PMC patient severity scale: derivation and validation - Patient Management Category

Health Services Research, August, 1994 by Wanda W. Young, Susan Kohler, Jeanne Kowalski

The difficulty arises when one tries to quantify these severity differences, especially across different diseases. An additional complexity arises when two or more clinical conditions occur simultaneously. That is, does the additional disease and/or complication make the patient more severe? To what extent is the probability of continued morbidity or imminent death increased by the presence of more than one disease or clinical manifestation of the disease(s)?

In addition to these disease-specific issues are other aspects of a severity index that should be identified before it is selected for use. For example, is the severity index attempting to predict the probability of death, the potential for organ failure, the risk of permanent impairment, the total impact of the disease process on the patient's long-term survival probability, or some combination of morbidity and mortality? Does the index refer to disease severity or patient severity at a point in time? What is the methodology (i.e., statistical analyses versus psychometric methods) used to derive final severity scale values? Similarly, it is important to define the reference disease(s) or clinical condition(s), the applicable time interval, and the extent to which treatment and its interaction with the disease are included in the conceptualization of severity.

A severity index designed to standardize hospital lengths of stay and death rates should incorporate the patient's acute illnesses during the hospital stay as well as chronic or coexisting conditions that have the potential to influence the patient's overall probability of continued morbidity or death (Charlson et at. 1987; Dubois et al. 1987; Greenfield et at. 1988). Complications that are not preventable and/or are part of the disease process should also be included in the construction of such a relative severity score. In most cases, both the severity of a patient's illness and the types, quantity, and intensity of effective care rendered to the patient are factors that influence the outcomes of care, especially the probability of long-term survival.

This article reports on the development of the PMC Severity Scale, a seven-level ordinal scale that quantifies the severity of a patient's clinical condition(s) and/or the patient's clinical manifestation(s) of disease (not necessarily the disease per se) during the hospital stay. This clinical severity level is then adjusted upward for the effect of specific comorbid conditions and complications, yielding one overall patient severity level (PMC Severity Score) for that hospitalization.

DATA AND METHODS

The methodology used to develop the PMC Patient Severity Scale was based primarily on an empirical analysis of more than a half million patients discharged from acute care hospitals in Maryland during calendar year 1989. Patient discharges from the same Maryland hospitals during calendar years 1988 and 1990 were used to assess the reliability of the PMC Severity Scale over time. A six-month patient database from California (1990) was also used to determine the validity of the PMC Severity Scale across geographic regions.