be continuous beginning with CHIS 2011.
Source of Data: Population interviews with persons living in households
Mode of Data Collection: Random digit dialing landline and cell phones
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Topics vary from CHIS cycle to cycle, but age-appropriate questions related to the chronic conditions listed above have been administered to the adult, adolescent, and child samples.
Information Obtained: Prevalence; functional health outcomes including disability (physical and mental); risk factors, including diet, physical activity, alcohol/drug use, smoking, health screening, interpersonal violence, suicide ideation, falls in elderly, mental health, sexual behaviors, etc.; clinical care information, access to health care, usual source of care, and medical home and care coordination; demographic characteristics, including age, race/ethnicity, geographic region, income, education level, occupation, health insurance status, as well as immigration status, sexual orientation, and other characteristics.
Cost of Survey/Registry: Varies from cycle to cycle.
Who Pays for Data Collection?: Data collection is funded by several sources, including private, state, and federal agencies. Sources of funding vary from cycle to cycle; however, there is a core set of large and consistent supporters of CHIS, including the California Department of Health Care Services, the California Department of Public Health, the California Department of Mental Health, the National Cancer Institute, Kaiser Permanente, The California Endowment, and First 5 California.
Dissemination of Data: Statewide pubic-use files are available online at no charge. For convenience, files are provided in several data formats, including SAS, SPSS, and STATA. Detailed documentation for the data files and their content is also provided. Confidential data files with geographic identifiers and sensitive variables are availability for legitimate research through remote access.
CHIS has an online query system called AskCHIS that is a user-friendly and easily accessible tool for generating data estimates on many health indicators at the state and county levels and at sub-county levels for Los Angeles and San Diego counties. Results can be adjusted and displayed to the user’s preferences and output can be displayed as charts, exported to Excel, etc. The data are weighted to the population and rounded to the nearest thousand. The system has more than 20,000 registered users who have made nearly 500,000 queries of the data.
Data access is available to the general public, researchers, county local health departments, federal, state, and community agencies and funders. Those interested in obtaining access to confidential data must apply to use it through a secure data center at the UCLA Center for Health Policy Research. The wide dissemination of the data to achieve usability and application is a priority of CHIS.
Comment: CHIS data are widely used in policy analysis, policy making, seeking funding for health services, research, and news stories and educational efforts. Some of the key impacts of CHIS data are listed in “CHIS Making an Impact” (available at http://www.chis.ucla.edu/chis-impact.html and http://www.chis.ucla.edu/pdf/chis_making_impact.pdf). CHIS data are used in many peer-reviewed research studies, which are listed in http://www.chis.ucla.edu/peerpubs/.
Main Purpose of Study: The Hawaii Health Survey is a continuous statewide household survey that gathers information on health and socio-demographic conditions.
Sample: In 2004, 6,769 adult respondents in households aged 18 and older; 19,699 household members surveyed. Geographic scope is the state of Hawaii, with indicators that include county, island, ZIP code as reported by the respondent, and telephone prefix.
Frequency of Collection and Sources of Data: The survey is conducted annually.
Mode of Data Collection: Population interviews random telephone dialing; computer-assisted telephone interviewing (CATI).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Ever been told to have asthma, diabetes, high blood cholesterol, hypertension?; Health data on asthma, diabetes, high blood cholesterol, hypertension, heart, and lung disease.
Information Obtained: Asthma, diabetes, high blood cholesterol, hypertension, health-related quality of life; mental and physical health scores, suicide ideation, smoking, physical activity, intimate partner violence, orphans, housing, hunger, incarceration, age, gender, income, race, education, household size, insurance status, ethnicity by parents, marital status, employment and jobs, poverty status, food insecurity, military, migration.
Dissemination of Data: Summary data tables by condition are available online. The survey does not include households without telephones, Niihau, group quarters, and the homeless. The Office of Health Status Monitoring (OHSM) at the Hawaii Department of Health is responsible for compilation and analysis of data from vital statistics records and a statewide survey.
Main Purpose of Study: This is an ongoing data project that has spawned over 1,000 research articles. The purpose is to collect hospital, ED, and ambulatory surgery (AS) administrative (claims) data for a wide range of research applications. Full information can be found at www.hcup-us.ahrq.gov. HCUP includes a census of records from participating states, as well as samples of hospital discharges and a national ED sample that provide the ability to make national estimates.
Sample: HCUP currently contains over 90% of all hospital discharges in the United States. The HCUP inpatient data include virtually all discharges from nearly all hospitals in each state that participates—currently 43 states participate in HCUP. Fewer states provide data for ED and AS use. National estimates can be produced for the four census regions; selected state data are also available. Data are available at the local, state, regional, and national levels. Minorities are not oversampled; the data are a census of inpatient discharges, ED visits, and ambulatory surgery encounters from participating states. All age groups and both men and women are included. Response rate—HCUP is not a sample; it contains all records from participating states.
Frequency of Data Collection: Annual, with data since 1988. The number of states has grown over time, so earlier data are based on fewer states. National estimates are possible since 1993.
Source of Data: Administrative (claims) data on hospital discharge records, ED visits, and ambulatory surgery encounters.
Mode of Data Collection: Administrative data collected by state governments, hospital associations, and state data organizations directly from the hospitals. These statewide organizations participate in a voluntary arrangement with AHRQ and provide their data to HCUP for research purposes.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Contains all records with ICD-9-CM codes for diagnoses and procedures to enable selection of particular populations of interest.
Information Obtained: Data include records of discharges/visit/encounters that can form the numerator of rates using population-based denominators. For most conditions/procedures, this will represent a discharge or visit rates, not prevalence or incidence rates. Functional health outcomes are not collected, and risk factors are not reliably collected using ICD-9-CM codes. The clinical care information that is gathered includes procedures that are coded using ICD-9-CM codes. Demographic characteristics collected include age, gender, patient county, health insurance status, and median household income of patients’ ZIP code (available across all data sources); race/ethnicity is available from 38 of 43 states.
Cost of Survey/Registry: This is an ongoing data collection project with data available for over 20 years—not a survey or registry.
Who Pays for Data Collection?: Funded by AHRQ.
Dissemination of Data: Data are not available online but can be obtained through the HCUP Central Distributor: http://www.hcup-us.ahrq.gov/tech_assist/centdist.jsp. Data are available for purchase (states set the prices of the state databases, thus there is a lot of variation across states in the purchase price; the nationwide databases are
available for purchase at prices set by AHRQ). Purchase of all data requires a data use agreement to be signed by the researcher.
The website http://hcupnet.ahrq.gov/ provides access to descriptive statistics from the nationwide data and many states (their participation is voluntary). The online query system is freely available; the data must be purchased. Researchers can purchase the raw data.
Main Purpose of Study: Several states perform broad-based health interview surveys of their states’ populations to more accurately identify health risks and to more fully explore ways to address threats to health. These surveys are designed to provide information that meets the needs of policy makers—data that are (a) local, (b) timely, (c) accurate, (d) adequate descriptive, and (e) informative at each intervention level (state, community, family, individual).
In Illinois, the Illinois County Behavior Risk Factor Survey (ICBRFS) has long served this need for adults. But there is no similar health status survey for children and adolescents. Several key institutions in Illinois are working together to build the Illinois Health Survey for Youth (IHSY)—a health survey that would expand what is known about the health of Illinois children and adolescents and support better public health and educational planning.
Sample: The Illinois BRFSS currently collects data from 17,000 adults in Illinois each year, about 5,000 for the statewide strata, and 12,000 for the county-level strata. After three years, each county has a sample of 400 adults. The Illinois Department of Public Health has cooperated in developing a sampling plan that coordinates these two sampling efforts and that would leave open sample for the youth survey. Although there is general agreement on principals, it is not clear that it will be possible to integrate the youth and the Illinois BRFSS survey. When fully operational, a sample of 20,400 surveys of parents on behalf of children and adolescents over a three-year period is expected. Additionally, adolescents would be given a supplemental survey. About 5,000 completed adolescent supplements over a three-year period are expected.
Illinois County BRFSS samples the geographic regions of 95 local health departments (these are typically county level, with a few exceptions). Counties are preferred because that is where local public health policy is made in Illinois. The 51 strata are used to assure that large complex populations have adequate sample and to preserve at least some of the local health department focus of the Illinois BRFSS.
The IHSY will have 51 geographic strata (using PUMAs and SuperPUMAs). In rural regions of the state, PUMAs combine 3–4 counties. In more densely populated areas of the state, SuperPUMAs divide up counties into subsections. For example, Chicago will have 5 geographic strata; Suburban Cook County will have an additional 5.
Minorities are not oversampled. The Illinois County BRFSS samples adults aged 18 and up; IHSY will sample children aged 0–17. Both men and women are included.
The pilot in Chicago yielded an interview completion rate of 28% for the BRFSS and of 55% for the IHSY. Parents were much more willing to talk about the health of their children than to answer questions about themselves.
Frequency of Data Collection: When fully operational, every three years.
Source of Data: Population interviews
Mode of Data Collection: Random digit dial landline; will use a cell phone supplement.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The Illinois County BRFSS includes the standard BRFSS questions for these. IHST asks about conditions relevant for children and their development. These include asthma and diabetes, health behaviors (PA and nutrition).
Information Obtained: Demographics: Age, gender, race ethnicity, primary language, marital status, household composition, education level, county of residence, ZIP code, number phone lines per household, veteran status, attended school last week. Chronic health conditions: General health, asthma, diabetes, heart attack, angina, coronary artery disease, prostate cancer, birth weight, ADHD (screening, treatment), Autism/Asberger’s (treatment), seizures (treatment), hay fever/breathing, skin allergies, food allergies. Acute health conditions: Frequent headaches, stomach pains, growing pains, serious injury, infectious disease (past 12 months), bed time, wake time, snore frequency, nap frequency, days school missed due to health problem.
The survey also collects information on disability, mental health, women’s health, disease prevention, injury
prevention, health behaviors, interpersonal violence, healthcare utilization and access, employment, household income and program participation, child care, adult supervision, neighborhood environment.
Cost of Survey/Registry: About $1,000,000 per year for the Illinois County BRFSS. About $800,000 per year for IHSY.
Who Pays for Data Collection?: Illinois County BRFSS is paid for through the Preventive Services Block Grant. IHSY is not yet funded. Negotiations are under way with the Illinois Governor’s Office (though it will not be funded this year considering the budget crisis). It has been determined to be eligible for Medicaid matching funds, which will cut the cost to the State of Illinois to $400,000 per year.
Dissemination of Data: Data files are not yet available online (survey not funded). Only the pilot has been conducted. There is a planned online query system which is expected to be public.
Main Purpose of Study: The 2005 IHHS was the second comprehensive, statewide effort to evaluate the health status, access to health care, and social environment of children in families in Iowa. The primary goals of the IHHS were to: (1) assess the health and well-being of children and families in Iowa, (2) assess a set of early childhood issues, (3) evaluate the health insurance coverage of children in Iowa, and (4) assess the health and well-being of racial and ethnic minority children in Iowa.
Sample: The 2005 IHHS was conducted using population-based telephone interviews with a sample of 3,669 families with children in Iowa. It included a targeted oversample of African American and Hispanic children. 331 parents of children from a targeted telephone sample who were identified by a parent as African American (170) or Hispanic (161) were added to the original sample for the purposes of this health disparities study. Two Spanish-speaking interviewers conducted the telephone surveys in Spanish for 105 families who chose to do the interviews in Spanish. The age range was 0–4 yrs., 5–9 yrs., 10–14 yrs., and 15–17 yrs.
Frequency of Data Collection and Sources of Data: Household Health surveys are completed every five years. Interviews were completed with the parents of 3,669 children throughout the state of Iowa.
Mode of Data Collection: Phone numbers dialed included a combination of random digit dial (22%) and phone numbers targeted toward families (78%) and were obtained from a private vendor. Targeted lists came from a variety of resources, including white pages and other lists (e.g., voter registration, magazine subscriptions, and warranty cards).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: In this study, asthma was chosen as a chronic health condition of emphasis. About one of four African American children (23%) have been diagnosed with asthma at some time in their life. In contrast, only 4% of HSI children have been diagnosed with asthma, followed by HEI children (7%), and white children (9%).
- Have you ever been told by a doctor or other health professional that [CHILD] had asthma?
- Does [CHILD] still have asthma?
- How old was [CHILD] when a doctor or other health professional first said [HE/SHE] had asthma?
- How long has it been since you last talked to a doctor or other health professional about [CHILD]’s asthma? This could have been in a doctor’s office, the hospital, an emergency room, or urgent care center. Would you say … ?
- How long has it been since [CHILD] last took asthma medication?
- Symptoms of asthma include coughing, wheezing, shortness of breath, chest tightness or phlegm production when [CHILD] did not have a cold or respiratory infection.
a. How long has it been since [CHILD] last had any symptoms of asthma?
b. Has a doctor or other health professional EVER given you or [CHILD] an asthma action plan?
c. During the past 12 months, how many different times did [CHILD] stay in any hospital overnight or longer because of [his/her] asthma?
Information Obtained: The % of respondents with asthma diagnosis, child’s weight, physical activity, overall eating patterns, parenting stress, child’s screen time (TV/computer), medical care and access to health care, pre-
ventive care, age, race/ethnicity of child/parent, education level of child/parent, health insurance status, household income.
Cost of Survey/Registry: Approximately $75,000
Who Pays for Data Collection?: Funding was provided primarily by the IDPH, with additional funding from the U.S. Department of Health and Human Services Maternal and Child Health Bureau (MCHB) and the Centers for Disease Control and Prevention (CDC).
Dissemination of Data: Final report data are presented to the Iowa Legislature, IDPH, and posted to the public website. The final report is available online, but not the data files. The only people with access to the data files are the researchers on the project who have IRB approval.
Main Purpose of Study: Since 1997, the Los Angeles County Health Survey (LACHS) has functioned as a primary vehicle for gathering information about access to health care, healthcare utilization, health behaviors, health status, and knowledge and perceptions of health-related issues among the LA County population. The overall objective of each survey is to update key health indicators, including health status, health conditions, health-related behaviors, health insurance coverage, and access to care among adults and children living in LA County. Data collected by the survey provide the Department of Public Health and other county and city agencies, along with policy makers, community leaders, academic researchers, and the public itself with information about the health and healthcare needs of LA County residents, in an effort to continuously improve their health status. To address the root causes of poor health, the survey looks beyond individual risk factors for disease to factors in the physical and social environment that influence health, such as availability of fresh fruits and vegetables, availability of community resources for exercise and play, and neighborhood safety.
Sample: The LACHS collects local data from a representative sample of noninstitutionalized LA County residents. Adult Survey: Data collected for the Adult Survey has ranged between 7,200 and 8,648 randomly selected LA County adults (18 years or older). The adult in the household with the most recent birth date is chosen to participate in the survey. The 2007 Adult LACHS surveyed a total of 7,200 individuals, while the 2010 survey will include 8,000 individuals. Adult surveys have also included 7 or 8 subsample sections, each administered to approximately 1,000 adult survey respondents. Child Survey: The Child Survey has collected data from a range of 5,728–6,032 parents or legal guardians of randomly selected children aged 17 years and younger.
Data collection is stratified to provide stable population estimates across the county’s 8 service planning areas (SPAs). For most indicators, stable data are also available for the county’s 26 health districts, which are subsets of the SPAs. Data are collected regarding respondents’ street address or cross streets, city, and ZIP code. Approximately 80% of respondents are geocoded to a Census Tract. Data can also be analyzed by geopolitical regions, including Supervisorial District, Assembly District, State Senatorial District, and Congressional District.
Minorities are not oversampled.
The adult survey includes LA County residents aged 18 and older. Data are routinely analyzed by age groups including: 18–24, 25–29, 30–39, 40–49, 50–59, 60–64, and 65 and over.
The child survey is administered to parents of children and adolescents 0–17 years old, with a majority of questions focusing on children 0–5.
Respondents include males and females, with each group comprising approximately 50% of the sample.
The 2007 LACHS response rate was 18% and 15% for the Adult and Child surveys, respectively, calculated by AAPOR RR3 formula. The response rates achieved in the 2007 LACHS were lower than those achieved in earlier LACHS cycles. (For 2002–2003, RR = 31.1% adult, 33.9% child, and for 2005, RR = 22.8% adult, 26% child.)
The overall cooperation rates for Adult and Child surveys were 40% each (AAPOR COOP3), meaning that 40% of the people successfully contacted by phone were willing to complete the interview with the surveyor.
Frequency of Data Collection: The LACHS has been conducted every 2–3 years since 1997. The 2010 LACHS follows surveys conducted in 1997, 1999, 2002, 2005, and 2007.
Source of Data: Population-based telephone survey of LA County adults (18 years and over) and the parents
or legal guardians of LA County children (17 years or younger). Previous cycles of LACHS included only landline telephones, but 10% of 2010 sample will consist of cell phone users.
Mode of Data Collection: RDD landline and RDD cell phones
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Adult Survey: Questions from the 2007 Adult LACHS pertaining to the above topics included: “Have you ever been told by a doctor or other health professional that you have:” heart disease; diabetes; high blood pressure or hypertension; high blood cholesterol. Furthermore, among those responding in the affirmative to being ever diagnosed with diabetes, follow-up questions included, “How old were you when you were told you have diabetes?” “Are you now taking insulin?” and “Are you now taking diabetes pills?” Past cycles have included a variety of other follow-up questions for adults diagnosed with cardiovascular disease.
Although questions pertaining to asthma did not appear in the 2007 LACHS, they were asked on the 2005 Adult Survey. The Adult Survey also includes many questions related to behavioral risk factors for cardiovascular disease, pulmonary disease, and diabetes, such as items assessing physical activity level, consumption of fruit and vegetables, and smoking status.
Child Survey: Child survey questionnaires assess diagnosis of diabetes and asthma among children. The child survey also includes questions related to cardiovascular disease risk later in life, such as frequency of fast food consumption, frequency of soda or sweetened drink consumption, and physical activity levels.
A copy of the Adult and Child questionnaires and a link to a complete description of the survey’s methodology—for the 1999 though 2007 Los Angeles County Health surveys—may be viewed or downloaded from the following LA County Department of Public Health (DPH) Health Assessment Unit’s website: http://publichealth.lacounty.gov/ha/hasurveyintro.htm.
Information Obtained: The LACHS collects data on physical activity levels, frequency of fast food consumption, frequency of soda or sweetened drink consumption, fruit and vegetable consumption, access to fresh fruit and vegetables, smoking status, alcohol use, perceived safety of neighborhood, ease of access to parks/playgrounds, mental health status/depression, etc. The LACHS collects data reflecting adults’ and children’s health and dental insurance status, type of health insurance, availability of regular source of care, difficulty accessing care when needed, and the potential reasons for this difficulty. Data pertaining to care received specifically for cardiovascular disease are not currently collected, but past survey cycles have assessed treatment and clinical services for adults diagnosed with hypertension, diabetes, high cholesterol, etc. Some of the demographic data collected by the LACHS are: age, race/ethnicity, income and education levels, health insurance status, employment status, disability status, place of birth, language spoken at home, and sexual orientation.
Cost of Survey/Registry: The cost of the 2010 LACHS is $1,375,959.
Who Pays for Data Collection?: Funding for the LACHS is derived from programs within LA County’s DPH and other programs within the county and/or community. More specifically, funding for the upcoming 2010 survey was received from DPH programs, including Alcohol and Drug Program Administration, Emergency Preparedness and Response Program, Tobacco Control and Prevention Program, and Environmental Health. Other funding was contributed by LA County’s Department of Mental Health and First 5 LA.
Dissemination of Data: Although data tables based on the LACHS data are available on the Health Assessment Unit website (http://publichealth.lacounty.gov/ha/hasurveyintro.htm), the actual data sets are available through a research collaboration process, which requires submission of an analysis plan and forms acknowledging adherence to HIPPA regulations and IRB review. The online query system can soon be accessed at the following website: http://publichealth.lacounty.gov/ha/hasurveyintro.htm.
Data tables and publications based on the LACHS data are available to the public at the following Health Assessment Unit website: http://publichealth.lacounty.gov/ha. In addition, customized data requests for community groups, researchers, and other health care providers and advocates are routinely performed. Researchers interested in using the LACHS data set to conduct more in-depth original analyses may obtain the full data set by submitting a research collaboration application.
Comment: Additional limitations of the LACHS data include the following: Data are based on self-report, which can introduce potential inaccuracies in the reported data. Moreover, for all previous survey years (1997–2007), only individuals with landlines were included in the sample, so data exclude those who were currently homeless
or residing in cell-phone-only households. In addition, individuals within “institutionalized” settings (e.g., prison, college dormitories, nursing homes) were not sampled, a factor that can limit the generalizability of the observed data. Finally, given the time constraints of the survey, the content of the questionnaires is necessarily limited.
Main Purpose of Study: Since 1996, MEPS, a set of large-scale surveys of families and individuals, their medical providers (doctors, hospitals, pharmacies, etc.) and employers across the United States, collect information on utilization, cost, and payment sources of healthcare services and the cost, scope, and breath of health insurance held by and available to U.S. workers. MEPS contains two components: the household component and the Insurance component. The household component gathers data from a sample of families and individuals drawn from prior year’s National Health Interview Survey (NHIS). The insurance component gathers data from a sample of private and public employers on the health insurance plans offered to employees. Only the household component contains information on chronic diseases.
Sample: 32,577 individuals of all ages were included in 2006. Data are collected at the national level.
Mode of Data Collection: In-person interviews
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Priority Conditions (Quality Supplement) and Priority Conditions Enumeration include diabetes, asthma, hypertension, coronary heart disease, angina, heart attacks, other heart disorders, strokes, emphysema; Questions include whether adult respondents have ever been diagnosed as having a particular condition and identification of condition code associated with medical events reported during the reference period; Medical conditions and diagnoses not validated.
Information Obtained: Prevalence information is collected. Health status information collected includes limitations in activities of daily living and instrumental activities of daily living, disability days. Demographic information collected includes age, race, ethnicity, income level, insurance status.
Dissemination of Data: Information is online and available to the public.
Main Purpose of Study: The Minnesota Heart Survey began in 1979 as an ongoing surveillance project of cardiovascular disease in the Minneapolis/St. Paul metropolitan area (population 2.6 million, 2000 Census). The origins of the study are found in the NHLBI Conference on the Decline of Cardiovascular Disease and the search for reasons for that decline. The project has been funded in multiple R01s principally from the NHLBI but also from NINDS. It collects mortality, morbidity and population risk factor data at regular intervals in the metropolitan area. The most recent survey was completed in 2009.
Sample: Mortality surveillance is comprehensive, including all deaths in the Minneapolis/St. Paul metropolitan area plus data from citizens dying out of state. Morbidity surveillance includes acute myocardial infarction, stroke, and congestive heart failure. Surveys are at five-year intervals. Surveillance is comprehensive of all metropolitan area hospitals with a 50–100% sample of the target metropolitan population. From 1980–1995 morbidity was collected on hospitalized patients aged 30–74. After that time, the upper age limit was removed and elderly oversampled. Risk factor surveillance is with random population samples are drawn at five-year intervals. Sample size ranges from 4,000–6,000 for each survey. From 1980–1995 the adult population aged 25–74 was surveyed. After 1995, children aged 5–18 were included. At the same time, the upper age restriction was removed. MHS is comprehensive for the Minneapolis/St. Paul metropolitan area. MHS is comprehensive for the Minneapolis/St. Paul metropolitan area. In 1986, an African American cohort aged 30–74 of 1,000 were surveyed. In 2007–2009, Hispanics were oversampled.
Response Rate: Mortality: Death certificate data are comprehensive with over 98% of deaths available. Morbidity data are also comprehensive in MHS based on access to hospital records. There are two exceptions. First, a very small hospital did not allow access for several survey years. That hospital constituted less than 2% of the admissions for the target diagnoses. Second, for one year, a larger hospital required a direct signed consent from hospitalized patients to participate in the congestive heart failure survey. This led to a reduction in participation at
that hospital. Risk Factor Survey: The home interview portion has a response rate of 70–90%. The entire survey, including the clinic visit, has a response rate of 60–70%.
Frequency of Data Collection: Mortality data collection is continuous, with the most recent data complete to 2008. For morbidity data, acute myocardial infarction was collected in 1980, 1985, 1990, 1995, and 2001–2002; stroke data were collected in 1980, 1985, 1990, 1995, and 2000; congestive heart failure data were collected in 1995 and 2000. Risk factor data in the population were collected in 1980–1982, 1985–1987, 1990–1992, 1995–1997, 2000–2002, and 2007–2009.
Sources of Data: Mortality: Minnesota Death Index, National Death Index. Morbidity: abstraction of hospitalization records. Risk factor Survey: Face-to-face interviews and clinic visit.
Mode of Data Collection: Mortality: Computerized death certificates from the state of Minnesota and the National Death Index. Morbidity: Nurse abstraction of hospital records. Risk Factor Survey: Random samples are selected based on a cluster household sampling design and a home interview is followed by clinic visit.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Mortality: Complete death certificate data are available, including information from next of kin on some sub-studies. Morbidity: Comprehensive data are collected on pre-hospital characteristics, hospital course, and discharge plans. Risk Factor Survey: Extensive data are collected on demographic characteristics, health behaviors, medication use, health knowledge and biological data, including phlebotomy, blood pressure measurement, ankle-brachial index, and other characteristics
Information Obtained: Incidence data are available at a population level. Prevalence data are available at a population level. A sub-study of patients with congestive heart failure and stroke was performed in conjunction with a large HMO in the metropolitan area. The study included multiple indicators of functional health outcomes following hospitalization.
Traditional risk factors are measured, including lipids, hemoglobin 1Ac, smoking with biochemical validation, and others. Questions about stress were asked at certain surveys. Extensive data were gathered on individual subjects from nurse abstraction of hospital records in MHS. All major demographic characteristics were collected, including occupation and health insurance status.
Cost of Survey: 1979–1984 $5.2 million; 1984–1989 $6.3 million; 1989–1994 $9.4 million; 1995–1999 $7.5 million; 1999–2005 $10.2 million; and 2006–2011 $7.1 million. Totals include direct and indirect costs.
Who Pays for Data Collection: Data collection is RO1 supported by NHLBI and NINDS.
Dissemination of Data: Data are available to investigators and data are shared with collaborators around the world. There are multiple sub-studies that have been part of MHS over the years, including hot pursuit of acute myocardial infarction, autopsy study of sudden death, factors associated with women and cardiovascular disease treatment, and many others. These are found in the multiple publications associated with this study.
Main Purpose of Study: The National Ambulatory Medical Care Survey (NAMCS) is a general purpose survey of nonfederal office-based physicians and community health centers (CHCs) in the United States. Its purpose is to provide accurate, relevant, nationally representative data annually about visits to these settings. Primary areas of interest include use of healthcare services and resources for different conditions; quality of care, including disparities among diverse populations; and monitoring diffusion of technologies, including drugs, medical procedures, and electronic health record (EHR) systems.
Sample: A substantial proportion of the NAMCS sampling frame is ineligible, for example, because physicians may no longer be in practice or may fall into ineligible categories (see below). In 2010, the expected number of survey responses is 35,000 visits to 1,200 physicians in office-based practice and 280 physician or mid-level providers from about 95 community health centers.
NAMCS uses a multistage probability design that involves probability samples of geographically based primary sampling units (PSUs), physicians within PSUs, and patient visits to physicians. A PSU consists of a county, a small group of contiguous counties, or a metropolitan statistical area (MSA) from the 50 states and the District of Columbia. Prior to sampling 112 PSUs with probability proportional to size, PSUs were stratified into 4 geographic regions (Northeast, South, Midwest, and West), which correspond to those used by the
U.S. Bureau of the Census. Within each of the regions, PSUs were further divided into areas located within or outside MSAs.
The second stage of the NAMCS sampling consists of a probability sample of practicing physicians selected from the master files maintained by the American Medical Association (AMA) and American Osteopathic Association (AOA). Anesthesiologists, radiologists, and pathologists are excluded. Within each PSU, all eligible physicians are stratified into specialty groups. The core NAMCS has 15 strata: general and family practice, osteopathy, internal medicine, pediatrics, general surgery, obstetrics and gynecology, orthopedic surgery, cardiovascular diseases, dermatology, urology, psychiatry, neurology, ophthalmology, otolaryngology, and all other specialties. In some years, external funders have paid for additional specialty groups to be sampled separately. For example, in 2010, CDC’s National Center for Chronic Disease Prevention and Health Promotion funded the inclusion of a special stratum of oncologists.
In order to improve the precision of CHC physician estimates, starting in 2006, a dual-sampling procedure has been used to select CHC physicians and other providers. First, the traditional NAMCS sample of physicians is selected using established methods and sources. Second, a sample of 104 CHCs is selected, and within each CHC, up to three physicians, physician assistants, nurse midwives, or nurse practitioners are selected for survey participation. After selection, CHC providers follow the sampling procedure used by NAMCS traditional physicians in selecting patient visits. The list of CHCs is from the Health Resources and Services Administration and the Indian Health Service. To ensure that CHC physicians are included only once, all CHC physicians selected in the traditional NAMCS sample are omitted from the survey response and subsequent weighting.
The final stage is the selection of patient visits within the practices of sampled clinicians. This involved two steps. First, the total physician sample is divided into 52 random subsamples of approximately equal size, and each subsample is randomly assigned to 1 of the 52 weeks in the survey year. Second, a systematic random sample of up to 30 visits is selected during the assigned week. The sampling rate varies for this final step from a 100 percent sample for very small practices to a 10 percent sample for very large practices as determined in a pre-survey interview.
NAMCS is nationally representative and also provides some data by region and location within or outside an MSA. In 2010, the Office of the National Coordinator for Health Information Technology (ONC) funded NCHS to collect data about EHR adoption and use from a supplemental mail survey using a state-based sample of physicians. The state-based EMR supplement has a different modality and includes less content than the core survey, but it illustrates the flexibility of adding more geographic detail.
As noted above, data are gathered about visits to physicians in office-based practice. Therefore, the data are representative of physician visits, not the general U.S. population. Moreover, information is available only for people who are treated in a physician’s office or community health center. People who do not seek medical care in one of these settings have no opportunity to have their data included in the survey. All age groups, sexes, and racial and ethnic groups are included. Minorities are not oversampled, but because PSUs are sampled with probability proportional to size, adequate representation is anticipated of populations that are concentrated in large metropolitan areas.
Response rate: In 2007, the most recent year for which final response rate information is available, the final unweighted response rate and the response rate weighted to account for the probability of selection were both 61 percent.
Frequency of Data Collection: Annual
Source of Data and Mode of Data Collection: For the core survey, data about clinicians and their practices are gathered using a paper and pencil questionnaire during an in-person interview. Data from a sample of visits are obtained by abstracting information from the medical record for the sampled visit. Initially, NAMCS was designed for clinicians to complete the forms in real time from visits as they occurred. In reality, when clinicians complete these forms, they do so by abstracting data from their own records. To improve response rates, each sampled clinician is given the option of having forms completed by a member of his or her office staff or by a trained NAMCS field representative. In 2007, more than half of the abstractions were performed by NAMCS field representatives, with clinicians completing fewer than 10 percent of the forms. Mail surveys are used for the state-based supplemental survey on EMRs.
Demographic Composition of National Ambulatory Medical Care Survey, 2007
|Unweighted Frequency||Unweighted Percent|
|Less than 15 years old||5,161||14.9|
|15–24 years old||2,852||8.2|
|25–44 years old||7,160||20.6|
|45–64 years old||10,408||30.0|
|65–74 years old||4,498||13.0|
|75 years or older||4,598||13.3|
|Race and Ethnicity|
|Native Hawaiian/Other Pacific Islander||60||0.2|
|American Indian/Alaskan Native||273||0.8|
|Missing race or ethnicity||13,407||38.6|
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: For each sampled visit, the patient record form includes a series of check-boxes to indicate whether the patient currently has asthma, cerebrovascular disease, congestive heart failure, COPD, diabetes, hyperlipidemia, hypertension, or ischemic heart disease. The 2010 patient record form is available at http://www.cdc.gov/nchs/data/ahcd/NAMCS_30A_2010.pdf.
Information Obtained: The incidence of health events that trigger a visit to physician offices or CHCs may be estimated for specific diagnoses (as determined by ICD-9-CM coding). The data collection form distinguishes visits for flare-ups of chronic conditions, such as asthma attacks, from routine visits for chronic conditions.
The visit data that NAMCS collects cannot determine the prevalence of a particular chronic condition in the population, since a patient may well visit multiple providers. The prevalence of a specific chronic condition among a clinician’s patients may be estimated. The form collects data on selected chronic conditions that the patient has, regardless of the reason or diagnosis for the sampled visit. That information could be coupled with the number of visits a patient has had to that provider during the past 12 months to estimate the number of patients with a specific condition that that provider has.
The survey does not collect information on functional health outcomes. Information is available on some risk factors when this information is recorded in the medical record. For example, body mass index is available only when the patient’s height and weight were recorded for the sampled visit. A series of check-boxes indicate whether the patient currently has asthma, chronic renal failure, diabetes, hyperlipidemia, hypertension, ischemic heart disease, or obesity. Other risk factors for which information is gathered include current tobacco use and blood pressure at the time of visit. Up to 8 medications ordered, administered, or continued may be recorded. The results of laboratory results being collected from 2010—namely total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glycohemoglobin, and fasting plasma glucose for tests performed in the 12 months prior to the sampled visit by the sampled provider—will provide additional information on risk factors.
The survey is designed to focus on clinical care. One can, for example, estimate total use of physician office care for patients with COPD or diabetes. Similarly, one can examine the content of services delivered to patients by diagnosis or blood pressure levels, including the specific medications that are being prescribed.
Patient record forms include the following:
- verbatim reason for visit and up to three visit diagnoses;
- regardless of visit diagnosis, a series of check-boxes indicate whether the patient has asthma, cerebrovascular disease, chronic renal failure, congestive heart failure, COPD, diabetes, hyperlipidemia, hypertension, ischemic heart disease, or obesity;
- major reason for visit (new problem, chronic problem, routine; chronic problem flare-up, pre-/post-surgery; preventive care);
- the names of up to 8 medications ordered, supplied, administered, or continued during the visit (which are coded using the Multum system);
- other services ordered or provided, including health education, and blood and imaging tests; and
- continuity of care, including whether the physician is the patient’s primary care provider, whether the patient is new or has been seen before, and how often the patient has been seen in the past year.
In 2010, NAMCS is gathering data on the results of laboratory tests of total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glycohemoglobin, and fasting plasma glucose, if the clinician ordered these tests in the 12 months prior to the sampled visit. The plan is to continue to collect these data in future years.
Data are collected on patient’s date of birth, sex, ethnicity, and race. The specific ethnicity categories collected are Hispanic or Latino and Not Hispanic or Latino. The specific race categories collected are as follows: White, Black/African-American, Asian, Native Hawaiian/Other Pacific Islander, American Indian/Alaskan Native, and more than one race. Data are also collected on the expected source(s) of payment for the visit. Patient ZIP code data are gathered and linked to Census socio-demographic variables for the patient’s Zip Code Tabulation Area.
Cost of Survey/Registry: $6.1 million for FY 10, excluding the central NCHS infrastructure.
Who Pays for Data Collection?: NCHS pays for collection of core data elements. Sponsors pay for developmental work, additional data elements, and supplemental modules associated with their topics of interest. For example, ONC is sponsoring the state-based EMR mail supplement.
Dissemination of Data: All public-use files are available online, free to anyone. The files have undergone disclosure review to minimize potential for disclosure of organizations and individuals included in the survey. These files are usually released within 15 months of the end of data collection period, which is about 19 months after the end of the calendar year. There is no online query system. The data are available at http://www.cdc.gov/nchs/ahcd/ahcd_questionnaires.htm.
- NAMCS also collects data about the clinician and his or her practice. For example, information is collected on ownership; practice size; patient volume; physician specialty; medical technologies on-site; and availability, use, and features of EHR systems. These data make it possible to examine associations among characteristics of the clinician, the practice organization, patient, and clinical management of the patient’s care.
- The NAMCS PSU sample has not been updated since 1988 because of resource constraints.
- High rates of missing data on race and ethnicity (above 30%) were noted in the 2006–2008 samples. These increases were concurrent with the rapid rise of the proportion of data collection forms completed by NAMCS field representatives. In 2009, efforts were made to reduce the amount of missing data through better training of field representatives and more reminders about the importance of these data to researchers. NCHS will continue to reinforce field representatives’ training about the methods and importance of these data. There is optimism that efforts under way will help address the missing problem, but there is not yet a final data set for use in assessing the full impact. To the extent that missing data remain a problem for these key variables, NCHS is also developing procedures to conduct multiple imputation for these critically important data fields.
- NAMCS is still a paper and pencil activity. Conversion to computerized data collection and more general conversion to collect data from electronic data sources must be addressed as the healthcare information technology infrastructure changes. Preliminary estimates from 2009 NAMCS data indicate that only 6% of office-based physicians, especially those in large practices, have fully functional EHR systems. Therefore, it is not yet possible to reply on electronic systems for nationally representative data. As recent legislation may change this landscape dramatically, NAMCS is planning for that eventuality.
- Other important/promising future activities include increasing sample sizes to produce state estimates and collection of longitudinal data. These activities would require additional resources. To better assess patients’ risk factors and the appropriateness of care, NAMCS could expand the data collected on clinical management and risk factors during the 12 months before the sampled visits. For patients with hypertension, hypercholesterolemia, or prior stroke, for example, the survey could collect the number of visits, medications, prescribed, changes in medications, and family history.
Main Purpose of Study: The National Diabetes Surveillance System is a comprehensive assembly of diabetes-related data from national and state-based surveys, including household surveys such as the National Health and Nutrition Examination Surveys (NHANES), telephone surveys such as the Behavioral Risk Factor Surveillance System (BRFSS), and other data sources. There are several important attributes or characteristics of the NDSS:
- It is the only diabetes surveillance system at the national level that collects, analyzes, and disseminates national, state, and county data on the growing public health burden of diabetes and its complications.
- It is dynamic, responding to changes in scientific knowledge and public health priorities.
- Unlike many surveillance systems, which typically rely on one or two data collection systems, the NDSS is a complex system that uses a variety of survey and data systems to describe and monitor the public health burden of diabetes and its complications.
- It adapts national and state health survey systems to allow the development and monitoring of key indicators, including national health objectives. Examples include the addition of the diabetes and prediabetes modules to the BRFSS. Also, in collaboration with the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), the diabetes and pre-diabetes component of NHANES was developed.
- It serves as a model for the diabetes surveillance activities of states and Division of Diabetes Translation provides diabetes surveillance technical assistance internal to CDC and to other federal agencies (e.g., CMA, VA, IHS) and the international community.
- Unlike many surveillance systems, which only describe the burden of disease in the U.S. population, the NDSS also describes the burden of the disease in a specific population—the population with diabetes. For example, it describes health care utilization for complications in the diabetic population, preventive care practices received by this population, and risk factor reduction in this special population.
- It uses a variety of mechanisms (fact sheets, publications, Internet) to disseminate diabetes surveillance data to a variety of audiences (e.g., policy makers, professional organizations, state health department). One example is the National Diabetes Fact Sheet (available at http://apps.nccd.cdc.gov/DDTSTRS/FactSheet.aspx).
- Information from the NDSS enables CDC to lead a consensus process among a dozen or more public and private agencies to derive national estimates and information on the burden of diabetes and its complications in the United States.
Because diabetes is a complex disease that affects nearly every organ system of the body, the scope of the NDSS is broad. Although the NDSS first began assessing diabetes prevalence, mortality, and long-term complications, it progresses to assessing preventive care practices, risk factors, and risk behaviors and is beginning to move more upstream to the surveillance of at-risk populations.
Sample: 3,141 counties or county-equivalence, 50 U.S. states, and the District of Columbia. Historically, the NDSS contained only national- and state-level estimates, but recently county-level estimates of diabetes and obesity prevalence derived from Bayesian multilevel modeling techniques have been included (http://apps.nccd.cdc.gov/DDT_STRS2/NationalDiabetesPrevalenceEstimates.aspx). Generally, disaggregation of trends by demographic characteristics is more likely to be feasible at the national level than at the state level because of sample size issues, and the NDSS contains no data for geographic units smaller than counties. Minnesota, Montana, New Mexico, North Carolina, and Oklahoma have conducted surveys with an oversample of American Indians. The survey includes all ages.
Frequency of Collection and Sources of Data: Annual. County-level diagnosed diabetes prevalence derived from the Behavioral Risk Factor Surveillance System (BRFSS) and the U.S. Census Bureau’s Population Estimates Program.
Mode of Data Collection: Vital statistics, ongoing national, state, and local data collection systems, billing data collected by Medicare, Medicaid, managed care organizations, survey data from NCHS (e.g., BRFSS), hospital inpatient data, the U.S. Renal Data System (USRDS); preventive care practices data include physical exam results from hemoglobin A1c testing, lipid tests, foot and eye exams, self-monitoring of blood sugar, tests of kidney function (Desai, 2003).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Dilated eye exam, daily self-monitoring of blood glucose, foot exam, doctor visit for diabetes, daily self-exam of feet, A1c tests, ever attended diabetes self-management class
Information Obtained: Incidence and prevalence data are collected. Functional health outcome information includes poor physical health, poor mental health, inability to do usual activities, limitations in mobility. Risk factor information is collected about physical inactivity, overweight, obesity, current smoking, hypertension, high blood cholesterol. Clinical care information is collected about mortality; diabetes-related use of healthcare services, disability, and preventive care practices, and mortality and health service use related to complications of diabetes. Where feasible, trends are examined by demographic characteristics (age, race/ethnicity, education, sex, and geographic level).
Who Pays for Data Collection?: BRFSS, state health departments
Dissemination of Data: Data are available online.
Main Purpose of Study: The National Health and Nutrition Examination Survey is the largest and longest-running national source of objectively measured health and nutrition data on children and adults across the United States.
Sample: NHANES surveys about 5,000 people from counties across the United States annually. The survey includes civilian noninstitutionalized men and women of all ages in the United States. NHANES oversamples those who are 60 and older, African Americans, and Hispanics
Frequency of Collection and Sources of Data: NHANES is conducted annually and includes a series of questionnaires that are used in both the home and the mobile center, a physical examination, and a laboratory component.
Mode of Data Collection: There are six instruments used for data collection. These are:
- Screener module administered on the doorstep;
- Family questionnaire that collects household and family level information, including demographics and occupation;
- Sample Person questionnaire that includes questionnaire hand cards and target topics that include blood pressure, cardiovascular disease, demographics, diabetes, and respiratory health and disease;
- Audio computer assisted personal interview (ACASI) and computer assisted personal interview (CAPI) questionnaires that are administered in the mobile examination center (MEC);
- Special follow-up questionnaires that include the flexible consumer behavior survey and Hepatitis C follow-up; and
- Examination and laboratory components that include the dietary recall, body measurements, and the environmental health profile.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The NHANES questionnaire collects information on chronic conditions, including blood pressure and cholesterol, cardiovascular disease, diabetes, and respiratory health and disease.
Information Obtained: Prevalence, risk factors, and functional health outcomes
Main Purpose of Study: National Health Interview Study data are used widely throughout the Department of Health and Human Services (HHS) to monitor trends in illness and disability and to track progress toward achieving national health objectives. The data are also used by the public health research community for epidemiologic and policy analysis of such timely issues as characterizing those with various health problems, determining barriers to accessing and using appropriate health care, and evaluating federal health programs.
Sample: The theoretical targeted sample size is 35,000 households and 87,500 persons; however, in recent years, budget limitations have restricted the sample size. In 2008, for example, about 29,000 households and 74,000 persons were in the NHIS sample, including about 9,000 sample children and 22,000 sample adults. All states and the District of Columbia are included in the sample; geographic stratification is a factor in the sampling design. Although all states are included in the NHIS, the sampling design is primarily aimed at making national and regional estimates. Larger states (currently approximately 20) have sufficient coverage in their samples so that reliable estimates can be made. Black, Hispanic, and Asian populations are oversampled for the family questionnaire. Persons over 65 years of age and in one of these three minority groups also have a greater probability of being selected as the sample adult. Persons of all ages are included in the NHIS sample. Persons of both sexes are included in the NHIS sample. The response rate varies by section of the questionnaire; the response rate for the first section (the family questionnaire) was 84.5% in 2008, the sample child final response rate was 72.3%, and the sample adult final response rate was 62.6%.
Frequency of Collection and Sources of Data: Data are collected annually and continuously, with a different cross-sectional sample each year. NHIS is conducted primarily by personal interviews in households with interviewers from the U.S. Census Bureau. Following the family questionnaire, which covers everyone in the family, one sample adult and (if any are in the family) one sample child are randomly selected. The sample adult answers for him/herself and a “knowledgeable” adult answers for the sample child.
Mode of Data Collection: NHIS is an in-person survey. After the initial in-person interview, completion of sections such as the sample adult questionnaire can sometimes be done by telephone.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: A standard set of questions are asked each year of the sample adult and respondent for the sample child to obtain data on these conditions; in some years, additional questions may be asked about these conditions in the supplemental questionnaires, which vary by year. Additionally, persons who experience limitation in a wide variety of activities are asked which conditions cause the limitation; some of the categories are related to these diseases.
Information Obtained: Incidence estimates can be roughly calculated for adult diabetes; not for other mentioned conditions. For conditions causing limitation, incidence can also be roughly calculated. Prevalence information can be calculated for the conditions listed under specific questions that include hypertension/high blood pressure, coronary heart disease, angina or angina pectoris, heart attack, stroke, asthma, chronic bronchitis, diabetes, emphysema.
Limitations caused by most of these conditions is available; however, if a person has multiple limitations caused by multiple conditions, specific limitations cannot be attributed to specific conditions. Some related stressor data are collected in the NHIS including height/weight, body mass index, smoking and drinking behaviors, serious psychological distress, poverty, and unemployment. Depending on what is considered a “stressor,” there may be other data available.
The care data collected in the NHIS are relatively general and, in almost all cases, not tied to specific conditions. There are standard questions about the quantity of various kind of care obtained from various sources. Not infrequently, NHIS supplemental questionnaires collect more detailed data on care that is more likely to pertain to specific conditions. A wide variety of demographic characteristics are collected in the course of the NHIS, including all of those mentioned above. The survey also collects information on personal earnings (as opposed to family income), country of birth, and citizenship status for all family members, employment status for all family members aged 18 or older, and industry and occupation data for a randomly selected adult in the family (e.g., the sample adult). In addition, information on health insurance coverage and race/ethnicity for each family member is quite detailed (although not all of this information is available on the public-use data files).
Cost of Survey: The NHIS costs about $30 million per year. NCHS covers the cost of the core survey; supplement sponsors cover the cost of the supplements. Sponsoring agencies are almost always other federal government agencies and, most commonly, NIH.
Dissemination of Data: Public-use files are released annually (and, fairly recently, back through 1963). Except for some variables that might pose a disclosure risk, all variables are on these files, with some collapsing of categories. All public-use files are available to anyone with Internet access. Additional access to variables with some disclosure risk or variables that can be linked to other sources of data (e.g., geocoding) can be obtained through a Research Data Center with approval of a proposal; there is a charge for this service.
Main Purpose of Study: The National Hospital Ambulatory Medical Care Survey is a general purpose survey of nonfederal noninstitutional hospitals with outpatient departments (NHAMCS-OPD) and emergency departments (NHAMCS-ED) in the 50 states and the District of Columbia. In addition, NHAMCS includes freestanding and hospital-based ambulatory surgery centers (NHAMCS-ASC). NHAMCS’s purpose is to provide accurate, relevant, nationally representative data annually about visits to these settings. Primary areas of interest include use of healthcare services and resources for different conditions; quality of care, including disparities among diverse populations; monitoring diffusion of technologies, including drugs and medical and surgical procedures and EHRs; and ED crowding.
Sample: In 2010, expected numbers of survey responses are 100,000 patient visits to 360 hospitals with OPDs, EDs, or ASCs, and 200 freestanding ASCs. Sample design and geographic stratification: NHAMCS uses a multistage probability design that involves probability samples of geographically based primary sampling units (PSUs); samples of hospitals within PSUS; samples of OPDs, EDs, and ASCs within hospitals; and samples of visits within each setting. A PSU consists of a county, a small group of contiguous counties, or a metropolitan statistical area (MSA) from the 50 states and the District of Columbia. Prior to sampling 112 PSUs with probability proportional to size, PSUs were stratified into 4 geographic regions (Northeast, South, Midwest, and West) that correspond to those used by the U.S. Bureau of the Census. Within each of the regions, PSUs were further divided into areas located within or outside MSAs.
Within PSUs, hospitals were stratified by hospital class, type of ownership (not-for-profit, nonfederal government, and for-profit) and hospital size. The original sample for the NHAMCS frame was compiled as follows. Hospitals with an average length of stay for all patients of fewer than 30 days (short-stay) or hospitals whose specialty was general (medical or surgical) or children’s general were eligible for NHAMCS. Excluded were federal hospitals, hospital units of institutions, and hospitals with fewer than six beds staffed for patient use.
The original NHAMCS sample was drawn from the universe of 6,249 hospitals in the 1991 SMG Hospital Database that met the survey’s eligibility criteria. Of the eligible hospitals, 5,582 (89%) had emergency departments (EDs) and 5,654 (90%) had outpatient departments (OPDs). Hospitals were defined to have an ED if the hospital file indicated the presence of such a unit or if the file indicated a non-zero number of visits to such a unit. A similar rule was used to define the presence of an OPD. Hospitals were classified into four groups: those with only an ED, those with an ED and an OPD, those with only an OPD, and those with neither an ED nor an OPD. Hospitals in the last class were considered as a separate stratum, and a small sample (50 hospitals) was selected from this stratum to allow for estimation to the total universe of eligible hospitals and the opening and closing of EDs and OPDs in the sampled hospitals.
A fixed panel of 600 hospitals was selected for the NHAMCS sample; 550 hospitals had an ED and/or an OPD and 50 hospitals had neither an ED nor an OPD. To recognize the possibility of seasonality in healthcare delivery, the sample of 600 hospitals was randomly divided into 16 subsets of approximately equal size. Each subset was assigned to 1 of the 16 4-week reporting periods, which rotate across survey years. Therefore, the entire sample does not participate in a given year, and each hospital is inducted approximately once every 15 months. Over time, hospitals close and new ones open. Therefore, the NHAMCS sample is refreshed every 3 years. The sample was refreshed most recently in 2009 using the SDI Market Profiling Database to construct the sampling universe.
Within each hospital, either all outpatient clinics and emergency service areas (subsets of EDs) or a sample
of such units is selected. Outpatient clinics are in scope if ambulatory medical care is provided under the supervision of a physician and under the auspices of the hospital. Clinics are required to be “organized” in the sense that services are offered at established locations and schedules. Clinics where only ancillary services were provided or other settings in which physician services were not typically provided are out of scope. An ED is in scope if it is staffed 24 hours a day.
For freestanding ASCs, a list sample was constructed by combining information from a commercial Freestanding Outpatient Surgery Center Database and Medicare-certified facilities included in the CMS Provider-of-Services (POS) file. Facilities specializing in dentistry, podiatry, abortion, family planning, or birthing were excluded. Prior to sampling, ASCs were stratified by facility specialty and geographic region.
Within each sampled setting, patient visits are systematically selected over a randomly assigned 4-week reporting period. A visit is defined as a direct personal exchange between a patient and a physician or a staff member acting under a physician’s direction for the purpose of seeking care and rendering health services. Visits solely for administrative purposes, such as payment of a bill, and visits in which no medical care was provided, such as visits to deliver a specimen, are out of scope.
Geographic levels included in sample (e.g., local, state, national): NHAMCS is nationally representative. The sample is stratified into 4 geographic regions (Northeast, South, Midwest, and West) that correspond to those used by the U.S. Bureau of the Census. Within each of the regions, PSUs are further stratified by whether or not they are located within or outside MSAs. With additional resources, NHAMCS has the capacity for state-based sampling and estimation.
Demographic composition of the sample: As noted above, data are gathered about visits to hospital ambulatory units, including OPDs, EDs, and ASCs, as well as freestanding ASCs. The data are representative of healthcare visits to these settings and do not represent the general U.S. population. Moreover, information is available only for people who are treated in these settings; people who do not seek medical care in one of these settings have no opportunity to be included in the survey. All age groups and both males and females are included in the sample. Minorities are not oversampled, but because PSUs are sampled with probability proportional to size, adequate representation is anticipated of populations that are concentrated in large metropolitan areas.
Response rates: In 2007, the most recent year for which final response rate information is available, the unweighted response rates were 84% for NHAMCS-ED and 72% for NHAMCS-OPD. Response rates weighted to account for the probability of selection were 86% and 61% for NHAMCS-ED and NHAMCS-OPD, respectively. Response rates for NHAMCS-ASC are not yet available because hospital-based ASCs were added in 2009 and freestanding ASCs were added in 2010.
Frequency of Data Collection: Annual
Source of Data and Mode of Data Collection: Data about ambulatory care facilities are gathered using a paper and pencil questionnaire during an in-person interview. Data from a sample of visits are obtained by abstracting information from the medical record for the sampled visit.
NHAMCS was designed with the intention that hospital staff within the sampled department would complete the patient record form. To improve response rates, each sampled unit is also given the option of having forms completed by a trained NHAMCS field representative. In 2007, more than half of the abstractions were performed by hospital staff. Field representatives completed about one-third of the abstractions. The remainder was completed by a combination of hospital and survey staff (7%), or the person who completed the form was not noted (2%).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Content varies by setting.
- NHAMCS-OPD. For each sampled visit, the patient record form includes a series of check-boxes to indicate whether the patient currently has asthma, cerebrovascular disease, congestive heart failure, COPD, diabetes, hyperlipidemia, hypertension, or ischemic heart disease.
- NHAMCS-ED. For each sampled visit, the patient record form includes a series of check-boxes to indicate whether the patient has cerebrovascular disease or history of stroke, congestive heart failure, or diabetes. Vital signs at the time of triage are recorded. These include blood pressure, heart rate, respiratory rate, pulse oximetry, and whether the patient was on oxygen.
- NHAMCS-ASC. No specific questions relate to these conditions, though the form collects diagnoses and procedures.
Demographic Composition of National Hospital Ambulatory Medical Care Survey, 2007*
|Unweighted Frequency||Unweighted Percent|
|Less than 15 years old||12,500||17.9|
|15–24 years old||10,175||14.5|
|25–44 years old||19,506||27.9|
|45–64 years old||17,506||25.0|
|65–74 years old||4,791||6.8|
|75 years or older||5,411||7.7|
|Race and Ethnicity|
|Native Hawaiiann/Other Pacific Islander||290||0.4|
|American Indian/Alaskan Native||239||0.3|
*The 2007 sample contained ED and OPD data only.
Information Obtained: The survey does not collect information on functional health outcomes. The incidence of events that trigger a visit to an OPD, ED, or ASC may be estimated for specific diagnoses (as determined by ICD-9-CM coding). The ED data may be one of the best nationally representative sources for the incidence of acute events or conditions, such as AMI, for which the patient is expected to come to the hospital. Admissions to the hospital, by principal discharge diagnosis, are also recorded. ED data would, of course, exclude people who died and were not brought to a hospital. In concert with NHLBI and CDC’s National Center for Chronic Disease Prevention and Health Promotion, methods are being developed, for example, using troponin results that are intended to more accurately identify AMIs than has been possible with ICD-9-CM codes alone. Though developed for inpatient discharges, the methods may be applicable to improve estimates of incidence using ED data as well.
The data collection form distinguishes visits for flare-ups of chronic conditions such as asthma attacks from routine visits for chronic conditions. The visit data that NHAMCS collects cannot determine the prevalence of a particular chronic condition in the population, however, since a patient may well visit multiple providers. For OPDs, the prevalence of a specific chronic condition among a clinic’s patients may be estimated. The form collects data on selected chronic conditions that the patient has, regardless of the reason or diagnosis for the sampled visit. That information could be coupled with the number of visits a patient has had to that provider during the past 12 months to estimate the number of patients with a specific condition that that provider has.
Demographic data: Data are collected on the patient’s date of birth, sex, ethnicity, and race. Data are also collected on the expected source(s) of payment for the visit. Patient ZIP code data are gathered and linked to Census socio-demographic variables for the patient’s ZIP Code Tabulation Area.
Risk factors/stressors: For NHAMCS-OPD, information is available on some risk factors when this information is recorded in the medical record. For example, body mass index is available when the patient’s height and weight were recorded for the sampled visit. A series of check-boxes are available to indicate whether the patient currently has certain conditions, including asthma, cerebrovascular disease, COPD, diabetes, hyperlipidemia,
hypertension, ischemic heart disease, or obesity. Other risk factors for which information is gathered include current tobacco use and blood pressure at the time of visit. Up to eight medications ordered, administered, or continued may be recorded. From 2011, the plan is to gather laboratory results for total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glycohemoglobin, and fasting plasma glucose for tests performed in the 12 months prior to the sampled visit by the sampled provider.
For NHAMCS-ED, for each sampled ED visit, a series of check-boxes are available to indicate whether the patient has cerebrovascular disease or a history of stroke, congestive heart failure, diabetes, or a condition requiring dialysis; pulse oximetry; and whether the patient is on oxygen. Up to eight medications administered or prescribed may be recorded. Vital signs include initial blood pressure.
For NHAMCS-ASC, other than diagnoses and procedures, the abstraction does not collect data on risk factors.
Clinical care: For NHAMCS-OPD, information is obtained on clinical care information including verbatim reason for visit; up to three visit diagnoses (ICD-9-CM coded); regardless of visit diagnosis, a series of check-boxes indicate whether the patient has asthma, cerebrovascular disease, chronic renal failure, congestive heart failure, COPD, diabetes, hyperlipidemia, hypertension, ischemic heart disease, or obesity major; reason for visit (new problem, chronic problem, routine; chronic problem flare-up, pre-/post-surgery; preventive care); the names of up to eight medications ordered, supplied, administered, or continued during the visit; other services ordered or provided, including health education and blood and imaging tests; and information about continuity of care, including whether the clinician is the patient’s primary care provider, whether the patient is new or has been seen before, and how often the patient has been seen in the past year. One can, for example, examine all primary care visits for patients with high blood pressure and analyze the medications and other services prescribed for these patients by characteristics of the patient and the hospital.
For NHAMCS-ED, data elements collected include the patient’s initial vital signs at triage assessment (temperature, blood pressure, heart rate, respiratory rate, pulse oximetry, patient on oxygen, coma scale, pain scale); verbatim reason for visit; up to three visit diagnoses (ICD-9-CM coded); a series of check-boxes indicate whether the patient has cerebrovascular disease or history of stroke, congestive heart failure, diabetes, or conditions requiring dialysis; the names of up to 8 medications ordered or administered; other services and procedures performed (including BUN/creatinine, cardiac enzymes, glucose, cardiac monitor, ECG/EKG); information on previous care in the same ED in the past 72 hours and 12 months, and discharge from a hospital within past 7 days; information about patient flow, including length of time to initial treatment by a physician and total visit length; and visit disposition, including admission to the hospital, the type of unit to which admission occurred, and final discharge diagnosis. Visit times may be especially interesting, as there is some suggestion that hospital inpatients admitted following prolonged stays in the ED may have worse outcomes.
For NHAMCS-ASC, data elements collected include up to 5 visit diagnoses (ICD-9-CM coded); up to 7 surgical procedures; time in the operating room before, during, and after surgery; up to 12 prescriptions and over-the-counter medications and anesthetics administered during the visit or provided or prescribed at discharge; and details about anesthesia administration, including the type(s) of anesthesia and the training and specialty of the provider. Information about surgical outcomes, including symptoms present during and after the procedure, discharge to the ED or hospital, and medical care sought at 24 hours is also gathered.
Cost of Survey/Registry: $5.9 million, not including the central NCHS infrastructure
Who Pays for Data Collection?: NCHS pays for collection of core data elements. Sponsors pay for developmental work, additional questions, and supplemental modules associated with their topics of interest. For example, in 2010 NCI and CDC are collaborating to develop a colonoscopy supplement to the NHAMCS-ASC module.
Dissemination of Data: Public-use files are available online, free to anyone. The files undergo disclosure review to minimize the potential for disclosure of hospitals and individuals included in the survey. These files are usually released within 15 months of the end of data collection period, which is about 19 months after the end of the calendar year. A prototype for an online query system to include NHAMCS-ED data is currently under development. It is hoped that other data sets will be included in the future. The data are available at http://www.cdc.gov/nchs/ahcd/ahcd_questionnaires.htm.
Comments: The biggest challenges are lack of timeliness, lack of longitudinal data to provide better information on transitions and outcomes, and inadequate sample sizes to produce state estimates. All of these barriers
could be overcome with increased resources. To better assess patients’ risk factors and the appropriateness of care, like the potential described above for NAMCS, NHAMCS-OPD could expand the data collected on clinical management and risk factors during the 12 months before the sampled visits. For patients with hypertension, hypercholesterolemia, or prior stroke, for example, the survey could collect the number of visits, medications, prescribed, changes in medications, and family history.
NHAMCS is still a paper and pencil activity. Conversion to computerized data collection and more general conversion to collect data from electronic data sources must be addressed as the healthcare information technology infrastructure changes. Although there are still too few hospitals with electronic systems to gather the representative data through health information exchange, recent legislation is expected to change this landscape dramatically. Options are being explored for gathering data electronically through a single integrated National Hospital Care Survey to cover OPD, ED, ASC, and inpatient care. In addition to ultimately providing data in a more timely fashion, the integrated survey would permit patient linkage for care across these hospital settings in the future permit linkage with the National Death Index and CMS databases.
Main Purpose of Study: The National Hospital Discharge Survey is a continuing general purpose survey of inpatient hospital care in the United States, ongoing since 1965. Its purpose is to provide accurate, relevant, nationally representative data annually about hospital discharges. Primary areas of interest include monitoring trends in the use of hospital care for different conditions, describing demographic characteristics of patients receiving hospital care for different conditions, and monitoring diffusion of new medical and surgical procedures to treat specific conditions.
Sample: About 150,000 inpatient discharges from 239 hospitals annually for 2008 through 2010. In 2007 and prior years, sample sizes were approximately twice as large. Sample design and geographic stratification: The scope of NHDS encompasses discharges from noninstitutional hospitals, excluding federal hospitals, located in the 50 states and the District of Columbia. Only hospitals having six or more beds staffed for inpatient use are included in the survey. General (medical and surgical) and children’s general hospitals are included, regardless of length of stay, but other hospitals are included only if they have an average length of stay fewer than 30 days. The 1988 sample was selected from a frame of short-stay hospitals listed in the 1987 SMG Hospital Market Data Base. The NHDS redesign in 1988 implemented a modified 3-stage design that employed stratification by creating both primary sampling unit (PSU) and hospital strata defined within the four major Census regions and based on characteristics such as metropolitan statistical areas (MSAs)/non-MSA status and data collection type (manual or automated). Units selected at the first stage of sampling consisted of either hospitals or geographic areas (i.e., PSUs), such as counties, groups of counties, or MSAs. Within the sampled PSUs, additional hospitals were selected with probabilities proportional to their annual number of discharges. Finally, at the last stage, discharges within hospitals were selected using systematic random sampling. Third-stage sampling rates were determined by the hospital’s sampling stratum and the data collection method used. The target sample for manual system hospitals was 250 discharges annually, whereas discharges from some automated hospitals were oversampled, depending on whether NCHS received a sample or a census file of their discharges. The hospital sample has been updated approximately every 3 years to allow for hospitals that opened or changed their eligibility status since the previous sample update. Hospitals that were no longer eligible for NHDS have been deleted.
Geographic levels included in sample: NHDS is designed to make national, as well as regional estimates, where regions are defined to be the four major Census regions: Northeast, Midwest, South and West. With additional resources, NCHS has the capacity for state-based sampling and estimation of information about inpatient hospital care.
Demographic composition of the sample: As noted above, data are representative of inpatient hospital discharges, where “inpatient” is defined by the hospital. Because the survey samples discharges, not patients, a small number of sampled records may be readmissions; however, the survey does not have a mechanism to track these cases. Patients of all ages, sexes, races, and ethnicities are included in the sample.
Response rate: In 2007, the most recent year for which final response rate information is available, the unweighted response rate was 88%. The response rate weighted to account for the probability of selection was 82%.
Frequency of Data Collection: Annual. Data are gathered throughout the year. 2010 is the last year that NCHS plans to conduct the current NHDS (see Comments below).
Source of Data and Mode of Data Collection: Two data collection procedures are used for the survey. The first is a manual system of sample selection and data abstraction, which was used for approximately 55% of the responding hospitals in 2007. The second is an automated method, which was used for approximately 45% of the responding hospitals in 2007. The automated method involves the purchase of computerized data files from abstracting services or state data systems, or transmissions from the hospitals themselves.
In the manual system, the sample selection and the transcription of information from the hospital records to abstraction forms are performed at the hospitals. Of the hospitals using this system in 2007, about 23% had the work performed by their own medical records staff. In the remaining hospitals using the manual system, personnel of the U.S. Bureau of the Census did the work on behalf of NCHS. The completed forms, along with sample selection control sheets, were forwarded to a contractor for coding and data entry, and then to NCHS for editing and weighting.
For the automated system, NCHS purchased files containing machine-readable medical record data from which records were systematically sampled by NCHS. All the information comes from Uniform Billing (UB)-04 forms, the format in which hospitals submit claims for payment.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: See medical record abstraction form at http://www.cdc.gov/nchs/data/hdasd/nhds_form_updated.pdf.
Information Obtained: Frequencies and rates of inpatient hospitalizations for specific diagnoses (determined by ICD-9-CM coding) are useful in describing disease burden. To the extent that acute events, such as acute myocardial infarction, result in hospital admissions, and are properly diagnosed, national hospital discharge data may represent the best opportunity for monitoring incidence nationally. Improvements to these data are possible, and as noted below under Comments, NCHS has been working with NHLBI and CDC to better realize that potential.
The survey does not collect information on the prevalence of conditions or functional health outcomes.
Demographic data: Data are collected on the patient’s date of birth, sex, ethnicity, and race. Data on ethnicity are missing at such a high rate that the missing rate, and the data themselves, are not reported and are not used routinely. For example, for 2007, 49% of the discharge records in NHDS data set had ethnicity missing. Data are also collected on the expected source(s) of payment for the visit.
Risk factors/stressors: Limited to those coded as diagnoses or procedures. For example, some discharge diagnoses, such as diabetes, may be risk factors for others, such as heart disease or stroke.
Clinical care: Information is obtained on up to 7 discharge diagnoses and 4 procedures performed. Data are also collected on admission and discharge dates and discharge status. Beginning in the 2001 survey year, two additional items were included in the medical abstract form: type of admission and source of admission. In late 2007, an admitting diagnosis and a present on admission check-box for all of the diagnosis variables were added to data collection because hospitals were required to collect this information for billing effective at that time.
Cost of Survey/Registry: $2.5 million for FY 10, excluding the central NCHS infrastructure
Who Pays for Data Collection?: NCHS pays for data collection.
Dissemination of Data: Public-use files are available online, free to anyone. The files undergo disclosure review to minimize the potential for disclosure of hospitals and individuals. Currently, development is a prototype for an online query system that will include NHAMCS-ED data, and NHDS hopes to include other data sets in the future. The data are available at http://www.cdc.gov/nchs/nhds/nhds_questionnaires.htm.
Comments: Historically, to minimize the burden on hospitals and to ensure complete and representative data, NHDS has been tied to the UB-04. Unfortunately, this has limited the flexibility of the survey to gather more detailed information. Over the past 5 years, NHDS has tested methods to selectively sample discharged patients with specific characteristics (especially diagnoses) and obtain additional information from the medical records of these discharged patients. Part of this process has involved testing an acute coronary syndrome (ACS) module in collaboration with NHLBI and CDC’s National Center for Chronic Disease Prevention and Health Promotion. The purpose of this module is to gather laboratory data to improve the validity of AMI discharges identified solely on the basis of discharge diagnosis. The ACS module tested the collection of troponin data from discharges with a diagnosis of acute myocardial infarction (AMI) or other acute and subacute ischemic heart disease (ICD-9-CM:
Demographic Composition of National Hospital Discharge Survey, 2007
|Unweighted Frequency||Unweighted Percent|
|Less than 15 years||62,392||17.1|
|15–44 years old||100,047||27.4|
|45–64 years old||83,881||22.9|
|65 years or older||119,328||32.6|
|Native Hawaiian/Other Pacific Islander||198||0.1|
410-411). These tests illustrated that it was possible both to systematically oversample specific discharge diagnoses and to obtain information on troponins from patients with a discharge diagnosis of AMI or other acute and subacute ischemic heart disease. Further developmental work is needed to evaluate the usefulness of this module in monitoring the incidence of AMI, from hospital discharges with and without a discharge diagnosis of AMI. Important next steps include (1) testing the ACS module in a larger sample of discharges from a more varied set of hospitals and (2) testing the feasibility of gathering similar data from patients with other discharge diagnoses, such as old myocardial infarction; angina pectoris; other forms of chronic ischemic heart disease; hypertensive heart disease; cardiac dysrhythmia; heart failure; and acute edema of the lung, unspecified.
A new hospital survey with the flexibility to include such modules as well as linkages with the National Death Index and CMS databases is planned for the future. That integrated National Hospital Care Survey would permit linking patients’ care across OPD, ED, ASC, and inpatient hospital settings. The integrated survey would rely on electronic transmission of UB-04 data and clinical data from EHR systems, as the capabilities in the new national sample of hospitals permit.
Main Purpose of Study: The Community Health Survey (CHS) is conducted in order to inform public health policy and practice in New York City, making data use and dissemination cornerstones of the project. The CHS data have three main roles in public health work: instrumental, that is, influencing health program decisions; conceptual, that is, increasing the understanding of the relationship between health behavior and health status; and ersuasive, that is, supporting health policies
Sample: Approximately 9,500 interviews each year since 2002. It includes men and women aged 18 and older. The CHS uses a stratified random sample to produce neighborhood and citywide estimates. Strata are defined using the United Hospital Fund’s (UHF’s) neighborhood designation, modified slightly for the addition of ZIP codes added since the initial UHF definitions. There are 42 UHF neighborhoods in NYC, each defined by one or more adjoining ZIP codes. Response rates vary by year. For the most recent survey for which data are publicly available (2008 CHS), AAPOR Response Rate #3=33.3%, Cooperation Rate #3=80.7%
Frequency of Collection and Sources of Data: Annually since 2002; Population interviews
Mode of Data Collection: Telephone interviews using RDD landline sample (2002–2008) and RDD landline and cellular telephone sample (from 2009 on)
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The CHS contains multiple measures related to these conditions. Multiple health surveillance measures, varies by year. Questionnaires for each year can be downloaded at http://www.nyc.gov/html/doh/html/survey/chsdata.shtml.
Information Obtained: Multiple health surveillance measures. Not all measures are asked each year. Most are designed to provide prevalence estimates. Risk factor information is limited self-report (e.g., asked if taking medications for hypertension). Multiple demographic variables are collected, including age, ethnicity, race, ZIP code location, household income, education level, employment status, health insurance status, and others.
Who Pays for Data Collection: The cost of running the survey is approximately $1 million per year. It is funded with a city tax levy. External (grant and private) funds have supported some modules of the CHS in the past.
Dissemination of Data: SAS-formatted data sets 2006–2008 are available at http://www.nyc.gov/html/doh/html/survey/chsdata.shtml. Data sets for 2002–2005 are available directly from the DOHMH. Along with data from the 2009 CHS, these will be posted to the public website in the future. Visit EpiQuery at http://www.nyc.gov/html/doh/html/epi/epiquery.shtml. Data sets are publicly available.
Main Purpose of Study: The primary objective of the CHS Heart Follow-up Study is to estimate a baseline population sodium intake for NYC; Secondary objectives are to: estimate mean sodium intake by race/ethnicity; estimate the prevalence of awareness, treatment, and control of hypertension; assess the relationship between blood pressure (both self-reported and physical measurement) and sodium intake; and assess the relationship between potassium intake and blood pressure.
Sample: Approximately 1,800 interviews from the CHS 2010 will be completed, resulting in about 1,500 usable urine samples. The surveys will be conducted citywide. Men and women aged 18 and older are included.
Frequency of Collection and Sources of Data: April through August 2010. A follow-up study is planned for 2014, if funding is available. Sources of data are population interviews with some clinical measures.
Mode of Data Collection: Telephone interviews using RDD landline and cellular telephone sample, plus 24-hour urine collection and in-person clinician measurements of height, weight, waist circumference, and resting blood pressure
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: In addition to the CHS 2010 questions, participants in the Heart Follow-up Study complete a brief questionnaire including questions on nutrition and personal and family history of CVD.
Information Obtained: Measures will include items from the 2010 CHS as well as a separate series of questions related to dietary sodium consumption. Multiple variables, including age, ethnicity, race, ZIP code location, household income, education level, employment status, health insurance status, and other will be collected.
Who Pays for Data Collection?: The OPTS cost of the 2010 CHS Heart Follow-up Survey is approximately $1.2 million. The Robert Wood Johnson Foundation provided about $566K, Kellogg Foundation provided $98K, New York State Health Foundation provided $181K, and the Centers for Disease Control and Prevention provided $88K in support of the study. NYC DOHMH has funded about $22K through city tax levy funds.
Dissemination of Data: When the study has been completed, data will be accessible via EpiQuery and used for a range of agency and city activities. The data set will be publicly available.
Main Purpose of Study: Specific objectives of NYC HANES are to: estimate the number and percent of persons in the NYC population with selected diseases and risk factors; estimate citywide awareness, treatment, and control of selected diseases; estimate prevalence, awareness, treatment, and control of selected diseases among a limited set of demographic subgroups identified by race/ethnicity, gender, and broad age bands; monitor prevalence and
magnitude of environmental exposures in NYC; analyze risk factors for selected diseases in NYC; and establish a population-based serologic repository that can be used to explore emerging public health issues in NYC.
Sample: Noninstitutionalized adult New York City residents aged 20 years or older. 3-stage cluster sampling design (geographic segment, household, individual). Of the 4,026 households randomly selected for the survey, eligibility screening questionnaires were completed for 3,388 households and 3,047 study participants were identified. Of those identified, 1,999 individuals participated in the survey, yielding a response rate of 55%. Post stratification weighting based on age group, sex, race/ethnicity, borough was applied to minimize the impact of component and item nonresponse.
Frequency of Collection and Sources of Data: The survey was conducted once in 2004. It will repeat as serial cross-sectional at regular intervals, if funding is available. Sources of data are population interviews with clinical measures.
Mode of Data Collection: face to face computer-assisted interviews; audio computer-assisted self-interview (drug use, sexual behavior, incarceration, violence); and physical exam with blood draw, urine collection, and anthropometry (10% conducted in home; 90% in clinic).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Yes. The questionnaires can be downloaded at http://www.nyc.gov/html/doh/html/hanes/datasets.shtml.
Information Obtained: Prevalence and incidence of selected diseases and risk factors for selected diseases; awareness, treatment, and control of selected diseases; and the prevalence and magnitude of environmental exposures. Demographic information collected includes multiple variables, including age, ethnicity, race, household income, education level, employment status, health insurance status, and others.
Cost of Survey: For HANES, personnel is a costly due to fieldwork. In 2003–2005, total cost was $3,397,273. Lowball estimate for a second NYC HANES with same design is $5 million. This includes a contract for survey operations, but no contract for technical support for sampling, training, or IT.
Who Pays for Data Collection?: 2004 NYC HANES funded by NYC DOHMH almost exclusively through city tax levy funds. Most field staff were reassigned from their regular jobs at DOHMH, and school nurses were paid overtime. No funds have been identified for a second NYC HANES. DOHMH contracted with NCHS to provide IT and other technical support. NYC DOHMH is seeking external funding in collaboration with academic partners.
Dissemination of Data: Variable lists, documentation, codebooks, and data sets are available at http://www.nyc.gov/html/doh/html/hanes/datasets.shtml. Visit EpiQuery: https://a816-healthpsi.nyc.gov/epiquery/EpiQuery/NYCHANES/index.html. Data sets are publicly available.
Main Purpose of Study: The YRBS data have three main uses: instrumental, that is, the data influence health education program decisions; conceptual in that they increase the understanding of the health status of New York adolescents; and persuasive in that they support health policy positions. The Bureau of Epidemiology Services uses YRBS data in Vital Signs publications on health behaviors of NYC public high school students.
Sample: 1,500–10,000 annually. From 1993 through 2001, the NYC YRBS was designed to provide data on a citywide level. In 2003 the survey was expanded to provide borough-level data. It was further expanded in 2005, 2007, and 2009 to data for the three District Public Health Office (DPHO) areas in the South Bronx, North and Central Brooklyn, and East and Central Harlem, as well as borough and citywide estimates. It includes public high school students (grades 9–12) in New York City. The YRBS does not include students in juvenile detention centers. English as a Second Language (ESL) and special education classes in otherwise-eligible high schools are also excluded from the sample. The overall response rate for each YRBS stratum needs to be at least 60% to be used.
Frequency of Collection and Sources of Data: Data are collected every other year since 1993 through population interviews.
Mode of Data Collection: Self-administered, paper questionnaire
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: The YRBS is designed to monitor priority health-risk behaviors that contribute to the leading causes of mortality, morbidity, and social problems among youth in New York City. The questionnaire measures tobacco, alcohol and drug use; behaviors that contribute to
unintentional injury and violence; sexual behaviors; dietary behaviors; and physical activity. It also monitors the prevalence of obesity and asthma.
Information Obtained: Prevalence and incidence of selected diseases and risk factors for selected diseases; awareness, treatment, and control of selected diseases; and the prevalence and magnitude of environmental exposures. Multiple demographic variables are collected, including age, ethnicity, and race.
Who Pays for Data Collection?: The OPTS cost of the 2009 YRBS was $180,000. It is funded with CDC dollars and city tax levy.
Dissemination of Data: Variable lists, documentation, codebooks, and data sets are available at http://www.nyc.gov/html/doh/html/episrv/episrv-faq-yrbs.shtml. Data sets are publicly available. Visit EpiQuery: https://a816-healthpsi.nyc.gov/epiquery/EpiQuery/NYCHANES/index.html.
Main Purpose of Study: Tracking of medical care indicators for improving the delivery of recommended preventive services (list limited to cardiovascular related indicators): Aspirin/antithrombotic therapy, blood pressure control, cholesterol control, smoking cessation intervention, diabetes care indicators (A1c test, A1c control, lipid screening, lipid control), asthma symptom assessment and control (ages 5–56, adult and pediatric indicators can be stratified).
Sample: Nearly 1,800 providers are using the EHR system, representing over 100,000 encounters per month. The data base will increase in the next two years as an estimated additional 2,500 providers will be transmitting information to PCIP. The geographic stratification is citywide. The geographic levels included are provider, practice, ZIP code, borough, and citywide. The age ranges are currently based on quality metric collected (e.g., blood pressure control measure age range includes patients 18 to 75).
Frequency of Collection and Sources of Data: All providers with EHR have the capacity to transmit the data. Estimated 80% of practices with EHRs are transmitting information. Data are collected monthly.
Mode of Data Collection: Automated medical chart abstraction facilitated by electronic health records maintained by independent primary care practices (this currently includes small practices, community health centers, and outpatient hospitals) Automated electronic transmission of extracted data fields from the medical charts to a secure data warehouse maintained by PCIP.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Smoking status, smoking cessation intervention, A1c control (< 7%), A1c testing, antithrombotic tx (IVD or DM), body mass index, BP Control in DM (130/80), BP Control in HTN (140/90), BP Control in IVD (140/90), cholesterol control (genl pop), cholesterol screen (genl pop), LDL control (high risk), LDL testing (high risk), asthma symptom assessment, asthma control (18–56 yrs), depression screening, depression follow-up, depression control, alcohol use screening
Information Obtained: CVD disease and risk factor prevalence among all patients seen by providers transmitting data to PCIP; control among those with select CVD risk factors as per accompanying quality indicators. Limited patient characteristics as information is aggregated to the provider or practice. Provider and practice demographics currently collected include ZIP code, % Medicaid patients, % Uninsured patients, encounters, number of providers, and FTE of providers.
Who Pays for Data Collection: Costs incurred include staffing, associated hardware and software purchases, and maintenance. Estimated $1.2 million per year. It is funded by a New York City tax levy; HEAL5 and HEAL 10 grants from New York State; Regional Extension Center funding from the Office of the National Coordinator for Health Information Technology; Communities Putting Prevention to Work—Tobacco Cessation—Centers for Disease Control; Centers of Excellence—Centers for Disease Control; Agency for Healthcare Research and Quality; Robin Hood Foundation
Dissemination of Data: The data are currently used for internal program operations (technical assistance, coaching on quality improvement). Summarized trends and population averages will be made available in the future. Currently, only PCIP staff and providers can access their own data. In the future, summarized information will be made publicly available.
Main Purpose of Study: CDC Division for Heart Disease and Stroke Prevention fund Arkansas, Kansas, Oklahoma, and Washington to collect data on measured levels of blood pressure and blood cholesterol; compare data between priority and general populations; provide information to guide development, implementation, evaluation of cardiovascular health promotion and risk factor control strateiges; reduce the burden of heart disease and stroke-related death, disabilities, and cost; eliminate health disparities. Based on the findings, states will develop hypertension and cholesterol control strategies.
Frequency of Collection and Sources of Data: Started in 2005. The source of data is questionnaire.
Mode of Data Collection: Interview and health exam
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Arkansas: blood pressure knowledge/history, cholesterol knowledge/history, cardiovascular disease knowledge/history, diabetes knowledge/history; Kansas: blood cholesterol, LDL/HDL, triglycerides, blood sugar, HbA1C, Medical history (MI, Angina/CHD, Stroke, TIA, Atrial fibrillation, deep venous thrombosis, pulmonary embolus, peripheral arterial disease, left ventricular heart hypertrophy, other heart disease, hospitalization on selected heart conditions, symptoms of angina, congestive heart failure, peripheral arterial disease, heart attack and stroke symptoms, personal history of blood pressure, cholesterol, diabetes, family history of blood pressure, heart attack, high cholesterol, diabetes, apnea); Oklahoma: actions to control high blood cholesterol, history of cholesterol checked, actions to control high cholesterol, history of CHD/angina, stroke, other cardiovascular diseases, diabetes, aspirin use, rehab following heart attack, stroke, family health history; Washington: HDL, LDL, VLDL, triglycerides, glucose, blood pressure, pulse
Information Obtained: Smoking and tobacco use, drinking, weight history, diet behavior and nutrition, supplement use, physical activity, stress aspirin use. Arkansas: age, gender, ethnicity/race, education, income, marital status, employment/occupation, household composition (number of adults, number of children), housing
Cost and Who Pays for Data Collection?: Arkansas: $1.1 million in cash, $250,000–$300,000 in kind; Kansas: $800,000 in cash, $250,000 in kind. Funded by the Centers for Disease Control and Prevention’s Division for Heart Disease and Stroke Prevention.
Dissemination of Data: Data are not publicly available.
Main Purpose of Study: To monitor priority health risk behaviors among youth.
Sample: Varies by component; National Survey ~15,000; In 2007, the national YRBSS received 14,103 questionnaires across 157 schools; state, territorial, tribal, and local surveys ~2,000 per site. National, state, territorial, local, tribal levels have separate surveys. The national survey oversamples black and Hispanic students; measure of size was modified so that it would be more likely to select schools with disproportionately high minority enrollment; two classes per grade were selected in schools with large number of minority students.
High school students in grades 9–12; There is a middle school questionnaire available as well, but it is done by only 10 to 15 sites per cycle and not at the national level. Response rate varies by survey. In 2007, the school response rate was 81%, the student response rate was 84%, and the overall response rate was 68%.
Frequency of Collection and Sources of Data: Every other year, usually during the spring semester of odd numbered years. Students complete the self-administered paper and pencil questionnaire during one class period and record their responses directly on a computer-scannable booklet or answer sheet.
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Ever been told by a doctor or nurse that they had asthma?; Ever told by a doctor or nurse that they had asthma and still have asthma?
Information Obtained: Asthma, obesity, health risk behaviors. Behaviors that contribute to unintentional injuries and violence; tobacco use; alcohol and other drug use; sexual behaviors that contribute to unintended pregnancy and sexually transmitted diseases (STDs), including the human immunodeficiency virus (HIV) infection; unhealthy dietary behaviors; and physical inactivity. Grade level/education level, sex, race/ethnicity.
Cost of Survey and Who Pays for Data Collection?: For the entire surveillance system, the cost is about $6 million per cycle. CDC has funding available across 50 state education agencies and a small number of territories, tribal governments, and large urban school districts during each five-year funding cycle.
Dissemination of Data: The national YRBSS are available in ASCII, SPSS, SAS data, SAS formats, and Access formats; go to www.cdc.gov/yrbs; click on National Data Files and Documentation. The public may contact YRBSS for state and district-specific data access. The latest report of 2007 national, state, and local results: http://www.cdc.gov/mmwr/PDF/ss/ss5704.pdf.
The 2009 YRBSS results were released in June 2010.
Main Purpose of Study: To evaluate performance of health plans; HEDIS Health Plan includes HMOs and PPOs
Sample: Mandatory or voluntary (Medicare Advantage) reporting of 702 HMO products and 277 PPO products. A product is an insurance plan at a state level, segmented by Medicare, Medicaid, and Commercial. The total number of covered lives in the plans reporting HEDIS data is estimated at 116 million. Information is collected by state, reported by region. Geographic levels are national, state, regional, and individual plan. Age groups are dependent on the measure; 0–18 for child-related measures, 18–85 for adults, with older ages variable with measure. Response rate: About 85% of HMOs and 40% of PPOs report data—missing plans are mostly smaller Medicaid and commercial plans.
Frequency of Collection and Sources of Data: Annual data collection reported by plans in May of the year following services
Claims data (diagnoses, procedures, laboratory and pharmacy claims) augmented in some cases by clinical chart reviews (with specified sample size) and patient surveys (random sample of a specific sample size).
Mode of Data Collection: Health Plan Consumer Assessment of Healthcare Providers and Systems (CAHPS), a public–private initiative to develop standardized surveys of patients’ experience with ambulatory and facility-level care (HEDIS Measures 2010, NCQA website).
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: Clinical measures for cardiovascular disease, chronic pulmonary disease, asthma, and diabetes; 2010 HEDIS summary measures include appropriate testing for children with pharyngitis, appropriate treatment for children with upper respiratory infection, avoidance of antibiotic treatment in adults with acute bronchitis, use of spirometry testing in the assessment and diagnosis of COPD, pharmacotherapy of COPD exacerbation, persistence of beta-blocker treatment after a heart attack, comprehensive diabetes care, relative resource use for people with diabetes, relative resource use for people with asthma, relative resource use for people with cardiovascular conditions, relative resource use for people with uncomplicated hypertension, and relative resource use for people with COPD (HEDIS Measures 2010, NCQA website).
Information Obtained: Incidence—indirectly; prevalence—indirectly; functional health outcomes—only with HOS survey used for Medicare advantage plans; risk factors, including stressors—depending on measure but not stressors; clinical care information—with measures; demographic characteristics: Patient level data ONLY in HOS survey (Medicare Advantage plans)
Who Pays for the Data Collection?: Plans pay for data collection at MD and patient level; they then submit data in aggregate to NCQA via a standard data collection tool.
Dissemination of Data: Processed data available via the NCQA Quality Compass. Anyone can purchase a license to use data within set of restrictions dealing with commercial use of data. Special arrangements can be made for research use. Health plans submit data in aggregate to NCQA via a standard data collection tool.
Main Purpose of Study: To evaluate physician office practices
Sample: Open to qualified medical practices (mostly primary care). Practice level data; reported by practice within state and national. All ages are included. About 15,000 physicians in 1,000 practices are recognized.
Frequency of Collection and Sources of Data: Every 3 years. Medical record review.
Mode of Data Collection: Standardized web-based data collection tools; Provision for direct reporting from some EMR installations
Specific Questions Related to CVD, COPD, Asthma, and/or Diabetes: All clinical data, no surveys from patients. 2010 HEDIS measures for physicians include appropriate treatment for children with upper respiratory infection; controlling high blood pressure; cholesterol management for patients with cardiovascular conditions; comprehensive ischemic vascular disease; comprehensive diabetes care; use of appropriate medications for people with asthma; use of spirometry testing in the assessment and diagnosis of COPD; and pharmacotherapy management of COPD exacerbation (HEDIS Measures 2010, NCQA website). Practice-level demographics; language diversity of membership; race/ethnicity diversity of membership (NCQA website).
Dissemination of Data: Researchers can apply to use the data.
Bild, D. E., D. A. Bluemke, G. L. Burke, R. Detrano, A. V. Diez Roux, A. R. Folsom, P. Greenland, D. R. Jacobs, Jr., R. Kronmal, K. Liu, J. C. Nelson, D. O’Leary, M. F. Saad, S. Shea, M. Szklo, and R. P. Tracy. 2002. Multi-ethnic study of atherosclerosis: Objectives and design. American Journal of Epidemiology 156(9):871-881.
Desai, J., L. Geiss, Q. Mukhtar, T. Harwell, S. Benjamin, R. Bell, and E. Tierney. 2003. Public health surveillance of diabetes in the United States. Journal of Public Health Management and Practice 9:S44-S51.
Go, A. S., D. J. Magid, B. Wells, S. H. Sung, A. E. Cassidy-Bushrow, R. T. Greenlee, R. D. Langer, T. A. Lieu, K. L. Margolis, F. A. Masoudi, C. J. McNeal, G. H. Murata, K. M. Newton, R. Novotny, K. Reynolds, D. W. Roblin, D. H. Smith, S. Vupputuri, R. E. White, J. Olson, J. S. Rumsfeld, and J. H. Gurwitz. 2008. The Cardiovascular Research Network. Circulation: Cardiovascular Quality and Outcomes 1(2):138-147.
Hagan, P. G., C. A. Nienaber, E. M. Isselbacher, D. Bruckman, D. J. Karavite, P. L. Russman, A. Evangelista, R. Fattori, T. Suzuki, J. K. Oh, A. G. Moore, J. F. Malouf, L. A. Pape, C. Gaca, U. Sechtem, S. Lenferink, H. J. Deutsch, H. Diedrichs, J. Marcos y Robles, A. Llovet, D. Gilon, S. K. Das, W. F. Armstrong, G. M. Deeb, and K. A. Eagle. 2000. The International Registry of Acute Aortic Dissection (IRAD) new insights into an old disease. The Journal of the American Medical Association 283(7):897-903.
Magid, D. J., J. H. Gurwitz, J. S. Rumsfeld, and A. S. Go. 2008. Creating a research data network for cardiovascular disease: The CVRN. Expert Review of Cardiovascular Therapy 6(8):1043-1045.
Mokdad, A. H. 2009. The behavioral risk factors surveillance system: Past, present, and future. Annual Review of Public Health 30(1):43-54.