Improving Adherence to Mental Health Treatment in a Low-Income Clinic

The increasing prevalence of mental illness in the United States presents significant challenges for primary care providers particularly in low-income settings. Integrated Behavioral Health (IBH) programs have resulted in improved general health for low-income participants; however, managing appointment adherence, in which the patient attends appointment as scheduled, is particularly challenging. The purpose of this pilot project was to implement bundled interventions at a low-income primary care clinic in a Mid-Atlantic state to improve patient adherence to behavioral health treatment using a collaborative, community-based approach. Interventions were delivered in a bundle format and included patient education, warm patient handoffs, and follow-up phone calls by behavioral health counselors. In the 4-month postintervention time frame, the average number of patients who did not come to their appointment decreased by 60%, and the average number of patients who canceled decreased by 15%. These differences were significant (χ2 = 9.263, df = 2, p < .01). This pilot project showed that patients who became engaged as active participants in a bundle of interventions were more likely to keep their appointments.


List of
Total number of seen, no-show, and cancelled visits v Individuals benefit from evidence-based, collaborative Integrated Behavioral Health (IBH) care; however, many low-income adults and families do not receive beneficial mental health treatment (Santiago, Kaltman, & Miranda, 2013). Nationally, the unmet need for mental health services increased from 4.3 million in 1997 to 7.2 million in 2010 (Roll, Kennedy, Tran, & Howell 2013). The Behavioral Health Barometer, Virginia (2014) reports that: 1) 59% of adults reported improved functioning from treatment received through the Virginia public mental health system as opposed to 70% nationwide, and 2) among adults served in Virginia's public mental health system in 2013, 60.5% of those aged 18-20, 53.4% of those aged 21-64, and 89.1% of those aged 65 or older were not in the labor force.

List of Figures
The need for behavioral health services exists locally in Rockingham County and Harrisonburg, Virginia. A Community Needs Assessment acknowledged that behavioral health related hospitalizations are an important indicator of community health status (Community Health Needs Assessment, 2015). In the assessment, 703 per 100,000 patients discharged from the local hospital had a behavioral health diagnosis as compared to 680 per 100,000 statewide. The leading diagnoses for these discharges were affective psychoses (55%), schizophrenic disorders (13%), and depressive disorders (6%) (Community Health Needs Assessment, 2015).
The Affordable Care Act in 2010 emphasized and promoted the use of integrated primary and behavioral health care services. Primary care clinics that are integrated provide behavioral health services in addition to primary care. Integrated programs help organizations improve outcomes by increasing access to mental health services and improving collaboration between specialties (National Committee for Quality Assurance, 2016). Accustomed to collaborating with various health care systems to meet the health needs of underserved clients, low-income clinics (free, low-cost, and sliding scale) are well suited for integrated programs.
This pilot project enabled a low-income clinic to test and evaluate the use of a bundle of interventions to improve adherence to mental health treatment. Pilot studies are a first step in the development of complex interventions because they help avoid duplication of efforts in assessing the feasibility of interventions for future research (Thabane, 2010). Results from this pilot can help facilitate the implementation of bundled interventions in a larger organization or inform the design of future research projects.

Problem Statement:
A new Integrated Behavioral Health program at a low-income clinic had high rates of missed appointments (43%). Over half of the patients counseled never returned for a second session. To promote the mental health of low-income individuals, new interventions were needed to improve behavioral health appointment adherence of clinic patients.

Specific Aims:
The aim of this pilot project was to demonstrate the effectiveness of bundled interventions at a low-income clinic to increase appointment adherence leading to improved overall health, decreased costs, and increased access to mental health services.
The project objectives were to: 1. Increase the number of patients receiving mental health treatment.
2. Reduce the number of missed appointments from cancellations and no-shows.
3. Decrease patient dropout rates after the initial appointment.
4. Decrease the number of ED visits of patients in mental health treatment.

Literature Review:
The databases MEDLINE, CINAHL, and PsycINFO were searched using the following key words: mental health, behavioral health, integrated behavioral health, integrated care, collaborative care, low-income, adherence, appointment compliance, and interventions. The terms "mental health services" and "behavioral health" were used in searches in an attempt to ensure that all documents that examined mental health needs were located. The terms "integrated", "embedded", and "collaborative care" were included since these words are used interchangeably in behavioral health literature.
Appointment adherence is particularly challenging in the long-term management of both chronic and episodic disorders since individuals with serious mental illness are more likely to miss appointments and show poor compliance with the prescribed plan of care (Defife al., 2010). Using scales to determine the severity of mental disorder and level of social disorganization, Killaspy, Banerjee, King and Lloyd (2000) found that those who miss psychiatric follow-up outpatient appointments are more unwell, more poorly socially functioning, and have a greater chance of dropout from clinic contact and subsequent hospital admission than those who attend. They concluded that appointment adherence is especially important for those with severe mental illness, since those who drop out after their first contact may experience significant deterioration in their mental state. Primary care patients that have a high propensity to no-show will have suboptimal clinical outcomes and higher rates of acute care utilization compared to those with a lower propensity to no-show (Hwang et al., 2015).
Clinicians who use a bundle or combination of interventions that utilize available resources appear to have higher rates of success. A literature review by Lefforge, Donohue, and Strada (2007) demonstrated that attendance improvement interventions were shown to be particularly effective when they employed multiple, empirically derived intervention strategies. Interventions they reviewed included a combination of transportation vouchers, orientations, letters, home visits, patient contracts, and prizes.
Bundles appear to have a greater impact than single interventions but no research points to one particular bundle or specific combination of interventions that work well together.
Research has addressed the importance of improving mental health literacy levels through education and insight. Mental health literacy embodies having sufficient knowledge to aid patients in the recognition, management and prevention of mental disorders (Jorm, 2012 improves attitudes and willingness to be treated. Nose, Barbui, and Tansella (2003) revealed in a systematic review that in 13 of 81 (16%) studies, insight (understanding about treatment and medication) had a positive association with adherence. Lack of insight was associated with non-adherence in 14 of the 81 studies (17%). Poor adherence with mental health referrals in the elderly was associated with a lack of perceived need (Mojtabai, 2005). Bonabi et al. (2016) concluded that mental health literacy, positive attitudes to help seeking, and perceived need for treatment, significantly predict the use of psychotherapy over time.
Patients with early follow-up (a follow-up phone call or visit with a counselor or care manager within three weeks of treatment initiation) were less likely to drop out of behavioral health care and more likely to receive appropriate pharmacotherapy (Bauer et al., 2011). In an underserved area, Clouse, Williams, & Harmon (2016) found that telephone engagement by a Psychiatric Nurse Practitioner which included an introduction and discussion of the behavioral health treatment plan reduced the rate of no-show rates from 27% the previous year to 20% in a three-month period.
The goal of warm handoffs (immediate, in-person referrals between primary care provider and mental health specialist) is to ensure that individuals will feel comfortable and not judged by healthcare providers during visits (Manoleas, 2008). Davis, Moore, Meyers, Mathews, and Zerth (2016) concluded that as little as five minutes of contact with a primary care mental health specialist led to a statistically significant increase in the likelihood of completing a referral when compared to the absence of contact with a provider. Horevitz (2013) however, found that not all warm handoff referrals are experienced as "warm" to patients, and that the strength of the patient-provider relationship is a key component affecting patients' experience of the referral, and subsequent decision to engage in depression treatment.
Theoretical Framework: The project followed the Standards for Quality Improvement Reporting Excellence (SQUIRE) guidelines to provide a framework and guide for project reporting.
SQUIRE guidelines are intended for reports that describe system level work to improve the quality, safety, and value of healthcare. It guides the use of methods to establish that observed outcomes were due to the intervention(s) (Standards for Quality Improvement Reporting Excellence, 2015).
In addition, the Plan-Do-Study-Act (PDSA) was incorporated for evaluation of the interventions. It is derived from the Deming Quality Model and has been effectively applied in health care settings, including low-income clinics. It uses easily adaptable techniques to analyze data and measure compliance to expectations that have already been proven to improve patient outcomes (Baker, 2014). The four stages of the PDSA cycle (see Appendix I) can be repeated as part of a cycle of continual improvement. The use of the PDSA model encourages learning, reflection and validation throughout implementation of the project (Institute for Healthcare Improvement, 2016). It was chosen as a framework for this project after proving to be successful with other quality improvement approaches in this organization.
The theory of planned behavior guided the intervention focused on education.
This theory, developed by Ajzen, (1991) links beliefs and behavior and provides useful information for the development of communication strategies (See Appendix II). It is frequently used in evaluation studies. Ajzen believes that the best predictor of behavior is intention. Intention is the cognitive representation of a person's readiness to perform a given behavior, and it is considered to be the immediate antecedent of behavior. Intention is determined by the following three things (Ajzen, 1991): 1. Attitude: Only specific attitudes toward the behavior in question can be expected to predict that behavior.

Subjective norm: an individual's perception about the particular behavior,
which is influenced by the judgment of significant others (e.g., parents, spouse, friends, teachers).

Perceived behavioral control: Influences intentions. Perceived behavioral
control refers to people's perceptions of their ability to perform a given behavior.
The Theory of Planned Behavior was utilized in this project to design interventions that target mental health program adherence. Education and follow-up conversations with patients on the benefits of mental health services and integrated care can promote positive attitudes and improve motivation to pursue healthy behaviors. In this project, brochures, posters, and discussions with clinic staff were designed to portray IBH care as a positive measure that contributes to overall well-being (see Appendix III, IV). In addition, ideally, the discussion that occurs during the provider/patient follow-up phone call will raise awareness that subjective norms are favorable towards counseling.
The belief that mental health problems are a sign of weakness and treatment socially unacceptable will be negated and patients will believe that they can improve health by attending sessions.

Context:
The setting of the project was a low-income clinic in rural, southeastern United

Study Population:
The researcher anticipated that a minimum of 25 client records would be reviewed for the pilot project. The population was low-income adults over age 18 but less than 65 with mental health needs who met eligibility criteria. A retrospective chart review proposed to look at six months of information on all patients scheduled for one or more counseling sessions. The counseling session did not need to take place for the records to be included. Patients who spoke a language other than English or Spanish were excluded from the study, as the educational materials were only available in English and Spanish.

Interventions:
The project design was a longitudinal, descriptive, pilot project. A collaborative project team consisting of the researcher, administrator, nurses, nurse practitioners, counselors, and social worker convened and developed interventions based on: 1) clinic data that showed deficiencies in the program (% missed appointments, # appointments/patient) 2) research on best methods for evidence-based practice, and 3) the motivation and intent to reschedule and continue with treatment 3) concerns related to treatment and/or social stigma. This step was intended to identify stressors and increase motivation to adhere to a treatment plan.
A pre and posttest measurement of data was chosen as the approach used to establish whether the improved adherence was due to the bundle of education, handoffs, and phone calls. Knowing it would not be possible to determine the effect of individual interventions, the study looked at the impact of intervention results collectively using Chi Square for statistical analysis.

Measures:
The researcher, with the assistance of clinic staff and volunteers, was responsible for gathering and analyzing the data. As a volunteer and former employee of the clinic, the researcher had the support of the Board and administration to implement this project.
Data was gathered and measured over a six-month period. A retrospective chart review provided pre-intervention baseline data followed by post-intervention data.
Sources of data included: 1. Clinic Electronic Health Record (EHR) for: i) Number of mental health counseling sessions.
ii) Number of patients receiving counseling.
iii) Number of missed appointments (cancelled and "no-show").
iv) Dropout rates (number of actual visits per patient).
2. Local Emergency Department: It was planned that the number of ED visits of those patients have attended counseling would be compared before and after interventions to determine if there was a relationship between the number of ED visits and counseling sessions resulting in improved health plus cost-savings to the community. However, due to the inability to obtain data from the local ED, this data was not collected and/or analyzed.

Analysis:
Data analysis included information collected from the clinic's EHR. A retrospective chart review (pre-intervention) provided baseline data and consisted of visit information on all patients who were scheduled for one or more counseling sessions for two months between the dates of 01/01/17 through 2/28/17. The second time frame for data collection (post-intervention) lasted four months and was from 03/01/17 through 06/30/17. It consisted of the following data: 1. Number of mental health counseling sessions. This information was downloaded from the EHR using a "mental health chart notes report". The researcher, who has licensed access to the EHR and patient data, downloaded the report and manually entered it into an excel spreadsheet titled, Behavioral Health Appointment Summary (see Appendix VIII).

Number of missed appointments. This information was not available in an EHR
report. The number of canceled and "no-show" appointments were counted manually by the researcher and entered in the spreadsheet, Behavioral Health Appointment Summary, that had columns for: 1) date of scheduled appointment, 2) whether the missed appointment was a no-show or cancelled, 3) reason provided for the missed appointment. At the initiation of the interventions, this data was recorded weekly in the Excel spreadsheet based on the missed mental health appointments for that week. No patient information was included in the spreadsheet.

Quality:
To ensure quality of the analysis, as much information as possible was downloaded directly from the EHR. Data that was manually entered was cross checked three times by the researcher. A SPSS and quantitative data consultant reviewed excel data and the accuracy of analysis.

Ethical Considerations:
This pilot project held minimal risk for the patient and health care workers.
Patients in the project received three bundled interventions of education on the benefits of IBH care, warm handoff referrals, and a follow up phone call from a provider. The risks of harm or discomfort anticipated in the proposed research were not greater, considering probability and magnitude, than those ordinarily encountered in daily life. Identifiable, private information was not collected on any patient and no names were included on the data reports. Instead, a unique identifier assigned by the researcher was used. The code for linking patient names with the unique identifier was stored on a private server accessible only by the researcher. These Excel spreadsheets were safely stored on the clinic's private server in a drive accessible only to the Executive Director, Office Manager, Accountant, and the researcher.
Approval for the project was obtained from the Institutional Review Boards (IRB) at the local hospital and university. Since the purpose of the patient education and phone calls were meant to refine the interventions and increase adherence, not identify personal stressors, informed consent was not needed; however, a cover letter was given to all new patients at the clinic and those receiving mental health treatment (see Appendix IX).

Results
Over the course of the study, 33 new patient records that met criteria were reviewed. This exceeded the expected number of patients for the pilot (25)   A modification was made to data collection from what was initially planned.
Data from December 2016 was going to be included in the pre-intervention phase but it was excluded from the study. On January 1, 2017, the clinic began an incentive program that allowed patients in all appointments (not just mental health) to obtain a month of free medication for going one full year without a "no-show" visit. To prevent this contextual element from interacting with the intervention, data from December was excluded. This kept the impact of the new incentive program element consistent throughout the entire project.
Seventy-two patients were seen in the IBH program during the study period and the number of visits analyzed. Pre and post data was compared for statistical significance.
Data was entered in SPSS Statistics version 24 and Excel. Chi-Square was used for statistical analysis. The average number of mental health sessions that were attended by a patient pre-intervention was 30 and post-intervention, 34. Patients were 13.3% more likely to adhere to the appointment after the bundle of interventions was introduced (See Figure 1).

Figure 4: Number of patient sessions with a Mental Health Counselor pre and post intervention
Of the thirty-three patients who were new to the clinic during the six months of data collection, two out of 13 (15%) in the pre-intervention group who received treatment with a counselor continued treatment after 1-2 sessions. Five out of 20 (25%) stayed in treatment in the post-intervention group, showing a 66.6 % increase for patients staying in treatment after the bundle was introduced (see Figure 2).

Figure 5: Percentage of new patients who stayed in treatment >2 sessions
To determine the effect of the bundle of interventions on appointment status (seen, no-show, cancelled), the number for each was calculated pre and post intervention (see Table 2). The mean for each group and percentage change was then determined (see Figure 3). After the introduction of the bundled interventions at the clinic, the average number of patients who kept their scheduled appointments and were seen by a mental health specialist increased by 22%. The average number of patients who no-showed for their appointment decreased by 60%, and the average number of patients who cancelled decreased by 15%. These differences were significant (x 2 = 9.263, df = 2, p < 0.01).
After the intervention, patients were more likely to keep their appointments and less likely to no-show or cancel.  After the project data was analyzed and the weakness discovered, the collaborative team met to review challenges, improve sustainability, and guide future practice. Team discussion exposed the following barriers to the process for follow-up calls: 1) Counselors were calling only patients who no-showed for an unknown reason, not those who had notified the front office that they were going to be absent 2) The procedure and form for documenting calls had been moved to a location distant from the counseling rooms 3) new counselors rotating into the clinic were not being updated on the purpose and procedure for follow-up calls. Barriers to the warm handoffs included: 1) lack of an easy way to document the encounters 2) patient privacy issues and 3) lack of consistent personnel for process. A second PDSA worksheet for cycle 2 was developed to meet the barriers and incorporate new methods to improve and sustain the process (See

Appendix X)
This pilot project aimed to address cost savings associated with mental health appointment adherence by examining local ED data. The hypothesis was that adherence to mental health treatment would provide a cost savings by decreasing the number of ED visits since mental health treatment improves the overall health of individuals (Defife et al., 2010). It was planned that the records of patients who were in counseling services would be examined pre and post intervention to determine if the number of ED visits dropped after consistent mental health counseling. Unfortunately, during the data collection stage, the local hospital underwent an extensive EHR update that restricted the clinic's access to ED visit information. Despite multiple attempts to retrieve this information by both the researcher and Executive Director, it remained unavailable and the impact on health and cost-savings associated with reduced ED visits was unavailable for analysis.

Discussion
Findings of the pilot project validated the benefit of using a bundle of interventions to improve mental health appointment adherence. 71 patients participated in the IBH program during the study period and showed improved adherence with the 292 visits that were scheduled. 72% of scheduled appointments were kept after being introduced to education, follow-up phone calls, and warm handoffs as opposed to 59% who were not exposed. Patients were also more likely to remain in treatment after 1-2 visits. The implication of this is that patients felt more comfortable with counseling sessions and were more motivated to adhere to a behavioral health treatment plan.
The project had several strengths. First, the IBH program and procedures were already in place, providing an existing framework for improvement. Improving adherence enhanced the safety, efficiency and effectiveness of the established program. Second, staff and volunteers were successful in other projects at the clinic and were open to evaluation and change. Third, the interventions were not costly to implement, requiring only minimal resources for the printing of brochures, flyers, and posters.

Limitations:
The primary weakness of the program was that there was no way to determine if one intervention was more effective than another. Overall results were positive, but it is unknown which of the individual interventions of education, warm-handoffs, or followup calls had greater impact within the bundle, if any. This was complicated by the fact that an unknown number of participants received follow-up phone calls and handoffs.
Although the health care providers acknowledged that these interventions took place, the imprecision in method collection resulted in an inaccurate count of those who participated. As a result, one could surmise that the education intervention was the most effective and the usefulness of handoffs and phone calls questionable.
Additional insight on the relationship between the cause of missed appointments and demographics such as age, race, mental health literacy level, and socioeconomic status would have been beneficial in understanding why the bundle worked for this lowincome population. Demographics in this particular clinic will differ from others and could impact the replication and results of the program in other settings.
Low-income clinics are seldom part of a larger hospital systems and therefore lack the ability to acquire data needed for analysis of research. This dependence on others (i.e. ED data in this project) limits the extent for what is known regarding interventions and the improvement of health and cost savings associated with new processes.

Conclusion:
This project was useful because it piloted the implementation of an inexpensive bundle of interventions that could be well-suited to clinics and other low-income settings where resources are limited. The bundle was easy to incorporate into practice, consisting of simple educational materials and easy procedures for phone calls and handoffs. In this pilot, results were impressive. The occurrence of missed appointments dropped significantly and overall adherence improved by 22%. The challenge lies in the capacity of small numbers of staff and volunteers to enact multiple interventions. This challenge must be acknowledged and understood ahead of time. Since the outcome was positive even though the handoffs and phone calls underperformed, additional research on utilizing the educational intervention alone would be useful.
The Plan-Do-Study-Act (PDSA) ensured sustainability of the initiative. The team met after six months of data analysis to celebrate the initial results, recognize the work that had been done, and address the low performance of the two interventions; follow-up phone calls and warm handoffs. Steps were identified to reduce future barriers and procedures were updated. These steps guided future practice; however, sustainability also depends on having an on-site leader or manager who is in charge of the process to continuously promote the interventions and to sustain excitement for the project. In this particular clinic, the Clinical Director who is responsible for ensuring clinic protocols are followed, will take over this responsibility from the researcher.

Suggested Next Steps:
The researcher delivered results of the project with the collaborative team in presentations at two clinic meetings; one for the volunteer counselors and another for the Clinical Services Committee (clinic committee responsible for clinic oversight and the implementation of clinical protocols). Clinic patients were informed of results through the monthly patient newsletter. Future plans for dissemination include submission to a professional journal for publication and presentation at a professional conference.
Success of an Integrated Behavioral Health and Primary Care program depends on a well-planned model that identifies appropriate, attainable, and positive outcomes for the population. This pilot project highlighted the benefit of using multiple interventions to address adherence. Incorporating the use of education, follow-up phone calls, and warm handoffs was successful in improving attendance rates at mental health appointments. Low-income clinics with limited resources can easily replicate this program to improve mental health literacy, decrease stigma, and improve motivation, allowing vulnerable populations access to needed behavioral health treatment.

Funding:
This work was supported by resources and the use of facilities within the clinic, which provides free services to patients primarily through volunteers and donations. Scoring: PHQ-4 total score ranges from 0 to 12, with categories of psychological distress being: None 0-2 Mild 3-5 Moderate 6-8 Severe 9-12 Anxiety subscale = sum of items 1 and 2 (score range, 0 to 6) Depression subscale = sum of items 3 and 4 (score range, 0 to 6) On each subscale, a score of 3 or greater is considered positive for screening purposes The PHQ scales were developed by Drs. Robert L. Spitzer, Janet B.W. Williams, and Kurt Kroenke and colleagues. The PHQ scales are free to use. For research information, contact Dr. Kroenke at kkroenke@regenstrief.org b. Referrals for positive screens: Patients who have a positive screen will be encouraged to attend a counseling session. If they aren't interested in therapy, the Eligibility worker will make a note on the PHQ-4 form stating, "Counseling services offered but declined". c. Scheduling appointments: If patient is eligible and agreeable to a counseling session, an appointment will be made by the front desk with "Mental Health Counselor" in the "Mental Health" calendar at check out.

Front Deskreminder calls and documentation:
a. Patient will be given a reminder call by front desk prior to apt. b. Patient will check in with front desk, front desk will flag as in lobby 3. Mental Health Specialistdocumentation: a. Review the schedule, when EHR shows in lobby, escort patient from lobby to the counseling room. b. Open the patient encounter from the scheduling screen by clicking on view encounter. c. In encounter details change note type from SOAP note to Mental Health Note. d. Make sure the date is today. e. In the Chief Complaint section, click edit and make a brief note stating purpose of visit with plan and recommendations for Primary Care followed with name of counselor and degree. f. When visit is completed, change appointment status to seen on the schedule. g. At end of day, Mental Health Counselor will print a schedule and place it in "Carol's" box in the office.

Medical Provider Referral:
a. Patient will be identified as a candidate for counseling by care team. b. Medical provider will introduce patient to counselor via warm handoff which is a brief introduction to the counselor and benefit of services. If counseling rooms aren't available for the handoffs, they will take place on the second floor in a location that can guarantee privacy of patient information.
5. Scheduling Future/Follow-up appointments: a. Follow up appointments will be scheduled by the Front Desk during check-out or by phone call. b. Counselor completes appointment slip including how many weeks for next visit and with which counselor.
6. Follow-up for Missed appointments: a. Mental Health Specialists will call all those who missed appointments either from "no-show" or "cancellation" even if they conveyed a reason for missing. b. Counselors will document the phone-call in the patient encounter note in the EHH. They will explore: i. reason for missing ii. any acute needs iii. motivation to continue with follow-up visits