Perhaps the biggest priority of modern healthcare is to find ways to improve population health, as measured by improved outcomes and lower overall costs, aligned with the goals of value-based care.
But effectively managing population health is rife with challenges, including lack of access to care, health disparities, and the amount of work required to manage patients outside the four walls of the hospital.
We also have to be honest about the capacity of our beleaguered healthcare system to take on such an ambitious challenge amid shrinking hospital margins, crippling labor shortages and a diminished public health infrastructure.
Waiting for something to change is no longer an option. There is broad recognition that technology must play a leading role in addressing population health challenges by automating key parts of the process, driving efficiency and lowering costs, particularly outside the hospital.
Fortunately, there is a device that’s already in your pocket that can monitor health vitals, make cognitive assessments, and encourage medication reconciliation/adherence: your smartphone.
Ubiquity and advanced technology
Now in use by greater than 90% of Americans, smartphones are equipped with a variety of sophisticated sensors, such as GPS, gyroscope, accelerometer, magnetometer, proximity, ambient light, microphones and high resolution cameras.
When paired with advancements in machine learning (AI) models, these sensors can measure nearly every physiological metric that you can get from remote patient monitoring (RPM) devices, including vitals, brain health and medication scanning.
The dream of the handheld Star Trek Tricorder data analyzer has arrived, as researchers predicted in a 2019 study: “The smartphone and its embedded sensors coupled with present-day information and communications technologies have opened a new window of opportunity for cost-effective remote healthcare services.” The authors added, “the incredible improvements in the processing and data storage capabilities in the modern-day smartphones may allow for faster, real-time and onboard execution of complex predictive algorithms and/or artificial intelligence (AI) technologies using the high-volume of raw data measured by the smartphone sensors.”
Since then, AI healthcare models have made major advances. Google said last year its Med-Gemini models achieved 91.1% accuracy on the MedQA benchmark, outperforming Open AI’s GPT-4 in understanding and analyzing medical text, images and real-time data. Meanwhile, Google’s Articulate Medical Intelligence Explorer (AMIE), a diagnostic AI chatbot, recently matched or outperformed human clinicians in multi-visit disease management consultations in a randomized study.
Limits of RPM
Lowering costs was the impetus behind RPM, which relies on a collection of connected health devices to monitor key vitals for the patients at the highest risk of a serious health crisis requiring hospitalization or readmission.
RPM can be part of the solution. At my former company we saw average overall care costs drop by more than 50% and significant decreases in mortality when we paired at-risk patients with our RPM kits and clinicians with our portal. The platform kept patients healthy at home and out of the hospital.
But RPM comes at a significant cost and provides limited views into the overall health of each patient. Devices are limited to measuring vitals, promoting a reactionary approach to remote care, and can cost up to $1,000 per patient, with ongoing support and logistic expenses exceeding $50 per month. Additionally, nurses who should be focused on clinical tasks find themselves instead managing lost or malfunctioning devices.
Because of these high costs and logistical complications, no one has been able to scale RPM to address population health needs.
Proactive, not reactive
By contrast, AI models can now listen to smartphone recordings of a user speaking to detect mild cognitive impairments, stress, anxiety, depression, and even early signs of dementia, Alzheimer’s and Parkinson’s. Models are now being trained to accurately recognize medications from a simple photo, heralding enormous potential for medication management, adherence and the many beneficial downstream health effects.
All of this leads to proactive, rather than reactive, population health management. It eliminates the high cost of devices, and it also broadens the lens of what clinicians can observe.
While these AI models can detect patterns from millions of data points almost instantly, they can also sift through the reams of clinical “health signal” data they create to surface important clinical insights, using the world’s knowledge of best practice health care to highlight the next best actions for improved outcomes at lower costs. That’s another significant advantage over traditional RPM, which relies on overburdened nurses to parse data and determine whether patients are stable or need intervention.
Together, these new health signals can work in concert to unearth critical insights. For example, imagine a senior citizen who is at risk for falling because her depression medication has side effects of high heart rate and dizziness. A brain health signal might indicate she no longer exhibits signs of depression and may no longer need the medication, reducing her frequent falls and visits to the ED.
The implications
Taken together, these technological advances hold great promise for population health management.
- We can enable proactive, preventive care by eliminating barriers to access, especially for patients who live in underserved rural or low-income urban areas. Through continuous, passive data collection, the technology can identify subtle physiological changes that might be indicative of more serious health issues, like early diabetes, neurodegenerative disease, or heart disease. The AI models can then relay those to clinicians to encourage earlier intervention, potentially preventing more costly complications or hospitalizations.
- We can personalize health management at scale as the models analyze the rich datasets of clinical actions paired with population health outcomes to tailor insights and recommendations to individual patients within a population.
- We can strengthen patient engagement and adherence to treatment regimens by reaching patients where they already are, without asking them to use separate devices, unless acutely or medically necessary.
- Finally, we can generate valuable reams of real-world evidence at scale to help fuel research, drug discovery and effective public health strategies.
The rapid advances in AI and smartphone technologies hold promise for many stakeholders in the healthcare system — providers, payers, pharmaceutical drugmakers and public health agencies — that need to understand what’s happening with the patient in real time.
We now have the opportunity to achieve what RPM can deliver on an individual basis, and more, at the scale of population health.
Photo: JanWillemKunnen, Getty Images
Eric Rock, CEO and Co-Founder of Peripio Health, is a veteran healthcare technology entrepreneur and innovator. He has founded, scaled and exited three software companies, including Vivify Health, a remote patient monitoring platform that was acquired by UnitedHealthGroup’s Optum division in 2019.
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
