Introduction

Behavioral health disorders, such as depression, are among the most prevalent health conditions in New York State and throughout the country, disabling many and impairing successful control of or recovery from co-existing medical disorders, including diabetes, asthma, cardiovascular and lung diseases, cancer, and neurological illnesses. Although safe and effective treatments for depression exist, the great majority of people in need are not being detected or receiving adequate care due to: how they are managed (not managed, in effect) in primary care; access problems to specialty mental health care: a shortage of mental health specialists; and stigma [1].

As health care systems undergo transformation, care is becoming more consumer-centered and measures are being put in place to drive down overall costs [2]. In this health care environment, there is increasing recognition of the vital role of “integrated care” programs, and commitment to providing them (Appendix 6). Integrated care aims to provide both medical and mental health care in one setting, most often within primary care. Accessibility to depression treatment in primary care is convenient for consumers, can help to reduce the stigma associated with the treatment for mental disorders, builds upon existing doctor–patient relationships, improves care and outcomes for patients who have both depression and co-occurring medical disorders, and over time can reduce costs. Evidence also shows that patient satisfaction with integrated care systems is high [3, 4].

Collaborative Care (CC), as we use the term here, refers to an evidence-based model for delivering quality depression care in a primary care setting. Developed at the University of Washington and based on principles of chronic illness management, CC focuses on detecting depression in primary care using a specific validated screening test, then medical diagnosis of the disorder, followed by tracking those with the illness through a registry, with the use of a measurement-based depression care path that identifies needed changes in treatment if a patient does not improve; in addition, there is training of clinical and administrative staff in the practice, and educating and activating patients.Footnote 1 Collaborative Care has now been tested in more than 70 randomized controlled trials in the USA and in other countries, in a variety of treatment settings, in both urban and rural environments and with diverse patient groups (Appendix 1) [1].

Evidence suggests that collaborative care for depression not only improves consumer outcomes for depression but also for common co-occurring general medical conditions such as diabetes, hypertension and hyperlipidemia [5]. It has been shown to lead to better patient and provider satisfaction. In addition, CC has demonstrated cost savings in long-term studies when compared to conventional care [6].

Despite its robust evidence base, large scale implementation of CC has been very limited. This is largely because CC requires practice changes on multiple levels—it is tantamount to a new way of practicing medicine. However, with this amount of evidence of its effectiveness, with improved patient and provider satisfaction, and with the need to reduce unnecessary spending, its adoption has been increasing and needs to scale-up further.

The New York State Collaborative Care Initiative (NYS-CCI)

New York State has committed its medical policy and practice goals in integrating behavioral (mental) health care into primary care. In what may be the largest state government behavioral health effort, through the New York State (NYS) Hospital Medical Home Program, the NYS Department of Health (DOH) and Office of Mental Health (OMH) have partnered to implement CC for depression across the state [7]. The NYS Collaborative Care Initiative (CCI) for depression has been part of a 2.5 year, Federal Hospital-Medical Home (HMH), Graduate Medical Education (GME) grant-funded project designed to advance primary care practices, including the integration of mental health, throughout New York State. The NYS-CCI is specifically the integration of depression care into ambulatory, primary care resident training sites, using selected Academic Medical Centers (AMCs). This project began in July of 2012 and ended in December 2014 when the grant terminated.

As the largest state sponsored implementation of integration of behavioral health in primary care to date, questions as to whether CCI implementation is feasible, successful, scalable and sustainable in NYS (and across the country) are critical clinical and policy questions that call for answers. We report on our experience and lessons learned here.

What are the Essential Elements of CC?

CC in a primary care setting has explicit requirements for what constitutes a clinical team and the essential elements of care that must be provided (Appendix 3). CC is delivered by a depression care team. This team approach includes: (1) training primary care providers in screening for and treating common mental health conditions, in this case depression; (2) employing in the primary care setting care managers who engage, educate, and provide basic counseling and medication support for patients diagnosed with depression and entered into the registry and treated; and, (3) psychiatrists who provide caseload consultation as well as consult on those patients who may need changes in treatment or more intensive, specialty services to care managers and primary care physicians, principally by telephone or video.

The CC approach also requires a very particular set of tools: a standardized screen for depression [the Patient Health Questionnaire-9 (PHQ-9)] [8] (Appendix 7) to detect and track the progress of depressed patients using a registry (similar to diabetes and asthma registries) (see footnote 1). The monitoring and tracking allows primary care doctors and care managers to adjust and/or intensify treatment if clinical improvements are not achieved as expected, much as tracking a person’s hypertension would lead to changes in treatment. Referrals to specialty mental health care are typically also reduced as effective care is delivered in the primary care setting, thereby sparing specialty mental health resources for those with the most significant mental health conditions.

The NYS CCI Project: Implementation

OMH contracted with the AIMS center at the University of Washington to provide the technical assistance needed by participating clinics to implement CC through the HMH project. The AIMS center subcontracted with the Institute for Family Health (IFH) to provide training and technical assistance in NYS. A core team developed the training program specific for the New York state sites. In all, 19 academic hospital centers across NYS chose to implement CC at one or more of their primary care clinics (32 clinics serving 1 million patients).

The technical assistance was designed to be delivered in two tracks: “Innovator” and “PCMH grantee” sites. Six medical centers were chosen as innovator sites and received intensive training and technical assistance, including in-person workshops, webinars, and regular ongoing consultation. The other sites, the PCMH grantees, received webinars and information packages on implementation. This distinction fell away over time as many sites engaged their own technical assistance in addition to what this project provided.

All sites began seeing patients by the requirements of the grant by July 2013. They reported quarterly on the project deliverables (specified in Appendix 2). The project deliverables were established to help ensure fidelity to the core aspects of the CC model and thus its likely success (Appendix 4).

The NYS CCI Project: Results

Nineteen Academic Medical Centers implemented CC in 32 of their primary care clinics. Over time, all clinics worked to improve their implementation of CC based on continuous information feedback and technical assistance; data collected during the project (by DOH) indicate almost all practices succeeded in delivering CC.

In general, with the exception of screening yield among those screened, which was relatively stable around 10–13 %, all other measures, including rates of screening, depression diagnosis given positive screens, enrollment in depression care given positive depression diagnosis, and improvement among those in treatment for 16+ weeks showed improvement over the course of the grant (Fig. 1).

Fig. 1
figure 1

Average rate at all sites, by quarter of the grant. Outcomes across all sites (n = 32)

Full size image

At the beginning of the implementation, many clinics reported they had some form of depression screening protocol in place; however, in fact, on average across clinic less than half (46 %) of patients served were screened for depression at the beginning of the grant. Over the 2 years of the project, participating clinics steadily increased their screening rate, with an end of grant average screening rate across clinics of 85 %, with many clinics near 100 %. Of the 32 clinics, 23 (72 %) met or exceeded the original goal of screening 85 % of all patients screened at least annually. Training and new practice protocols were put into place to ensure that depression screening became standard practice, much like measuring blood pressure or HgA1c (Table 1).

Table 1 Average proportions for all sites, by quarter of the grant
Full size table

The number of patients with a positive screen who were then diagnosed with depression also increased over the course of the grant, indicating that fewer patients with this illness were falling through the cracks. We saw evidence of better follow-up on positive screens and, with training, PCPs becoming more comfortable making a diagnosis of depression and treating the condition in their practices. The rate of diagnosis of depression among those with a positive screen increased from an initial rate of 44 to 66 % by the end of the grant, on average across clinics.

In terms of enrollment, the number of those diagnosed with depression who were subsequently enrolled in CC has increased from a low of 32 to 43 %. This has in part been due to increased staffing (especially of care managers), the use of Electronic Health Record fields or spreadsheets to enroll patients in CC, and the enthusiasm that comes from experiencing success in implementing an effective treatment that patients like. During the final quarter of the project, nearly 6000 patients were enrolled in CC, a threefold increase from the previous year, suggesting that as clinics become familiar with the model they can improve their patient engagement and retention in care.

Finally, CC was effective in reducing the burden of depression among a large proportion of those retained in treatment. At the end of the grant (Q7–Q8), 45–46 % of patients in treatment for at least 16 weeks showed improvement of their PHQ-9 scores to less than 10 up from only 17 % at the end of the first year (Q4); this is indicative of significant clinical improvement as scores <10 are generally not consistent with a diagnosis of clinical depression.

Provider and Consumer Experience

The participating clinics reported increased satisfaction with their implementation and the use of this model. Primary Care providers (PCPs) remarked that it is a pragmatic approach, appreciate the psychosocial support for their patients, and wish to see it sustained in their clinics. Anecdotal feedback from consumers was also very positive (Appendix 5).

One patient, who had a PHQ-9 that went from 16 (moderate depression) to 6 over the course of 5 weeks, was a 55 year old man with many chronic medical conditions who had recently moved to NYC and had little support. His treatment included sessions of problem solving therapy, which helped him organize his scheduling and time. His attention to his medical illnesses also improved as he became an active participant in his own self-care.

Another success story is that of a 58 year old patient who originally had a PHQ-9 score of 19 (moderate to severe). She had a history of recurrent major depressive disorder. Through her participation in a program of self-care as well as receiving better medication dosing and care manager support she was able to reduce her PHQ-9 score to 9, as well as cut her smoking down to half within a few weeks. She reported new found optimism for her future.

Across the sites, both clinicians and patients commonly reported enthusiasm about delivering CC.

Because the fully realized adoption of CC was at most 1½ years, less in most sites, we did not attempt to determine if medical costs were reduced in the population receiving CC. Ongoing work, supported by a Medicaid supplemental payment, will attempt to do so.

Lessons Learned and Barriers to Sustainability

Overall, there was considerable performance improvement by primary care practices, over 2 years, in the implementation of CC. With the right training and support, CC is both feasible and effective. In fact, we anticipate the results to date to increase over time, as long term studies of the CC model show that gains accrue—especially in years 3 and 4, both in terms of cost effectiveness and reduced medical morbidity among patients with co-existing depression.

However, there have been challenges to the large sale implementation of CC in NYS. Trying to change the attitudes of physicians and creating a radical shift in the way medicine is practiced often initially prompts many clinicians (and administrators) to resist. Primary care physicians have traditionally been reluctant to treat depression in primary care, and psychiatrists have been reluctant to manage care through a caseload model of consultation (without face to face evaluation). Integration, thus to date, has not been a standard of primary care practice. Because Collaborative Care is a fundamental departure from usual care, it requires practitioners to orient to the model and learn new roles—an often underappreciated aspect of implementing Collaborative Care.

Another challenge was that there have been other demands on practices related to other aspects of health care transformation, leaving many providers overwhelmed by new practice demands, the introduction of additional regulatory and payment requirements, and almost constant change. However, as clinics adapted to the model (with ongoing technical assistance) we received positive feedback that as primary care physicians and practices felt supported they became able to detect and treat depression in patients whom they had known to be ill for years but had never screened or diagnosed.

A second and substantial challenge involved the way the project was funded. Funding was provided centrally to the AMCs, not to the actual primary care practices; some, as a result, encountered barriers to receiving the money they needed from their central offices to implement the model properly. Many AMCs were initially reluctant to hire the additional staff required for such a model, concerned about the end of the funding period and how they would pay for such staff or bill for the new care methods required whose expenditures had been covered by the grant. Many sites reported hiring freezes or significant delays in obtaining approval to hire additional staff with no clear, future funding stream to support staff time.

A third challenge was that practices seemed reluctant to fully invest in the training and quality improvement of a model that itself came with a variety of regulatory and licensing burdens. Along with insecure funding, regulatory barriers added to the reluctance of practices to fully commit to implementation of a model whose sustainability remained uncertain.

Another challenge worth noting is the difficulty in obtaining standardized performance reporting. Even well operationalized metrics may not be reported in the same fashion across providers without built in quality checks. Given the large number of practice transformation projects typically underway in primary care practices today, provider capacity to respond to multiple third party quality improvement data requests is limited.

Recommendations

The following are recommendations to sustain CC based on the experience of the CCI in NYS.

  1. 1.

    There must be a clear and credible path to state level payment mechanism(s) beyond grant funding.

  2. 2.

    Clinics must be able to implement CC without undue regulatory and licensing burdens; for example, meeting both the requirements of the departments of health and mental health.

  3. 3.

    There needs to be continued support for training and supervision in integrated care and attention to recruiting and retaining the staff needed to deliver CC, including the hiring and supervision of care managers and the presence of psychiatrists needed for consultation in collaborative care.

This is a remarkable time in health and mental health transformation—perhaps the greatest changes we have seen in the country since the 1960s. We are seeing an historic push toward truly integrated care. We believe the NYS-CCI project offers experience, knowledge, and hope to propel health care systems forward in delivering integrated mental health care. What we have achieved can be scaled-up further in NYS, and throughout this country.