Patient Matching Sets Providers and Patients Up for Success — If Done Right

The pressure on behavioral health providers to improve outcome measures may make patient-matching technology a crucial part of their tech stack.

The hope is that providers and patients alike are set up to succeed at the outset of treatment, accelerating the development of the oft-vaunted therapeutic alliance in behavioral health. If behavioral health organizations successfully employ patient matching at the outset of treatment, a whole list of benefits could follow, according to industry experts.

Several of the largest outpatient mental health providers in the U.S. have significant interest in developing better patient-matching technologies.

“We can reasonably assume that we can get better clinical outcomes, which creates a whole flywheel of impacts not only for patients but also for the larger system,” Colleen Hilton, a former licensed marriage and family therapist and founder and CEO of patient matching technology provider Alli Connect, told Behavioral Health Business. “It really creates a beautiful feedback loop.”

Alli Connect uses machine learning to match patients to providers across 12 different domains and over 150 endpoints. The machine learning platform then tracks patient engagement and clinical outcomes and learns which aspects of the match leads to better outcomes.

Patients and providers have to make inputs into the system via an onboarding process. Patients go through an intake process that identifies their preferences, especially their distastes, by teasing out their needs.

Providers answer a series of questions meant to “build a really robust profile telling us all about themselves and their practice,” Hilton said.

Alli Connect’s initial machine learning patient matching algorithm was based on a system developed at Hilton’s previous company — Acuity Counseling, which was acquired by Scottsdale, Arizona-based outpatient mental health giant LifeStance Health Group Inc. (Nasdaq: LFST) in 2021. However, Hilton noted that Acuity’s process was manual and could be subject to human biases.

She said she hopes that adopting the matching process developed at Acuity into a machine learning-based system and generating more data will allow for further refining of the matching process. It focuses especially on increasing patient satisfaction. Patient satisfaction is tied to increased care completion and “superior clinical outcomes,” according to one study.

Patient matching plays a vital role for LifeStance as well. Last year, LifeStance rolled out a new online booking and intake experience (OBIE), including a patient matching element.

“We can reasonably assume that we can get better clinical outcomes, which creates a whole flywheel of impacts not only for patients but also for the larger system.”

Colleen Hilton, CEO of Alli Connect

LifeStance is the largest provider of outpatient mental health services in the U.S., with 5,631 mental health clinicians and 619 centers as of the end of 2022, according to its latest annual financial disclosure.

LifeStance declined to comment for this article. However, in a previous interview, company executives described key elements of the matching process in OBIE. These include helping patients find providers based on clinician availability, type of insurance accepted, clinician specialty and location, Danish Qureshi, LifeStance president and chief operating officer, told BHB.

LifeStance Chief Digital Officer Pablo Pantaleoni described the matching algorithm as proprietary. He added this matching algorithm is part of the value-add the company offers providers, saying it enables providers to “refocus on the care that they want to deliver.”

LifeStance aims to use the technology to help retain its providers, a challenge for the company over the past two years as it has sought to integrate dozens of acquired and de novo therapist offices.

“From outside of LifeStance, the idea is you’re calling randomly down lists of providers hoping someone answers the phone and is willing to book you and you can afford,” Qureshi said in a previous interview. “If you show up to that appointment and it’s not a good fit, and you have to start all the way over at the beginning of that process, … we can all intuitively know that that’s a very demoralizing path for a patient that is already making a very difficult decision in the first place.”

The critical view

David Kraus, president of Outcome Referrals and a clinical psychologist, maintains that most matching systems don’t make scientific assessments of mental health providers and patients. They are, therefore, availability filters that have no bearing on the likelihood of good outcomes.

While important to find providers that are accessible, holding up such systems as patient matching puts barriers in the way of pairing the right patients with the right providers.

Kraus said that many patient matching tools require patients to identify their clinical needs before seeing a provider.

“In no other field of medicine does the patient need to diagnose themselves before they get matched,” Kraus said. “The other problem is the therapist is then guessing what their strengths are. And we already know that the correlation between what therapists think they’re good at and what they’re actually good at is zero. And so those two problems create nothing in the middle that’s of value.”

To-be-released data from Kraus and his research partners show that therapists’ perceptions of their skill were no better a predictor of measurable skill than chance.

Several points of research find weak to no correlation between a mental health provider’s assessment of their strengths and what can be assessed objectively as their strengths. Mental health providers routinely overestimate their abilities.

Kraus developed the Treatment Outcome Package (TOP), a standardized assessment. It assesses 12 functional domains: depression, quality of life, mania, panic or somatic anxiety, psychosis, substance misuse, social conflict, sexual functioning, sleep, suicidality, violence, and work functioning. Outcome Referrals can generate a clinician report card after 15 patients begin and end treatment and complete a TOP.

A randomized clinical trial of Outcome Referrals patient matching and outcomes measurement protocols found that using objective measures of clinicians’ abilities and patients’ needs led to outcomes that are twice as good compared to typical patient assignment practices.

The trial was conducted in clinics now owned by LifeStance in Cleveland. Kraus added that the clinics’ existing system was used as the control group in the trial.

“Notably, the good fit in this study came not from changing what the therapists did in their treatment, but rather who they treated,” a report based on the trial states. “Capitalizing on whatever it is that a therapist historically does well when treating patients with certain mental health problems, the current data indicate that our match system can improve the effectiveness of that care, even with neither therapist nor patient being aware of their match status.”

The trial also finds that, on average, therapists have at least one clinical domain in which they are effective and one in which they are ineffective.

While improved patient outcomes are the hoped-for impact, Kraus also expects improved patient matching to improve clinician satisfaction.

“If you do matching scientifically, it makes a world of difference to everyone,” Kraus said. “And even therapists say that this is the way treatment should work. They go home happier because they’re seeing patients that they do really well with.”

The randomized trial was funded by the Patient-Centered Outcomes Research Institute (PCORI) and led by the University of Massachusetts Amherst. A pending roll-out, backed by a $4.6 million PCORI grant announced in April, will track the Outcomes Referral system across 50 mental health care sites operated by Refresh Mental Health in more than 30 states. 

This phase of the trial will assess an additional 700 clinicians and 500,000 patients, PCORI said.

Vincent Bellwoar, a psychologist and senior advisor of the East region at Refresh Mental Health, told BHB that patient matching helps avoid harmful care or instances where clinicians and patients don’t work well.

“This is based on stuff that shows that if you get a patient matched to the therapist that deals best with these conditions, you’re going to double the outcomes,” Bellwoar said. “And that’s something insurers should certainly be aware of and interested in.”

Patient matching via new technology and science upends a decades-old process of matching patients to therapists based on provider availability and preferences, Bellwoar said. He saw and participated in similar processes during his 25 years of practice.

The practice of simply slotting patients where they are available is still the industry norm.

He is also skeptical of other health care companies’ claims that they are using patient matching meaningfully. Specifically, he said LifeStance’s claim that they use patient matching in any scientific way was “baloney.”

“That’s all these companies are doing; they’re using the same old way that people would match, which isn’t anything scientific,” Bellwoar said of other systems.

Patient matching and other parts of the system

Patient matching could have a significant impact on value-based care models. Payers and providers struggle to establish payment models tied to positive outcomes. In some cases, providers and payers are teaming up to establish initial grounds for what the basic standards of certain behavioral health specialties are and how they should be measured.

Patient matching, especially when tied to measuring patient outcomes, could give payers and providers a means to reach an industry consensus on what is valuable and help providers have some degree of assurance that their care will be successful.

“In a fee-for-service model, I get paid the same regardless of which patients I’m treating and regardless of how effective I am in providing that treatment,” Hilton said. “But under a value-based care system, we’re incentivized to provide the best treatment. You can only determine what that is if you have data to back that up. And through the use of our technology, we can provide that [data], which actually enables providers to be paid more.”

Something like this is already at play at Refresh Mental Health. Highmark Health Plan, one of America’s largest Blue Cross Blue Shield insurers and a part of the larger payer-provider network Highmark, pays Refresh a 70% higher rate for its weekly bundled rate for using Outcome Referrals.

“That moves the needle because they see the outcomes. … They’re like, ‘I want to purchase that. I want to purchase really good care, not just average or mediocre care,'” Kraus said.

Patient matching is also at play in the tech stack of Talkspace Inc. (Nasdaq: TALK), one of the nation’s largest digital-only mental health care providers.

Talkspace uses a matching algorithm that looks at patient outcomes and assessments of providers’ competence. The company’s data includes over 6 billion words sent via asynchronous messages, 4 million completed psychological assessments, about 700,000 diagnoses and 1 million progress and therapy notes.

“Our matching model concurrently gathers client and therapist data and screens the therapists’ population to match the patient’s characteristics, clinical needs and preferences,” Talkspace said in its 2022 financial report. “Our machine learning technology also enables us to track the frequency and quality of clinical interactions, allowing us to provide a better therapist match should the patient request a new clinician.”

Talkspace declined to comment for this article. However, its CEO, Dr. Jon Cohen, teased additional information to come about the company’s use of machine learning and other forms of artificial intelligence to improve care and care outcomes earlier in the year.

Payers and providers alike — including Talkspace — have looked to machine learning-backed assessments of therapist quality and patient outcomes to create new standards for assessing care outcomes instead of assessing a condition’s episodic severity.

Plus, the faithful adoption of scientific assessment of patients and providers brings the long-sought objectivity that behavioral health lacks when compared to the rest of health care.

“We’re talking about the shift that the rest of medicine did just over 100 years ago when they invented the medical lab,” Kraus said. “It was only then that medicine became science as opposed to witchcraft.”