Premium
Application of the STOPP/START criteria to a medical record database
Author(s) -
Nauta Katinka J.,
Groenhof Feikje,
Schuling Jan,
Hugtenburg Jacqueline G.,
Hout Hein P.J.,
HaaijerRuskamp Flora M.,
Denig Petra
Publication year - 2017
Publication title -
pharmacoepidemiology and drug safety
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.023
H-Index - 96
eISSN - 1099-1557
pISSN - 1053-8569
DOI - 10.1002/pds.4283
Subject(s) - medicine , pharmacoepidemiology , medical record , database , pediatrics , pharmacology , medical prescription , computer science
Purpose The STOPP/START criteria are increasingly used to assess prescribing quality in elderly patients at practice level. Our aim was to test computerized algorithms for applying these criteria to a medical record database. Methods STOPP/START criteria‐based computerized algorithms were defined using Anatomical‐Therapeutic‐Chemical (ATC) codes for medication and International Classification of Primary Care (ICPC) codes for diagnoses. The algorithms were applied to a Dutch primary care database, including patients aged ≥65 years using ≥5 chronic drugs. We tested for associations with patient characteristics that have previously shown a relationship with the original STOPP/START criteria, using multivariate logistic regression models. Results Included were 1187 patients with a median age of 75 years. In total, 39 of the 62 STOPP and 18 of the 26 START criteria could be converted to a computerized algorithm. The main reasons for inapplicability were lack of information on the severity of a condition and insufficient covering of ICPC‐codes. We confirmed a positive association between the occurrence of both the STOPP and the START criteria and the number of chronic drugs (adjusted OR ranging from 1.37, 95% CI 1.04‐1.82 to 3.19, 95% CI 2.33‐4.36) as well as the patient's age (adjusted OR for STOPP 1.30, 95% CI 1.01‐1.67; for START 1.73, 95% CI 1.35‐2.21), and also between female gender and the occurrence of STOPP criteria (adjusted OR 1.41, 95% CI 1.09‐1.82). Conclusion Sixty‐five percent of the STOPP/START criteria could be applied with computerized algorithms to a medical record database with ATC‐coded medication and ICPC‐coded diagnoses.