Implementing Collaborative Care for Older People with Comorbid Hypertension and Depression in Rural China

Abstract

Background.

Depression often coexists with other chronic conditions in older people. The COACH study is an ongoing RCT to test the effectiveness of a primary care-based collaborative care approach to treat comorbid hypertension and depression in Chinese rural elders. In the COACH model, a team—village doctor (VD), aging worker (AW), and psychiatrist consultant—provides collaborative care to enrolled subjects in each intervention village for 12 months. This study examines how COACH was implemented, and identifies facilitators and barriers for its more widespread implementation.

Methods.

Five focus groups were conducted, two with VDs, two with AWs, and one with psychiatrists, for a total of 38 participants. Transcripts were analyzed using qualitative content analysis.

Results.

COACH care team members showed shared understanding and appreciation of the team approach and integrated management of hypertension and depression. Team collaboration was smooth. All members regarded COACH to be effective in reducing depressive symptoms and improving patient health. Facilitators to implementation include training, leaders’ support, geographic proximity between VD and AW pairs, preexisting relationships among care team members, comparability of COACH activities and existing practices of VDs and AWs, and care team members’ caring about older members of their villages. Barriers to sustainability include frustration of some VDs related to their low wages and feelings of overload of some AWs.

Conclusions.

COACH was positively perceived and successfully implemented. The findings offer guidance for planning primary care-based collaborative depression care in low- and middle-income countries.

Introduction

Hypertension affects more than two-thirds of people after age 65 and depression is common in people with hypertension (Chobanian et al., 2003; Sandstrom et al., 2016). A systematic review suggests that 26.8% of hypertensive patients suffer from depression (Li et al., 2015; Zhang et al., 2008). The nature of the relationship between hypertension and depression is unclear. Possible explanations include shared genetic risk for both conditions (Scherrer et al., 2003), sympathetic nervous system reactivity that exacerbates both HTN and depression (Scalco et al., 2005), and the associations of depression with non-adherence to antihypertensive medications and poor self-care (Krousel-Wood and Frohlich, 2010; Rubio-Guerra et al., 2013; Wang et al., 2002). Comorbidity of hypertension and depression has larger effects than each condition alone on health status, health care utilization and cost, hence addressing depression in hypertensive patients is a public health priority (Moussavi et al., 2007).

Effective treatments for depression exist, but they are not available to many people living in low- and middle-income countries (LMICs), including older adults in rural areas of China. Rural China has a 3-tier health care system. Village clinics provide primary care (Wagstaff et al., 2009). Every village, which is the basic production and social unit of rural China, has a village clinic staffed by a village doctor (VD). VDs have mixed qualifications. The ‘new’ generation has received structured medical training and passed the Ministry of Health exam, though they do not have a medical degree. The ‘old’ generation is ‘barefoot’ doctors who usually have less than high school education. VDs receive no training in diagnosis or treatment of mental disorders. Township health centers and county hospitals provide secondary and tertiary care. Psychiatric services are only available in county hospitals, which are in central locations and not easily accessible to many older villagers due to transportation barriers. In addition, the number of psychiatrists is small relative to mental health care needs of the population they serve, and priority is given to patients with severe psychiatric problems.

The COACH Study

In 2013, our team began a randomized controlled trial (RCT, ClinicalTrials.gov ID: NCT01938963) to test the effectiveness of using a primary care-based collaborative care management approach—Depression/Hypertension in Chinese Older Adults Collaborations in Health (COACH)—to treat older patients with comorbid hypertension and depression in rural China. A protocol of the trial has been published (Chen et al., 2018). The primary objectives of the COACH study were to determine whether COACH was more effective than Care as Usual (CAU) in treating depression and hypertension.

The current practice to treat hypertensive patients with suspected depression in rural China constitutes CAU. Currently, VDs are expected to follow guidelines for management of hypertension developed by the Chinese Center for Disease Control and Prevention (CDC; Liu, 2011). However, no depression treatment is available in village clinics. When depression is suspected by VDs, current practice involves suggesting to patients (or family members) that they consult a psychiatrist at the county mental hospital. It is uncommon for patients to take the initiative.

The COACH intervention involves forming a care team composed of a VD and an aging worker (AW) in each intervention village, and a consultant psychiatrist in the county mental hospital which is about 30-minute driving distance from the villages on average. AWs are staff of the village committee which is the body that governs the village’s public affairs. In their official roles, they are responsible for implementing either family life education (women worker) or helping older villagers to participate in community life (health liaison). As local residents, AWs know the villagers well.

Key elements of COACH include: (a) VD was trained to screen and monitor blood pressure and depressive symptoms with the PHQ-9, and use evidence-based practice guidelines for management of both hypertension and depression. (b) AW was trained to conduct systematic assessments of patients’ social stressors and support through home visits, educate the patient and family, and support patients’ treatment adherence and adoption of healthy behavior. After study participants were enrolled in the study, the VD and AW separately met with each patient and then got together to develop a care plan for their shared patients. (c) Psychiatrists travelled to villages to conduct diagnostic assessment, and in consultation with the VD initiated antidepressant treatment as indicated. The VD continued to meet with the patient at the intervals designed by the practice guidelines, the AW visited the patient weekly for the first 2 months, then bi-weekly for 2 months, then monthly. The AW and VD met weekly to review their shared caseload. The psychiatrist joined these meetings monthly by telephone. Between monthly meetings, the VDs could call the psychiatrists for additional consultation. The COACH intervention required VDs, AWs, and psychiatrists to work as a team for 12 months.

VDs participated in a small-group, four-day training workshop that covered: (a) depression management, (b) hypertension management, (c) case management and (d) psychoeducation and communication skills. AWs received three days of training that included: (a) overview of depression, hypertension and treatments, (b) social environment as risk and protective factors, (c) ways to support patients’ self-management, (d) communication and problem-solving skills, and (e) ethical standards. Throughout their 12-month COACH participation, AWs received continuous support from project staff via WeChat, which is a popular messaging app in China. In addition, VDs, AWs and psychiatrists participated together in a one-day training about how to collaborate in care management.

The Present Study

The present qualitative study was conducted in the 3^rd year of the 5-year COACH study. Objectives were to understand the implementation of COACH and to identify factors affecting its implementation. The intention was to prepare for upscaling COACH should it be found effective. Collaborative depression care has been extensively tested in Western countries, including among older adults (Archer et al., 2012; Unutzer et al., 2002), demonstrating both clinical effectiveness and savings in total healthcare costs (Unutzer et al., 2008). But they have been rarely applied in LMICs. Our findings may shed light on preferred ways to implement such approach in LMICs and other low resources settings.

To be described further in the Methods section, a qualitative approach was used to obtain perspectives of the care team members about the COACH model. We analyzed and interpreted the qualitative data based on the framework of the Normalization Process Theory (NPT) which suggests that the embedding of a new practice in its social context involves four generative mechanisms—coherence, cognitive participation, collective action, and reflexive monitoring (May and Finch, 2009). Coherence refers to participants’ (i.e., people who are involved in implementing the practice) understanding and conceptualization of the new practice. Cognitive participation focuses on participants’ enrollment and engagement. Collective action is enactment of the new practice and its interaction with existing practices. Reflexive monitoring refers to participants’ appraisal of the new practice. The NPT postulates that factors that promote or inhibit participants’ coherence, cognitive participation, collective action and reflexive monitoring shape the work of implementation.

The COACH study concluded in 2018, with 1,232 subjects in the intervention group and 1,133 in CAU. Preliminary analysis shows that COACH participants had significantly greater improvements in depression and better blood pressure control than those in CAU. A report of the COACH findings is in progress.

Methods

Village was the unit for random assignment (COACH or CAU). Villages were recruited in four phases (consecutive years). When the present study was conducted, 41 villages (i.e., 41 pairs of VD and AW—20 in cohort 1 and 21 in cohort 2) had completed the intervention. The COACH study coordinator recruited focus group participants among these 41 pairs. She was instructed to achieve variation in age, gender, and education within each group. We conducted two focus groups with VDs and two with AWs, each group consisting of eight members with the same role (AWs or VDs) from the same cohort (See Table 1). Six of the eight psychiatrists who were involved in COACH participated in a focus group. In all, we included five focus groups with a total of 38 participants.

Table 1.

Demographic Characteristics of Focus Group Participants

	Cohort 1		Cohort 2
	VD1 (n = 8)	AW1 (n = 8)	VD2 (n = 8)	AW2 (n = 8)	Psychiatrist Consultants (n = 6)
Age
25-34	1	0	8	5	1
35-44	1	3	0	0	3
45-54	1	5	0	0	2
55+	5	0	0	3	0
Sex
Men	4	0	0	5	1
Women	4	8	8	3	5
Education
Elementary	1	1	0	1	0
Middle school	4	2	0	2	0
High school	1	4	0	1	0
College	2	1	8	4	6
Years of medical practice
3-10	2	Not applicable	8	Not applicable	4
11-20	1		0		2
21+	5		0		0
Official role	Village doctor	Women worker	Village doctor	Health liaison	Psychiatrist

Note. VD = Village Doctor, AW = Aging Worker

An interview guide was used for the focus groups. The guide included four sections: (1) engagement and understanding of COACH (How did you get involved in COACH? How do you think about the COACH model?), (2) perceptions of new roles/tasks (what were you asked to do in COACH? How do you feel about these new tasks?), (3) collaboration with other team members (During your participation in COACH, did you feel like working as a team with the [other two members]? Please elaborate?) and (4) changes in your patients (What was the most significant change you saw in your patients who participated in the COACH study?) All groups were held within two months, each lasted for about 90 minutes. All were audio-recorded and later transcribed. Each participant received $10 as compensation. The IRBs of the authors’ universities approved the study.

Analysis of the qualitative data was an iterative process. The following three major steps were undertaken. First, three authors (LL, XJ and QY) and two note takers read the transcripts and identified meaning units independently, and then met as a group to share their thoughts. Second, the first author then completed a first round of coding which was reviewed by XJ and QY separately and discussed jointly in subsequent meetings. Any disagreement was resolved through consensus. Third, the first author sorted the revised codes into categories and sub-categories, mapped them to the NPT framework, and identified emerging themes by searching for repetition and pattern, and comparing and connecting the categories. The themes were discussed among LL, XJ and QY. We describe the themes and use direct quotes to illustrate, if appropriate, in the Results section. Demographic data of participants were extracted from the COACH study database.

Results

Participants

Our 38 participants included 16 VDs, 16 AWs, and 6 psychiatrists (Table 1). The first group of VDs and AWs (i.e., VD1 and AW1) completed their participation in COACH about 22 months, and the second cohort (i.e., VD2 and AW2) about 10 months, before our interview. Psychiatrists were involved in the COACH study until the end. Sociodemographic characteristics of participants of each focus group are shown in Table 1. For the total sample, on average, the VDs, AWs, and psychiatrists were 40, 44 and 42 years old, respectively. About 75% (n=12) of VDs, 69% (n=11) of AWs and 83% (n=5) of psychiatrists were women. Most VDs (63%, n=10) had some college and most AWs (56%, n=9) had middle to high school education. All psychiatrists had a college degree.

Coherence

The COACH team members identified two core characteristics of the COACH model - the team approach and integrated behavioral health and medical care. There was a consensus among the team members that a team approach made good sense, as illustrated by a VD:

Since psychiatrists have more professional knowledge and AWs know much more about the family and living condition of patients than we do, working as a team can help us to have a whole-person understanding of the patient. (VD2H)

VDs, AWs and psychiatrists all said that each of the team members had unique strengths. Specifically, they perceived that VDs had medical knowledge and skills to care for patients, AWs provided organizational and logistical assistance, and psychiatrists offered professional guidance and consultation.

Even though AWs were not medical professionals, they were seen as representing the support of village leadership, which was vitally important in getting collaboration of patients. A psychiatrist said:

We certainly need the support of village leaders.… The village leaders have authority and credibility, and the villagers listen to them. When we need to see the patients, we contact the AW. (PSY E)

Having AWs in the team also helps to reduce resistance of older patients toward mental health treatment. A VD commented:

About mental health problems, I think it is a taboo subject in rural villages. If the AW goes to home visit with us, it’s more like usual chatting and the interaction is smooth. If only we village doctors visit the patients, the atmosphere would be tense. (VD2E)

In addition to a shared appreciation of the team approach, VDs in particular perceived integrating management of hypertension and depression as efficient and effective. As one VD said:

It is our responsibility to manage villagers’ chronic conditions such as hypertension and diabetes. Now we also screen older patients for depression, and that helps us manage their hypertension. I think it is like killing two birds with one stone, a strength of the project. (VD2D)

Cognitive Participation

All VDs, AWs and psychiatrists were asked by a county hospital administrator to participate in the COACH study. Participants felt that their leaders, including those in the county hospital and their villages, cared about the COACH study. They wanted to support their leaders by being a part of the care team. They began by attending training sessions offered by the COACH research team. The training was credited for cementing their engagement and enabling their collaboration. A VD said,

Three years ago we were ignorant about mental health and did not consider depression at all when treating patients. The training has increased our professional knowledge and raised the management standard. (VD1G)

Likewise, AW2H said:

We really didn’t know about depression in the past. We came to understand depression only after the training. After learning more about depression and its relationship with blood pressure control, we were able to collaborate with village doctors and psychiatrists. (AW2H)

While all VDs felt that their COACH participation was voluntary, some VDs from cohort 1 had grievances. They talked in great length about the demands of their job and the sacrifices they made. For example:

We can’t get off from work… when it comes to closing time, some patients come here to ask for a transfusion, I can’t refuse their request. (VD1B)

Others complained of low reimbursement:

Village doctors are protectors of the community. That is true. But our salary is lower than the swineherds…The government provides us much training. That is good and improves our effectiveness. But how about rewards? (VD1G)

AWs from cohort 1 also expressed misgivings about their COACH participation, particularly feelings of role overload. In their official role as women workers, they are responsible for a wide range of village affairs to which COACH added an extra burden. However, they saw a need to help older villagers. One AW expressed this dilemma:

We all do it from the heart and do not ask for payment. We walk more miles and spend more time at work due to participating in COACH. ..Sometimes when we visit an older person, it is not easy to leave even when someone else is calling us to attend another matter. The older person would hold your hands and want you to sit down and talk more. The scene is quite touching… To be honest, it is tiring. (AW1G)

VDs and AWs from cohort 2 did not echo the sentiments of cohort 1 participants described above. VDs of cohort 2 said that the COACH participation did not increase their workload. AWs of cohort 2 expressed a sense of duty:

We’ve got to continue this COACH model, regardless of the change in personnel. Many of our older villagers have hypertension. It’s our responsibility to do our best to help. (AW2B)

Collective Action

VDs, AWs, and psychiatrists reported smooth team collaboration in that each member did their part in a coordinated fashion. A VD described the operation this way:

The AWs arranged venues and logistics; we did the depression screening. Then the psychiatrist saw the patients and made the diagnosis. After that, we followed up according to the guidelines of hypertension and depression management. That’s how we operated. (VD1E).

The critical role played by village doctors was underscored by VDs, AWs, and psychiatrists alike. The VDs coordinated the teams, partly because they had frequent contact with both psychiatrists and AWs. A high level of integration that involved joint planning and problem-solving was reported by care team members from cohort 2, as illustrated by an AW:

We met periodically to discuss the patients in private. Just three of us. When we met, we informed each other about the characteristics of the patients; for example, this patient was very concerned about ‘saving face,’ so we had to be careful about framing his problems. That means paying attention to the individual patient’s personality and situation. (AW2H)

Cohort 1 participants also reported frequent and direct communication between VDs and AWs, and between VDs and psychiatrists. However, there were few direct interactions between AWs and psychiatrists. A cohort 1 AW remarked that women workers and psychiatrists were not in the same circle (AW1A).

Two factors enabled the teams to work well together: geographic proximity and preexisting relationships. Ordinarily VDs and AWs had frequent contact with each other because they lived and worked in the same village, and some had known each other for decades. Their well-established personal and professional relationships, and common understanding of the village context, facilitated collaboration. Likewise, VDs said that they knew the psychiatrists well before COACH and communications with the psychiatrists were easy:

It’s very easy for us to get professional guidance. I call the psychiatrist whenever I have a problem and get his reply almost immediately. If there is something we can’t talk clearly over the phone, it’s also easy to meet face to face. They are in the hospital, quite convenient. (VD2A)

A psychiatrist noted that the collaboration between VDs and psychiatrists was smooth because village clinics were under the supervision of the county hospital (Psy C).

Reflexive Monitoring

All team members appraised the COACH model positively. VDs, in particular, reported witnessing positive changes in their patients and the reciprocal influence of mood and hypertension.

After their mood improves, they sleep well and their blood pressure gets better. (VD2A)

The most significant change in patients is their health conditions.…when the patients think rationally and feel better, other symptoms are also reduced. (VD1E)

Both VDs and AWs felt that their positive interactions with patients have contributed to the patients’ positive change. VDs said that “talking and listening” was more effective than medicine.

After we talk for a little while, she would feel warm and secure. I think this is much more effective than taking medicine and is good for the family. (VD1G)

AWs echoed the healing power of care and concern and said that their home visits increased patients’ positive mood which led to improved health.

He knows that you care about him and you ask him questions with genuine concerns. He knows it. Because he knows that someone cares about him, he smiles. He smiles beautifully even though he has no teeth. He feels good. (AW2B)

No focus group participants, however, mentioned the use of antidepressants as a cause of positive change in patients’ symptoms and health. Both VDs and AWs claimed that patients confided in them because they were able to establish trusting relationships with patients. Such trust, both said, was built on continuous contact and small talk.

We follow up regularly… So the patients trust us. They know that we will protect their privacy. They tell us their thinking and feeling and whatever bothers them. (VD1E) We visit them [patients] regardless of the day of the week. …They tell us problems that they cannot tell their children. (AW1G)

The family is an important support system for older patients. VDs and AWs felt that older patients’ family support increased as a result of the COACH intervention, due partly to reduced depressive symptoms and partly to the education of their family members. As explained by an AW and a VD:

I ask adult children of older patients to visit their parents as frequently as possible. Older people need care and concern from their children more than anything. It’s more effective than medicine. So we talk to the adult children. (AW1B)

Due to [the older person’s] depressive symptoms, family members find it hard to communicate with them. After our intervention, their mood elevates and the communication with family members is improved. The whole family atmosphere is much better. I feel very good seeing that. (VD2A)

Overall, we found that the COACH team members—VDs, AWs, and psychiatrists—demonstrated a shared understanding and appreciation of two core characteristics of the COACH model: the team approach and integrated behavioral health and medical care. Their leaders’ support of the COACH study motivated team members’ participation, and training received from the COACH research team cemented their engagement. However, some VDs and AWs from cohort 1 had grievances which may deter their commitment. Coordinated teamwork and frequent communication between team members were reported. Geographic proximity, pre-exiting relationships, and organizational structure helped the functioning of the team. Team members appraised the COACH model as effective in improving patients’ health, and they attributed the effectiveness to the care and concern they provided to patients.

Discussion, Implications, and Limitations

It has been recognized that community mental health services provided by a multidisciplinary team is the appropriate approach to meet the challenge of a growing population of older persons with psychiatric and physical ill-health in the developed world and in China (Tse et al., 2013; Wilberforce et al., 2013). Yet, problems related to the functioning of multidisciplinary teams, including low engagement of primary care providers (PCPs) and breakdown in inter-professional communication, are common (Overbeck et al., 2016; Wood et al., 2017). These problems were infrequent in COACH. Our findings suggest that several factors may have influenced the implementation of COACH. We depict these factors and their relationship with the four NPT mechanisms in Figure 1, which may serve as a conceptual model for planning collaborative depression care in LMICs.

Figure 1.

Conceptual Model for Implementing Collaborative Depression Care in LMICs

Facilitators and Barriers

First, similar to that reported in developed nations (Wood et al., 2017), sufficient training is essential. The training that VDs and AWs received from the COACH study emphasized teamwork in addition to knowledge and skills. Based on feedback collected after each training session (not reported above), we knew that the training was well received. The training enabled a common understanding of the COACH model and facilitated smooth team collaboration. Also, the training has empowered the team members to intervene, and positive results of their intervention further strengthen their beliefs in the team approach and integrated care.

Second, geographic proximity and pre-existing relationships among team members made COACH-focused communication and collaboration easy and natural, enabling collective action. Many VDs and AWs know each other well, and in many villages have adjacent offices. Western studies have reported that co-location of PCP and case manager facilitates the implementation of collaborative depression care (Overbeck et al., 2016; Wood et al., 2017). In addition, existing organizational structure supports collaboration among COACH care team members. For example, village clinics are administratively under the county hospital where psychiatrists work.

Third, the COACH activities were appropriate to the skills of the participants and can be easily integrated into their already existing practices. For example, VDs found it convenient and useful to administer the PHQ-9 during patients’ clinic visits. Likewise, the COACH activity of visiting and supporting older patients was just an extension of the existing role of AWs whose official responsibility included paying home visits to villagers.

Unlike some coordinated depression care models in the U.S. (Ciechanowski et al., 2004; Unutzer et al., 2002), the AWs—the care managers in COACH—were not expected to provide counseling or therapy. But they were trained to provide emotional support and address social factors affecting patients’ treatment adherence and response. A study in England reports a proliferation of support workers without professional certifications in community mental health teams, partly because support workers added values to patient care by establishing personal relationships with patients and providing ongoing emotional and social support (Wilberforce et al., 2013). The AWs in COACH performed a similar role and were valued by other team members.

Fourth, leaders’ support is important to participants’ enrollment and engagement. Before the study began, the COACH research team secured support from village leaders and top officials of the county hospital. All training activities were conducted in the county hospital, which legitimized the COACH study for those involved. Having the administrator of the county hospital to recruit team members also reinforced the perception that those in authority approved and supported the COACH study. Western studies have suggested that organizational buy-in is a facilitator to set up collaborative care (Wood et al., 2017).

Finally, subjective factors also contribute to team members’ participation. Notably, VDs and AWs genuinely cared about older people in their villages and wanted to help. They received a small stipend for the tasks associated with COACH interventions and reporting requirements, but it was not enough to compensate for the time and effort they devoted to their patients. Their positive attitude toward older people appeared to help sustain their commitment to the COACH study.

However, the frustration and misgiving expressed by some VDs and AWs reveal potential barriers for sustaining and upscaling COACH. Cohort 1VDs complained about their low wages. Most of these VDs were from the ‘old’ generation with lesser training and lower pay. Similar complaints were not voiced by the second group of VDs, who tended more often to be of the ‘new’ generation that has received structured medical training and commensurate salary. Although the number of ‘old’ generation village doctors is declining in China, they are still the primary care providers of rural residents in many villages, especially the remote ones where the ‘new’ generation of village doctors are often not willing to go. The poor compensation may cause these ‘old’ generation VDs to resent any new practice such as COACH.

Cohort 1 AWs, whose official role as women worker makes them responsible for multiple and diverse village affairs, felt burdened by the extra work from COACH. In contrast, cohort 2 AWs, whose official role was health liaison responsible for promoting older villagers’ well-being, were motivated to continue the COACH model. Their different experiences may be related to the alignment of the COACH mission and their official responsibility. Whom to recruit as AWs in the local rural community is an important decision. Preferably, the COACH activities are integrated with the “official” responsibility of whoever plays the role of AWs.

The added workload for psychiatrists may also be a barrier. The psychiatrists in the focus group did not make workload an issue, only one mentioned that he/she worked in weekends to manage the extra work. Nonetheless, travelling to rural villages to conduct diagnostic assessment is very time consuming. Future applications of the COACH model may consider using technology, such as videoconferencing, to achieve a more efficient use of psychiatrists’ time (Ramos-Rios et al., 2012).

Limitations and Implications

Only a small number of VDs and AWs from the first two phases of the larger COACH study participated in the focus groups. They may not represent the perspectives of all VDs and AWs. Also, the sample may be biased if the COACH study coordinator excluded those who had a negative experience with COACH. Self-selection is also possible as those who had positive experience with COACH may have been more motivated to join the focus groups. Second, some study participants, especially those from cohort 1, may have faded memory of the COACH project, due to the relatively long time lag between the focus group interviews and their completion of COACH participation. Third, the findings are based on data from focus groups, a relatively short encounter. Finally, the voice of patients is not included in this study as our focus was the COACH care team. Future studies to understand patients’ experience will be necessary to improve the COACH model further.

The generalizability of the COACH model to other LMIC settings deserves discussion. On the one hand, the rural areas where COACH was implemented are located in a relatively ‘progressive’ province in China where the local health department is open to new ways to deliver health services. Their receptivity may be atypical relative to other areas of China or other LMICs. As well, some LMICs have virtually no psychiatrists, and many have even fewer per population than rural China or greater distances between the villages and the psychiatric hospitals/clinics (Patel and Hanlon, 2018). While the model might still apply, it will likely require adaptations that spread the psychiatrist support more thinly, for example, by telepsychiatry and telementoring (Komaromy et al., 2016; Ramos-Rios et al., 2012).

It was estimated that 173 million adults in China have mental disorders and 158 million of them have never received treatment (The Lancet, 2015; Tse et al., 2013). These numbers speak to the urgency of developing community mental health services in China, as hospital- and institutional-based services are neither affordable or desirable (Tse et al., 2013). The COACH model is based in primary care clinics which are more accessible and less stigmatized than mental hospitals. By utilizing a collaborative care approach and strengthening the capacity of primary care providers, the COACH model allows patients with mental illness to receive mental health treatment in the community. China is not alone in the challenge of providing mental health services. Many LMICs are faced with an increased number of people having comorbid physical and mental health problems, due partly to the growth of the older population (von Humboldt, 2016). Given the multiplicative adverse effects of comorbidity on health, disability and health care costs (Ho et al., 2014; Moussavi et al., 2007), it is imperative to develop scalable models of collaborative care that integrate mental health into primary care in LMICs (Tse et al., 2013).

Conclusion

This qualitative study reports and synthesizes the perspectives of team members who participated in testing the COACH model, identifying factors that facilitated the implementation of COACH as well as potential barriers. These findings offer guidance for how a primary care-based collaborative care model for comorbid depression and other medical conditions can be more broadly disseminated in China and other LMICs.