Landsman P, Berger ML, Straus W; Academy for Health Services Research and Health Policy. Meeting.

Abstr Acad Health Serv Res Health Policy Meet. 2002; 19: 7.

Merck & Co. Inc, WP39-166, PO Box 4, West Point, PA 19486-0004; Tel: (215) 652-7492; Fax: (215) 652-0860; E-mail: pamela_landsman@merck.com

RESEARCH OBJECTIVE: Variability in coding of diagnoses limits the reliability and generalizability of medical claims analyses. We hypothesized coding of claims for prevalent conditions increases the likelihood of coding for other conditions such that the comorbidity estimates would approach those identified in a nationally representative health survey. STUDY DESIGN: Individuals with conditions predictive of adverse cardiovascular outcomes HTN, diabetes, stroke, CHD, and CHF were identified by presence of 2 claims with a corresponding ICD-9-CM or 1 claim with a relevant CPT-4 code. In these populations, persons with osteoarthritis (OA) and rheumatoid arthritis (RA) were identified using ICD-9-CM codes. Using a national survey, comorbidity rates were determined utilizing both self-reported health status and where applicable HTN & diabetes results from medical examination. Data were weighted appropriately. Point estimates of prevalence and comorbidity rates were compared between the two data sources. POPULATION STUDIED: 1999 Medstat MarketScan: beneficiaries and their dependents continuously enrolled in employer-sponsored health plans using both the Commercial Claims and Medicare Supplemental databases; coprevalence for RA was determined in persons > 20 y/o (n=2,491,467) and OA in those> 40 y/o (n=1,748,086). The Third National Health and Nutrition Examination Survey (NHANES III) weighted household questionnaire, medical and laboratory examination data were utilized to determine the comorbidity rates of these conditions within the same age groups for the US population. PRINCIPAL FINDINGS: The prevalence of conditions associated with adverse CV outcomes was similar in Medstat and NHANES. In individuals > 20 y/o, the respective rates are HTN (17.5%, 18.6%), diabetes (7.3%, 5.8%), CHD/CHF/stroke (7.0%, 9.4%). In individuals > 40 y/o, the respective rates are HTN (22.2%, 30.7%), diabetes (10.8%, 8.1%), CHD/CHF/stroke (10.8%, 10.6%). For individuals > 40 y/o, the prevalence of these conditions in OA patients was similar in Medstat and NHANES: HTN (23.7%, 30.1%), diabetes (8.1%, 7.9%), CHD/CHF/stroke (13.3%, 15.8%). For individuals > 20 y/o, the prevalence of these conditions in RA patients differed between the two datasets: HTN (16.1%, 36.1%), diabetes (6.8%, 9.1%), CHD/CHF/stroke (10.0%, 28.1%). CONCLUSIONS: For OA, because comorbidity with conditions associated with adverse CV outcomes was similar to NHANES III, assessment of healthcare utilization from the Medstat MarketScan claims database may be generalizable to the US population; the modestly higher diabetes comorbidity observed in Medstat may reflect the increasing prevalence of the disease and the time-frames covered in the two sources (Medstat 1999, NHANES III 1988-94); the lower comorbidity of HTN may reflect undercoding of HTN in claims and the inclusion of undiagnosed patients in NHANES. Subsequent analysis of only the diagnosed population in NHANES showed more similar comorbidity rates compared to Medstat. For RA, comorbidity was not comparable to that found in NHANES III. Thus estimates of healthcare resource use for RA from Medstat are not generalizable to the US population. In part this may reflect the debilitating nature of RA- more severe cases may leave the workforce. These analyses reinforce the importance of understanding both the reliability of the coding for the condition of interest and potential selection bias inherent in claims databases when evaluating the generalizability of results. IMPLICATIONS FOR POLICY, DELIVERY OR PRACTICE: National prevalence and health care utilization projections PRIMARY FUNDING SOURCE: Merck & Co. Inc

Publication Types:

• Meeting Abstracts

Keywords:

• Arthritis, Rheumatoid
• Comorbidity
• Databases as Topic
• Delivery of Health Care
• Diabetes Mellitus
• Diabetes Mellitus, Type 2
• Ethnic Groups
• Federal Government
• Health Surveys
• Humans
• Hypertension
• Nutrition Surveys
• Osteoarthritis
• Population Groups
• Prevalence
• Questionnaires
• Stroke
• epidemiology
• utilization
• hsrmtgs

Other ID:

• GWHSR0002354

UI: 102274030

From Meeting Abstracts