A Comparison of Patient and Provider Perspectives on an Electronic Health Record-Based Discharge Communication Tool: Survey Study

Abstract

Background: Hospital discharge for older adult patients carries risks. Effective patient-provider communication is crucial for postacute care. Technology-based communication tools are promising in improving patient experience and outcomes. However, there is limited evidence comparing patient and provider user experiences on a large-scale basis, hindering the exploration of true patient-provider shared understanding.

Objective: This study aimed to evaluate an electronic health record-based discharge communication tool by examining and comparing patient and provider perspectives.

Methods: This study comprised a cross-sectional self-administered staff survey and a pre-post cross-sectional patient survey. Physicians, nurses, and older adult patients aged 65 years and older discharged from 4 public hospitals were included. Patient-provider comparison items focused on 3 aspects of the design quality of the tool (information clarity, adequacy, and usefulness) and overall satisfaction with the tool. In addition, patients' experience of discharge information and their medication-taking behaviors before and after the program implementation were compared based on a validated local patient experience survey instrument. Providers' perceived usefulness of this tool to their work and implementation intentions were measured based on the technology acceptance model to enhance understanding of their experiences by conducting structural equation modeling analysis.

Results: A total of 1375 and 2353 valid responses were received from providers and patients, respectively. Patients' overall satisfaction with this communication tool is significantly higher than providers', and patients rated the information clarity and usefulness presented by this tool higher as well (P<.001). However, patients rated information adequacy significantly lower than providers (P<.001). Meanwhile, patients reported a significant improvement in their experience of discharge medication information, and fewer patients reported side effects encounters after the program implementation (126/1083, 11.6% vs 111/1235, 9%; P=.04). However, providers showed inconsistent implementation fidelity. Providers' perceived quality of the tool design (β coefficient=0.24, 95% CI 0.08-0.40) and perceived usefulness to their work (β coefficient=0.57, 95% CI 0.43-0.71) significantly impacted their satisfaction. Satisfaction can significantly impact implementation intentions (β coefficient=0.40, 95% CI 0.17-0.64), which further impacts implementation behaviors (β coefficient=0.16, 95% CI 0.10-0.23).

Conclusions: A notable disparity exists between patients and health care providers. This may hinder the achievement of the tool's benefits. Future research should aim for a comprehensive overview of implementation barriers and corresponding strategies to enhance staff performance and facilitate patient-provider shared understanding.

Keywords: EHR; acceptance; adoption; aged; aging; attitude; communication; cross-sectional; design; discharge; elderly; experience; geriatric; gerontology; medication information; old; older; older adult; opinion; patient-provider comparison; perception; perspective; portal; post-acute care; questionnaire; record; satisfaction; survey; technology acceptance model; technology-based intervention; user experience.

Figure 1.. Conceptual framework of the factors impacting implementation fidelity. AB: actual behavior; BI: behavior intention; DQ: design quality; PU: perceived usefulness; “ε1”, “ε2”, “ε3”, and “ε4” were the residual errors.

Figure 2.. Comparison of patient and health care providers’ postdischarge information summary experiences. P value was obtained from the Mann-Whitney U test. *P<.001.

Figure 3.. Comparison of the qualitative comments on the postdischarge information summary between patients and health care providers. PDIS: postdischarge information summary.

Figure 4.. Structural equation modeling for factors impacting the providers’ implementation fidelity. AB: actual behavior; BI: behavior intention; DQ: design quality; PU: perceived usefulness; “ε1”, “ε2”, “ε3”, and “ε4” were the residual errors.