Previous research cannot account for the discrepancy between registered nurse (RN) reports of understaffing and studies showing slight improvement. One reason may be that "adjusted patient days of care" (APDC) underestimates patient load. Using data from all Pennsylvania acute care general hospitals for the years 1994 through 1997, we found that APDC is underestimated by two hours. After adjusting APDC, we examined the difference in nurse staffing over the period 1991-2000 before and after the adjustment. We found a significant difference between unadjusted and adjusted measures. However, when applied to the changes in nurse staffing between 1991 and 2000, the difference was not enough to account for the discrepancy between reports and data. Other measurement and conceptual problems may exist in terms of patients' increasing acuity levels, patients' declining lengths of stay and the associated greater proportion of nurse time devoted to admission and discharge, and lack of recent data in some empirical studies.