As a physician who transitioned into health technology leadership, my journey over the past few years has revealed a fundamental truth: The future of health care delivery isn’t confined to hospitals or clinics—it’s in patients’ homes as well. I am not just talking about patients who have to be discharged early from our hospitals, but also about patients who don’t have easy access to health care and patients who have other socioeconomic challenges that predispose them to end up in hospitals.
Yet, as we rush to digitize health care delivery and to integrate AI in health tech, there is not enough focus on home-based health care. This gap risks leaving behind the very populations who might benefit most from technology, AI or otherwise. The digital divide in health care isn’t just about access to technology; it’s about whether that technology is designed to serve all populations equitably.
The hidden inequities in home health care
When I first ventured into building an EHR specifically for home health care, I was struck by how traditional systems failed to account for social determinants of health (SDOH)—those non-medical factors that influence up to 80 percent of health outcomes. In home settings, these factors become glaringly obvious: the empty refrigerator, steep staircase, messy apartment, empty liquor bottles, embarrassment, and sometimes anger as well.
Yet our documentation systems rarely capture this critical context. A patient who misses appointments isn’t simply “non-compliant”—they might lack transportation or childcare. The diabetic patient with worsening A1C levels might live in a food desert. These social factors create and perpetuate health inequities, particularly among racial minorities, rural populations, and those with lower socioeconomic status.
AI as an equity enabler
Artificial intelligence, often viewed with suspicion in health care, has remarkable potential to address these inequities when thoughtfully deployed in home health care settings. Here are some examples:
1. Identifying hidden SDOH patterns. Traditional screening questionnaires for social needs are often inadequate—patients may feel stigmatized or simply not recognize certain factors as health-relevant. AI systems can analyze patterns in home visit documentation, patient demographics, vital signs, medication adherence, and even ambient conversation (with appropriate consent) to identify potential social barriers that might otherwise go undetected.
In practice, natural language processing can identify SDOH mentions in clinical notes with 87 percent accuracy—far exceeding what manual reviews could accomplish at scale. These insights allow for targeted interventions before health deterioration occurs.
2. Breaking language and literacy barriers. Nearly 36 million U.S. adults have limited English proficiency or low health literacy, creating significant barriers to effective home health care. AI-powered translation and communication tools can bridge this gap in real time.
We need systems that not only translate conversations but adapt educational materials to appropriate literacy levels, using visual aids and simplified language. The technology also identifies when patients may not understand instructions, prompting clinicians to use teach-back methods or alternative explanations.
3. Enabling community health workers. AI can serve as a force multiplier for community health workers (CHWs), who often come from the communities they serve and understand local needs intimately. By automating routine documentation and providing clinical decision support, technology can free CHWs to focus on relationship-building and addressing complex social needs. The same goes for skilled nursing providers, therapists, and medical social workers who visit patients at their homes.
Caution is necessary
The dark side of health care AI is its potential to perpetuate existing biases. Many algorithms are trained on datasets that underrepresent minority populations or contain historical biases in treatment patterns. When developing home health care EHR systems, it’s essential to implement fairness audits for all predictive models, continuously monitor for drift in model performance across populations, and involve diverse stakeholders in algorithm development. This vigilance is essential—technology that works perfectly for one population may fail catastrophically for another.
The economic imperative
Addressing health inequities isn’t just morally right—it’s economically essential. Studies consistently show that unaddressed social needs drive up health care costs through preventable emergency department visits and hospitalizations. One analysis found that addressing SDOH could save $1.7 trillion in health care costs over ten years.
The challenge lies in aligning financial incentives. Value-based payment models are beginning to recognize the importance of addressing social needs, but progress remains slow. Technology companies must design sustainable business models that make advanced solutions accessible to safety-net providers and underserved communities.
The promise
As we continue developing AI solutions for home health care, several principles must guide our work:
Community co-design: Technology must be developed with underserved communities, incorporating their lived experiences and needs.
Data equity: We must ensure our datasets represent diverse populations and continuously monitor for bias in both data collection and algorithm outputs.
Digital accessibility: Solutions must work across varying levels of technology access, including options for low-bandwidth environments and older devices.
Human-AI collaboration: Technology should augment, not replace, human caregivers, preserving the essential human connection in health care.
The promise of AI in home health care isn’t about futuristic robots or replacing human caregivers—it’s about using technology to see patients more completely, understand their needs more deeply, and connect them to resources more effectively. When designed with equity at the center, these tools can help us build a health care system that truly serves everyone, regardless of zip code, language, or socioeconomic status.
The digital transformation of home health care presents a once-in-a-generation opportunity to address longstanding inequities—but only if we approach it with intention, humility, and a commitment to justice.
Sreeram Mullankandy is a physician executive.
