Robert E. McGrath, PhD

Just about every article or conference on health care reform these days treats “integration” and “ending silos” as essential requirements for improving the U.S. health care system. In this article, I have no intention of diverging from the general cheerleading. However, I think it is important that we be clear about what is involved in integration and what is involved in ending silos. In the next two sections I will address each of these questions, followed by a description of a program I direct that illustrates the issues involved.

Continuing Education Information

The Meanings of Integration

The term “integration,” when referring to the relationship between biological and mental/‌behavioral health care (including substance abuse services), is often used in three very different ways. The first is when a traditional specialty mental health service is placed in physical proximity to a primary care or other medical service. This proximity encourages referral from one to the other, usually from the medical service to the mental health service. With the possible exception of requests to the mental health service for consultation/‌liaison on individual cases, though, the two often operate in relative isolation. Though practitioners in these settings like to think of themselves as integrated, for purposes of clarity it is probably better to call them co-located.

The second model emphasizes integration in treatment planning, and emerges out of the work of Wayne Katon (e.g., Katon et al., 1995):

  • A care team, which often includes a mental/‌behavioral health provider, develops an integrated treatment plan.
  • The treatment plan is overseen by a care manager who, for example, facilitates appointments with primary and specialty care providers.
  • This model has typically focused on pre-identified medically complex patients who are high utilizers of health care services. These individuals tend to demonstrate a mix of both psychological and medical difficulties, so the resulting treatment plan typically involves both types of interventions.

This model has been referred to as the collaborative care model.

Finally, there is a variant that emphasizes integration of mental/‌behavioral health services in the context of primary care (Butler et al., 2008):

  • The behavioral health provider is a member of the primary care team. There is no firewall between behavioral and primary care services.
  • Behavioral interventions are offered in a format consistent with primary care. That is, the nature of the problem may be unknown prior to initial evaluation during the primary care visit, immediate brief and targeted interventions are employed, and the goal is to improve functioning quickly.
  • Like other members of the primary care team, the behavioral health provider is available to see any patient at any time (warm handoffs from the primary care provider), discuss a case with another provider (curbside consults), collaborate on the development of a treatment plan, or provide brief assessments and targeted interventions.

Though the term integration is frequently applied to both the collaborative care and integrated primary care model, they are different in several important ways. The former emphasizes the coordination of specialty care, the latter the expansion of primary care. Mental/‌behavioral health services are not as central to the former model as the latter, though the high rate of co-morbid psychological disorders among medical complex patients means that mental/‌behavioral health providers should be included in the treatment planning team. One of the most important differences is that, because the collaborative care model focuses specifically on high utilizers of medical services, it has been easier to demonstrate cost savings associated with the model (e.g., Katon et al., 2012). In fact, very few of the studies that ostensibly demonstrate a costs offset for integrated care involve the integrated primary care model.

The potential contribution of psychologists can also differ substantially in the two models. Because of the variety of problems faced and the lack of prior information about the patient’s condition, the mental/‌behavioral health provider in the integrated primary care setting has to demonstrate a fairly high level of training, and psychologists can play a particularly important role as care providers. In contrast, the collaborative care manager is usually a registered nurse, social worker, or even a paraprofessional. Some settings do incorporate psychologists in the treatment planning team, but this role is often filled by a psychiatrist or social worker, with the result that psychologists are often excluded from collaborative care completely.

In summary, the term integration has been used to refer to specialty mental health care co-located with primary care (a particularly poor use of the term), collaboration in an interdisciplinary treatment planning team that is often managed through primary care, and embedding of mental/‌behavioral health providers in primary care. Any discussion of integrated care must be clear about what form is being discussed at any time.

The Meanings of Silos

When psychologists talk about the end of silos, they tend to mean the lack of coordination between specialty and primary care providers in general, and the lack of coordination between specialty mental/‌behavioral health providers and primary care providers specifically. It is not surprising that this is how psychologists tend to think about siloing, since it is about us, but focusing on this issue alone underestimates just how fragmented our health care system has become, and what will need to be fixed before fragmentation ends.

In fact, there are at least three other sources of fragmentation in U.S. health care that must be addressed before we can even hope to create an effective system of care. One is fragmentation in the communication of information across providers and health care settings. In fact, there a community of information technology specialists within health care that uses the word siloing specifically to refer to the lack of coordination between information systems. A second fragmentation is the lack of coordination between outpatient and inpatient care. The final one is the division between the haves and have-nots within the health care system. Addressing these three factors is just as important to the future of integration as is the fragmentation of primary and specialty care, and each will shape the movement towards integration in important ways.

The Primary/Specialty Care Silo

As already noted, it is the first of these four that has received the greatest attention among psychologists. The lack of coordination between primary and specialty services is widespread in health care, but particularly acute in the case of specialty mental health care. At the administrative level, the growth of this silo can be attributed to the decision to carve out mental/‌behavioral health services from the rest of health care, a decision it has been estimated affected almost 1 in 3 Americans by 1996 (Oss, 1996). The carve-out for mental/‌behavioral health occurred for several reasons. Several studies found that management of these services reduced costs by as much as 40%, mainly through reduced use of hospitalization (Frank, McGuire, & Newhouse, 1995; Ma & McGuire, 1998), and the formation of separate managed care entities specializing in mental/‌behavioral health that served multiple insurers allowed for a broader distribution of risk.

...very few of the studies that ostensibly demonstrate a cost offset for integrated care involve the integrated primary care model.

Though the strategy proved to be good business, it was bad practice. It encouraged treating mental/‌behavioral health as something distinct from the rest of health care. It interfered with the potential for coordinating behavioral and medical health services. In particular, it interfered with efforts to create a health care system in which medical and psychological conditions are treated holistically. Health care reform cannot proceed without reimbursement reform, with the mental/‌behavioral health carve out representing one of the most egregious factors in system fragmentation.

That said, nefarious insurors are not the sole contributors to the fragmentation between mental/‌behavioral and medical health. The specialty mental health system is still largely dedicated to the use of relatively long-term individual psychotherapy, which is a relatively labor-intensive treatment modality compared to other modern treatments in health care. As a result, primary care providers consistently report that mental health is the most difficult of the specialties to access (Cunningham, 2009). It is not surprising to find then that about half of individuals who receive treatment for psychological disorders did so in primary care (Kessler et al., 2005), a setting biased towards the use of medications over psychological interventions.

The Information Silo

The fragmentation in information sharing in U.S. health care is similarly largely the result of forces outside our control, and partly of our own making. In the early stages of health care reform, it was thought that the use of electronic health records (EHRs) would create the opportunity for sharing information across health care settings, which would in turn provide the foundation for seamless health care across settings and specialties. It was this belief that motivated passage of the Health Information Technology for Economic and Clinical Health Act (HITECH) in 2009, which rewarded physicians and other medical providers for the “meaningful use” of EHRs. Unfortunately, the reality has so far been underwhelming. EHRs have failed in terms of interoperability (e.g., United States Government Accountability Office, 2015), i.e., their ability to share information across systems. Though this was the primary justification for an investment of billions of dollars in these systems, this outcome was almost preordained when the decision was made to allow free market forces to determine the selection of EHR vendors across health care systems. Physicians note a number of negative consequences of EHRs as well, including their time demands, interference with the patient relationship, and the number of alerts generated in normal use (Friedberg et al., 2013). It is not surprising then to find participation in meaningful use of EHRs is actually declining (Conn, 2015).

Not one of the measures has to do with substance abuse, a remarkable omission.

Mental health providers have been particularly skeptical about the use of EHRs, driven by concerns about the potential sensitivity of information relevant to mental/‌behavioral health treatment and the need for confidentiality. I have found other health providers have little sympathy for this argument given the amount of sensitive information they also receive about patients, though the potential dangers should not be underestimated for mental health providers working with minors (Nielsen, Baum, & Soares, 2013; Smolyansky, Stark, Pendley, Robins, & Price, 2013). It is reasonable to anticipate that within the next decade standards will have evolved to the point at which mental health providers are often expected to store notes in a manner that will allow other providers access to information about participation in mental/‌behavioral health care.

The Inpatient/Outpatient Silo

There is nothing new to the lack of coordination between inpatient and outpatient providers; it is a problem that has bedeviled efforts at continuity in medical care for years. What is new is the growing recognition that the operation of outpatient and inpatient facilities as separate corporate entities means that no system-wide mechanism exists to incentivize outpatient providers to provide preventive services that reduce hospitalization rates. In fact, hospitals generally have little incentive to participate in systemic efforts to keep people out of the hospital.

This situation is slowly starting to change. One of the most important results of the Affordable Care Act was the formation of accountable care organizations (ACOs) responsible for the comprehensive care of Medicare patients (Fisher, Staiger, Bynum, & Gottlieb, 2007). For those unfamiliar with the model, the ACO accepts responsibility for comprehensive care of a certain number of Medicare lives in return for a per-member/‌per-month payment. If at the end of the year the ACO has used less than the amount allocated, the savings are divided between the ACO and the government according to a formula that considers performance on a series of 33 quality measures (RTI International, 2014). One of the major concerns faced by mental health providers within this system is that only one of the 33 measures has to do with a mental health issue, and that measure is procedural rather than outcome-based: patients should be consistently screened for depression, and when the screen is positive a follow-up plan should be implemented. Not one of the measures has to do with substance abuse, a remarkable omission. A number of measures do address issues frequently addressed by health psychologists (e.g., controlling blood pressure), or issues that are frequently exacerbated by mental/‌behavioral health problems, but at this point the measures demonstrate little concern about the psychological health of Medicare recipients.

The Access Silo

The final issue I will raise here is the one of access. Research suggests 60% of people with a behavioral health disorder and 40% of those with a severe mental illness do not receive behavioral health treatment of any kind (Kessler et al., 2005), with more than 2% of the entire adult population indicating they did not receive any mental health services in the past year despite the perceived need for such services (Substance Abuse and Mental Health Services Administration, 2014). The largest contributor to this lack of care would seem to be access due to financial limitations.

Psychologists and other health care providers have acquiesced to a system of inequitable access for years. Our failure to grapple with how to create a mental health system that allows for access for all has barely registered on the radar of psychology, but the physician Paul Farmer has raised it is a key ethical failure of all the health care professions (e.g., Farmer & Campos, 2004). It is hard for me to think of psychology as a truly ethical discipline unless we pursue the redesign of health care to allow access to mental health care regardless of economic resources. Unfortunately, the labor-intensive model of mental health treatment that is the norm in specialty mental health care—a model I believe in deeply, by the way—is an important contributor to inequities in access to care. Even without considering the cost of care, there is a large portion of the population that is incapable of setting aside 45-50 minutes per week (plus travel time) for mental/‌behavioral health care no matter how much they need it.

Integration and Tearing Down Silos

The integrated primary care model described earlier by definition requires tearing down the silos of primary care and mental/‌behavioral health care. For that reason, some psychologists are very skeptical of integration, fearing the new model will supplant specialty mental health care and offering only short-term interventions. In fact, the two are complementary. Many patients will not show sufficient improvement in response to brief interventions. In those cases the patient will ultimately have to be referred for more intensive treatment. Some have suggested that the exposure to mental health providers in the primary care setting actually improves the completion rate for specialty mental/‌behavioral health care referrals. That is, rather than seeing integrated care as a replacement for traditional specialty care, it should be perceived as a component of a future in which all care is stepped care, where initial interventions are of relatively low intensity, but failure to improve results in increasing levels of treatment intensity (McGrath & O’Donohue, in press). This contrasts with the traditional mental health model in which there are two levels of intervention: none or intensive psychotherapy. A system that titrates the level of intervention to the level of need is far more efficient. Given evidence that overall improvement rates are often no different between less and more demanding mental health interventions (Cuijpers, Donker, van Straten, Li, & Andersson, 2010; Cuijpers et al., 2009), the expectation of matching level of care to level of need will probably eventually be forced on us if we do not embrace it.

Effective integration and collaboration will require overcoming obstacles to the sharing of information. Unfortunately, the failure to emphasize seamless sharing of information in the first generation of EHR adoption in the U.S. has resulted in widespread inefficiencies with limited added benefit. Competition between EHR providers and between health care provider entities will obstruct the process. It will probably be another decade before true interoperability is achieved and the potential of EHRs is achieved. Ultimately, though, pressure from reimbursers—particularly the federal government—will overcome the obstacles and force the adoption of systems that operate as intended.

It is hard for me to think of psychology as a truly ethical discipline unless we pursue the redesign of health care to allow access to mental health care regardless of economic resources.

Psychologists need to prepare themselves for a future of shared medical records. Whether or not we ultimately become eligible for meaningful use incentives, a point will be reached where referrals for mental/‌behavioral health care will in part be determined by ability to share data. This means better training in focusing on assessment and plan in writing notes, and greater sensitivity about entering data. Unfortunately, we have so far been shaped by a culture in which we often only write to other mental health providers. This is one of the more subtle aspects of the culture of society that is likely to change with increased involvement in interdisciplinary settings.

Reimbursement models will also start to evolve to overcome the lack of incentive for hospitals to participate in efforts to reduce hospitalization rates, especially now that the ACO model has highlighted the flaws of the existing system. If integrated care (whether involving collaborative care or integrated primary care) proves to be an effective means of reducing hospitalization, this creates a tremendous opportunity for expanding integrated care services. Currently, reimbursement systems simply do not exist that will support integrated care (Kathol, deGruy, & Rollman, 2014). A survey conducted in 2011 revealed that 78% of primary care agencies with integrated behavioral health services were paying for them—at least in part—with grants, and on average only 21% of costs were covered by revenue (Colorado Health Foundation. (2012). This will only change with changes in reward systems in U.S. health care from fee for service to fee for outcomes, a goal we have only barely started to approach even in the context of ACOs.

This will create an opportunity for mental health providers, in that both collaborative and integrated care models require participation of mental health professionals. Whether those providers will be psychologists, though, versus psychiatrists (at least as supervisors in the collaborative care model), social workers, or counselors is a question that will have important implications for the future of the profession. Given the substantial withdrawal of psychology from the community mental health system (Cypress, Landsberg, & Spellman, 1997), similar failure to engage with integrated primary care, and to create a rationale for using psychologists in such settings, can ultimately mean a loss of a key referral source in the future. This is one of several reasons why McGrath and Sammons (2011) saw prescriptive authority and involvement in primary care as complementary contributions to the future viability of the profession, because a mental health prescriber familiar with mental/‌behavioral health services will be attractive to primary care settings.

The involvement in primary care is not just about a future for the profession; it is also about a future for those individuals who have never had access to mental/‌behavioral health care but who could benefit from it. This is particularly likely to happen if integrated care becomes available in federally qualified health centers, the network of primary care centers maintained for the indigent.

Out Here in the Field

Change is scary, but change is an opportunity. The students I am now training to work in integrated care find it exciting, challenging, and rewarding: exciting to be testing out a new approach to care different than anything they have done before; challenging to master new clinical skills, the intricacies of the EHR, and how to work closely with providers who have little understanding of mental/‌behavioral health care; and rewarding when a patient starts crying because no provider has ever asked them before about their history of abuse or their depression. Our integrated primary care program is in a federally qualified health center where more than 90% of patients are Hispanic and poor, and local mental health services usually report a 4-5 month wait list. For many of the patients we see, we are the only option. As the medical providers become more comfortable with our skills, we are becoming involved in efforts to reduce hospitalization through collaborative care, and to improve population health by increasing human papillomavirus vaccination rates. All of the students I work with envision a future in which they will provide intensive psychotherapy in a specialty mental health care setting. However, they also see a future in which they work closely with traditional medical providers, treating a variety of conditions through brief interventions, serving people who would never appear in specialty mental health care. For the first time, they envision a future in which mental health is truly part of health care, and in which they are involved in ensuring the mental and behavioral health of the entire population.


McGrathRobert McGrath is a Professor of Psychology at Fairleigh Dickinson University in Teaneck, NJ. He is also currently the Director of Integrated Care for the Underserved of Northeastern New Jersey, a program funded by the Health Resources and Services Administration to train doctoral students in clinical psychology in integrated primary care at a federally qualified health center in West New York, NJ. He is also a Senior Scientist with the VIA Institute on Character.


Butler, M., Kane, R. L., McAlpine, D., Kathol, R. G., Fu., S. S., Hagedorn, H., & Wilt, T. J. (2008). Integration of mental health/substance abuse and primary care (AHRQ Publication No. 09-E003). Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from

Colorado Health Foundation. (2012). The Colorado Blueprint for Promoting Integrated Care Sustainability. Denver, CO: Author.

Conn, J. (2015, August 12). Meaningful use of EHRs by doctors fell in 2014. Modern Healthcare. Retrieved from‌20150812/NEWS/150819976

Cuijpers, P. P., Donker, T. T., van Straten, A. A., Li, J. J., & Andersson, G. G. (2010). Is guided self-help as effective as face-to-face psychotherapy for depression and anxiety disorders? A systematic review and meta-analysis of comparative outcome studies. Psychological Medicine, 40, 1943-1957.

Cuijpers, P. Marks, I. M., van Sraten, A., Cavanagh, K., Gega, L., & Andersson, G. (2009). Computer-aided psychotherapy for anxiety disorders: A meta-analytic review. Cognitive Behaviour Therapy, 38, 66-82.

Cunningham, P. J. (2009). Beyond parity: Primary care physicians’ perspectives on access to mental health care. Health Affairs, 28, w490-w501.

Cypress, A., Landsberg, G., & Spellman, M. (1997). The impact of managed care on community mental health outpatient services in New York State. Administration and Policy in Mental Health, 24, 509-521.

Farmer, P., & Campos, N. G. (2004). Rethinking medical ethics: A view from below. Developing World Bioethics, 4, 17-41.

Fisher, E. S., Staiger, D. O., Bynum, J. P. W., & Gottlieb, D. J. (2007). Creating accountable care organizations: The extended hospital medical staff. Health Affairs, 26, w44-w57.

Frank, R. G., McGuire, T. G., & Newhouse, J. P. (1995). Risk contracts in managed mental health care. Health Affairs, 14, 50-64.

Friedberg, M. W., Chen, P. G., Van Busum, K. R., Aunon, F. M., Pham, C., Caloyeras, J. P., … & Tutty, M. (2013). Factors affecting physician professional satisfaction and their implications for patient care, health systems, and health policy. Santa Monica, CA: RAND Corporation.

Kathol, R. G., deGruy, F., & Rollman, B. L. (2014). Value-based financially sustainable behavioral health components in patient-centered medical homes. Annals of Family Medicine, 12, 172-175.

Katon, W., Russo, J., Lin, E. H., Schmittdiel, J., Ciechanowski, P., Ludman, E., Peterson, D.,… & Von Korff, M. (2012). Cost-effectiveness of a multicondition collaborative care intervention: A randomized controlled trial. Archives of General Psychiatry, 69, 506-514.

Katon, W., Von Korff, M., Lin, E., Walker, E., Simon, G. E., Bush, T, … & Russo, J. (1995). Collaborative management to achieve treatment guidelines. Impact on depression in primary care. JAMA, 273, 1026-1031.

Kessler, R. C., Demler, O., Frank, R. G., Olfson, M., Pincus, H. A., Walters, E. E., … & Zaslavsky, A. M. (2005). Prevalence and treatment of mental disorders, 1990 to 2003. New England Journal of Medicine, 352, 2515-2523.

Ma, C.-t., & McGuire, T. G. (1998). Costs and incentives in a behavioral health carve-out. Health Affairs, 17, 53-69.

McGrath, R. E., & O’Donohue, W. T. (in press). The importance of stepped care protocols for the redesign of behavioral health care in patient centered medical homes. In W. T. O’Donohue (Ed.), Integrated primary and behavioral care: Role in medical homes and chronic disease management. New York: Springer.

McGrath, R. E., & Sammons, M. T. (2011). Prescribing and primary care psychology: Complementary paths for professional psychology. Professional Psychology: Research and Practice, 42, 113-120.

Nielsen, B. A., Baum, R. A., & Soares, N. S. (2013). Navigating ethical issues with electronic health records in developmental-behavioral pediatric practice. Journal of Developmental and Behavioral Pediatrics, 34, 45-51.

Oss, M. (1996, October 9). Change is the only constant for MCO rankings. Open Minds.

RTI International. (2014). Accountable care organization 2014: Program analysis quality performance standards narrative measure specifications. Waltham, MA: Author. Retrieved from

Smolyansky, B. H., Stark, L. J., Pendley, J. S., Robins, P. M., & Price, K. (2013). Confidentiality and electronic medical records for behavioral health records: The experience of pediatric psychologists at four children’s hospitals. Clinical Practice in Pediatric Psychology, 1, 18 -27.

Substance Abuse and Mental Health Services Administration (2014). Results from the 2013 National Survey on Drug Use and Health: Mental health findings (NSDUH Series H-49, HHS Publication No. (SMA) 14-4887). Rockville, MD: Author.

United States Government Accountability Office. (2015). Electronic health records: Outcome-oriented metrics and goals needed to gauge DOD’s and VA’s progress in achieving interoperability (GAO-15-530). Washington, DC: Author.