This blog is based on a recent webinar with Microsoft titled “AI-Enabled Hospital at Home – Considerations for Payers and Providers.” You can watch the full webinar here.
Hospitals can be a scary place for patients during a pandemic. Last year, around 40% of patients delayed or avoided care out of fear of COVID-19. Fortunately, the pandemic has also shown that many patients can be treated from the comfort and safety of their own home.
The Hospital at Home movement has been slowly gaining steam over the past two decades since it was pioneered at Johns Hopkins in 1995. Then the pandemic hit. Almost overnight, virtual care technology that had been incubating for years saw rapid adoption as a way to treat patients at a safe distance. Remote patient monitoring had its breakout moment.
The good news is that patients don’t have to sacrifice the quality of their care at home. Research has shown that home care improves patient outcomes, lowers the costs of care, and enhances the patient experience. In other words, it achieves the Triple Aim of Healthcare.
Here are a few of the benefits shown from preliminary data on home care:
- Up to 20% reduction in 6-month mortality
- Up to ⅔ reduction in 30-day readmissions
- Up to 38% reduction in admission costs
- Fewer complications (pressure injuries, infections, etc.)
- Very high patient satisfaction scores
On top of that, family members can also visit more easily at home, relieving stress on caregivers. Dementia patients have less delirium when treated in the familiarity of their home.
And providers can provide a more person-centered approach to care at home, addressing the social determinants of health (SDOH) that caused patients’ health to decline in the first place.
There’s a lot of promise behind the Hospital at Home movement. But to reach its full potential, we need the help of AI.
The Role of AI in Hospital at Home
The rise of virtual care has been made possible by the convergence of three trends. First, consumers have come to expect on-demand services that aren’t tied to a specific location. That extends to healthcare. Second, the arrival of the cloud enabled the storage of vast amounts of data from health devices in the patient’s home. Finally, advances in AI made it feasible to process this data, separating the signal from the noise.
The Hospital at Home approach depends on countless data streams to monitor the patient remotely. To start, there’s biometric monitoring technology, including connected devices such as blood glucose monitors, O2 monitors, EKGs, smart beds and more. Then there’s clinician data collected from telehealth sessions and in-person visits. Plus, to treat patients holistically you have to somehow integrate data on the patients’ social determinants of health as well.
Inevitably, Hospital at Home providers end up with more data than they know what to do with. Clinicians don’t have time to parse through the endless data for one patient, let alone dozens of patients they may be caring for. To make sense of all this data, you need AI to detect patterns and alert providers to changes in the patients health trajectory.
To that end, there are two key ways AI will support the Hospital at Home movement. The first is triaging patients for home care, and the second is identifying patients’ risk of deteriorating at home and driving preventative measures.
Triaging Patients for Home Care with AI
The Hospital at Home process usually follows a similar path. A patient arrives at the ED, receives treatment, and is triaged to determine if they are eligible for home care. From there, they’re transported home, where they are on-boarded and assessed by nurses and physicians to establish a plan of care. That’s followed by regular in-person and virtual touch points until the patient is discharged from the program.
Triaging in the ED is usually driven exclusively by clinical factors. But that leaves out details about the patients’ behavior, lifestyle, and living conditions that are critically important to the success of home care. AI can bring in SDOH and behavioral data to see patients more holistically, providing insight on what patients will want to participate in and what additional steps should be taken to address any barriers to care at home.
While it might be tempting to think that a person can look at SDOH data and draw clinically relevant conclusions from it, the reality is that any single SDOH data point in isolation tells you very little. You need AI to interpret the data in the context of other data.
As an example, consider a data point indicating a patient has had a long-term residence. This could mean the patient has a stable home. It could also mean they are socially isolated, left behind by friends that have long since left the neighborhood. There could be environmental health hazards from a lack of maintenance. You need to look at other data — marital status, income, environmental health hazards, etc. — to get the full picture.
To give you an example of how AI can provide that full picture, take a 68-year-old patient with pneumonia. AI can reveal her SDOH risk factors: high likelihood to lack digital fluency, lack of other adults in the household, education limited to high school. All of these factors need to be considered when she goes home. Going one step further, prescriptive AI-enabled solutions, like Jvion’s, can provide guidance on how to address these risk factors: for example, having an in-person aide visit the patient to provide social and technological support.
AI Helps Identify Patient’s Risk of Deterioration at Home
Once a patient is set up at home for long-term care, there will be a continuous stream of data from their remote patient monitoring devices back to their care providers. With dozens of patients to manage, clinicians don’t have time to look at all the data and predict which patients may need additional attention. Without AI to parse the data in real time, it’s very difficult for remote clinicians to know when a patient is deteriorating.
AI can ingest data from thousands of patients in real time and provide care teams with alerts when the data suggests a patient might be on a trajectory for preventable harm. This gives providers time to act to change the patient’s trajectory. But with most AI-enabled predictive analytics, clinicians may not have a clear direction on how to change the patients’ trajectory.
This is why Hospital at Home clinicians need prescriptive analytics, not predictive analytics. Jvion’s prescriptive analytics solution not only identifies which patients are on a modifiable risk trajectory, but it also provides clear guidance on the modifiable risk factors driving that risk and the steps clinicians can take to address them. Although the clinician is ultimately still responsible for deciding the course of action, prescriptive analytics can ease their cognitive burden.
We’re currently demonstrating the impact of prescriptive analytics on home care with AccentCare, a nationwide leader in post-acute healthcare services with over 240 locations in 29 states. AccentCare is currently using our solution to understand how social determinants drive patients’ risk of readmission and gain recommendations for how to address these risk drivers.
The Future of Hospital at Home
There are still many challenges to overcome for Hospital at Home to reach its full potential. We need to make sure patients have the digital fluency, connectivity, and trust to use this technology effectively. Data quality, privacy and cybersecurity protections must be in place.
On the legislative side, we need to ensure Hospital at Home can be reimbursed. The Acute Hospital Care at Home program, implemented to increase hospital capacity during the pandemic, is an important step forward, but so far only covers 116 hospitals. We need sustainable reimbursement beyond the pandemic for providers to adopt Hospital at Home.
Culturally, we need to overcome the perception that patients get worse care at home, that hospitals are only doing this to spend less money. This can be addressed by providing patients with high quality experiences at home. Operationally, we need to make sure that technology can be easily used by both the consumer and the clinician.
Challenges aside, the Hospital at Home movement can provide a lot of value for clinicians and patients alike, and should be an integral part of care delivery long after the pandemic. As America’s population ages, we will need Hospital at Home to ease the burden on hospitals and keep the healthcare system sustainable. The technology enabling home care, including AI, will only improve. We need legislative barriers to come down and clinical leaders to step up and lead the charge on adoption.