Premium
Are rates of pathology test ordering higher in general practices co‐located with pathology collection centres?
Author(s) -
Studdert David M,
Britt Helena C,
Pan Ying,
Fahridin Salma,
Bayram Clare F,
Gurrin Lyle C
Publication year - 2010
Publication title -
medical journal of australia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.904
H-Index - 131
eISSN - 1326-5377
pISSN - 0025-729X
DOI - 10.5694/j.1326-5377.2010.tb03819.x
Subject(s) - odds ratio , logistic regression , confounding , test (biology) , poisson regression , medicine , statistics , demography , odds , pathology , mathematics , biology , paleontology , population , environmental health , sociology
Objective: To determine whether rates of pathology test ordering by general practitioners in general practices co‐located with pathology collection centres (PCCs) are higher than those of GPs in practices located apart from PCCs. Design, setting and participants: We identified all practices in the Melbourne and Sydney metropolitan areas that were co‐located with PCCs (same or immediately adjacent suite) and the date co‐location was established. This information was merged with the Bettering the Evaluation and Care of Health database to identify samples of GP–patient encounters in co‐located practices ( n = 31 700) and practices located apart from the nearest PCC ( n = 289 700) over the period 2000–2009. Using Poisson regression analysis and logistic regression analysis, we compared GP test‐ordering rates across the two types of practices, controlling for a range of potential confounders. Main outcome measures: Numbers of tests ordered per encounter; likelihood of ordering one or more tests per encounter. Results: In unadjusted analyses, GPs in co‐located practices ordered more pathology tests than GPs in practices located apart from PCCs (40.3 v 37.0 tests per 100 encounters, P = 0.01) and had a higher likelihood of ordering one or more tests (16.8% v 15.5% of encounters, P < 0.01). After adjusting for other predictors of test ordering, however, neither test‐ordering rate (rate ratio, 0.98; 95% CI, 0.93–1.05; P = 0.56) nor likelihood of ordering one or more tests per encounter (odds ratio, 1.01; 95% CI, 0.95–1.07; P = 0.79) differed significantly by co‐location status. Sub‐analyses within specific test groups and types showed few systematic differences. Conclusions: Pathology test‐ordering rates are not higher in practices co‐located with PCCs. To the extent inappropriate commercial influences and relationships exist in the pathology sector, GPs’ test‐ordering behaviour may be unaffected.