Inclusion of Telemedicine in Behavioral Health Quality Measures

One in four adults in the United States has a mental health or substance use disorder, yet >30% of these individuals are not receiving treatment (1, 2). The most frequently cited barriers to treatment for behavioral health conditions include cost, availability of services, transportation, inconvenience, and stigma (2, 3). Telemedicine modalities, such as video conferencing, telephone communication, and web-based technologies, present an opportunity to address barriers to accessing mental health and substance use treatment. A survey of primary care physicians rated telepsychiatry as the second most useful telemedicine service after primary care (4).

Telehealth is defined as “the use of electronic information and telecommunications technologies to support long-distance clinical health care, health-related education of both patients and professionals, public health, and health administration.” (5) Telemedicine refers specifically to remote clinical services, whereas telehealth can refer also to remote, nonclinical services such as provider training, administrative meetings, and continuing medical education, in addition to clinical services. Approximately two million primary care video consultations took place in the United States in 2015, a number that is expected to more than to five million by 2020, with total video consultations projected at 27 million (6) and the current COVID-19 pandemic likely accelerating this increase.

As of September 2018, 39 states and the District of Columbia had enacted laws governing private payer reimbursement for telemedicine services. However, only a few states require equal reimbursement for telemedicine and in-person visits (7). Moreover, claims codes or modifiers to bill telemedicine services are limited, especially for asynchronous modalities such as web-based services. Limited payment and claims codes may contribute to limited documentation of telemedicine services. Quality measure reporting by health care organizations helps monitor and improve health care quality, and detailed measurement specifications support fair comparison across entities. Typically, diagnoses and services that form measure components have been captured solely from in-person visits. With the growth in telemedicine and the greater use of quality measures in quality reporting programs, it is important to consider how telemedicine fits into quality measurement.

Evaluating the potential inclusion of telemedicine in quality measures is part of the responsibility of the measurement steward, the National Committee for Quality Assurance (NCQA), to update these measures periodically. The current study was conducted as part of NCQA’s measure-updating process. When a visit was required for diagnosing and managing a condition, we assessed whether a telemedicine service could count as the eligible “visit.” This study evaluated available evidence and used expert opinion (a standard process used by NCQA to update quality measures) to determine whether telemedicine services are comparable to in-person visits for diagnosing and managing behavioral health conditions found in seven Healthcare Effectiveness Data and Information Set (HEDIS) behavioral health quality measures (8) (descriptions can be found in an online supplement to this article): follow-up care for children prescribed attention-deficit hyperactivity disorder (ADHD) medication, use of first-line psychosocial care for children and adolescents on antipsychotics, antidepressant medication management, follow-up after hospitalization for mental illness, mental health care utilization, initiation and engagement of alcohol and other drug dependence treatment, and identification of services for alcohol and other drug abuse or dependence. These measures capture behavioral health care provided to children and adults with a diagnosis of ADHD, depression, bipolar disorder, schizophrenia, and substance use disorder, as well as to children on antipsychotic medications. We organized the evidence into two areas: diagnosis and management of behavioral health conditions.

Methods

We conducted a rapid literature search in three electronic databases: PubMed, Google Scholar, and the Cochrane Database of Systematic Reviews. We used the following keywords: telemedicine (or tele-rehabilitation, telepsychiatry); ADHD, depression, bipolar disorder, schizophrenia, psychotic disorder, alcohol abuse, or drug abuse; diagnosis (or assessment, evaluation) or treatment (or therapy, rehabilitation, psychosocial care); and children or adults. We checked the reference list of studies to identify additional studies. Studies published in English were included, with no restrictions on the countries in which these studies were conducted.

Of 79 studies identified, we excluded 47 because patients receiving telemedicine services did not meet our age requirements (28 studies with patient populations that did not match the HEDIS measure age requirements) or behavioral health condition/diagnosis requirements (19 studies that did not describe specific conditions or diagnoses) specified in the measures (descriptions of these measures can be found in the online supplement). The 32 remaining studies were included in this study; they were published between 1997 and 2016 and addressed synchronous (telephone and video conferencing) and asynchronous (online assessments, and computer-assisted therapy) telemedicine modalities (studies are listed in the online supplement). Of these articles, 11 were systematic reviews, 13 randomized controlled trials (RCTs), two experimental studies, two pre-post studies, and four program evaluations. The studies covered both developing and developed countries, with most being from the United States.

We also sought input from expert panels on the appropriateness of potential telemedicine inclusion in quality measures. Feedback from external expert panels through consensus-building discussions is a key component of measure development (9). We built consensus by identifying major themes that emerged from discussions with expert panels, as well as through voting by committee. Input was solicited from NCQA expert panels on telemedicine (11 members), behavioral health (11 members), and coding (seven members) and from the Committee on Performance Measurement (CPM) (17 members). The telehealth panel was assembled at the beginning of this study, and the others were existing panels of long standing. The panels and committee comprised external experts, including clinicians (i.e., general medical and behavioral health providers), researchers, consumers, and telemedicine and coding experts.

Each panel/committee met three times between August 2016 and May 2017 to discuss literature review results and provide input on whether it was clinically appropriate (on the basis of literature review findings) and feasible (according to availability of telemedicine claims codes) to include telemedicine modalities in HEDIS measures. At each meeting, we presented findings from the literature review and asked the experts to discuss the appropriateness and feasibility of including specific telemedicine modalities in the measures. The three expert panels discussed the issues, and we summarized the major themes at the end of the meetings. We also took meeting minutes and reviewed them after the meetings for additional details of identified themes (see Results). The CPM voted on whether to approve inclusion of specific telemedicine modalities in HEDIS behavioral health measures.

Results

Eleven studies (three systematic reviews, three RCTs, two experimental studies, and three program evaluations, listed under “Diagnosis of Behavioral Health Conditions via Telemedicine” in Table 2 in the online supplement) supported the use of specific telemedicine modalities for diagnosing behavioral health conditions. Several studies reported that telemedicine was as effective as in-person visits for diagnosis (10–15) (see additional references in online supplement); for example, two systematic reviews found good intermethod reliability in diagnosing psychiatric disorders among children and adults, and clinicians could identify similar patient symptoms in both telemedicine and in-person visits (4, 12).

Evidence varied by diagnosis and population. Eight studies (mostly systematic reviews and RCTs) supported video conferencing for diagnosing ADHD, depression, anxiety, bipolar disorder, schizophrenia, and substance use disorders among children and adults (see online supplement) (10–12, 14, 15). One RCT supported telephone visits for diagnosing depression among children and adults (13). Two program evaluation studies supported telephone visits and asynchronous telemedicine modalities for diagnosing substance use disorders in adults (see online supplement). We found no studies on using telephone visits or online assessments to diagnose bipolar disorder and schizophrenia in children, adolescents, and adults or to diagnose substance use disorders in children and adolescents. We found no empirical studies reporting that telemedicine services were less effective than in-person visits for diagnosing behavioral health conditions.

We found more studies on the use of telemedicine for treatment than for diagnosing behavioral health conditions. Twenty-one studies (eight systematic reviews, 10 RCTs, two pre-post studies, and one program evaluation, listed under “Treatment of Behavioral Health Conditions via Telemedicine” in Table 2 in the online supplement) supported use of telemedicine modalities for treating behavioral health conditions (telemedicine services were as effective as in-person visits in improving outcomes such as clinical symptoms and quality of life) (see online supplement). Eleven studies (mostly systematic reviews and RCTs) supported video conferencing for managing ADHD, depression, anxiety, bipolar disorder, schizophrenia, panic disorder, posttraumatic stress disorder, and substance use disorders among children and adults and for managing developmental disorders in children (see online supplement). Seven systematic review or RCTs reported that telephone visits were effective in managing depression in children and adults and substance use disorders in adults (see online supplement).

Several studies found that telephone consultations were as effective as or better than in-person visits in reducing substance use and in improving clinical symptoms of depression, adherence to medication treatment, and quality of life. Three systematic reviews and two RCTs suggested that asynchronous telemedicine modalities (e.g., computer-assisted therapy) were effective for managing depression in children and adults and substance use disorders in adults (see online supplement). We found no studies using telephone visits or web-interface programs to manage bipolar disorder, schizophrenia, or substance use disorders among children and adolescents or bipolar disorder and schizophrenia among adults. We found no empirical studies concluding that telemedicine services were less effective than in-person visits for managing behavioral health conditions.

The expert panels and committee from which we sought feedback supported inclusion of telemedicine in HEDIS behavioral health measures when there was evidence that the use of telemedicine was effective in diagnosing and managing behavioral health conditions. Many experts stated that allowing telemedicine services in quality measures could expand the utilization and accessibility of these services, particularly for people living in areas with provider shortages. The experts recommended that evidence on using telephone visits or web-interface programs to diagnose and manage substance use disorders among adults could be extrapolated to children and adolescents. The expert panels did not support telephone visits or web-interface programs to diagnose or manage bipolar disorder or schizophrenia among children, adolescents, and adults, citing a lack of evidence for any age group. Some experts expressed concerns about availability of reimbursement for telemedicine services. Providers might not code telemedicine services without a clear payment mechanism, which might result in variations in claims data available to capture telemedicine services in measures.

Because telemedicine use was informed by the literature review, supported by expert panels, and approved by the CPM, NCQA added it as a modality that providers can use to diagnose or manage conditions relevant to the seven behavioral health measures in HEDIS, beginning with the reporting year 2018. Video conferencing (using the GT claims or 95 code, denoting service delivered “via interactive audio and video telecommunications systems”) was added to all seven measures; telephone visit (using six claims codes for telephone visits) was added to depression and substance use disorder measures; and online assessment modalities (using two online assessment claims codes) were added to substance use disorder measures for children and adults (measure-specific additions are available in Table 1 of the online supplement). Specific telemedicine claims codes or modifiers were added to the measure denominator, numerator, or both where a visit was required. NCQA will update inclusion of telemedicine modalities as new research evidence develops.

Discussion

The evidence examined here suggests that the use of specific telemedicine modalities (video conferencing, telephone call, or web-based services) is as effective as in-person services for diagnosing and managing some behavioral health conditions among children and adults. We found more evidence that supports telemedicine use for managing behavioral health conditions than for diagnosing these conditions. Video conferencing was supported by more studies than were telephone or web-based services. We found no empirical studies indicating that telemedicine is less effective than in-person care.

Expert panels supported inclusion of select telemedicine services to diagnose and manage behavioral health conditions in most of the seven HEDIS behavioral health measures. Codes for documenting telemedicine services are limited, and reimbursement may be lower than for in-person services or may be unavailable in some states. A recent study using the Health Care Cost Institute database of claims from UnitedHealthcare, Humana, and Aetna for ≥50 million beneficiaries found that commercial insurance and Medicare Advantage plans reimburse telemedicine services for mental health at half the fee of in-person services (7). Telemedicine services, indicated by the telemedicine modifier GT, accounted for only 3% of all mental health and substance use disorder service claims from 2009 to 2013 (10). Limited payment and few available codes may contribute to a lower documentation of telemedicine services. Inclusion of telemedicine services in quality measures may have only limited impact on performance rates initially but may encourage better documentation and increase professional and policy maker awareness of telemedicine services in behavioral health care. Increased awareness could ultimately encourage broad access to telemedicine treatment modalities.

Research on the effectiveness of telemedicine services remains limited, which we mitigated by seeking advice from an array of experts. Future research should examine the effectiveness of using telephone visits or web-based services for diagnosing and managing bipolar disorder and schizophrenia among children and adults and substance use disorders among adolescents.

Conclusions

We found that specific telemedicine services are similar to or better than in-person services for diagnosing and managing some behavioral health conditions. Video conferencing was supported by more evidence than were telephone or web-based services for diagnosis and management of these conditions. The inclusion of telemedicine services in national quality measures may facilitate broad adoption of telemedicine services and increase access to behavioral health services.