Results Summary
PCORI funded the Pilot Projects to explore how to conduct and use patient-centered outcomes research in ways that can better serve patients and the healthcare community. Learn more.
Background
Half of primary care patients from low-income or minority populations live with anxiety, depression, or other mental health problems. Community health centers provide primary care for many of these patients. However, only about 70 percent of community health centers offer mental health services.
When community health centers do offer mental health services, it still can be difficult for patients to get the care they need. For example, health centers may not have enough staff to serve all of their patients, or staff may not speak the same language as the patients. Health centers may not have systems in place to make sure medical staff can easily and quickly share information about the services patients need.
Project Purpose
The researchers used a type of research called a discrete choice experiment to learn if it would be helpful for getting information from patients about how to make mental health care better.
With this method, participants answer questions that have two choices. For each question, people make hard choices about what is more important—for example, daytime appointments in Spanish versus evening appointments in English only. By studying what people choose, researchers can see what matters most. This method is often used in consumer product design research.
Methods
The study took place at a rural community health center near the United States–Mexico border. Most of the patients were Mexican American.
First, the research team worked closely with patients, staff, and health center directors as partners to find out what aspects of mental health services were most important to patients. They trained community health workers to bring together groups of patients and learn about their ideas. The community health workers spoke with 64 primary care patients in eight groups. The researchers also interviewed 15 health center staff to get their ideas about ways to provide mental health services. The researchers studied the results of these group discussions and interviews to learn what patients and health center employees thought was most important for helping people get mental health care.
Using the findings from the group discussions and interviews, the researchers created a survey with different options for getting mental health services. Patients had to choose between important aspects of care. For example, the survey asked if patients would prefer to get care from a highly trained health professional in another location or a less experienced professional with an office within the clinic.
In the study, 604 primary care patients took the survey. The researchers looked at all the answers to see which features of mental health care were most important.
Findings
The researchers found that patients and staff agreed on what made it hard to get mental health services.
The researchers found that patients wanted their health center to help more people understand what mental illness is. The patients wanted to know what they could do on their own to help with daily stress. They suggested creating support groups for people to talk about stress.
When researchers reviewed the answers to the survey, they found that two features of mental health services were important to patients, but they hadn’t shown up in the group discussions. First, patients wanted the services to be available at their own health center. Second, they wanted a provider who spoke their language and understood their culture.
Limitations
Everyone in this study was from the same community health center. The patients were mostly Mexican American. The results might be different at other health centers or with different patients. However, this project took a lot of time and work. It might be hard to do this kind of study at other health centers.
The researchers did not test whether the suggestions for improving mental health care helped patients get better.
Conclusions
The researchers found that group discussions and interviews with patients and community health center staff were useful for getting information about what matters to patients and planning ways to deliver care that matches those priorities.
Adding the survey helped researchers learn what aspects of mental health services are most important to patients. This information can help community health centers focus on the most important services for patients.
Sharing the Results
The research team shared the results with patients, community health workers, other healthcare team members, and the community health center leadership. The health center added several mental health providers who are able to provide care on a regular basis at the health center. The researchers also published articles about the study in medical journals (see below).
Professional Abstract
PCORI funded the Pilot Projects to explore how to conduct and use patient-centered outcomes research in ways that can better serve patients and the healthcare community. Learn more.
Background
Half of primary care patients from low-income and minority populations experience anxiety, depression, or other mood disorders. Although 70 percent of community health centers—the backbone of primary care systems for medically underserved populations—provide behavioral health services, these services are limited due to high patient volume, communication barriers, and lack of care coordination. A health center without mental health services approached the researchers to ask them to help address such barriers by implementing pragmatic and feasible patient-centered outcomes research (PCOR) methodologies.
Project Purpose
The goal of this study was to evaluate two methods for responding to patient-centered behavioral health needs for mainly Mexican-American patients living near the US–Mexico border. The project had three main goals:
- to address a methodological gap by comparing two promising approaches to systematically soliciting patient preferences about their care
- to address a clinical gap by aiming to reduce disparities in mental health services among a medically underserved population
- to assess the feasibility and usefulness of adding a consumer-products design method to community-based participatory research (CBPR)-guided qualitative methods
CBPR is a collaborative method that involves patients, providers, staff, and administrators across all stages of the research, although, typically, CBPR has not integrated advanced quantitative techniques such as consumer products design methods in its execution. The consumer products method used in this study was a discrete-choice conjoint experiment that modeled the patients’ preferences for mental health service components that they would likely encounter within in a real-world, resource-constrained, healthcare environment.
Study Design
The main CBPR-guided qualitative components included eight focus groups with patients as well as open-ended interviews with provider staff. Community health workers were responsible for facilitating focus groups. The subsequent product-choice method, a discrete-choice experiment (DCE), was built upon the prior qualitative findings via the technique of conjoint survey construction, implementation, and analysis. The DCE survey was administered by community health workers in waiting rooms and was operationalized through 15 choice tasks, for which each patient chose among three mental health services options using attributes generated by focus groups. As an example, patients might have to choose if they would prefer a more highly trained health professional in another agency or a professional with less expertise but who has an office within the primary care clinic. A random series of many such forced choice scenarios were created via computer algorithms, and advanced statistical techniques employed to identify the most critical issues to patients among plausible options.
Data Sources
Focus groups included 64 general primary care patients; 604 primary care patients completed the DCE; and 15 employees were interviewed. Primary outcomes included feasible patient preference–based alternatives to deliver mental health services, and secondary outcomes were center leadership’s views about the research process and the utility of the findings, which were ascertained through stakeholder interviews.
Data Analysis
This study used qualitative and quantitative analysis. Qualitative analyses included creating a “conceptual map” of patient preferences, consisting of six attributes of care from patient perspectives, with several dimensions defining each attribute. Two researchers coded the focus group and interview data using NVivo 10 software. Based on an interrater agreement level of 80 percent, researchers then discussed and recoded all attributes about which there was disagreement. Conjoint analysis was executed using Sawtooth software, which simulated how patients would most likely respond to alternative behavioral care packages that could plausibly be offered by the health center.
Findings
The initial qualitative findings confirmed that patients and providers shared an understanding of barriers to accessing behavioral health services in this rural community. The researchers also found that patients wanted the center to be active in informing the community about the nature and importance of mental illness and in trying to reduce stigma within their patients and, more broadly, the community. The qualitative findings also suggested that patients wanted to be active in their own care through selecting coping strategies they could use to address their daily stress, and that the center could create support groups for people to share and address stressors. The conjoint methods identified patterns not apparent from the prior qualitative findings. The attributes that had the largest influence on whether patents would likely receive behavioral health treatment were where patients could receive services and the language and cultural awareness of the provider delivering treatment. Simulations indicated that the center could substantially improve its patients’ welfare with one major change: offering behavioral health services on-site.
Limitations
Although the methodology is transferable, the generalizability of the specific behavioral health recommendations may be limited as the study was conducted at only one site. Another limitation is the substantial commitment in time and resources required to conduct focus groups and interviews, compile data, and translate findings into deliverable behavioral health services. Further, the even more resource-intensive conjoint method may require a larger clinical setting or a larger healthcare problem to be justified. Finally, within the project time frame, it was not feasible to test patients’ clinical improvements associated with different behavioral health services packages to be offered; thus, only simulation-based inferences could be made.
Conclusions
The CBPR-guided qualitative components yielded informative patient perspectives and served as a tool for primary care providers to develop patient-centered care protocols. The addition of the DCE was feasible and provided further unique information about patient preferences. Findings suggest that combining DCE and CBPR-guided qualitative methods may mimic patients’ decisions in real-world, resource-constrained environments.
Dissemination
Through presentations (academic and local) and individual consultations, findings were shared with patients, community health workers and other members of integrated healthcare teams, and the health center leadership. The study team also published several journal articles on these findings. The health center has now added behavioral health providers to its core on-site services.
Future Research
Next steps include testing whether, within primary care settings, the different care packages informed by the PCOR methods result in long-term clinical health improvements.