Intended for healthcare professionals

Endgames Statistical Question

Understanding the ecological fallacy

BMJ 2015; 351 doi: (Published 21 September 2015) Cite this as: BMJ 2015;351:h4773

This article has a correction. Please see:

  1. Philip Sedgwick, reader in medical statistics and medical education1
  1. 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
  1. Correspondence to: P Sedgwick p.sedgwick{at}

Researchers investigated whether an association existed between antibiotic prescribing by general practices and antibiotic resistance. A cross sectional study design was used. The setting was 28 general practices in Scotland. The participants were patients registered with a general practitioner between January 1995 and December 1996 who provided a urine sample for analysis during this period. In total, 8833 patients were included in the sample. For each patient in each of the practices the number of dispensed prescriptions for trimethoprim was obtained from databases of health records.1

Data were aggregated at the general practice level. For each general practice, antibiotic prescribing was measured as the number of dispensed trimethoprim prescriptions per 100 practice patients. Antibiotic resistance was defined as the proportion of patients in the practice with a urine sample that contained trimethoprim resistant Gram negative bacteria. Rates of trimethoprim prescription varied from 67 to 357 prescriptions per 100 patients between practices. The prevalence of trimethoprim resistance in Gram negative bacteria isolated from urine specimens varied between 26% and 50%. A weak negative association was seen between trimethoprim prescription and antibiotic resistance across the 28 general practices (Spearman rank …

View Full Text

Log in

Log in through your institution


* For online subscription