Covid-19 has changed the course of healthcare for the foreseeable future. Healthcare workers, doctors’ rooms, and equipment inventories were stretched thin even before the pandemic took hold, with patients waiting weeks or months to get a doctor’s appointment or book a surgery. Today, these resources are nearing the breaking point.
With physicians’ offices and testing facilities closed for in-person appointments, and elective surgeries put on hold, it has become incredibly complicated for patients to find the care they need for pre-existing or non-Covid-19 health issues. They might not be able to see their usual doctor or maybe they’re simply trying to find a provider for the first time. In these instances, patients need as much guidance and insight as they can get in order to achieve their desired outcome, particularly in less than optimal conditions. But traditional resources, such as web reviews, personal recommendations, and media directories, are imperfect solutions.
By combining electronic health records and artificial intelligence, it is now possible to pair patients with the available caregivers most likely to produce the best result—based on data for patients and outcomes across providers. Such approaches offer improved care outcomes, satisfaction, access and affordability.
Breakdown of the five-star system
The five-star review system is a popular tool for rating and selecting the best restaurants, films, hotels, and refrigerators. But when it comes to helping people choose the best doctor, it has significant limitations.
The internet has enabled people to share opinions about doctors, hospitals, and nursing homes on many ratings sites. Web reviews of healthcare providers are enormously popular, but the data these ratings provide are not personalized for each individual’s needs. They may also be influenced by factors other than objective long-term outcomes, such as preventing future hospital admissions or avoidable emergency room visits.
The problem with the five-star system with regards to healthcare boils down to this: you are not like your best friend, your neighbor, or the woman who posted a glowing review online about her cardiologist. You are unique. You should select the doctor who can best address your personalized healthcare needs and attain the best outcomes.
The healthcare characteristics of each individual can greatly vary. Finding the best doctor suited to handle such nuanced health complexities poses a far more daunting and consequential challenge than asking about which restaurant has the best deep-dish pizza in town.
The system is further complicated when the doctors that patients research are not available to treat them, either in-office or via a telehealth solution. Because Covid-19 can lead to a scarcity of caregivers especially when looked at within geographies, the system is even more complicated while yet the need for highly qualified doctors, especially for vulnerable members of society who may have multiple comorbidities, has never been more urgent. Needs for care outside of Covid-19 have not dissipated at all.
Choosing healthcare has been a game of chance
In recent years, studies have concluded that popular rating systems have the potential to confuse consumers on the selection of hospitals. One reason is that despite existing data which tracks patient outcomes for each physician, it is nearly impossible for the public to access, compare, or understand that data in a simple way. Then consider that some doctors have been deployed to the front lines of the Covid-19 fight while others are forced to the side lines, and patients are truly at a loss for accurate provider information.
Picking a provider was simple when there were few options; patients relied on whatever doctor or hospital was nearest. As the number of options has grown however, provider selection has increasingly relied on recommendations from friends and family, or the advice of trusted third parties like family physicians or perhaps a publication. US News & World Report started publishing its annual list of the country’s top hospitals in 1990. It remains influential in patient decision-making to this day.
Yet, when it comes to helping people find outcomes-focused healthcare, studies show that ratings, rankings, and personal recommendations leave a lot to be desired. As part of the hospital-ranking process performed by US News, the magazine considers some objective outcomes measures, such as each facility’s mortality rates. However, it also factors in much more subjective data, according to a 2017 study by researchers from Ohio State University.
In the study, the group notes that US News rankings are “disproportionately influenced by the subjective reputation measure” and that these subjective measures have “a more significant influence on total US News score than its objective counterparts.”
“Allowing such a subjective measure of care to influence the determination of America’s best hospitals,” the team further observes, “can affect the provision of care, introduce gaming in healthcare and lead to misinformed consumer decision making.”
In studies of other approaches, patient reviews haven’t fared much better in determining care quality. This is likely because patients are also subjective in their reviews and rating. Every patient brings his or her own experience, priorities, perspectives and biases to rating.
As one example, researchers from Cedars-Sinai Medical Center set out to learn if online patient ratings corresponded with more objective measures of physician performance. They discovered that the public appears to place a lot of trust in web ratings, with about 75% of patients willing to trust consumer ratings alone to make decisions on healthcare providers. Nonetheless, in their 2017 paper, published in the Journal of the American Medical Informatics Association, the researchers found “no significant associations” between consumer ratings and a doctor’s clinical performance.
This should surprise no one. Most of the approaches that have been traditionally available to consumers are not fundamentally grounded in objective outcomes. Multiple studies show online reviews typically focus less on treatment and more on service, such as the demeanor of front-office staff, billing issues, or how easy it is to book an appointment.
In 2016, researchers from New York University examined Yelp reviews to determine the public’s level of satisfaction with the nation’s radiology centers. Of the 1,009 patient reviews, and 2,582 accompanying comments, only 14% were related to radiologists—the other 86% focused on service aspects such as receptionist professionalism, billing, wait times, scheduling and physical office conditions.
Data to the rescue
It is now possible, however, to remove this level of subjectivity and use electronic medical data to have a much deeper understanding of each individual in terms of their healthcare problems and needs. Moreover, vast quantities of electronic medical data make it possible to track physician and hospital performance over time. However, these huge volumes of electronic medical data are simply too vast and complicated to be understood by any one individual.
We’re on the threshold of a new age, with artificial intelligence able to comb vast quantities of longitudinal patient data to spot patterns, make comparisons, and predict outcomes in ways that can dramatically improve provider selection and patient care.
This makes it possible for individuals to leverage the power of computing and AI to assist in choosing the right doctor by factoring in the unique characteristics of each individual and finding doctors with experience in taking care of patients like that individual. As an example, take an individual seeking a knee surgeon. Knowing the individual seeking care is important in selecting a surgeon who may either be practicing in the realm of sports medicine, and thus used to seeing a mix of young active and athletic patients, versus a doctor whose practice is focused on knee surgeries for the elderly. The latter is likely have a mix of comorbidities and will require a different approach both during the procedure and post-procedure rehabilitation.
AI-based matching systems are particularly well suited for the current climate because they can enable people to pick providers in a way that is far more scalable and flexible, given Covid-19 constraints and future patient backlogs. New approaches can also be distributed over a larger catchment or allow for selecting providers from coast to coast through greater telehealth options. On top of this, AI systems and machine learning can spot where normal referral patterns might be broken to step in and recommend a better option.
While AI-based systems may appear to be something still “in the future,” the Covid-19 pandemic has dramatically accelerated the adoption of healthcare technologies out of necessity. Given current conditions and the ripple effects they will have over the long-term, we can expect to see greater adoption of AI driven provider-patient matching now, and in the near future.