The role of evidence in population based health outcome assessments by health authorities

Article type
Authors
McColl A, Roderick P, Gabbay J
Abstract
Introduction/Objective: To identify the use and perceived strengths and weaknesses of population based health outcome indicators in health outcome assessments and the implications for their further development. Population based health outcome indicators using annual routine mortality and utilisation data are produced for health authorities in England. Findings relevant to the role of evidence will be presented.

Methods: 1) A structured telephone interview with representatives of 91 of the 100 English health authorities - 61 of these were Directors of Public Health; 2) Editing 19 case studies from these health authorities which were explicitly linked with use of population health outcome assessments or indicators. Many of these focused on common conditions for which there was high quality evidence on interventions, and where if implemented would result in a beneficial effect on important health outcomes.

Results: 1) The main perceived constraints in the use of population based health outcome indicators were data validity and timeliness and the attributability of these health outcomes to the quality of health care. 23 interviewees (25%) advocated the increased use of process indicators as proxies for health outcome. Evidence based drug therapies were being used as process indicators to monitor the delivery of health care. 21 interviewees were using the rates of use of thrombolytic agents after myocardial infarction, five the administration of steroids for preterm deliveries, and seven prescription data (e.g. the ratio of inhaled steroids to inhaled brochodilators in the treatment of asthma). 2) The cases studies highlighted the difficulties in using evidence on interventions. For some relatively well researched areas such as the use of steroids in preterm labour, the literature did not provide specific or precise answers. Despite no clear evidence for interventions such as diabetic day centres clinicians and patients regarded them an essential part of a service. Despite the evidence on thrombo-embolic prophylaxis there was resistance from some orthopaedic surgeons to routinely use this for hip fracture patients. There were examples of where evidence had been used successfully in districts by improving the uptake of aspirin in people at high risk of heart disease and stroke, and in the purchasing of carotid endarterectomy. Special data collection was usually required to determine how well practice reflected evidence.

Discussion: To be more useful at a district level, population based health outcome assessments and indicators have to be linked to evidence based process indicators, be linked to contracting and audit and have to be meaningful to clinicians. Standardised definitions in such indicators are required to encourage comparability. Evidence should be applied at a population level and not just to those receiving care within a provider unit. Population based health outcome assessments can highlight inappropriate health service delivery or poor access.