Qual Saf Health Care 11:137-143 doi:10.1136/qhc.11.2.137
  • Original Article

Scope and nature of prescribing decisions made by general practitioners

  1. P Denig1,
  2. C L M Witteman2,
  3. H W Schouten3
  1. 1Department of Clinical Pharmacology, Faculty of Medical Sciences, GUIDE/NCH, University of Groningen, 9713 AV Groningen, the Netherlands
  2. 2Institute of Information and Computing Sciences, Faculty of Mathematics and Computer Science, Utrecht University, the Netherlands
  3. 3Psychonomics Department, Faculty of Social Sciences, Utrecht University
  1. Correspondence to:
 Dr P Denig, Department of Clinical Pharmacology, Faculty of Medical Sciences, GUIDE/NCH, University of Groningen, Ant Deusinglaan 1, 9713 AV Groningen, the Netherlands;
  • Accepted 12 October 2001


Background: This study describes cognitive processes of doctors who are deciding on the treatment for a patient. This helps to uncover how prescribing decisions could benefit from (computerised) support.

Methods: While thinking aloud, 61 general practitioners made prescribing decisions for five patients with urinary tract infections or stomach complaints. The resulting 305 transcripts were analysed to determine the scope and nature of the decision processes. Differences in the process were related to case or doctor characteristics, and to differences in the quality of prescribing behaviour.

Results: The decision processes were not extensive, particularly for patients with a urinary tract infection. The doctors did not actively consider all possible relevant information. Considerations referring to core aspects of the treatment were made in 159 cases (52%) and to contextual aspects in 111 cases (36%). Habitual behaviour, defined as making a treatment decision without any specific contemplation, was observed in 118 cases (40%) and resulted in prescribing first choice as well as second choice drugs. For stomach complaints, second choice drugs were often prescribed after considering other treatments or in view of specific circumstances. Experience of the doctor was not related to the type of decision process.

Conclusions: The processes observed deviate from the decision theoretic norm of thoroughly evaluating all possible options, but these deviations do not always result in suboptimal prescribing. Decision support is useful for bringing pertinent information and first choice treatments to the prescriber's attention. In particular, information about relevant contraindications, interactions, and costs could improve the quality of prescribing.