Engaging Suicidal Patients in Mental Health Treatment

Leveraging Behavioral Economics and Implementation Science to Engage Suicidal Patients in Mental Health Treatment

The investigators will identify characteristics of suicidal patients who do or do not attend a first mental health visit following referral using administrative data. Then, the investigators will apply established approaches to contextual inquiry to identify barriers and facilitators to mental health treatment attendance for individuals at risk of suicide. Using established procedures from implementation science and behavioral economics, the investigators will then leverage the insights gleaned from Aims 1 and 2, relevant theories and frameworks, and the extant literature to develop preliminary strategies to support attendance at first mental health visit. Strategies will be developed in collaboration with a team of experts in suicide, implementation science, and behavioral economics. These preliminary strategies will then be iteratively tested and refined. The investigators also will assess putative mechanism using behavioral tasks and self-report tools.

In 2017, there were more than 800,000 deaths by suicide worldwide. Patients at high risk for suicide are less likely to die by suicide if they engage in psychotherapy. However, despite the development of evidence-based practices (EBPs) for suicide prevention, rates of suicide in the U.S. have increased by approximately 30% over the past two decades. One way to lower these rates would be to increase treatment initiation among those at risk. Suicidal individuals have difficulty engaging in mental health services, yet no studies have systematically developed and tested strategies to increase treatment initiation for suicidal patients. Nearly two-thirds of people who die by suicide interact with a primary care clinician in the year prior to death, making primary care an optimal setting in which to identify individuals at risk for suicide and connect them to mental health care. Yet only half of those referred by primary care providers attend an initial mental health visit, even when referred to care within the same practice. Various strategies to increase attendance at first appointment, including reminder calls and texts, motivational and informational interventions, and case management, have demonstrated small to moderate effects. Even when these strategies are implemented, approximately 40% of patients do not initiate treatment, suggesting that additional work is needed. The primary objective of this study is to develop acceptable, feasible, low-cost, and effective strategies that increase patients' treatment initiation (i.e., attendance at a first mental health visit) following identification of suicide risk in primary care. The investigators will partner with a large, diverse health system to rapidly prototype and test promising strategies to achieve this objective. Rapid prototyping involves a series of rigorous tests to optimize operations in the early-study stages. Industries outside of health care commonly use this approach to learn quickly and "de-risk" decision-making on a short timeline prior to a large roll-out. To maximize generalizability, the investigators will use rapid prototyping to develop strategies for increasing attendance in both collaborative care and outpatient specialty mental health. The strategies the investigators develop and test will be informed by behavioral economics and implementation science methods, leveraging the University of Pennsylvania's P50 ALACRITY center. In accordance with an experimental medicine approach to behavior change, the investigators will also identify and target mechanisms of action that impede treatment attendance. The specific aims are to: Aim 1. Identify characteristics of suicidal patients who do or do not attend a first mental health visit following referral. The sample will include adults 18 years and older, including Medicare, Medicaid, and commercial insurance enrollees, who are referred for mental health services after an initial screening in primary care identified suicidal ideation. Using medical records, the investigators will compare characteristics of patients who initiate treatment to those patients who do not. When possible, the investigators will utilize insurance claims data to identify any mental health services utilization that occurred outside the health system. The investigators also will explore data from patients who report suicidal ideation in primary care and are not referred to mental health services. Aim 2. Apply established approaches to contextual inquiry to identify barriers and facilitators to mental health treatment attendance for individuals at risk of suicide. The investigators will use direct observation and brief interviews with key stakeholders, including leaders (n = 12), primary care providers (n = 12), behavioral health providers (n = 12), mental health intake coordinators (n = 5), and patients who do (n = 12) and do not (n = 12) attend a first mental health visit following referral from primary care to understand key structural and behavioral barriers, facilitators, and mechanisms to engaging patients at risk for suicide in mental health services. The investigators also will use behavioral tasks and self-report measures to assess putative mechanisms of action in patient stakeholders. Aim 3. Rapidly prototype and test strategies to optimize engagement. Using established procedures from implementation science and behavioral economics, the investigators will leverage the insights gleaned from Aims 1 and 2, relevant theories and frameworks (e.g., EAST, Science of Behavior Change), and the extant literature to develop preliminary strategies to support attendance at first mental health visit. Based on the literature and their previous work, the investigators anticipate that strategies that target temporal discounting (e.g., incentives) and foster perceived social support (e.g., Caring Contacts) will be needed. Strategies will be developed in collaboration with a team of experts in suicide, implementation science, and behavioral economics. The investigators will then iteratively test and refine these preliminary strategies. Throughout this process, the investigators expect to uncover additional barriers and facilitators that will allow us to further refine and optimize implementation strategies. The investigators will assess putative mechanism using behavioral tasks and self-report tools. The primary output will be a menu of promising and feasible implementation strategies that directly address barriers to the initiation of mental health services for patients at risk of suicide. These strategies will be tested in a subsequent multisite R01. The long-term goal is to reduce deaths by suicide by increasing engagement in evidence-based mental health services for individuals at risk of suicide.

We anticipate that approximately 25 patients will receive an engagement strategy (~5 patients per strategy; only 1 strategy per patient). The exact strategies will be developed during the course of this R21 and cannot yet be specified. This is a non-control, pilot feasibility trial.

We will generate strategies for facilitating treatment initiation that address barriers and leverage facilitators, in partnership with a team of experts in behavioral science and implementation science. We will favor simple, established strategies with freely available resources (e.g., Caring Contacts templates) to promote scalability, but will also consider other potential strategies as indicated. We will explore both low- (e.g., manualized text messaging) and high-tech (e.g., electronic health record integrated patient messaging) strategies to maximize relevance across contexts. Next, prototype-strategy designs will be validated through rapid-cycle tests in this real world context and in a manner that is inexpensive and allows us to fail fast and learn quickly. After testing a given implementation strategy, the team will debrief and decide whether to refine or adjust it.

A reliable and valid 4-item tool to assess perceptions of the acceptability of engagement strategies. Each item is rated using a 5-point ordinal response options, ranging from 1= "completely disagree" to 5="completely agree" and scores are averaged across the four items to yield a mean scale score. Higher scores indicate greater acceptability.

A reliable and valid 4-item tool to assess perceptions of the appropriateness of engagement strategies to the context. Each item is rated using a 5-point ordinal response options, ranging from 1= "completely disagree" to 5="completely agree" and scores are averaged across the four items to yield a mean scale score. Higher scores indicate greater appropriateness.

A reliable and valid 4-item tool to assess perceptions of the feasibility of engagement strategies. Each item is rated using a 5-point ordinal response options, ranging from 1= "completely disagree" to 5="completely agree" and scores are averaged across the four items to yield a mean scale score. Higher scores indicate greater feasibility.

Derived from the electronic health record, defined as patient did or did not attend first mental health visit following referral. The tool uses a 5-point ordinal response option, ranking from "completely disagree" to "completely agree". The higher the score, the better the outcome.

Inclusion Criteria: 18 years and older Elevated suicidal ideation per item 9 of the Patient Health Questionnaire (PHQ item score ≥ 1) completed during a primary care visit Able to read and understand English Exclusion Criteria: Current psychotic episode requiring emergency services and/or precluding ability to provide informed consent Documented diagnosis of dementia in the past 2 years. The rationale for this criteria is that these individuals may have difficulty understanding the study information provided to them, including how to opt out of the study No patients will be excluded on the basis of sex, race, or ethnicity Primary care provider notes that participation is not indicated Already received a study engagement strategy following a different primary care visit.

All study personnel will have Patient Oriented Research certification from the University of Pennsylvania. Interview personnel meeting with primary care providers, behavioral health providers, leaders, and patients will hold a Bachelor's, Master's or doctorate degree. The PI will oversee the additional training and ongoing review of interviewers. Interviewers must display research reliability in collecting data and will work under the close supervision of the PI.

Khazanov GK, Jager-Hyman S, Harrison J, Candon M, Buttenheim A, Pieri MF, Oslin DW, Wolk CB. Leveraging behavioral economics and implementation science to engage patients at risk for suicide in mental health treatment: a pilot study protocol. Pilot Feasibility Stud. 2022 Aug 13;8(1):181. doi: 10.1186/s40814-022-01131-y.