The current study attempted to ascertain whether the use of data from either the Veteran's Administration (VA) or Medicare systems could accurately reflect the diagnoses and illnesses of patients who use both systems for health care. Relative risk scores (RRS) were calculated for veterans using VA and Medicare systems but not a Medicare HMO during fiscal years 2000 to 2002.
- Patients who used both VA and Medicare systems had more diagnoses noted in the Medicare system than the VA.
- The average number of common diagnoses for patients who used both systems was two.
- Using just Medicare data yielded 80 percent of patients' illness burden. Comparatively, the use of only VA data accounted for one-third of illness burden.
- When using only Medicare data, the RRS score was 2.4 times the RRS score for using just VA data. During the 2000-2002 fiscal years, patients who used both systems had increasingly more diagnoses noted in the Medicare system.
- Illness burden as represented by the RRS score will be underestimated by 20-66 percent using data from either the Medicare or VA system alone.