Doctor characteristics and prescribing antibiotics for urinary tract infections: the experience of an Asian country

Abstract
Rationale, aims and objectives This study aims to explore the relationship between
doctor characteristics and prescribing behaviour for patients with urinary tract infections
(UTIs) using a 2-year population-based data set in Taiwan.
Methods This study used data from the Taiwan National Health Insurance Research
Database. Our study sample consists of first-time ambulatory care visits for treatment of
UTIs among female patients between 2005 and 2006 (n = 45 934). The primary outcome
studied was ‘whether a broad-spectrum antibiotic was prescribed’, and the key independent
variables were ‘doctor characteristics’. Doctor characteristics included gender, age (<41,
41–50, >50), specialty, type (hospital-based vs. office-based) and practice location. Multivariate
logistic regression analysis using generalized estimated equations was performed
to assess the adjusted odds ratio of the doctors using broad-spectrum antibiotics.
Results Among the sampled patients, 13.5% were prescribed broad-spectrum antibiotics
at their first visit for treatment of UTIs. The adjusted odds of prescribing second-line
antibiotics for doctors aged between 41–50 years and >50 years were 0.80 (P < 0.001) and
0.90 (P = 0.007) times, respectively, that of doctors aged <41 years. Doctors specializing in
family medicine were 1.10 (P = 0.006) times more likely than doctors specializing in
obstetrics and gynaecology to prescribe broad-spectrum, and office-based doctors were
1.41 (P < 0.001) times more likely than hospital-based doctors to do so.
Conclusions We conclude that there is variation in doctor prescribing behaviour of antimicrobial
therapy for UTIs, after adjusting for possible confounding factors. Continuing
medical education and intervention should be designed for doctor groups with undesirable
performance in prescribing antibiotics.