discussions are critical to strengthening the appropriate use of all of the above types of research in building the evidence base on effective treatments for M/SU illnesses.
In general health care, routinely collected administrative data (e.g., claims or encounter data) that are generally produced each time a patient is admitted to a hospital or makes a visit to an ambulatory heath care provider are widely used for health services research, epidemiologic studies, and quality assessment and improvement initiatives (Iezzoni, 1997; Zhan and Miller, 2003). While these datasets have limitations with respect to their completeness, accuracy, and level of detail (AHRQ, 2004a; Iezzoni, 1997), administrative data remain a preferred and routinely used source of information for multiple quality-related purposes because they are readily available, inexpensive, and computer readable (AHRQ, 2004b; Zhan and Miller, 2003). For example, analysis of administrative data revealed the now well-known and sizable variations that exist in clinical care within the United States, an analysis that continues today (Mullan, 2004; Wennberg, 1999). Consequently, administrative data produce a variety of clinical quality indicators for hospital care (AHRQ, 2004b), underpin many of the quality measures found in the National Committee for Quality Assurance’s (NCQA) Healthplan Employer Data and Information Set (HEDIS) performance measures (NCQA, 2004a), and are the data source for the Agency for Healthcare Research and Quality’s (AHRQ) new patient safety indicators (Zhan and Miller, 2003). Because of their utility, administrative data are viewed as a mainstay of health services research on quality of care (Iezzoni, 1997) and are likely to become even more so as the National Health Information Infrastructure is developed (see Chapter 6).
These inpatient and outpatient datasets typically contain standardized information on each individual’s diagnosis (using International Classification of Diseases [ICD] codes) and on the specific therapies and procedures performed for that diagnosis (using the American Medical Association’s [AMA] Current Procedural Terminology [CPT] codes, the Centers for Medicare and Medicaid Services’ (CMS) Healthcare Common Procedure Coding System [HCPCS] for outpatient care, and ICD, ninth revision, Clinical Modification (ICD-9-CM) procedure codes for inpatient care). However, these codes are less useful at present for the study of M/SU care than for the study of general health care for several reasons. Psychotherapy codes are few and imprecise and differ across inpatient and outpatient settings. Codes for other psychosocial services generally are absent, as are codes for the use of restraints. And the new CPT II codes for use in performance measure-