Lost in Transition: Discrepancies in How Physicians Perceive the Actionability of the Results of Tests Pending at Discharge

Background:

Physician awareness of tests pending at discharge (TPADs) is poor, a potential patient safety concern. Inpatient physicians who order these tests and primary care physicians (PCPs) who follow up the results may have differing opinions about whether a specific TPAD result is actionable. Lack of agreement may affect subsequent communication and follow‐up actions. We sought to determine how often inpatient physicians and PCPs agree on the actionability of TPAD results.

Methods:

We developed an automated system that identifies patients with TPADs. During a 6‐month period, this system randomly identified patients with TPADs discharged from inpatient medicine and cardiology services at Brigham and Women's Hospital. For each patient with at least 1 TPAD, the responsible inpatient physician and PCP were independently surveyed regarding the actionability of that patient's TPAD results. We identified all cases in which either physician thought at least 1 TPAD result was actionable and then asked them to identify which TPAD results were actionable.

Results:

The system identified 441 patients with TPADs discharged during the study period. We received completed survey responses from both the responsible inpatient physician and PCP on 98 patients. In 59 patients, both physicians thought none of the TPAD results were actionable. In 12 patients, both thought at least 1 TPAD result was actionable, and they identified the same actionable TPAD result in all 12. Overall, the responsible inpatient physician and PCP agreed on actionability in 72.5% (71 of 98; kappa, 0.29; 95% CI, 0.09–0.50). There were 9 patients for whom the inpatient physician alone thought at least 1 TPAD result was actionable; of these, subsequent inpatient physician‐initiated communication occurred in 77.8% (7 of 9). There were 18 patients for whom the PCP alone thought at least 1 TPAD result was actionable; of these, subsequent PCP‐initiated communication occurred in 77.8% (14 of 18). In instances of disagreement, inpatient physicians frequently reported microbiology results (e.g., bacterial cultures, viral serologies) as actionable, and PCPs reported imaging and laboratory tests for chronic conditions (e.g., vitamin D levels) as actionable.

Conclusions:

The study was limited by small sample size and low response rate. Overall, we found fair agreement between responsible inpatient physicians and PCPs with regard to actionability of patients' TPAD results. Responsible inpatient physicians often considered certain TPAD results actionable when PCPs did not and vice versa in a large percentage of patients. This disagreement may be a consequence of differing perspectives of providers caring for the patient. In these instances, reliable communication between the inpatient team and PCP is crucial but may be suboptimal. Additional testing, unnecessary procedures, and delays in diagnosis and treatment could be minimized with more seamless and reliable mechanisms of communication among patients and providers.