Happy Holidays from DHEC!Webinar Highlights, What We're Reading, & Speaker Call

Dear DHEC members,

Happy Holidays from all of us at the Digital Health Equity Collaborative! As we close out the year, we want to take a moment to thank you for your continued engagement and commitment to advancing digital health equity. We’re excited to share a recap of our recent webinar and introduce some new ways to connect and collaborate in the coming months.

In This Issue

• Webinar Highlights: Key insights from our September session on prioritizing digital health equity in industry and nonprofit organizations.
• What We’re Reading: Latest research on machine translation and ChatGPT in healthcare—exploring equity, safety, and representation.
• Get Involved: Share your favorite digital health equity resources and speaker recommendations via our updated survey.
• Looking Ahead: Upcoming webinars in March and September 2026—stay tuned for dates!

Here’s what we’ve been up to—and what’s coming next in digital health equity!

1 – Webinar Recap: Prioritizing Digital Health Equity in Industry and Nonprofit Organizations

September 11 DHEC Meeting: Prioritizing Digital Health Equity in Industry and Nonprofit Organizations

In case you missed it, watch a recording of our September webinar here.

On September 11, 2025, the Digital Health Equity Collaborative (DHEC) hosted a virtual learning meeting focused on “Prioritizing Digital Health Equity in Industry and Nonprofit Organizations.” This session topic speaks to DHEC’s ongoing mission to unite operational leaders, academic researchers, patient advocates, and all stakeholders to drive innovation, promote equitable digital health practices, and champion historically excluded populations.

The meeting featured two speakers:

• Alister Martin, MD, MPP, Emergency Medicine Physician at Massachusetts General Hospital and Founder/CEO of Link Health, GOTVax, and A Healthier Democracy.
• Michelle de Haaf, BS, MS, Assistant Director of Digital Health at the Stanford Mussallem Center for Bio Design.

Dr. Martin shared his groundbreaking work leveraging AI-powered programs to help physicians address poverty-related health challenges. He emphasized that while his ER shifts often involve treating pneumonia or stroke, the most common condition he sees is poverty. His initiatives, developed during his time at the White House, tackle barriers to benefit enrollment by creating systems that automatically connect patients to multiple programs—such as SNAP and other social services—closing critical resource gaps and empowering patients to access what they need.

Michelle de Haaf offered a compelling perspective on designing equitable digital health solutions. She highlighted the importance of shifting from a “tech push” mindset to a “need pull” approach—starting with a deep understanding of the problem before building technology. Her example of Malama, a culturally sensitive app for pregnant women with gestational diabetes, illustrated how focusing on resource access rather than just health tracking transformed outcomes. Michelle also showcased innovative programs like Upside and GOODR, reinforcing that health happens beyond the clinic and that democratizing technology is key to healthier communities.

This webinar sets the tone for DHEC’s commitment to collaboration, innovation, and equity in digital health. We look forward to continuing these conversations and driving meaningful change together every 4–6 months to explore key topics, share case studies, and engage in meaningful Q&A sessions. Our 2026 webinars will be held in March and September with newsletter updates between those sessions. Stay tuned for official dates!

2 – New Segment: What Are You Reading That Is Hopeful?

We’re excited to introduce a new section in our newsletter where we share readings that inspire hope and progress in digital health equity. This week’s feature will focus on machine translation and its potential to break down barriers in healthcare. We are also collecting digital health equity reading resources on our website, and we’d love for DHEC members to provide their own recommendations. You are welcome to send us your suggestions through our updated survey link found at the end of the newsletter. Please see the readings reviewed below.

• Human-in-the-Loop AI Translation for Discharge Instructions (Brewster et al.)
• ChatGPT-4 vs. Google Translate for Patient Instructions (Kong et al.)
• Pediatric Discharge Translation Performance (Brewster et al.)

Latest Evidence on Machine Translation & ChatGPT in Healthcare

Human-in-the-Loop AI Translation for Discharge Instructions (Brewster et al.)

Brief Summary:

This multidisciplinary study evaluated ChatGPT-4o, professional translations, and human-in-the-loop approaches (AI + human post-editing) for translating discharge instructions across six languages. It found that human-in-the-loop strategies achieved comparable or superior quality to professional translations while being significantly faster.

Key Takeaways:

1. Human-in-the-loop is optimal: Combining AI translation with professional linguist review achieved the best balance of quality, speed, and user preference.
2. Representation matters: ChatGPT-4o performed poorly for digitally underrepresented languages like Armenian and Somali, highlighting persistent equity gaps that require human oversight to address safely.

Link: https://pubmed.ncbi.nlm.nih.gov/41136708/

ChatGPT-4 vs. Google Translate for Patient Instructions (Kong et al.)

Brief Summary:

Comparing ChatGPT-4 and Google Translate for emergency department discharge instructions found both tools achieved ≥90% accuracy for Spanish and Chinese at the sentence level, but accuracy dropped for Russian, with cumulative risks increasing when evaluating full instruction sets rather than individual sentences. Click here for a video summary of this article!

Key Takeaways:

1. Sentence vs. instruction set risk: While potentially harmful mistranslations were rare at the sentence level, up to 6% of complete instruction sets contained at least one error with potential for harm, emphasizing the need to consider cumulative risk.
2. Limited use recommended: Results support cautious use of machine translation only for low-stakes English-to-Spanish communication when professional translators are unavailable, with all other scenarios requiring professional oversight.

Link: https://pubmed.ncbi.nlm.nih.gov/40633961/

Pediatric Discharge Translation Performance (Brewster et al.)

Brief Summary:

Evaluation of standardized pediatric discharge instructions revealed that both ChatGPT and Google Translate performed comparably to professional translations for Spanish and Portuguese but demonstrated significant quality deficits and safety concerns for Haitian Creole.

Key Takeaways:

1. Representation matters: Poor performance for Haitian Creole (a digitally underrepresented language) compared to Spanish and Portuguese reveals how AI systems risk worsening healthcare disparities for already vulnerable populations.
2. Regulatory urgency: Findings emphasize the need for transparent governance frameworks, diverse multilingual training data, and involvement of bilingual clinicians, interpreters, and patients in establishing safety standards for clinical machine translation use.

Link: https://pubmed.ncbi.nlm.nih.gov/38860299/

3 – Speaker Recommendations for Our Next Webinar

We’re planning our next deep-dive session and would love your input! If you have recommendations for speakers who are doing impactful work in digital health equity, please send us an email at [email protected] or submit your recommendations via our updated survey below.

4 – Updated Survey

Help us shape future webinars and stay connected! Complete our updated survey to share suggestions and subscribe to our newsletter:

DHEC Interest Survey

Thank you for your collaboration and commitment to advancing digital health equity. Wishing you a joyful holiday season and a healthy new year!

—The DHEC Team

If this message has been forwarded to you, you can sign up to receive future updates directly here.