| Summary: | <p>Abstract</p> <p>Background</p> <p>In-hospital case-fatality rates in patients, admitted for acute myocardial infarction (AMI-CFRs), are internationally used as a quality indicator. Attempting to encourage the hospitals to assume responsibility, the Belgian Ministry of Health decided to stimulate initiatives of quality improvement by means of a limited set of indicators, among which AMI-CFR, to be routinely analyzed. In this study we aimed, by determining the existence of inter-hospital differences in AMI-CFR, (1) to evaluate to which extent Belgian discharge records allow the assessment of quality of care in the field of AMI, and (2) to identify starting points for quality improvement.</p> <p>Methods</p> <p>Hospital discharge records from all the Belgian short-term general hospitals in the period 2002-2005. The study population (N = 46,287) included patients aged 18 years and older, hospitalized for AMI. No unique patient identifier being present, we tried to track transferred patients. We assessed data quality through a comparison of MCD with data from two registers for acute coronary events and through transfer and sensitivity analyses. We compared AMI-CFRs across hospitals, using multivariable logistic regression models. In the main model hospitals, Charlson's co-morbidity index, age, gender and shock constituted the covariates. We carried out two types of analyses: a first one wherein transferred-out cases were excluded, to avoid double counting of patients when computing rates, and a second one with exclusion of all transferred cases, to allow the study of patients admitted into, treated in and discharged from the same hospital.</p> <p>Results</p> <p>We identified problems regarding both the CFR's numerator and denominator.</p> <p>Sensitivity analyses revealed differential coding and/or case management practices. In the model with exclusion of transfer-out cases, the main determinants of AMI-CFR were cardiogenic shock (OR<sub>adj </sub>23.0; 95% CI [20.9;25.2]), and five-year age groups OR<sub>adj </sub>1.23; 95% CI [1.11;1.36]). Sizable inter-hospital and inter-type of hospital differences {(OR<sub>comunity vs tertiary hospitals</sub>1.36; 95% CI [1.34;1.39]) and (OR<sub>intermediary vs tertiary hospitals</sub>1.36; 95% CI [1.34;1.39])}, and nonconformities to guidelines for treatment were observed.</p> <p>Conclusions</p> <p>Despite established data quality shortcomings, the magnitude of the observed differences and the nonconformities constitute leads to quality improvement. However, to measure progress, ways to improve and routinely monitor data quality should be developed.</p>
|
|---|
