might not have a high score on a medical severity-of-illness algorithm but still could require a large amount of nursing care.

Nonetheless, Medicare data and a limited amount of state-specific hospital data support the observation that, beginning in the mid-1980s following implementation of the Medicare prospective payment system (PPS) for hospitals and continuing into the late 1990s, patients admitted to hospitals were increasingly more acutely ill. Data on all Medicare hospital admissions for 1985–1997 show an annual increase in the complexity of cases treated in acute care hospitals as measured by the Medicare case mix index (CMI),6 while a review of patient data for all payors and all acute care general hospitals in Pennsylvania during 1994–1997 revealed that the severity of illness of patients admitted to those hospitals increased in the aggregate by 4.5 percent over the 4-year period (Unruh, 2002b). The annual increases were highest in the early years just after implementation of the PPS and slowed fairly steadily until 1998, when a decline in severity as measured by the CMI was observed. This decline continued into 1999, the last year for which these data are available. It was determined that the CMI decrease of 0.5 percent in 1998 likely reflected changes in coding practices; however, this was not the case for the 0.4 percent CMI decline in 1999 (Medicare Payment Advisory Commission, 2001).

This increase in the severity of illness of hospital patients has had a ripple effect throughout all health care settings. Evidence indicates that patients receiving care in long-term care facilities, in their homes, and in other community-based settings are more ill and debilitated and/or require more technologically complex medical care than in the past. In nursing homes, the proportion of patients who are more frail (i.e., need assistance with three or more ADLs, such as bathing, dressing, eating, and toileting) and therefore need more skilled and/or specialized care increased from 72 percent in 1987 to 83 percent in 1996. As a consequence, over the last few years, nursing homes have developed specialized units to care for patients who need more extensive care, such as those with dementia, rehabilitation needs, ventilator dependency, or brain injury. Approximately 12.6 percent of all nursing homes in 1996 had units devoted to the specialized care of


A hospital’s CMI represents the average diagnosis-related group (DRG) relative weight for that hospital. It is calculated by summing the DRG weights for all Medicare discharges during a fiscal year, and dividing by the number of discharges (CMS, 2003). The Medicare CMI is calculated annually based on charges submitted to the Medicare program for all hospital patients. While the CMI is therefore a direct measure of costliness, it is often used as a surrogate indicator of severity of illness because more acutely ill patients typically are higher-cost patients. However, this is not always the case, especially because technology is often costly, but may not always be used by the most acutely ill. The CMI is therefore an imperfect indicator of severity of illness and patients’ need for nursing care.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement