Article Abstract:
Examining the controversy around hospital mortality statistics generated by Florence Nightingale and William Farr in 1863 may prove instructive in examining hospital mortality rates being used as measures of hospital quality today. Nightingale and Farr calculated mortality rates based on deaths over the year divided by bed occupants on a single day. This resulted in rates exceeding 90% for some hospitals. As today, statistics served political purposes. The two wanted to prod hospitals to improve care whereas hospitals, mostly charitable institutions, wanted to persuade benefactors that they offered quality care. Also as today, the two recognized that not admitting patients with certain serious illnesses and discharging dying patients or transferring them to another hospital could skew the accuracy of mortality rates. Nightingale's and Farr's statistical technique was criticized in the medical literature of the time. Today, statistical measures of mortality by insurers and others measuring hospital quality invite similar criticism.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
The type of assessment tool used to predict patient survival may influence the usefulness of hospital death rate analyses. Researchers compared the consistency and the accuracy of four patient survival assessment tools, including Medisgroups scores, physiology scores, Disease Staging, and All Patient Refined Diagnosis Related Groups (APR-DRGs), on 11,880 patients admitted for heart attack. They also evaluated the correlation between each assessment tool and symptoms thought to predict patient survival. Medisgroup and physiology scores correlated better with high death risk symptoms than Disease Staging and APR-DRGs. However, Disease Staging and APR-DRGs tended to predict death better than the Medisgroup and physiology scores. There was a strong correlation between each high-risk symptom and patient death. Two pairs of assessment tools gave similar survival predictions, Medisgroup and physiology scores, and Disease Staging and APR-DRGs.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
Administrative data sets collected in health care settings can provide a limited framework for studying treatment variation, cost-effectiveness, and health care quality. These data sets frequently include anonymous demographic and procedure profiles as entered by third party payers, government agencies, and health care providers. Data set analyses can point to general areas for improvement. Researchers can readily and inexpensively access volumes of data set information via computer networks.
User Contributions:
Comment about this article or add new information about this topic: