A method to quantify and compare clinicians' assessments of patient understanding during counseling of standardized patients

Objective: To introduce a method for quantifying clinicians' use of assessment of understanding (AU) questions, and to examine medicine residents' AU usage during counseling of standardized patients about prostate or breast cancer screening. Methods: Explicit-criteria abstraction was done on 86 transcripts, using a data dictionary for 4 AU types. We also developed a procedure for estimating the "load" of informational content for which the clinician has not yet assessed understanding. Results: Duplicate abstraction revealed reliability κ = 0.96. Definite criteria for at least one AU were found in 68/86 transcripts (79%). Of these, 2 transcripts contained a request for a teach-back ("what is your understanding of this?"), 2 contained an open-ended AU, 46 (54%) contained only a close-ended AU, and 18 (21%) only contained an "OK?" question. The load calculation identified long stretches of conversation without an AU. Conclusion: Many residents' transcripts lacked AUs, and included AUs were often ineffectively phrased or inefficiently timed. Many patients may not understand clinicians, and many clinicians may be unaware of patients' confusion. Practice implications: Effective AU usage is important enough to be encouraged by training programs and targeted by population-scale quality improvement programs. This quantitative method should be useful in population-scale measurement of AU usage.