By Sachin Patel
Given the significant amount of activity in healthcare over the past few years, everyone has heard numerous anecdotes related to a shortcoming of technology in healthcare. Instead of describing such a scenario, it seemed more appropriate to describe how Artificial Intelligence can improve patient experience and outcomes, via high-fidelity insights and recommendations provided at the point of care.
With the average patient encounter just 18 minutes long, physicians spend over 16 minutes using EHRs—a third of that on chart review. Simple math reveals that doesn’t leave much time for listening to the patient, assessing their needs, and devising a treatment plan, which is indeed where physicians’ hearts lie.
Pressed for time and lacking an efficient way to readily access patient information, providers may be forced to rely on patients to remind them of past conditions, any recent diagnostic tests performed, and associated results. Understandably, these sources of information are not always consistent or comprehensive.
The result: potentially missed diagnoses, fewer opportunities for proactive, preventative care, and an unsatisfactory experience for both the provider and patient.
Value-Based Care Opportunity
With the shift toward value-based care, clinicians need faster, easier access to comprehensive, accurate, and relevant data to make the best, most informed decisions. And they need more efficient tools to surface patient data insights and streamline workflows at the point of care to ensure a better patient experience and improve outcomes.
But in the current environment, this is extremely difficult for several reasons. First, data is scattered across multiple, disparate sources. Some are in the provider’s EHR, but may not include pharmacy or other records, diagnostic results performed at off-site clinics, or walk-in/rapid care visits.
Next, integrating those data sources is a challenge not only because of complex disparities between the systems but also because there isn’t a compelling reason for those data source entities to interact, especially when integration could be a complicated, cumbersome, and costly process. Data quality issues can also make integration challenging and introduce errors and omissions when content fields don’t mesh properly. Finally, with 70-80% of data residing in unstructured sources, such as handwritten notes or PDFs, digitizing and integrating that wealth of information is extremely difficult without the right technology.
The value-based care model places primary care physicians (PCPs) in a unique position. As the patient care “quarterback” bearing the full risk of managing the patient’s overall health, PCPs especially need access to the complete patient history to make the right choices for referrals, testing, medications, and overall health management. This is the right incentive to drive unification of data sources, which is significantly enhances the ability to develop and utilize artificial intelligence algorithms.
How AI Can Help in the Clinic
As data sources become more integrated and readily available, there is also the issue of data volume and diversity, where artificial intelligence can begin to shine. Using AI to sift through and understand large data sets to identify the highest-risk patients is a well-understood concept at this point. Thinking beyond this, implementing certain types of machine learning techniques can improve the performance of AI based on tuning to a provider group’s practice patterns, thus supporting more refined, efficient value-based decisions.
In a situation where a provider may have less interconnected data, especially when considering both structured and unstructured data types, AI can provide a benefit in a different way. Specifically, applying a sophisticated algorithm that has been trained on population types similar to that which the provider sees can serve as a powerful supplement to clinical expertise, allowing for identification of high-risk health conditions and associated value-based care pathways.
In either case, deploying AI to power a pre-visit workflow can give clinical staff quick access to a comprehensive view of the patient’s longitudinal medical record, allowing them to review key insights before the patient encounter, and potentially suggest treatment courses. AI can also handle administrative tasks in parallel, reducing this burden via automation. As a result, providers can spend more time focusing on optimal patient care during a visit, and less time reviewing and verifying history or keying in medical coding items.
Improved Patient Outcomes
With the tremendous amount of innovation in healthcare over the past several years, providers are focused on leveraging the right technology to maximize efficiency and improve patient outcomes. Improved patient outcomes benefit the overall healthcare ecosystem through reduced expenditures and more targeted care that has the highest impact on the right set of patients.
For example, using AI predictions based on practice patterns as noted earlier, a physician can determine that, instead of coming in multiple times for in-office treatments, a patient may benefit more from equally efficacious medication taken at home. Even better, the treatment can be delivered to their door alongside an in-home lab draw, which overcomes transportation issues, lowers the risk of exposure to infectious diseases, and enhances treatment adherence.
Implementing AI-based pre-visit solutions can enable up to five times faster chart review. Additionally, up to 20% more information about risks, conditions and treatment pathways can be made available to providers when and where it matters most—at the point of care. The results are better outcomes, better experience, and higher satisfaction for patients and greater efficiency, less friction, and lower costs for providers.