Distributing questionnaires about smoking to patients: impact on general practitioners' recording of smoking advice

BACKGROUND: Little is known about the impact of questionnaire-based data collection methods on the consulting behaviour of general practitioners (family physicians) who participate in research. Here data collected during a research project which involved questionnaires on smoking being distributed to patients before and after appointments with general practitioners (GPs) is analyzed to investigate the impact of this data collection method on doctors' documenting of smoking advice in medical records. METHODS: Researchers distributed questionnaires on smoking behaviour to 6775 patients who attended consultations during surgery sessions with 32 GPs based in Leicestershire, UK. We obtained the medical records for patients who had attended these surgery sessions and also for a comparator group, during which no researcher had been present. We compared the documenting of advice against smoking in patient's medical records for consultations within GPs' surgery sessions where questionnaires had been distributed with those which occurred when no questionnaires had been given out. RESULTS: We obtained records for 77.9% (5276/6775) of all adult patients who attended GPs' surgery sessions, with 51.9% (2739) being from sessions during which researchers distributed questionnaires. Discussion of smoking was recorded in 8.0% (220/2739) of medical records when questionnaires were distributed versus 4.6% (116/2537) where these were not. After controlling for relevant potential confounders including patients' age, gender, the odds ratio for recording of information in the presence of questionnaire distribution (versus none) was 1.78 (95% CI, 1.36 to 2.34). CONCLUSION: Distributing questionnaires about smoking to patients before and after they consult with doctors significantly increases GPs' recording of discussions about smoking medical records. This has implications for the design of some types of research into addictive behaviours and further research into how data collection methods may affect patients' and doctors' behaviours is warranted.