Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Time is scarce and a pre-requisite for effective information exchange between the principal and agent in a physician’s office practice. While physicians possess medical knowledge and skills, it is the patients who often know more about their illnesses and life circumstances. Highly regimented clinical schedules that allow only a limited amount of time with patients pose challenges to physicians’ ability to effectively address salient patient concerns because cogent thinking and clear communication cannot be conducted like a race being run. This is particularly true with physicians treating older patients who commonly have multiple chronic and complicated problems which can not be solved quickly. Physicians must decide at what time the visit needs to end, however, either because no more new information is forthcoming or can be admitted due to time constraint. Using an idea of viewing office visit as containing a sequence of 'topics,' we present a model of a physician deciding about concluding a visit or allowing a new topic to be discussed. The physician admits a new topic depending on the expected benefits and costs of doing so. The efficient allocation of the physician’s time can be determined by a threshold value, λ, such that any topic with value >λ is dealt with and any topic with value < λ is not. We hypothesize that physicians have a 'target' amount of time they desire to spend with each patient. We regard the physician as holding a rising shadow price of time around the target. Let τ be the time elapsed during a visit, the shadow price of time for a physician is λ(τ). Empirically, we estimate a model of whether a physician admits a new topic and test for the effect of τ on this behavior. Data come from 385 videotaped patient-physician interactions took place in an academic medical center, a private managed care group practice, and several solo practices. We conducted mixed level logit regression analyses to explore the probability of a topic being the last topic of the visit and its determinants. To examine the shadow price of time, the time when a topic was initiated (τ) was measured by 4 binary variables for topics introduced in less than 5 minutes, between 5 – 10 minutes, 10-15 minutes, and later than 15 minutes since the beginning of the visit. We also included the number of topics occurred prior to the current topic, and the natures (biomedical versus mental health and psychosocial issues) of the current and the topic immediately prior to it. Results suggest drastically increasing shadow price of time. Compared with topics introduced in the 1st 5 minutes, coefficients for later topics were 7.3, 27.8, and 145.4 (p<0.01 for all). Topics on depression, living will, and death were more likely to be the last topic than biomedical topics. Model fitness test results suggest a better fit when information on nature of prior topics was included. Policy and clinical implications will be discussed.