Collaborative Care Implementation: Lessons Learned

Am J Manag Care. 2022;28(8):e308-e311. https://doi.org/10.37765/ajmc.2022.89204

Takeaway Points

• We established core principles of collaborative care: evidence-based population management, timely communication, patient-centered care with patient-centered decision-making, and honoring primary care culture.
• We created new tools to aid staff in implementing collaborative care, including smart phrases for the electronic health record, scheduling templates, and a self-assessment tool.
• Using core principles to make decisions about what the work is, who should do the work, and how it should be done has enabled the core group to streamline our implementation; this framework may be applicable to other organizations involved in implementation of new models.

In US primary care, helping patients with depression is a daily crisis. Approximately 15.7 million adults in the United States (7%) had at least 1 major depressive episode in the past year that caused functional impairment and contributed to lost productivity, costing an estimated $23 billion in 2013.^1,2 However, at least one-third of primary care patients receive inadequate treatment for depression.^3-7 Over the past decade, collaborative care has emerged as the new standard for depression treatment in primary care.^8-10 Recommended by the US Preventive Services Task Force, this model speeds improvement through increased engagement, symptom monitoring, and rapid treatment adjustment.¹¹

Although collaborative care is a proven solution to this crisis, implementing it in any medical system exposes unexpected challenges.¹² As an implementation group shapes a collaborative care model to fit its local environment, there is no systematic method to determine which parts of the model can be modified or where flexibility might undermine success.¹³

The goal of this paper is to describe how we developed a serviceable framework for making implementation decisions during a collaborative care pilot. As implementation occurs, there are inevitable points of friction: Something that seems financially or logistically necessary conflicts with what the model recommends. Stakeholders agree that they want a successful implementation but disagree about how to accomplish this vision. When decision makers agree not only on goals but also on the fundamental precepts those goals are based on, all implementation decisions get filtered through a lens of underlying shared values. By asking whether the proposed suggestion aligns with these shared values, a group can then make 1 of 2 decisions: either (1) an idea has merit and flexibility should be allowed in the model, or (2) an idea undermines 1 or more of the shared values and fidelity to the original model must be enforced. By dissecting problems our group faced and lessons we learned, this work can help other institutions approach any implementation more effectively.

Setting

UW Health is a large public academic health system (the integrated health system of the University of Wisconsin–Madison) with 400,000 medically homed patients in 24 adult medicine and family medicine academic and community clinics. Starting in November 2013, the organization underwent a comprehensive primary care redesign that involved panel-based compensation, redefined care teams, and standardized workflows.^14-16 Our collaborative care pilot planning and implementation took place in this climate of revising how primary care was delivered.

In 2013, we created a hybrid model of collaborative care based on the Improving Mood—Promoting Access to Collaborative Treatment (IMPACT) study to integrate mental health treatment into primary care.¹⁷ The developed model was approved in 2014 for piloting. In 2015, an 8-person group, including stakeholders from psychiatry, primary care, and operational leadership, was charged with implementing the pilot. We hired the University of Washington’s Advancing Integrated Mental Health Solutions (AIMS) Center (the primary investigators of the IMPACT trial) to help with implementation,¹⁷ meeting in person twice and having weekly phone calls as we planned and piloted our program in a family medicine clinic and an internal medicine clinic.

Collaborative Care Implemented in Response to a Crisis

A 2013 gap analysis showed that primary care providers (PCPs) and patients were dissatisfied with mental health care delivery, particularly depression treatment. Provider satisfaction was low; working to meet primary care redesign metrics left little time to open a “Pandora’s box” of depression symptoms. Patients were dissatisfied; they wanted more accessible visits with competent providers who could treat depression at their familiar primary care clinic. UW Health’s Department of Psychiatry was overwhelmed with triaging the needs of thousands of patients, some of whom could have been treated in primary care.

With guidance from AIMS, our group began making decisions about roles, job descriptions, and workflows. Our initial model included 2 care managers: a primary care registered nurse (RN) and a licensed mental health provider. We called these care managers registered nurse care coordinators (RNCCs) and behavioral health clinicians (BHCs), respectively. Both the RNCCs and BHCs would maintain a patient registry and use the AIMS approach of structured problem-solving therapy to help patients meet self-identified goals in short, frequent visits over 3 to 6 months. The professionals in both roles reviewed cases with the consulting psychiatrist and communicated with the PCP.

Unexpected Obstacles Forced Reevaluation of the Implementation

Although patients were very pleased with our pilot program and made significant clinical improvement, our core group encountered unexpected obstacles that forced us to reevaluate many of our initial implementation decisions.

Both the RNCCs and BHCs were uncomfortable with their roles for different reasons. Although the BHCs philosophically agreed with a primary care model of psychiatric management, in practice, they tended toward a more traditional approach to psychiatric care of the patient. The primary care environment was fast paced, with PCPs and other providers dealing with complex issues in 15- to 20-minute visits. Some of the BHCs had difficulties adjusting to this aspect of primary care culture and struggled to fit collaborative care therapy interventions into short (20-30 minutes), problem-focused visits. BHCs routinely lapsed into traditional 50-minute visits, saw fewer patients during the day, and had smaller panels than the model recommended. In addition, our pilot initially focused on patients with depression and supported the use of only 1 type of psychotherapy: problem-solving treatment (PST), a behavioral form of cognitive behavioral therapy. This prescribed diagnostic scope frustrated the BHCs, who were used to treating more complex diagnoses. Being limited to use only the AIMS-recommended PST was too restrictive for the BHCs; they wanted autonomy to choose the psychotherapy treatment they thought would best serve the patient.

The RNCCs were part of a preexisting chronic care management program. We later learned that some of the RNs felt pushed to accept the new role of RNCC, which was required by primary care redesign, and they missed doing team nursing work that was more familiar to them. Although we provided the AIMS-recommended training for their new tasks, most of the RNCCs felt inadequately trained to treat patients who were depressed. Lack of stable infrastructure and the fact that most RNCCs did not enjoy their existing work or the pilot work made it clear that we needed to remove the RNCC role from our treatment team.

Some PCPs were quick adopters, eager to refer to collaborative care, whereas other providers struggled with changing their mindset and workflows. Some PCPs lacked confidence in the RNCCs and were reluctant to change their routine pattern of referring all mental health problems to specialty psychiatry. Some PCPs saw the more highly skilled BHCs as a safety net and referred any patient with mental health problems, ignoring the inclusion criteria of a depression diagnosis. Whereas some providers understood the therapeutic importance of introducing the BHC in person to make a “warm connection,” which invites patient engagement, others felt it was time-consuming and omitted that step, preferring to just order a collaborative care consult in the electronic health record.

After reviewing an unexpected large increase in referrals to specialty psychiatry, we discovered that (1) many of the referrals were for patients with mild or moderate depression who could have been treated in our collaborative care pilot and (2) if we expanded our pilot’s scope to include anxiety disorders, we could manage 50% of the recent referrals to specialty psychiatry in our collaborative care program.

It became clear that our current implementation plan was not sustainable with the personnel and approach that we had originally planned, and to proceed, our core group needed a systematic way to make myriad decisions around shared priorities. Although the AIMS Center had documented the core principles of collaborative care, these had not been presented as a decision-making tool. In our group, some members wanted to modify the AIMS model to fit our institution’s needs, whereas other members wanted to abandon that model altogether and find an easier model to try. Because we did not have a rubric of shared values that we all agreed on for collaborative care at our institution, decision-making was often cumbersome as we tried to find solutions that aligned with each member’s personal perception of what collaborative care was supposed to be.

Defining Shared Values Led to Improved Implementation Decisions

Although the AIMS Center had given us solutions to address specific problems, our core group needed to clarify and simplify our process of making numerous group decisions in a limited time frame. When one member suggested that we had not yet truly implemented our intended model, we began to clarify the key elements that would define fidelity for collaborative care in our specific context. We drafted a document, “Shared Values of Collaborative Care,” that outlined our fundamental principles of collaborative care: evidence-based population management, timely communication, patient-centered care with patient-centered decision-making, and honoring primary care culture. This document and other materials are available for free download online.¹⁸

The first shared value required fidelity to evidence-based population management. The IMPACT model used a population health lens to help as many patients as possible using short-term, evidence-based techniques and systematically focusing on patients who were not improving. We had the technology to track populations of patients on a “workbench,” but our psychiatrists and BHCs were using this tool only intermittently. Our group realized that it was worth the time and effort to make this training much more robust so psychiatrists and BHCs would rely on the patient registry workbench to consistently track how every patient moved through the treatment process. Using the workbench ensured that no patient fell through the cracks. We intensified training in population health management for the team psychiatrists and BHCs so they understood why it was critical to use depression assessment scores and other registry data, rather than just their general clinical impression, to drive their patient management decisions. We redefined an ideal panel size to 70 to 100 patients for each BHC and broadened their therapeutic scope to include anxiety because clinics had the evidence-based anxiety screening tools and BHCs had considerable skill with evidence-based anxiety treatments. We made sure that providers, medical assistants, and RNs understood when, why, and how to use the screening tools.

We also identified areas where we needed to be more flexible and allow changes that would better fit our particular institution. Although the AIMS Center recommended that we consider continuing with only PST, we decided that if the BHC was using short-term, evidence-based techniques, they could choose any treatment modality they felt was appropriate for the patient. This decision enabled BHCs to provide more kinds of evidence-based treatment. They were more satisfied with their work and more willing to use the registry workbench to track the outcomes of their interventions.

The second shared value required fidelity to timely communication. The IMPACT model encouraged BHCs to contact patients early after the original referral and frequently thereafter to maintain engagement. The treatment team, which included the PCP, BHC, and team psychiatrist, should behave like a primary care team and communicate efficiently and effectively among themselves and to patients. We reinforced that BHCs should be calling the patient within 3 days of the warm connection and scheduling a prompt intake appointment. We also empowered BHCs to identify unengaged patients and remove them from their panel to make room for other patients. The BHCs and consulting psychiatrists were struggling to get through the recommended 6 cases per hour during their weekly case review meeting, then had difficulty communicating care plans promptly to PCPs. We realized that psychiatrists who had never worked in primary care were asking BHCs for diagnostic details that were not relevant to the focused therapy that BHCs were providing. We revised our team psychiatrist training to stress how collaborative care required them to be comfortable making diagnoses with different information. We developed preparation tools to help BHCs use a crisp, concise presentation style when discussing cases with the consulting psychiatrist. We trained both the team psychiatrist and BHC to document in real time in the electronic health record using smart phrases and other documentation tools. We had BHCs shadow PCPs so they could understand the volume of decisions that PCPs make each day and why communication must be clear and brief with action items in bold type at the beginning of the message. These adaptations helped our teams maintain fidelity to the timely communication needed in primary care.

The third shared value was patient-centered care with patient-centered decision-making. This idea is central to collaborative care: Patients are treated in their familiar PCP’s office, set their own goals, and engage in their evolving treatment plan. Introducing the BHC through the warm connection is one of the most patient-centered ways to help patients with depression and anxiety because they have a chance to engage with the BHC and begin treatment immediately. We interviewed the PCPs individually about how they used the warm connection and reinforced how much better patients engage with treatment when providers introduce the BHC in person at the office. We created a brief self-assessment tool that enabled the provider to review how consistently they were remembering to bring the BHC into the room to meet the patient.

The fourth shared value was honoring primary care culture. BHCs saw patients for short, focused therapy visits. They also built a culture of trust by helping everyone in the primary care clinic understand why collaborative care was the most appropriate treatment for patients who were depressed and/or anxious. Despite this relationship building, some PCPs continued to refer all their depressed and anxious patients to psychiatry. To encourage a change in this behavior, we created a more rigorous psychiatry referral in the electronic health record that required PCPs to specifically state why they felt the patient could not be treated in collaborative care and empowered the psychiatry triage team to redirect patients who were appropriate for collaborative care back to the PCP. We created a schedule template for BHCs to help them maintain shorter appointment times. One of our implementation team members, a PhD psychologist who worked in collaborative care, made our interview process more rigorous and thorough. She hired BHCs who understood that the content and pace of primary care therapy is fundamentally different from that of traditional psychiatry. She developed her own training about working in primary care to ensure that their treatment was brief and nimble, thus helping them establish and maintain competency in the world of busy clinics. In 2021, we hired and trained 21 BHCs to serve our 24 primary care clinics. Our BHCs have collectively seen 7500 patients with depression and/or anxiety and helped 53% of those patients achieve remission in an average of 5 months.

When we started implementation, billing for collaborative care used therapeutic treatment codes that did not accurately capture the breadth of BHC work that occurred on phone calls. The adoption of specific collaborative care billing codes in 2020 enabled BHCs to bill for the actual work they do while co-pay costs remained low for most patients.

CONCLUSIONS

Using shared values as our primary decision-making tool streamlined our collaborative care implementation. When there was conflict or uncertainty about whether to adapt the IMPACT model to meet our needs or adapt our work to be faithful to the model, we relied on our self-defined shared values to make that decision. This framework may be applicable to other organizations involved in implementation of new models. UW Health leadership recently endorsed collaborative care as a top priority and has now implemented it in all 24 primary care clinics.

When someone proposes an implementation change, our core group no longer deliberates about whether we should stay true to the IMPACT model or be flexible and open to change. We use our shared values tool to decide whether a proposed change will compromise or support our goals, and we use these shared values to explain to others why their proposed change is feasible or not. Our group dynamics are more cohesive, and we have more confidence about making implementation decisions that bring collaborative care to our patients who need accessible, affordable, and effective treatment for depression and anxiety.

Author Affiliations: Department of Family Medicine and Community Health (EP, BA, JL, MAS), Department of Academic Affairs (WC), Division of General Internal Medicine, Department of Medicine (HH, MM), and Department of Psychiatry (HH), University of Wisconsin School of Medicine and Public Health, Madison, WI; Primary Care Academics Transforming Healthcare Collaborative, UW Health (EP, NP, WC, SD, BA, MM, JL, MAS), Madison, WI; Department of Family and Community Medicine, University of New Mexico Health Sciences Center (NP), Albuquerque, NM; University of Wisconsin Law School (SD), Madison, WI; Center for Patient Partnerships (SD), Madison, WI.

Source of Funding: Support was provided by the Primary care Academics Transforming Healthcare (PATH) writing collaborative, the Health Innovation Program, and the UW-Madison School of Medicine and Public Health from the Wisconsin Partnership Program.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (EP, NP, WC, HH, MM, MAS); analysis and interpretation of data (WC, HH, BA, MM); drafting of the manuscript (EP, NP, WC, HH, BA, JL, MAS); critical revision of the manuscript for important intellectual content (EP, NP, HH, SD, BA, MM, JL, MAS); administrative, technical, or logistic support (NP, SD, JL, MAS); program design and refinement (SD); and supervision (EP).

Address Correspondence to: Elizabeth Perry, MD, Department of Family Medicine and Community Health, University of Wisconsin School of Medicine and Public Health, 088 UW Clinic Fitchburg, 5543 E Cheryl Pkwy, Fitchburg, WI 53711. Email: Elizabeth.Perry@fammed.wisc.edu.

REFERENCES

1. The global burden of disease: 2004 update. World Health Organization. 2008. Accessed October 25, 2018. https://apps.who.int/iris/bitstream/handle/10665/43942/9789241563710_eng.pdf

2. Shao Z, Richie WD, Bailey RK. Racial and ethnic disparity in major depressive disorder. J Racial Ethn Health Disparities.2016;3(4):692-705. doi:10.1007/s40615-015-0188-6

3. Joo JH, Solano FX, Mulsant BH, Reynolds CF, Lenze EJ. Predictors of adequacy of depression management in the primary care setting. Psychiatr Serv.2005;56(12):1524-1528. doi:10.1176/appi.ps.56.12.1524

4. Kniesner TJ, Powers RH, Croghan TW. Provider type and depression treatment adequacy. Health Policy.2005;72(3):321-332. doi:10.1016/j.healthpol.2004.09.008

5. Pence BW, O’Donnell JK, Gaynes BN. The depression treatment cascade in primary care: a public health perspective. Curr Psychiatry Rep.2012;14(4):328-335. doi:10.1007/s11920-012-0274-y

6. Stein MB, Sherbourne CD, Craske MG, et al. Quality of care for primary care patients with anxiety disorders. Am J Psychiatry.2004;161(12):2230-2237. doi:10.1176/appi.ajp.161.12.2230

7. Young AS, Klap R, Sherbourne CD, Wells KB. The quality of care for depressive and anxiety disorders in the United States. Arch Gen Psychiatry.2001;58(1):55-61. doi:10.1001/archpsyc.58.1.55

8. Aregonès E, Palao D, López-Cortacans G, et al. Development and assessment of an active strategy for the implementation of a collaborative care approach for depression in primary care (the INDI.i project). BMC Health Serv Res.2017;17(1):821. doi:10.1186/s12913-017-2774-2

9. Katon WJ, Schoenbaum M, Fan MY, et al. Cost-effectiveness of improving primary care treatment of late-life depression. Arch Gen Psychiatry.2005;62(12):1313-1320. doi:10.1001/archpsyc.62.12.1313

10. McGough PM, Bauer AM, Collins L, Dugdale DC. Integrating behavioral health into primary care. Popul Health Manag.2016;19(2):81-87. doi:10.1089/pop.2015.0039

11. Siu AL; US Preventive Services Task Force. Screening for depression in adults: US Preventive Services Task Force recommendation statement. JAMA.2016;315(4):380-387. doi:10.1001/jama.2015.18392

12. Eghaneyan BH, Sanchez K, Mitschke DB. Implementation of a collaborative care model for the treatment of depression and anxiety in a community health center: results from a qualitative case study. J Multidiscip Healthc.2014;7:503-513. doi:10.2147/JMDH.S69821

13. Herrera S, Peccoralo L, Hinrichsen GA. Treating depression in an urban primary care setting: introducing an adapted therapeutic approach to improve behavioral health engagement in primary care. Soc Work Health Care.2018;57(8):607-619. doi:10.1080/00981389.2018.1474162

14. Caplan W, Davis S, Kraft S, et al. Engaging patients at the front lines of primary care redesign: operational lessons for an effective program. Jt Comm J Qual Patient Saf.2014;40(12):533-540. doi:10.1016/s1553-7250(14)40069-2

15. Koslov S, Trowbridge E, Kamnetz S, Kraft S, Grossman J, Pandhi N. Across the divide: “primary care departments working together to redesign care to achieve the Triple Aim.” Healthc (Amst).2016;4(3):200-206. doi:10.1016/j.hjdsi.2015.12.003

16. Zhong X, Williams M, Li J, et al. Primary care redesign: a simulation study at a pediatric clinic. In: 2014 IEEE International Conference on Automation Science and Engineering (CASE). IEEE; 2014:546-551. doi:10.1109/CoASE.2014.6899379

17. Unützer J, Katon W, Callahan CM, et al; IMPACT Investigators. Collaborative care management of late-life depression in the primary care setting: a randomized controlled trial. JAMA.2002;288(22):2836-2845. doi:10.1001/jama.288.22.2836

18. Perry CE, Wells S, Huang H. Shared values and tools for implementing collaborative care. University of Wisconsin Madison. 2021. Accessed February 1, 2021. https://www.HIPxChange.org/CollaborativeCare