Patients who've been hospitalized are particularly vulnerable during discharge, when suboptimal communication is known to contribute to an inadequate understanding of postdischarge care plans. The result can be preventable harms, such as medication errors, adverse events, and costly readmissions. Because empirical data from direct observations of the hospital discharge process are limited, researchers conducted a direct-observation quality improvement study to examine patient-practitioner interactions on the day of discharge, particularly whether key communication domains were addressed.

A total of 33 English- or Spanish-speaking patients were selected from the "discharge before noon" list on the internal medicine service of two academic teaching hospitals. A trained observer sat at the bedside of a patient from 7 AM until time of discharge and recorded the content and duration of staff communication with the patient. The observer then interviewed the patient after discharge.

Of 29 patients discharged with new or changed medication, eight (28%) weren't told the name and type of the medication, and 17 (59%) weren't told the purpose of the change. Of 11 patients whose prior medication was discontinued, six (55%) weren't counseled on the reason for the discontinuation. Of 31 patients discharged with follow-up appointments, 15 (48%) weren't told the purpose of the appointment. Eighteen of the 33 patients (54%) didn't receive counseling about self-management of their condition, and 24 (73%) weren't given any information about the type of symptoms to expect after discharge or the expected course of their illness after leaving the hospital. Only one patient was asked to teach back his understanding of the treatment plan. Twenty-eight patients (85%) weren't asked about potential barriers to the care plan.

Nurses spent the most time with patients on the day of discharge (a median of 22.5 minutes at the bedside, while performing any task). The median time for attending physicians, interns, and senior residents was three minutes, four minutes, and one minute, respectively.

The study was limited by the small sample size and possible selection bias. It isn't known whether the findings would apply to unplanned or simpler discharges, or to patients who don't speak English or Spanish.