How AI, Digital Health And Home-Based Services Can Help Prevent Hospital Readmission

How AI, Digital Health And Home-Based Services Can Help Prevent Hospital Readmission unknown

Dr. Darshak Sanghavi, Chief Medical Officer, Babylon.

getty

Hospital readmissions are a stubborn and complex problem; roughly 18% of all Medicare patients end up back in the hospital within a month of leaving, and many of these readmissions are preventable. There’s now a growing interest in automating these strategies through AI and digital technologies to provide better post-hospital care in the home. However, gaps in the infrastructure may contribute to miscommunication and result in delays in information sharing with patients. Yet, we must be careful to follow the evidence and not simply invest in digital solutions without exercising appropriate diligence.

The problem: People need help after being hospitalized.

Transitional care management (TCM) programs were developed to assist patients leaving the hospital but traditionally depend on highly manual processes. The most widely used approach, known as the "Coleman model," is a four-week process that often includes one hospital and one home visit and a series of follow-up phone calls with a designated care manager overseeing the transition plan. The four elements of the model include: medication management, developing and sharing patient-centered health records, completing follow-up visits with providers and specialists, and educating patients about when to suspect worsening conditions or drug reactions.

For some time, clinicians have explored ways to digitize and automate some aspects of TCM, particularly following 2012 when the Centers for Medicare and Medicaid Services began issuing financial incentives to prevent certain readmissions.

One puzzle piece: Texting is scalable and effective in preventing rehospitalization.

Can something as simple as texting help? The answer seems to be yes. A recent study by the University of Pennsylvania shows just how successful automated text messaging to TCM patients can be. About 2,000 patients at two academic primary care practices got texts asking if they had a follow-up appointment with a primary care doctor or specialist within two weeks. If patients replied with a "no," an alert was sent back via the electronic health record, and a care assistant reached out to the patient. Some options to check-in messages (e.g., “I don’t feel well”) would prompt a provider phone call. In the month after discharge, patients in the program had a reduced odds ratio of readmission to the hospital and ED by 41%, when compared to those not getting the messages.

The digital transformation of TCM is moving forward.

Increasingly, new technologies are being deployed to address each key aspect of traditionally manual care transition programs. Still, the majority of TCM occurs in the clinic and not at home. According to research conducted by the University of Washington School of Medicine, 87% of TCM services occur in physician offices—and only about 2% happen in a home setting. Thus, there is enormous potential to better meet patients where they are, and a growing array of solutions might be pieced together to augment in-person care.

For example, a hybrid model of care that reaches out to patients via AI-enabled voice calling, text and chat—like Care Angel—can close care gaps for patients who require a more personalized level of support. Taking advantage of increasingly common technologies for natural language processing and lifelike computer-generated speech, these platforms can interact with patients at scale and at home. Those calls drive data to provide alerts, and information is shared with the patient and family to reinforce care adherence. (Of note, my own voice was digitized and used when I served as a chief medical officer of a major national payer, and the outreach was deployed nationwide for millions of Medicare Advantage beneficiaries.) And to be sure, AI-enabled outreach raises the need for transparency to ensure patients are not inappropriately led to believe the voice is from a real person.

To better identify patient risk in real time, providers may also benefit from dynamic monitoring of data from recently hospitalized patients. For example, companies like Vivify Health (a cloud-based virtual platform for providers and payers), Ejenta, Biofourmis and many others are collecting data from patients through their mobile digital devices or at-home remote monitoring kits. This, in turn, provides actionable insights for clinicians, expands provider capacity and saves valuable clinical time. But to be fair, the evidence base on these programs’ effectiveness continues to evolve.

Post-discharge medical reconciliation offered via Cureatr includes medication management technology that pulls together patient and pharmacy information across networks and providers. Similar products increasingly use pharmacy claims data, electronic health exchange information and other sources to ensure up-to-date management of medications, and such tools are now integrated directly into EHRs like Epic (for example, DrFirst). Simply automating the flow of information and ensuring the right data gets to the right place at the right time is a core principle of the Coleman model.

Finally, digital tools can ensure patients themselves are better educated and connected to care once they depart a hospital. Curating care journeys and programs can be served using highly engaging, daily check-ins via digital platforms like DayToDay Health. Care teams provide timely communication, a dynamic patient-centric journey with interactive information, critical metrics and follow-ups all in a primarily digital-first model.

The future of post-hospital care is digital.

Moving forward, what lessons have been learned? For health system leaders, the following tips are helpful in determining their digital approach to TCM.

First, establish a reliable approach to track discharges and readmissions in real time in order to assess the trends and impacts of any interventions. Second, understand the key components of an effective TCM model and identify best-in-class solutions to address these—ideally, via a unified or complementary patient interface. Hold the vendors accountable to quantify and deliver data on each component. Third, reduce the spans and layers between management and the clinical workforce. Too often, the decision-makers are isolated from the on-the-ground realities of new technologies.

There is work to be done before a fully integrated and comprehensive evidence-based solution emerges that is widely adopted. With the right attention to evaluation and patient engagement, that day may not be too far away.

Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?