PHYSICIAN CHARACTERISTICS THAT PROMOTE OR IMPEDE CONTEXTUALIZED CARE PLANNING

Purpose: Adapting care plans to patients' individual needs--contextualizing care--is associated with better health outcomes and lower costs. The purpose of this study was to identify physician behaviors in the medical encounter associated with the contextaulization of care.

Method(s): Transcripts of audio files of 58 encounters with unannounced standardized patients by 15 physicians were purposively sampled to include five physicians who never probed contextual issues, five who probed consistently but never incorporated the context into their care plans, and five who probed consistently and incorporated the context into their care plans at least once. A constant comparative analysis was conducted to identify themes that led to contextual probing and planning. Three researchers worked with one transcript at a time, meeting regularly to discuss coding. After reaching saturation, the researchers independently coded to the encounters from audio and wrote memos on each physician. The association between the relative frequency of coded themes by physician (over encounters) and physician performance was also statistically confirmed within the sample using multinomial logistic regression.

Result(s): The better-performing physicians displayed a flexible approach to the interview. Their visits included a more relaxed pace and open-ended questions. In contrast, physicians who consistently controlled the agenda of the encounter set the pace without accommodating patient input, and were more prone to premature closure. Controlling behaviors were significantly associated with less probing and planning (p<.001); flexible behaviors were significantly associated with improved planning (p=.007).

Physicians using a systems- or checklist-based approach to history taking were least likely to probe contextual factors (p=.007). This approach was also associated with controlling the encounter. On the other hand, more flexible physicians used a hypothesis-driven approach to the history characterized by a natural flow of conversation rather than following a list of predetermined questions.

Finally, physicians struggling with the electronic medical record during the visit were less likely to probe for context (p<.001) and to incorporate critical context into the plan (p<.001).

Conclusion(s): Performance at contextualization of care is associated with physician behaviors that can be observed across multiple encounters. These behaviors are only detectable by listening in on the visit, emphasizes the need for direct observation of care as a source of performance measures.