Table 1.

Think Tank Topics

YearTopicMethodological ObstacleThink Tank Outcome
1994, 1995Nature of caringHow to measure and foster caring in physiciansAfter a pilot study using focus groups to establish the domain of caring, the study would use a continuous quality improvement approach to enhance caring behaviors in the clinical encounter and assess the impact on patient outcomes
1996Assessing complexity in practice patternsHow to measure complexity in clinical practiceThe use of 3 data sources to triangulate results and identify chaotic patterns via interstaff relations, congruence between activities and stated practice philosophy, and the impact of events on system stability
1997Epidemiology of bioethicsDefinition of what practice interaction needs informed consent and how to incorporate patients’ viewsAn initial grounded theory study would be the basis for a subsequent quantitative survey
1998Effectiveness of integrated complementary and alternative medicineHow to compare an integrated approach to traditional careUse a multilevel randomized case study design in which a matched cohort of patients with chronic back pain is assigned to either the traditional care of a spine clinic vs the care at the integrated care clinic after in- depth interviews about the meaning of pain
1999Diabetes and continuity of careCurrent concepts and methods for defining and easuring continuity were inadequateMultiple factors identified that contribute to the process of continuity of care at the level of the family, the individual, and the community. Developed a research plan for a cohort study to determine the temporal relationships between continuity, readiness to change self-care behaviors, and events that occur in the life of the patient and the family
2000Factors in smoking cessation in adolescentsMixed methods study with wealth of data but uncertain analysis planRelying heavily on visual time series plots of the data, analyses were developed that consisted of reflexively alternating between the qualitative and quantitative data. Visual plots would identify clusters of subjects that would then be subjected to further analyses, looking for common themes. The plan would call for identifying clusters based on one type of data (eg, quantitative) and then describing each cluster based on the alternative data (eg, qualitative)
2001Development of a typology for health habit adviceUncertain approach for developing a typology of health habit adviceThe approach derived in the think tank was to use the data from the Exemplar Study13 to develop an initial typology that would be reviewed and altered by the Exemplar physicians in the study itself. A panel of national experts in prevention would then review this typology; their input would be used with a review of the research and theoretical literature to produce a revised typology. This revision would be reviewed again by the expert panel. The resultant typological framework would then be tested using existing databases. If this typology appears robust, it would be used in clinical trials as a final validation
2002Interprofessional collaboration of primary care providers in managing patients with hyperlipidemiaCurrent measures of collaboration—self-reported style of relationships, counts of structural characteristics that promote collaboration, and measures of provider satisfaction with joint decision making—are poor surrogatesBefore developing an instrument to measure collaboration on a continuum that could be used to study the relationship between collaboration and outcomes, a pilot study was needed to identify typologies of collaboration. The research question for the pilot study would be, How do primary care clinicians and staff work together to provide patient care? This multimethod study would use a comparative case study design with 3 to 5 primary care practices. Data collection would include practice genograms, observation of practice behavior, in-depth interviews with clinicians and staff, and focus groups of staff across practices
2003Integration of visual and audio data on the clinical encounter to assess impact in depression and smoking cessationIdentifying a method by which visual and auditory data from the clinical encounter could be integratedPilot work was necessary to identify specifics about the doctor-patient interaction that were associated with important outcomes. It was decided to focus on development of an R03 in the area of smoking cessation; it was deemed to be “cleaner” than addressing management of depression. Baseline assessment will establish current smoking behavior and stage of change. The clinical encounter will use digital recording technology and multiple cameras to captive interactions. Postencounter assessment will address patient characteristics, postencounter stage of change, and patient perceptions of the encounter. Follow-up assessment will assess stage of change. All current smokers will be recorded until 15 patients report a baseline-followup improvement in stage of change. From the remaining subjects, a matched control group of 15 subjects matched on patient characteristics will be selected. The interactions of these 30 subjects will then be analyzed using conversational analysis