When Mental Health Center of Denver first launched an illness management program, it began by sending treatment data back to the patients, detailing their progress toward recovery, using as a basis the information contained on patient and therapist surveys. Patients received reports every three months, and the therapy center hoped the additional information would spark new conversations between patients and therapists.
Later, the center added a harm reduction initiative—it included questions intended to assess patients’ risk for suicide. Their answers would alert therapists if a patient might require intervention. Over time, other behavior risk factors were added to the exercise, such as whether a patient was arrested, lost a job, had substance abuse issues, committed a violent assault, tried to commit suicide or was hospitalized.
Now, the center is further enlarging its efforts to use such “non-clinical” information to improve care for patients, says Wesley Williams, vice president and CIO at the Mental Health Center of Denver.
Grow your practice with our weekly newsletter.
Healthcare organizations of all stripes are beginning to take steps to incorporate social determinants of health into patient assessments. It’s an acknowledgement that a variety of factors feed into the physical health of a patient. Clinicians see mounting evidence that a holistic view of patients is needed to first make them healthy and then to keep them that way.
Capturing and sharing such patient-specific, non-clinical information is new to providers, and current information technology is slowly adapting to new calls for clinicians to have access to this information. Many believe such patient-specific information will help improve treatment and achieve cost-effective results.
Historically, behavioral health providers have had little integration of their records with those of medical providers. That is changing, but progress is slow, says Mike Valentine, CEO at Netsmart Technology, a company that develops IT solutions for behavioral health providers.
Five years ago, the only communication of patient data between mental health and medical providers was achieved through faxes and phone calls to administer referrals. Now, thousands of mental health providers are electronically communicating with medical providers via continuity of care documents, he adds.
Some are striving to do more, and Mental Health Center of Denver is a case in point. After its initial foray into collecting data on patients, the next step was to gather the data in an electronic health record system, which was the start of moving toward the use of behavioral analytics to change patient behavior. “We are stratifying the populations that we serve to be able to play closer attention to them,” Williams says.
Over time, the Mental Health Center of Denver started licensing its recovery measure tools to other mental health centers. Then, it used in-house data analytics to develop tools to aid therapists and case managers in assessing a patient’s progress in mental health recovery by rating patient self-assessments.
Netsmart learned of the work that Mental Health Center of Denver has done and two years ago contracted to resell the center’s products, which now are part of the vendor’s portfolio. “Analytics offer the radar that providers need to track gaps in care,” Williams adds. “We need to work with other providers and create a longitudinal view of patients. Prior to this, we had no visibility in the outside world.”
Rise of connectivity
Interoperability of electronic health records systems is improving, giving rise to increased incorporation of social determinants of health with electronic health records, and there’s growing hope that such exchange will improve in the coming years.
For example, Netsmart has its own health information exchange that connects to the Sequoia Project’s CareQuality HIE, enabling data exchange among providers using EHRs from Epic, Cerner and Allscripts, and building out to more trading partners, helping to enable more behavioral information to be included in providers’ EHRs.
Better connectivity between mental health and medical health providers is critical to improving the lives of at-risk patients, Valentine contends, adding that while analytics better enable providers to track gaps in care, it is up to providers to work with each other and create longitudinal views of patient records.
Valentine notes that a study several years ago found that those with a serious mental illness die 20 years earlier because of physical illnesses, and when a patient has a mental health issue and a chronic medical condition, treatment costs are two to five times higher than just having a chronic condition because of the lack of coordination between mental health and medical providers.
Winds of change
In large part, pervasive use of health information technology has been the exception in mental health care. Change is coming, as data analytics capabilities are maturing, enabling deep-dive looks into patient lives using social determinants of health, and structured and unstructured data, says Elizabeth Marshall, director of clinical analytics at Linguamatics. The vendor sells natural language processing software to life sciences and healthcare providers.
Social determinants of health are factors that contribute to a person’s current state of health, according to the Centers for Disease Control and Prevention. These factors may be biological, socioeconomic, psychosocial, behavioral or social in nature. The CDC classifies social determinants into these categories—biology and genetics; individual behavior; social environment; physical environment; and health services.
Analyzing data using natural language processing helps to use social determinants of health to identify the core factor affecting a patient and prospects for a better outcome, Marshall explains. A patient may have diabetes and be taking medications, but physicians often don’t know important social determinants of health that could inform the level of support a patient will need for a successful outcome.
For instance, if the patient’s income is at the poverty level, he or she may eat frequently at fast-food restaurants, which may indicate that the patient needs to see a dietician or receive social services to afford healthier meals.
“The concept of nature vs. nurture has been around since the late 1600s, yet the use of social determinants is just coming into play now,” Marshall says. “Everyone is concerned with capturing data, but we need to use the data to optimally improve patient outcomes.”
Data and analytics can help improve care for patients and spot those at risk for life-threatening events. For example, the Mental Health Center of Denver has adopted the Columbia-Suicide Severity Rating Scale of Columbia University, a free product that uses a series of simple plain-language questions to identify those who are at risk of suicide, measuring the severity of risk and the level of help that the person needs. Answers given by patients are compared with the answers they gave when they were last assessed, generating a new score and enabling trending over time to see how a patient is faring.
To incent patients to participate in the Columbia project, the mental health center offered a $5 grocery store coupon for those who signed up for a personal health record account and completed a self-assessment; about 1,500 patients actually got accounts. PHRs have not been overly useful to patients, Williams says, so the participation level was a pleasant surprise, and it enabled the center to get more data to support analytics and provide more feedback on recovery progress to participating patients.
Still, there is unfinished work for the health center.
With help from Netsmart, Mental Health Center of Denver is using natural language processing technology to sift through years of progress notes and treatment plans for “phrases of relevance” and to capture these phrases, Williams explains. For instance, a patient may have told a physician about suicidal thoughts, but that information may be buried in notes in the patient’s records, and may not be communicated to other clinicians who take on the case. NLP technology can find key phrases of suicidal tendency in patients’ records and facilitate communication of that risk to the care team.
Williams has hoped to be able to expand to use of predictive modeling analytics to better pinpoint patients at highest risk and alert clinicians.
Social determinants of health, and the use of IT to enable behavioral health, are gaining wider attention from new treatment approaches. For example, TelaCare, a telemedicine vendor, recently added behavioral health services that enable individuals to speak to a U.S -licensed counselor. The consultations can be done through video conferencing, text messaging or mobile apps. The behavioral health services were made available this September at no additional charge to clients and with no out-of-pocket expense for an individual receiving a consultation.
Services are offered for such issues as depression, anxiety, relationship challenges and other stresses of life. “Every year, one in four people will deal with a mental health disorder, yet less than half of these individuals will actually receive treatment,” says Larry Jones, CEO at TelaCare.
And there’s growing recognition of the need to improve IT offerings to better support behavioral health and non-clinical influences on a patient’s well-being. For example, the Massachusetts eHealth Institute at the Massachusetts Technology Collaborative in August awarded nearly $200,000 to four electronic health records vendors to aid behavioral health providers in electronically submitting reportable data to the state’s Children’s Behavioral Health Initiative, which works to ensure children with mental health challenges get treatment.
The funding will enable vendors to build new electronic interfaces connecting 12 behavioral health providers that will submit reports electronically, replacing a manual process currently used for about 40,000 reports that they submit to the state each year.
The Children’s Behavioral Health Initiative is a court-mandated program requiring behavioral health providers serving children younger than age 21 to submit patient data so the state can better assess the needs of the patients.
In Massachusetts, behavioral health providers have had to gather patient data via an assessment tool that was tracked in their electronic health record system, but providers then had to manually re-enter the data into a state information system, a process that takes two hours of staff time per patient per quarter and is vulnerable to data submission errors.
“From what I can see already, this upgrade will have a tremendous benefit to how we handle patient records,” says Scott Turton, director of health information systems at Gosnold on Cape Cod, an addiction treatment center. “This new interface is much more streamlined, as the data is pulled directly from the patient’s existing electronic record, compiled by this new interface, and then submitted from the new interface directly to the state.”