In this study in an urban practice, the presence of a neutral observer at follow-up visits enhanced the extent to which practitioners recognized problems which patients had in a previous visit. This improvement was limited to those problems which initially had been mentioned by patients as requiring follow-up. Follow-up of problems initially mentioned by practitioners as needing follow-up was not improved by the observer unless the problem was also mentioned by the patient. Investigators whose information about practitioner-patient interaction depends upon the presence of an observer should be aware of this and possibly other effects. Although routine involvement of a neutral observer in patient-practitioner interactions is probably undesirable, selected deployment of observers or similar alternatives may be useful in situations where practitioner-patient communication is inadequate.