In early 2020, the field of complex care has seen at least two new robust additions to the evidence base. First, the field was immersed in pondering the implications of the recent New England Journal of Medicine publication of the Camden Coalition’s randomized controlled trial (RCT) results — results which, at a high level, revealed no impact on 180-day hospital readmissions, the primary measure of interest. And just last week, the American Journal of Managed Care reported the findings of a CareMore RCT in Memphis — in this case, with favorable reductions in total medical expenditure and inpatient utilization.

Whereas the Camden results sent health care stakeholders (including ourselves) into a flurry of nuanced self-reflection about how much we still have to learn in this field, the Memphis results might be seen as a sigh of relief — that indeed complex care has been onto something all these years. However, it’s the two studies in combination that present the real contribution to the field — and the real call to action for the rest of us in terms of where to go next with this work. Collectively, these two studies highlight distinct approaches to identifying and coordinating care for two very different populations — not to mention two distinct approaches to studying their impacts. As such, taken together, these studies underscore the imprecision of the term “complex care,” and the compelling need for the field to sharpen everything from its language to its identification algorithms, intervention approaches and measurement strategies. Without CareMore’s study, we might have grown overly skeptical about the potential for these programs based on the Camden RCT. However, without Camden, we would very likely have over-generalized the success of the Memphis experience — particularly given how much we all love a success story. The clearest call to action comes from interpreting them together, and considering what they tell us in combination.

Complex Care RCT Study Comparison

 Camden CoalitionCareMore
Inclusion CriteriaAt least 2 hospital admissions in prior 6 months; at least 2 chronic conditions; and at least 2 of the following traits or conditions: use of at least 5 active outpatient medications, difficulty accessing services, lack of social support, coexisting mental health condition, active drug habit, homelessnessAt least 1 of: top 5% of total medical expenditures in prior 12 months, top 5% of Chronic Illness Intensity Index score, or care team member nomination. Of those, at least 1 of: 2 or more inpatient admissions in prior 12 months, 3 or more ED visits in prior 12 months, or 2 or more chronic conditions
Baseline Utilization among Study Population 1.72 inpatient admissions in 0-6 months prior; 0.74 inpatient admissions in 7-12 months prior1.28 inpatient admissions in 12 months prior
Care Team CompositionRegistered nurses, social workers, licensed practical nurses, community health workers, and health coachesCommunity health worker, social worker, and primary care physician
Length of intervention90 days

12 months
Measures180-day hospital readmissionTotal medical expenditures, inpatient bed days, ED visits, inpatient admissions, care center visits, specialist visits
Key FindingsNo difference in 180-day readmissions between treatment and control groups$7,732 lower total medical expenditure PMPY, 3.46 fewer inpatient days PMPY, 0.32 fewer inpatient admissions PMPY, 1.35 fewer specialist visits PMPY

Call to Action: Priority Opportunities for the Complex Care Field

Below are high-level reflections in this vein, informed by conversations with participants in CHCS’ Complex Care Innovation Lab (CCIL). Since 2013, with support from Kaiser Permanente Community Health and the Robert Wood Johnson Foundation, CCIL has brought together a cadre of leading health systems and health plans to accelerate the pace of learning in complex care. CCIL’s members, which include the Camden Coalition, have been both active contributors to development of complex care best practices and eager adopters of promising approaches developed by others in the field.

  1. Rigorous research helps us learn and we need more of it. We heard a universal “Bravo!” from complex care stakeholders for these recent investments in RCTs, recognizing that such studies are hard and expensive to implement, but invaluable for advancement of the field. As our CCIL colleagues at Maimonides Medical Center stated, “We appreciate these contributions to the literature in an area where there has been an unfortunate dearth of evidence.” We need more rigorous studies — RCTs and otherwise — in order to keep building the evidence base in complex care. As the Camden team has noted, its approach has evolved significantly since the launch of the RCT, and all CCIL members reflected on how much more we have to learn about how to do this work well. We encourage more partnerships between the health services research community and complex care practitioners, and we welcome the continued publication of both positive and null results — as both have much to offer.

    It’s no longer sufficient to segment the population into those individuals who are most likely to have high costs; we need to segment by those whom we can help in partnership with other sectors, and build those partnerships to ensure their long-term recovery.

    David Labby, MD, MPH, Health Strategy Advisor and former Chief Medical Officer at Health Share of Oregon
  2. Complex care is still a blunt instrument that needs to be refined. Despite all that we’ve learned to date, complex care is still a relatively nascent endeavor and there’s still so much we don’t know. We need more precision in our language, identification methods, and intervention approaches — so that we don’t erroneously lump apples and oranges together when we talk about populations served, intervention strategies employed and for how long, and how we assess impact. In particular, we need to get better at identifying which subpopulations will benefit from which interventions. Care management will be impactful for some, but not all individuals with complex medical and social needs — and those that benefit may need services of varying duration and intensity. If we expect health care organizations to continue to invest resources in this area, we have to continue to sharpen our tools. And where health care systems don’t have the necessary tools, we need to partner with other systems that do.
  3. We must expand and refine our measures. While the value of ongoing research efforts cannot be understated, these recent studies highlight how far we have to go in coalescing around measures that accurately reflect the full value and impact of this work. Relative to Camden, the Memphis study had the benefit of exploring a broader array of measures and over a longer time period — both of which are helpful for identifying impacts and suggesting pathways through which positive outcomes are achieved. However, even the Memphis study is limited to measures that could be collected through administrative data — a limitation shared by most studies in this field to date. To truly understand the impact of these programs, we need to expand the array of measures that we capture — for example, more holistic and patient-centered measures that get at broader concepts of health and wellbeing, or measures that get at the benefits that may be generated to other sectors from this work. We also need to reconsider the timeframe for which it is reasonable to expect positive outcomes for individuals with such complex, multi-faceted challenges that are often tied to multi-generational poverty. The development of such measures is an area ripe for investment and collaboration across stakeholders in the complex care field, and hopefully this recent slate of RCTs will encourage funders and others in the research community that there is a hungry audience ready to adopt such measures as they emerge.
  4. We must invest in services that are lacking. CCIL members are quick to point to the extensive behavioral health needs of the individuals they serve — for example, over 40 percent of the Camden study population had a substance use disorder — and the lack of sufficient treatment capacity in our publicly financed mental health and addictions systems. These indicators point not just to the complexity of individuals served in complex care programs, but also the limitations of striving to treat them in our current siloed and under-resourced systems. There are limits to what care management programs can accomplish without sufficient access to the basic treatment services needed to address behavioral health needs that are often pervasive among the populations targeted for these initiatives. If health care payers are wondering where else to invest to improve outcomes for populations with complex needs, these are high impact places within the broader health care universe to start. We need both more capacity and better evidence for what works — particularly in the case of addiction treatment. Beyond health care, housing is another area where more resources are needed. As Maria Raven, MD, MPH, Chief of Emergency Medicine at University of California, San Francisco, states, “The impact of homelessness and housing instability on complex populations cannot be underestimated.” Recent examples of large health care systems such as Montefiore Medical Center, and insurers such as United Healthcare, Kaiser Permanente and CareOregon, investing in large-scale housing efforts point not only to the field’s increasing acceptance that “housing is health care,” but also to exciting new opportunities for building this capacity and fostering cross-sector collaboration.

Moving Forward

We’ve learned a lot in the 10 years since Hot Spotters was published – how to engage the hard to engage, the natural history of complex populations, subpopulations and predictors, measures of success beyond utilization/cost. I hope this work can continue with more tempered expectations and a wider set of outcomes.

Tracy Johnson, Medicaid Director, Colorado Department of Health Care Policy & Financing (formerly with Denver Health)

On the heels of these well-designed and implemented studies, we hear this call to action loud and clear, and are as motivated as ever to support the efforts of those in the field who try to do this work better every day. For CHCS’ part, we will continue to push for and support additional efforts — both through on-the-ground testing and shared learning opportunities, as well as through rigorous research (RCT and otherwise) — that grow our collective understanding of what works for whom, and just as importantly, what doesn’t. We will support the development of better measures, and we will redouble our efforts to increase access to high quality mental health and addiction treatment. We will keep our finger on the pulse of this field and share practical tools and insights as they emerge. And we’ll keep working upstream as well — aiming to identify and disrupt risk trajectories as they emerge.

If you’ve got something to share, please be in touch. As these RCTs show in combination, these are still early days for the complex care field and we’ve all still get plenty to learn.


This site uses Akismet to reduce spam. Learn how your comment data is processed.

Inline Feedbacks
View all comments
Christopher A Langston
5 months ago

Thank you for this comparison of the results of the two projects. However, I couldn’t help but notice that nothing in your lessons actually comes from the comparison of the models or results — except that Camden “didn’t work” and CareMore “did.” Presumably CareMore could look at 180 day readmission rates and we could more directly contrast results in this time frame. If CareMore also failed to influence readmissions in the 180 day frame, but did over the course of a year, it would add evidence to the speculation that impact takes more than 90 days of intervention or more… Read more »

Allison Hamblin, CHCS
5 months ago

Thank you for these thoughtful comments Christopher. Our intent was not to suggest that one intervention worked and the other did not, but rather — as you substantively contribute to above — to highlight how difficult it is to compare complex care interventions when there are so many important variations between them. As you suggest, it would be very helpful for the field to have more consistency in the measures employed across these studies and in the data captured to describe the populations served. We applaud the notion of including measures that help us understand the mediators of outcomes observed.… Read more »

5 months ago

As is too often typical, neither of these initiatives nor the studies assessing impact nor your blog post address the needs of children with medical complexity . Given the fact that Medicaid is the default coverage program for children, that half of all Medicaid enrollees are children, and that Medicaid is especially important for children with medical complexity, it would be awfully nice if we occasionally thought about their needs which are frequently distinct from those of adults. This is, I think, a particular and important gap in our understanding of how to make these programs work better.

Allison Hamblin, CHCS
5 months ago
Reply to  Josh

Josh, we very much agree that the needs of children with medical complexity (e.g., Title V-supported children) are distinct from those of adults, and the programs serving these children must be tailored accordingly. There are a number of organizations (e.g., Family Voices, Catalyst Center, Supporting Families) that are working explicitly on the issues facing these children and their families/caregivers. At CHCS, we’re also focused on different types of complexity among children — those that arise from involvement with the child welfare system, behavioral health service use, and the life-long negative health and social outcomes associated with trauma/adverse childhood experiences (… Read more »

Paul J Nelson
5 months ago

Left untouched by this dialogue regarding complex healthcare is a national strategy (1) to promote an agreement of what we understand as high-quality primary healthcare and (2) to form a means to offer enhanced primary healthcare that is equitably available and accessible within every community by their resident persons. Remembering the power law distribution curve for individual health spending, we should acknowledge that 50% of our citizens account for 5% of our nation’s health spending and 5% of our nation’s citizens account for 50% of our nation’s annual health spending. Accepting an argument, that a new Federal program to increase… Read more »

Allison Hamblin, CHCS
5 months ago
Reply to  Paul J Nelson

Paul, we agree that a strong primary care system is vital — particularly for low-income Americans who experience significant health disparities and obstacles to care. Recognizing the importance of primary care, state Medicaid agencies (collectively the largest health care payers in the country) are increasingly asking not only if an individual has a relationship with a primary care provider (PCP), but also whether that patient and PCP is adequately supported by a multidisciplinary care team, and how that team addresses the diverse needs of patients, including behavioral health and social needs. To help guide states on how to accelerate such… Read more »