Cost-Benefit Analysis of Providing Non-Emergency Medical Transportation (2005)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Chapter 4: Description of the Transportation- Disadvantaged Population That Misses Medical Care Due to a Lack of Access to Non- Emergency Medical Transportation In Chapter 3, the number of non-institutionalized people in the U.S. who are transportation disadvantaged and have missed non-emergency medical treatment at least once in the past 12 months was estimated to be approximately 3.6 million. As discussed, this “target population” differs from the remainder of the U.S. in terms of its socio-economic, demographic, and health characteristics. In this chapter, the characteristics of this target population are described and compared with the rest of the U.S. population to provide a better understanding of who is in the target population and to show what medical conditions are prevalent and important for members of this population. In later chapters these conditions are shown to be critical to further analysis of providing NEMT to the target population. 4.1 Demographic and Socio-economic Characteristics of the Population That Lacks Access to NEMT Before studying the health characteristics of the target population, it is important to confirm that this population of roughly 3.6 million people demonstrates the demographic and socio-economic conditions that the literature discussed in Chapter 2 suggests that is has. The NHIS and MEPS data are again used to for this analysis. The NHIS adult sample data were used to compare the demographic and socio- economic characteristics of the target population with the rest of the U.S. population. This analysis showed that, compared to the rest of the U.S. population, the target population: • Has relatively low income (54.6 percent have household incomes below $20,000 per year compared with only 17.7 percent for the remainder of the U.S. population) • Is disproportionately female (62.8 percent female versus 51.9 percent) and non-white (19.1 percent non-white versus 17.7 percent) • Has a higher minority representation (13.5 percent African American versus 12.6 percent; 16.7 percent Hispanic versus 13.2 percent) • Is roughly one-half as likely to possess a four-year college degree • Is older (16.3 percent are 70 or older compared with 11.5 percent) • Is distributed across urban and rural America much the same as the U.S. population as a whole, although children are slightly more concentrated in urban areas While the demographic and socio-economic characteristics of the target population are decidedly different from the remainder of the U.S. population, these characteristics match well with the descriptions of the transportation-disadvantaged population found in the literature review as described in Chapter 2. Final Report 24

4.2 Urban-Rural Split of the Target Population That Lacks Access to NEMT To determine the cost of providing the NEMT that this population needs, the target population of transportation-disadvantaged people that missed non-emergency medical care due to a lack of transportation must be disaggregated by location (urban and rural), transportation service needs, and medical condition. Transportation costs in rural locations can be higher due to distance traveled, the reduced ability to consolidate riders into one vehicle, and poorer access to fixed-route public transportation. The type of trip affects the need for specially equipped vehicles, such as ambulances for stretcher transportation service or wheelchair-accessible vehicles. Both characteristics of transportation are crucial for estimating transportation costs for the target population. Disaggregating the target population by location (rural versus urban) is straightforward using NHIS data. Although the 2002 NHIS currently lacks information pertaining to location, the 2001 NHIS data include this information for both adults and children. NHIS uses a different geographic categorization than is common in transportation data sets, such as NHTS, but the non-MSA (Metropolitan Statistical Area) field in the 2001 NHIS closely matches the rural field in the NHTS. Specifically, the 2001 NHIS indicates that 20.7 percent of adults and 20.0 percent of children reside in non-MSA locations, while the 2001 NHTS indicates that 21.9 percent of all persons represented live in a rural location. Given this close agreement between the two datasets, we decided to use the non- MSA category in the 2001 NHIS as the operational definition of rural in our analysis. This analysis showed that 22.2 percent (weighted) of the adults who reported missed trips due to transportation problems lived in rural (non-MSA) locations, as did 14.8 percent of target population children. Thus, while the percentage of adults in rural areas who are transportation disadvantaged by NHIS standards is slightly higher than is the percentage of rural adults overall, the percentage of children in rural areas who are transportation disadvantaged by NHIS standards is well below the overall percentage of rural children, indicating that children who lack access to medical care due to a lack of transportation are somewhat more concentrated in urban areas. Applying these urban and rural percentages to the estimate of roughly 3.6 million in the target population, and recalling that 74.4 percent of the target population are adults (2.76 million adults and 0.95 million children who missed trips from the 2002 NHIS), about 593,721 adults in rural areas are in the target population (meaning that 2,084,319 adults are in urban areas). Likewise, for children, approximately 136,727 of those in the target population live in rural areas (meaning that roughly 785,233 are in urban areas). These results are summarized in Table 4-1. Final Report 25

Table 4-1: Number of Persons Missing NEMT Trips per Year in Urban and Rural Areas Age Category Urban Rural Adult 2,084,319 593,721 Children 785,233 136,727 Total 2,869,552 730,448 4.3 Identification of Medical Conditions Faced by Those Lacking NEMT In our analysis to identify the most important medical conditions for which lack of transportation is a barrier to health-care access, we focused on prevalent conditions for the transportation-disadvantaged population. We used the 2002 NHIS data to examine the prevalence of disease conditions and cross-tabulated this information with the NHIS concept of target population status, as described in Chapter 3. The NHIS has separate data sets for adults, children, and immunizations, and we discuss each in the sections that follow. 4.3.1 Adult Disease Conditions The results of the adult analysis are shown in Table 4-2. This table presents a comprehensive list of medical conditions for individuals who also reported difficulties accessing care due to transportation problems. It reports the weighted percentage of these adults experiencing the condition in question. Table 4-2: Medical Conditions Experienced by Adults in the Target Population Condition Unweighted Frequency Percentage of Adults Pain/Aching Joints 304 55.8 Depressed 280 49.7 Recurring Pain 261 48.2 Insomnia 258 49.4 Arthritis 235 40.0 Hypertension 233 37.7 Vision Problems 219 37.5 Excessive Sleepiness 176 35.2 Heart Disease 167 29.6 Poor Circulation 158 26.8 Dental Problems 157 28.0 High Cholesterol 146 25.7 Medication Allergies 130 23.2 Skin Problems 120 21.4 Urinary Problems 119 20.8 Asthma 113 22.1 Ulcer 110 19.6 COPD 101 17.6 Final Report 26

Condition Unweighted Frequency Percentage of Adults Sinusitis 101 16.9 Severe Sprains 99 20.0 Food/Odor Allergies 97 17.8 Diabetes 96 16.0 Irritable Bowel Syndrome 84 12.9 Menstrual Problems 73 13.1 Hay Fever 72 12.5 Thyroid Problems 71 11.7 Cancer 60 11.3 Gynecologic Problems 55 8.8 Stroke 44 8.1 Seizures 42 7.5 Menopausal Problems 40 6.5 Renal Disease 37 7.2 Hearing Aid 35 6.8 Liver Condition 35 6.7 Neuropathy 18 3.1 Multiple Sclerosis 6 1.2 Prostate 3 0.5 Parkinson's Disease 2 1.4 Psychological Problems: Hopeless 30 5.8 Nervous 45 8.5 Restless/Fidgety 41 8.8 Sad 28 5.2 Worthless 26 4.8 Before defining a list of conditions for further study, two comparisons between adults in the target population and all other adults, as seen in the NHIS data, are especially instructive: 1. Viewing the phenomenon of multiple diseases suffered by the same person: (1) There is a very large difference in the percentage of adults experiencing multiple conditions from this list (92 percent for the target population versus 64 percent for non-target population); and (2) The difference in the percentage who experienced none of these conditions is also substantial (3 percent for target population versus 21 percent for non-target population). 2. For each condition (except for “no conditions”), the prevalence of the conditions is higher for target population adults. Examples for some of the highly prevalent conditions are shown in Table 4-3. Final Report 27

Table 4-3: Prevalent Conditions in Adults from 2002 NHIS Condition Target population Prevalence (Percentage) Non-Target population Prevalence (Percentage) Ratio of Prevalences Arthritis 40.0 20.5 1.9 Asthma 22.1 10.5 2.1 COPD 17.6 5.3 3.3 Cancer 11.3 6.9 1.6 Depression 49.7 15.2 3.3 Dental Problems 28.0 12.4 2.3 Diabetes 16.0 6.4 2.5 Heart Disease 29.6 15.5 1.9 High Cholesterol 25.7 20.5 1.3 Hypertension 37.7 24.0 1.6 Irritable Bowel 12.9 5.4 2.4 Medication Allergies 23.3 12.9 1.8 Pain/Aching Joints 55.8 29.1 1.9 Poor Circulation 26.8 8.3 3.2 Recurring Pain 48.2 17.7 2.7 Severe Sprains 20.0 8.1 2.5 Skin Problems 21.4 8.5 2.5 Vision Problems 37.5 16.1 2.3 Not only does this disadvantaged group suffer from insufficient transportation to conduct required health-related visits, they exhibit an exorbitant prevalence of a number of serious conditions. The picture that emerges is one of an especially unhealthy population. In addition to revealing conditions that should be included in the study, our empirical results also indicate one condition that should be dropped – substance abuse. Substance abuse was deleted from the list of conditions for two main reasons: (1) very low acknowledged prevalence in the NHIS; and (2) difficulty connecting it to required visits, and hence, missed visits. The NHIS data also demonstrate that mental health is a key condition for study. Not only do 52 percent of the disadvantaged adults indicate depression, close to one-third mention excessive sleepiness; nearly 50 percent note insomnia; and the psychological problem categories of “hopeless,” “nervous,” “restless/fidgety,” “sad,” and “worthless” each have a prevalence of 5 percent or more. Moreover, these related mental health conditions are highly correlated with other conditions. 4.3.2 Child Disease Conditions For children in the target population category, our analysis produces similar results as for the adults, as shown in Table 4-4. This table presents a list of the conditions for children in the target population category, including the weighted percentage of children citing this condition. The pattern of co-morbidities and condition prevalence Final Report 28

that we found for adults is also relevant. The results show a large difference both for the percentage experiencing multiple conditions from the list (32 percent for target population children versus 14 percent for non-target population children), and for the percentage experiencing none of the listed conditions (39 percent for target population children versus 57 percent for non-target population children). The prevalence is higher for each condition, when comparing across target population status. For the most prevalent conditions, the percentages are: ADHD/ADD (9.4 percent versus 5.5 percent), asthma (24.4 percent versus 12.4 percent), frequent headaches (12.8 percent versus 5.3 percent), colds (32.2 percent versus 20.5 percent), and learning disability (11.7 percent versus 6.5 percent). Even for low-prevalent conditions, the percentages are: stutters (6.7 percent versus 1.2 percent) and vision problem (7.8 percent versus 2.6 percent), again where the first percentage indicates target population children, and the second non-target population children. Table 4-4: Medical Conditions Faced by Children Who Lack NEMT Condition Unweighted Frequency Percentage of Children Head/Chest Cold, past 2 wks 58 32.2 Asthma 44 24.4 Freq Headaches/Migraines 23 12.8 Learning Disability 21 11.7 ADHD/ADD 17 9.4 Vision Problem 14 7.8 Stutters/Stammers 12 6.7 Ear Infections 7 3.9 Freq Diarrhea/Colitis 6 3.3 Eczema/Skin Allergies 5 2.8 Hay Fever 4 2.2 Heart Disease 4 2.2 Mental Retardation 4 2.2 Respiratory Allergies 4 2.2 Food/Digestive Allergies 3 1.7 Seizures 2 1.1 Anemia, Past 12 months 1 0.6 Arthritis 1 0.6 Autism 1 0.6 Cerebral Palsy 1 0.6 Depressed 1 0.6 Muscular Dystrophy 1 0.6 Other Dev Delay 1 0.6 Sickle Cell Anemia 1 0.6 Final Report 29

Table 4-4 provides strong evidence that a comprehensive focus on children would include the following conditions: • ADHD/ADD and learning disabilities in general (as a way to avoid educational delays to the extent possible) • Asthma • Frequent headaches/migraines 4.3.3 Discussion and Conclusions from Analysis of Medical Conditions Our analysis of these nationally representative healthcare datasets reveals that about 3.6 million Americans miss or delay non-emergency medical care per year due to transportation difficulties. Based on the 2002 NHIS data, we selected eleven critical medical conditions that are prevalent in the target population. These conditions include both chronic conditions amenable to disease management and those amenable to preventive care. Table 4-5 lists the conditions and their prevalence in the target population. Using a three-year merged NHIS data set (2001-2003), it has further been determined that approximately two-thirds of the target population is affected by at least one of the seven chronic conditions, thus demonstrating the importance of analyzing these conditions for cost-effectiveness in Chapter 7. Table 4-5: Critical Medical Conditions Affecting Transportation-Disadvantaged Persons Who Lack Access to NEMT Type of Care Medical Condition Prevalence in the Target Population (%) Chronic Depression or Other Mental Health Problem 50 Hypertension 37 Heart Disease 26 Asthma 20 Chronic Obstructive Pulmonary Disease 19 Diabetes 15 End-stage Renal Disease 7 Preventive Dental Problems 28 Cancer 12 Premature Births 2 Vaccinations N/A For both the adult and child samples, the analysis of the NHIS data reveals a much higher prevalence of conditions (and multiple diseases) for target population individuals compared with non-target population individuals. This general finding has important implications for the project. Transportation issues that result in missed trips will potentially exacerbate the diseases afflicting these individuals and may result in costly subsequent medical care (specialist visits, emergency room visits, and possibly hospitalizations). Even when this is not the case—i.e., the potential does not exist to decrease subsequent utilization by more prompt care of an existing condition—there are important quality-of-life concerns. For example, the prevalence of frequent headaches is more than twice as high for children in the target population Final Report 30

as it is for others. To the extent that visits are being delayed for these disadvantaged children, they could be subject to considerable pain and suffering that otherwise could be effectively treated. 4.3.4 The Prevalence of Multiple Medical Conditions Experienced by the Target Population To support many of the points made above, we conducted a special analysis of condition prevalence by merging the most recent three years of NHIS data, including 2003 results that were released in December 2004. In Table 4-6, we focus on nine conditions for which the specific survey questions aligned over the three-year period. These data clearly show that the target population not only has a higher prevalence of any single health condition, but also are more likely to have multiple medical conditions when compared to the general population. This holds true for all cross sections except for the end-stage renal disease (ESRD) population with hypertension. Accordingly, the proposition that the target population that misses visits is more chronically ill is well supported. The extended cost of care to society at large may be increased by the lack of adequate NEMT. Table 4-6: Prevalence of Co-Morbidities in the Target Population Compared to Everyone Else Percent with another health condition Î As th m a C O P D D ia be te s ES R D H ea rt D is ea se H yp er te n si on C an ce r C u rr en tl y P re gn an t Asthma 100% 45% 20% 10% 36% 42% 14% 2% 95% 91% 5% COPD 47% 100% 20% 14% 41% 47% 13% 1% 99% 97% 1% Diabetes 26% 25% 100% 16% 50% 69% 15% 1% 97% 95% 3% ESRD 26% 35% 33% 100% 45% 54% 20% 0% 98% 95% 2% Heart Disease 27% 29% 28% 13% 100% 59% 18% 1% 97% 94% 3% Hypertension 22% 24% 27% 11% 41% 100% 15% 1% 96% 92% 4% Cancer 23% 20% 19% 12% 40% 47% 100% 1% 97% 93% 3% Currently Pregnant 25% 11% 7% 0% 12% 17% 8% 100% 65% 46% 35% Asthma 100% 20% 8% 2% 18% 28% 8% 1% 80% 71% 20% COPD 37% 100% 13% 5% 30% 41% 14% 1% 94% 86% 6% Diabetes 13% 11% 100% 6% 35% 64% 14% 0% 91% 78% 9% ESRD 17% 19% 28% 100% 42% 62% 20% 0% 94% 86% 6% Heart Disease 15% 13% 18% 5% 100% 54% 16% 0% 91% 81% 9% Hypertension 12% 9% 17% 3% 28% 100% 12% 0% 83% 75% 17% Cancer 12% 11% 13% 4% 30% 44% 100% 0% 88% 78% 12% Currently Pregnant 12% 4% 1% 0% 4% 7% 1% 100% 43% 30% 57% Notes: *** The percent of this population that has only the one targeted condition H ea lth C on di tio n H ea lth C on di tio n Key Medical Conditions* ** Non-targeted conditions include: Arthritis, Hay Fever, Hearing Aid, Liver Condition, Nervous, Pain/Aching Joints, Poor Circulation, Restless/Fidgety, Sinusitis, Stroke, Ulcer, and Vision Problems. * Mental Health, Dental Problems, and Preventive care are not included in this table as targeted conditions due to data definitions. H as a t le as t on e ot he r ta rg et co nd iti on H as a t le as t on e ot he r no n- ta rg et ed co nd iti on ** N o ot he r ta rg et ed co nd iti on ** * Everyone else Transportation-disadvantaged population that missed care Final Report 31

4.4 Utilization of Health Resources by the Target Population The prevalence of multiple health conditions (co-morbidities) has a demonstrable effect on the utilization of health resources. Target population adults use far more healthcare services compared to the remainder of the U.S. population. In an analysis of the 2001 NHIS data, the target population required more trips for care than the remainder of the U.S. population. Table 4-7 shows utilization in terms of the mean number of visits for both the target population and non-target populations for emergency room visits, home care visits, routine provider office visits, and number of surgeries—all covering the 12 months prior to the date of the survey. We did not capture hospitalizations for this analysis. The utilization analysis uses 2001 instead of 2002 NHIS data, because the 2002 data do not yet include—and may never include—the data needed to enable one to assign a rural/urban breakdown (see the discussion section below). For 2001, 578 adult cases (1.37 percent or 2,801,152 of the weighted population) are designated as transportation disadvantaged according to our NHIS screen. Table 4-7: Aggregate Utilization Means for the Target Population and Non-Target Population (Prior 12 Months) Utilization Category Mean for Target Population Mean for Non-target Population Emergency Room (ER) Visits 1.31 0.35 Home Care Visits 0.54 0.15 Office Visits 6.78 3.97 Number of Surgeries 0.24 0.16 Additional utilization comparisons to consider: • 54.9 percent of target population had no ER visits compared with 80.1 percent for the remainder of the U.S. population • 10.9 percent of the target population had no office visits in the past 12 months, and 16.5 percent had 16 or more visits. The corresponding numbers for the non-target population are 19.4 percent who had no visits and 5.7 percent who had 16 or more visits. The differences across the four types of health services are quite large. The difference for office visits would be larger still but the NHIS utilization data capped the number of visits at 16 and, as noted above, a far greater percentage of disadvantaged individuals experienced at least 16 office visits in the preceding year compared to the other population. These results are substantiated in Section 6 using the more detailed and accurate MEPS data. Table 4-8 shows the average number of visits excluding hospitalizations for individuals with the seven chronic conditions analyzed in Chapter 7. Compared to an average of 3 outpatient visits per person for the total U.S. population, these numbers are extraordinarily high and, while not relating to missed trips, indicate the potential scope of the problem given the nearly 4 million transportation disadvantaged individuals. Final Report 32

Table 4-8: Condition-Based Visits from the 2001 MEPS Condition # of Cases Average Visits Asthma 1,347 8.8 Heart Disease (CHF) 161 18.9 COPD 1,105 9.9 Hypertension 3,829 11.1 Diabetes 1,555 13.0 Depression 1,427 14.8 ESRD 34 115.0 4.4.1 Disease Prevalence and Co-Morbidities These results amplify a theme discussed in Chapter 2, that is, the target population (as represented in the NHIS) has a higher prevalence of disease conditions, a much greater propensity for multiple, simultaneous diseases, and most likely has higher severity for individual conditions, as compared with the non-target population. Employing ten important conditions to represent the universe of diseases, 63.4 percent of the target population (versus 34.2 percent for the others) suffers from multiple conditions, while 18.4 percent of the target population (versus 41.1 percent for the others) experiences none of these conditions. Table 4-9 lists the ten selected conditions; they illustrate aspects of the co-morbidity findings and also serve to compare target population and non- target population utilization means at the condition level. Table 4-9: Condition Prevalence and Utilization Means at the Condition Level – Target Population Target Population Rest of U.S. Population Condition Weighted Percent with Disease Mean ER Visits Mean Home Care Visits Mean Office Visits Weighted Percent with Disease Mean ER Visits Mean Home Care Visits Mean Office Visits Arthritis 41.78 1.61 1.07 8.65 21.14 0.54 0.39 6.22 Asthma 16.74 2.37 0.75 7.51 10.79 0.66 0.18 5.33 COPD 18.35 2.04 0.25 8.44 6.34 0.78 0.39 6.62 Diabetes 15.47 1.89 1.19 8.68 6.25 0.73 0.62 7.13 Renal Disease 8.04 3.10 0.90 10.23 1.53 1.29 0.93 8.41 Heart Disease 22.03 1.77 1.69 8.64 11.37 0.74 0.62 6.74 Hypertension 33.31 1.32 1.06 8.59 23.39 0.54 0.35 5.83 Mental Health 15.35 2.64 0.87 9.79 2.81 1.11 0.48 7.24 Pain/Aching Joints 54.97 1.66 0.82 8.39 32.85 0.49 0.28 5.52 Vision Problems 31.31 1.82 1.06 8.23 9.70 0.60 0.53 5.81 Source: NHIS 2002 Final Report 33

For each of the ten conditions, the disease prevalence is considerably higher for the target population—transportation-disadvantaged persons who miss medical care due to a lack of access to NEMT—compared to the remainder of the U.S. population. For example, 18.35 percent of the target population, but only 6.34 percent of other adults, experience chronic obstructive pulmonary disease (COPD). With two exceptions (mean home care visits for COPD and renal disease), the mean utilization is higher for the target population than it is for other adults. The utilization in question is not specifically connected to each of the diseases. For example, target population adults who suffer from asthma had a mean of 7.51 office visits in the past 12 months, but we have no direct information that tells us to which conditions these visits pertain. What we know from this analysis is simply the total number of office visits, including those for any unrelated conditions. Even with comprehensive healthcare utilization information, such as the data available in the MEPS, one may know the primary reason for a physician visit, but it would be very difficult to determine what other conditions are treated during an encounter. Final Report 34