Estimation of numbers of British medical graduates working in the National Health Service: comparison of capture-recapture analysis and exhaustive tracing
Lambert T., Goldacre M., Davidson J., Parkhouse J.
OBJECTIVE: To compare the results of capture-recapture estimates and exhaustive follow-up in identifying the percentage of UK medical graduates who subsequently worked in the British National Health Service (NHS). METHODS: Postal questionnaire survey of all UK-trained doctors who graduated in 1974, with up to three reminders sent to non-respondents. Further exhaustive tracing of non-respondents to obtain whole population estimates of numbers in the NHS. Capture-recapture estimation of numbers working in the NHS, using survey data and Department of Health employment records, for comparison with results from exhaustive tracing. RESULTS: A total of 85.3% (1460/1712) of respondents were working in the NHS, but further exhaustive tracing of non-respondents revealed that only 75.9% (1758/2317) to 79.2% (1835/2317) of the whole cohort were in the NHS. By comparison, capture-recapture analysis estimated that 76.9% (1781/2317) of the cohort were in the NHS (95% CI: 75.9% to 78.0%). CONCLUSIONS: Survey data alone overestimated the percentage of the whole cohort in the NHS. Capture-recapture estimates were very similar to those obtained by exhaustive tracing but were obtained more quickly and for much less clerical effort.