William B Baine1, Sophia V Kazakova. 1. Center for Outcomes and Evidence, Agency for Healthcare Research and Quality, Department of Health and Human Services, 540 Gaither Road, Rockville, MD 20850-6649, USA. wbaine@ahrq.gov
Abstract
OBJECTIVE: A method to generate hypotheses about iatrogenic risk factors and complications from administrative data was developed and tested using hospitalization of the elderly for depression as a model. STUDY DESIGN AND SETTING: Hospital claims were selected for 30,998 elderly inpatients admitted for the first time for depression. Common principal diagnoses and procedures in hospitalizations within 90 days of the index depression admission were tallied. For each of these proximate clinical events, the ratio of how many happened before the index admission to how many occurred afterward was calculated. Ratios diverging markedly from unity were identified to generate hypotheses about possible risk factors associated with depression and complications associated with its management. RESULTS: Hospitalization for degenerative joint disease or back problems; abdominal pain or gastritis and duodenitis; coronary artery disease; or cerebrovascular disease was more common before an index depression admission than after it, as were coronary artery surgery, total knee replacement, and cholecystectomy. Admissions for fracture of the femoral neck--an established iatrogenic complication--were disproportionately likely after the index admission. So were admissions for aspiration pneumonia or acute respiratory failure. CONCLUSION: Proximate clinical event ratios provide a systematic approach to screening administrative data to identify candidates for further evaluation as possible iatrogenic risk factors or complications.
OBJECTIVE: A method to generate hypotheses about iatrogenic risk factors and complications from administrative data was developed and tested using hospitalization of the elderly for depression as a model. STUDY DESIGN AND SETTING: Hospital claims were selected for 30,998 elderly inpatients admitted for the first time for depression. Common principal diagnoses and procedures in hospitalizations within 90 days of the index depression admission were tallied. For each of these proximate clinical events, the ratio of how many happened before the index admission to how many occurred afterward was calculated. Ratios diverging markedly from unity were identified to generate hypotheses about possible risk factors associated with depression and complications associated with its management. RESULTS: Hospitalization for degenerative joint disease or back problems; abdominal pain or gastritis and duodenitis; coronary artery disease; or cerebrovascular disease was more common before an index depression admission than after it, as were coronary artery surgery, total knee replacement, and cholecystectomy. Admissions for fracture of the femoral neck--an established iatrogenic complication--were disproportionately likely after the index admission. So were admissions for aspiration pneumonia or acute respiratory failure. CONCLUSION: Proximate clinical event ratios provide a systematic approach to screening administrative data to identify candidates for further evaluation as possible iatrogenic risk factors or complications.
Authors: Sean Hennessy; Warren B Bilker; Charles E Leonard; Jesse Chittams; Cristin M Palumbo; Jason H Karlawish; Yu-Xiao Yang; Ebbing Lautenbach; William B Baine; Joshua P Metlay Journal: J Clin Epidemiol Date: 2007-05-10 Impact factor: 6.437