Objective:

To investigate how differences in patient portal message writing style and content affect clinician response rates, particularly among historically marginalized groups.

Approach:

• Data Analysis: Analyzed 3.6 million patient-initiated medical advice message threads from 511,020 adult patients using natural language processing.
• Sample Selection: Excluded non-English messages, resulting in 989,602 threads for text-based analysis.
• Outcome Measurement: Primary outcomes included response rates from care teams and target clinicians, with secondary outcomes measuring response times and expanded definitions of response.

Key Findings:

• Black patients had a 3.7-percentage point lower target clinician response rate compared to White patients, representing a 12% reduction relative to the 32% mean response rate.
• Message content accounted for 6% of the response gap among Black patients compared to White patients.
• Writing style influenced response rates significantly, accounting for 48% of the target clinician response gap among Black patients.
• Higher response rates were associated with messages that included the clinician's name, positive sentiment, and longer word counts.

Interpretation:

[Writing style was a potential explanation for approximately half of the observed disparities in care team response,] wrote lead study author Mitchell Tang, PhD, of the Harvard Kenneth C. Griffin Graduate School of Arts and Sciences at Harvard University, and colleagues.

Limitations:

• Data derived from a single academic health system, limiting generalizability.
• Excluded non-English messages and did not capture all aspects of message content or writing style.
• Cross-sectional design limits causal conclusions.

Conclusion:

[The study highlights the significant role of writing style in clinician response disparities, particularly among marginalized groups,] wrote lead study author Mitchell Tang, PhD, and colleagues.

Sources:

• JAMA Network Open