Consultation, waiting, prescribing and referral patterns: some methodological considerations.

This paper addresses the question of how large a sample is necessary before valid conclusions can be drawn with regard to average waiting and consultation times and prescribing and referral patterns for individual doctors. The statistical techniques of plotting cumulative means for these measures in relation to sample size, along with a formula using the maximum available sample size, and the mean/standard error of the measure at this sample size was used to determine 'stabilization values' (minimum sample sizes at which statistically valid conclusions can be drawn with a predetermined degree of certainty) regarding these measures. Using these techniques, stabilization values of 120 for consultation time, 200 for mean waiting time, 170 for prescribing pattern and 490 for referral pattern were determined using data collected by 136 experienced general practitioners during 40,363 consultations in the context of UK general practice. Confident conclusions cannot be drawn from studies examining these measures unless at least the sample sizes documented in this paper have been used.