In health care system, examining patients outcome, such as mortality rate, is central to evaluating quality and selecting preferred provides, especially as costs become increasingly constrained. Meaningful comparisons of outcomes across providers, however, usually required adjustment for baseline patient severity. Because some facilities treat sicker patient better than others, hospital comparison need to control for patients' risk. Since the early 1980s', various severity measures have been developed specifically for comparing hospitals or large patient groups. Articles, usually by the developers of the measures, described individual severity measures, but few studies by independent investigators involving multiple severity measures have been reported.
Quality of care could be evaluated in the three dimensions of healthcare - structure, process, and outcomes. Due to the relative easiness in the assessment and collecting data, mortality rates are commonly used as an indicator of outcomes of care. However, in the evaluation of outcomes of care, mortality rates should be adjusted by patients' baseline severity. Risk factors, such as clinical conditions of each patient, that could influence the results of care must be addressed.
This article aims to calculate the severity-adjusted length of stay and mortality rates of ICU patients with acute myocardial infarction. In the study, APACH-II, a risk adjustment tool, was used and and its performace as a predictor was examined. And then follows a comparative analysis of outcomes of care across providers. That involves a retrospective survey of medical records of 270 patients who received medical care from 11 doctors in two university hospitals.
Mortality rates had correlation with severity points, and the hospitals showed differences in actual mortality rates and expected mortality rates. Mortality rates of female patients and patients who were above age of 75 was significantlty high. Severity points which reflect the difference of clinical conditions at the time of hospital admission didn't show significant differences. Altough the analysis was restricted to small numbers of patients for individual doctors, on the whole, there were no statistically significant differences across doctors, except that internists and surgeons showed some difference in mortality rates. APACH-II had useful model performance for predicting death rate, as it scored 0.92 c-statistics. But in examining length of stay, the tool showed rather poor performance, scoring 0.03 R2.
In summary, we found that, for patients with acute myocardial infarction, APACH-II was reliable in predicting severity-adjusted mortality rates, and was poor in explaining length of stay. APACH-II was orginally developed for ICU patients or patients who are in their early stages of hospitalization, and that's why we suppose the tool shows excellent performances for patients with acute myocardial infarction.
Keywords : Severity-adjusted mortality, Process of care, Length of stay, Acute myocardial infarction, Intensive care unit