Abstract
Introduction
Behavioral health conditions, including mental health and substance use disorders, are common and costly in primary care settings.1-3 Approximately a quarter of adults in the United States meet the diagnostic criteria for a behavioral health condition annually, with a third of adults with a medical diagnosis reporting a behavioral health condition. 4 Patients with behavioral health issues are most likely to present in primary care rather than the specialty health system, yet most providers face challenges in addressing these issues. Providers are often limited in their knowledge and training and have to identify and treat behavioral health issues with limited time and resources. 1
Integrated behavioral health in primary care (IBHPC) is designed to overcome the barriers to patients accessing and providers offering behavioral healthcare. IBHPC combines medical and behavioral health services through integration, not simply the co-location of an embedded psychiatrist, psychologist, or social worker. 5 Studies conducted in various settings have demonstrated that IBHPC models are associated with improved health outcomes, including depression and anxiety symptoms, patient quality of life, disability, and management of comorbid chronic diseases.6-8 IBHPC is also associated with lower costs, increased treatment outcomes and adherence, enhanced patient engagement, improved patient and provider satisfaction, and reduced racial and ethnic disparities.5,7,9-14 IBHPC models offer an opportunity for continuity of care for patients who may not receive behavioral healthcare due to long waiting times for specialist referrals.
Although studies have documented high patient satisfaction with IBHPC post-visit, few studies have described patient preferences for IBHPC pre-visit.9-12,15-20 The available information on the experience with IBHPC documents a willingness to meet with an embedded behavioral health provider for a follow-up visit and that most patients prefer their mental health and medical providers co-located and in regular communication about their care.20,21 Patients may prefer to receive behavioral health services in primary care due to the potential for poor communication between primary and specialty care, the flexibility and comfort of care, immediate access to care, consolidated appointments, lower cost, and less perceived stigma.7,22 Alternatively, patients may have concerns about privacy and confidentiality and believe that primary care is for physical, not behavioral health concerns. There may also exist variation in preference by patient characteristics, such as race and ethnicity, sex, age, or prior experience with behavioral health treatment.23,24 If patients indicate a preference for receiving integrated behavioral healthcare in primary care settings, they may be more likely to reveal these concerns in future visits.
It remains unclear if patients prefer to have their behavioral health concerns addressed in primary care or value referrals to specialists. This question received considerable debate among primary care leaders in two clinics in California. This study addresses a major barrier around the implementation of IBHPC, whether patients prefer to receive behavioral healthcare services in primary care. A variation of primary care patients’ preferences for integrated behavioral health care was explored by several patient-level factors, including demographic characteristics, attitudes, and comfort regarding behavioral healthcare. The goal of this project was to understand patients’ preferences for IBHPC and help primary care providers understand how to serve patients’ needs best.
Methods
Design
The question of patients preferring their behavioral health issues addressed in primary or specialty care has undergone considerable debate within a local primary care system. In this rapid quality improvement study, cross-sectional data were collected from patients at two primary care clinics in an academic healthcare system, including an internal medicine and a family medicine practice. Both of these practices are located in the same building and staffed with medical residents, nurses, physician assistants, medical assistants, and attending faculty physicians. The clinics have a combined panel size of 21 960 patients, with approximately 1015 visits weekly. At the time of data collection, there were no integrated behavioral health services available other than referrals to a co-located social worker and external specialty care or community services. The use of standardized screening measures, such as the Patient Health Questionnaire (PHQ-9), was limited. Behavioral health screening was completed at the discretion of the individual provider. Neither patients nor providers had direct access to consulting psychiatry. Behavioral health needs were primarily addressed by primary care providers or referred to the health system’s psychiatry department or mental health providers in the community. Wait times for non-acute patients for mental health specialists ranged from 3 to 9 months.
Measures
A project-specific measure, the Integrated Behavioral Health Patient Preference survey, was developed (Table 1). The survey is an 8-item self-report measure that focuses on patient preferences for integrated behavioral healthcare. The authors constructed the survey with input and feedback from primary care staff, leadership, and patient representatives. The instrument’s psychometric properties were not investigated, and no scoring exists. The survey was designed to address the debate among primary care leaders to understand whether patients preferred IBHPC prior to implementation. The survey includes items on the patient’s general medical and psychiatric treatment history, behavioral health perceptions, comfort receiving behavioral healthcare, and behavioral health services preferences. The following behavioral health conditions and comorbid chronic health outcomes were included in the survey: diabetes, hypertension, chronic obstructive pulmonary disease, asthma, depression, anxiety or stress, and alcohol or drug use. These conditions were included as they are some of the most frequent presenting conditions in primary care and have been shown to improve in settings with IBHPC models.8,25,26 Two items were adapted from a patient satisfaction survey and related to comfort and preference for receiving mental health services in primary care. 18 Demographic information was collected and included age, sex, and race or ethnicity. The third item on the survey was the primary outcome of interest. It was operationalized as patient preferences for integrated behavioral healthcare: “If you had a behavioral, mental or emotional health concern, which of the following would you prefer?”
Items From the Integrated Behavioral Health Patient Preference Survey Administered Among Primary Care Patients.
Data Collection
All patients presenting for their primary care visit during a 3-week data collection period in May 2019 were invited to complete a voluntary, anonymous survey when presenting at the front desk for their appointment. After providing verbal consent, participants were instructed to complete the paper survey in the waiting room. Patients were not provided any compensation for their participation. Upon completion, front desk staff placed surveys in a bin behind the front desk, and surveys were collected at the end of the week by a quality improvement team member. The survey was in English and included no identifying information. The Stanford University School of Medicine Institutional Review Board designated this project as exempt as a quality improvement project. No identifying information was collected from participants.
Statistical Analyses
The analyses were conducted using Stata Version 16. Sex was included in the survey with the options of male, female, and transgender. Very few patients identified as transgender or other; thus, these categories were set to missing for analyses. Race and ethnicity were operationalized following aggregated United States Census categories. For the analysis, American Indian/Alaska Native, Native Hawaiian, or Pacific Islander were recategorized into other. Age was collected in 10-year categories, and for analyses, the youngest group was recombined as 30 or younger, and the oldest group as 60 or older. The importance of behavioral health to physical health was recategorized into two categories for the analysis: Not at all important to not important, and important to extremely important. Medical treatment history was categorized as: (1) chronic medical condition(s), such as diabetes, hypertension, COPD, and asthma; (2) behavioral health conditions, such as depression, anxiety/stress, and substance use problems (alcohol, prescription medications, cannabis or illicit drugs); and (3) any patients with comorbid behavioral health and chronic medical conditions. Preference for behavioral healthcare as the primary outcome was categorized as: (1) to get help in the clinic from a team member; (2) from my doctor; (3) from a behavioral health clinician; (4) it depends; and (5) to receive a referral to a specialist outside of the clinic. The categories related to receiving any form of IBHPC were combined to address the lower frequency of patients who preferred receiving help from a non-clinical team member. Thus, this outcome was categorized into receiving help in the clinic from any staff member, it depends, and receiving a referral from a specialist.
Descriptive statistics were calculated using cross-tabulations and frequency distributions. Participants’ demographic characteristics with missing data on the primary outcome were compared to those with complete data for IBHPC preferences, using Pearson chi-squared tests and Fisher’s exact test. Bivariate analyses were conducted using Pearson chi-squared tests to account for the categorical outcome with three levels of preferences for behavioral healthcare. All statistical assumptions were met for the bivariate analysis.
Results
Patient Respondent Characteristics
Of approximately 3045 patients who presented to the two clinics during the study period, a total of 752 participants completed the survey (24.8%). Table 2 presents the characteristics of the sample. A majority of participants (n = 420; 55.9%) identified as female. Most participants were White (n = 321; 45.0%), followed by Asian (n = 213; 29.8%), and Hispanic or Latino (n = 66; 9.2%). Patients’ medical history of treatment included: (a) chronic diseases (n = 180; 46.0%), including diabetes, hypertension, COPD, and asthma; (b) mental health conditions or substance use (n = 137; 35.0%); and (c) and the comorbid conditions (n = 74; 18.9%). The demographic characteristics that were missing included (4.3%) 32 patients on sex, 25 (3.3%) on age, and 38 (5.1%) on race and ethnicity.
Descriptive Statistics of Patients in Primary Care Settings.
Main Findings
More than half of patients reported that their behavioral health needs were mostly or completely met at the clinic (n = 322; 53%). The majority felt comfortable receiving help for behavioral health issues (n = 673; 90%). Patients generally preferred to have their concerns addressed within primary care (n = 301; 41.0%) by either a behavioral health clinician (129; 17.6%), their provider (100; 13.6%), or any team member (n = 72; 9.8%). A minority of patients desired a referral to a specialist (n = 55; 7.5%), and half of the patients reported, “it depends” (n = 369; 50.3%). Less than 3% of participants had missing data. Respondent missing data did not differ by patient characteristics.
Patient Factors and IBHPC Preferences
The chi-squared test results between patient-level factors and their preference for IBHPC are displayed in Table 3. There was no evidence of an association between patient preferences and patient demographic characteristics, such as age, sex, race or ethnicity, treatment history for chronic or behavioral health conditions, and behavioral health’s importance to their physical health.
Chi-Squared Tests of the Relationships Between Patient-Level Factors and Preferences for Integrated Behavioral Healthcare in Primary Care Settings.
Patient comfort levels with receiving help for behavioral health issues were significantly associated with patient preferences for receipt of behavioral health (P < .001). Similarly, there was an association between patients feeling that their needs were met at the clinic and preferences for receiving behavioral healthcare (P < .001).
Discussion
Summary of Findings
The goal of this quality improvement project was to assess patient preferences for integrated behavioral health services and explore whether these preferences varied by patient factors. Patients preferred receiving care for their behavioral health issues within the clinics rather than receiving a referral to an outside, offsite specialist. Most patients also reported that their behavioral health needs were met at the clinic, even in the absence of onsite formal integrated behavioral health services. On average, patients within the clinic reported a higher preference for having their concerns addressed by a behavioral health clinician than their primary care doctor or other staff members. Patients reported feeling comfortable being asked about and receiving help for their behavioral health concerns. Overall, the findings from this study satisfied concerns among primary care leaders about patient perceptions for IBHPC. Some primary care leaders questioned whether patients preferred receiving behavioral healthcare at the clinic and found addressing behavioral health issues in primary care appropriate. This quality improvement project significantly impacted these primary care clinics and addressed an important question among primary care providers in two clinics in California. As a result of the findings, IBHPC models were implemented in these clinics, and the programs continue to develop.
Even before patients engage in behavioral healthcare, however, differences in individual characteristics may contribute to the prevalence of behavioral health conditions and treatment preferences.19,27 African Americans and older patients tend to seek treatment for behavioral health concerns in primary care preferentially but are less likely to receive evidence-based practices. 28 Additionally, in many older patients, behavioral health conditions largely are left undiagnosed, and there is evidence that IBHPC increases patient engagement in this population. 29
Despite evidence of variation in patient satisfaction with IBHPC by patient characteristics,19,23,27 this study showed that preferences for receiving behavioral healthcare did not vary across patient demographic characteristics. However, there was substantial evidence that patients’ comfort level and the extent to which their behavioral needs were met at the clinic were associated with their preferences for receiving behavioral health services. These findings underscore that IBHPC is acceptable, if not preferred, by patients.
Limitations and Strengths
This project rapidly addressed a question in a local healthcare system by developing and deploying a brief, low-burden measure for patients to complete. In future iterations, some of the survey items should be improved. For example, 1 question included the response option “It depends” but offered no room to elaborate. A majority of participants selected this response category. Additional qualitative data from patient interviews could have augmented these quantitative findings to explore further what types of behavioral health problems patients feel comfortable addressing in primary care and how they would like them addressed. This information could ultimately support a patient-centered implementation of IBHPC.
Additionally, many of the items focused on hypothetical perceptions of services while experiencing a behavioral health concern. More evidence is needed on how reported preferences are related to actual engagement with IBHPC. Information on the survey’s psychometric properties would also inform reliability and validity and allow future use across a range of primary care settings. To date, no validated measure exists to assess patient preferences for IBHPC. 30
The findings may also not generalize beyond this setting. This recruitment method did offer a naturalistic sample to answer a localized question. This study’s response rate of nearly 25% was comparable to previous studies on post-visit patient satisfaction with behavioral health services.17,18,21 Still, no project with a survey focused on behavioral health in a primary care setting, pre- or post-visit, has featured a sample size of this magnitude. Critically, this sample may not have been large enough to establish hypothesized differences in patient preferences for IBHPC or racially and ethnically diverse to explore these health disparities.27,29,31
Future Directions
These findings suggest that relationships exist between patients’ comfort level and the clinic meeting patients’ behavioral health needs and preferences for IBHPC. The relationship of trust and confidence the patient has with their provider and the clinic may influence their openness to receive integrated behavioral health. Future work should investigate whether comfort level and patient needs being met may explain some of the relationships between patient demographic characteristics and preferences for IBHPC. The mechanisms that affect patients’ preferences for IBHPC should be further explored in larger and more diverse samples of primary care patients. Additionally, future studies should explore the differences in patient preferences for IBHPC based on additional factors, such as the role of clinical staff, age of patients, and treatment history across more diverse primary care settings.
Conclusions
The findings from this quality improvement study in 2 primary care clinics suggest that patients prefer to have their behavioral health concerns addressed in settings familiar and comfortable to them, from individuals they know, and where they already receive care for other medical conditions. These results support the expansion of IBHPC and alleviate concerns among institutional leadership and providers that patients may not wish to have behavioral health issues raised or addressed within the context of their primary care relationship.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
