The term “complexity” has many applications and shadings within the domain of health and health care. Medical clinicians commonly use the word “complex” to mean complicated or difficult, particularly as shorthand for a difficult patient-­clinician relationship. Complexity may also refer to multiple interacting medical conditions, such as those frequently encountered in care of chronic illness—sometimes termed “medical complexity.” In nursing, “complexity” can be a property of particular roles (and a staffing issue), referred to as “work complexity” or as social or mental health factors that elevate the risk to patients and staff in hospitals (Weydt, 2009). Similarly, “complexity” has also been discussed as a property of ambulatory care visits (Katerndahl, Wood, & Jaen, 2010). Others, especially those studying health care as a system and the applications of complexity science to health care, use “complexity” to mean the many possible ways that interconnected components of a health care system and its patients and communities can interact or adapt for successful fit in the environment (Zimmerman, Lindberg, & Plsek, 2008). A specific example is a model of “Complex Adaptive Chronic Care” for addressing chronicity in chronic care models that goes beyond disease-oriented thinking (Martin & Sturmberg, 2009; Peek, 2009). All these meanings and shadings of the versatile term “complexity” are important, but this chapter focuses on complexity as factors that interfere with standard care and decision-making (Peek, Baird, & Coleman, 2009).

While all clinicians instinctively relate to the concept of “complex patient,” there has been little shared definition in the field for what that means. For example, when encountering a patient they believe is “complex,” they may react with, “Oh my gosh!” or “Things aren’t going as I expected,” or “I don’t feel so good about this,” or “With these conditions, the patient should be getting better.” Recall that Grant et al. (2011) said that physician attribution of “complex” goes beyond what would be predicted by medical comorbidity alone. The field needs a more standard definition and vocabulary that allow clinicians to say more clearly for a given patient situation exactly what is complex and what to do about it—and then to incorporate complexity-linked interventions into the care plans already geared to diseases and conditions. The following paragraphs regarding definitions and domains of complexity are paraphrased or quoted from Peek, Baird, & Coleman, (2009).

One definition of patient complexity:

[A complex patient] is one for whom clinical decision-making and required care ­processes are not routine or standard. For complex patients, many recommendations from evidence-based medicine are unlikely to apply in a straightforward manner because of “exceptions” such as: multiple interacting chronic conditions, other comorbid conditions and socioeconomic factors such as homelessness or absence of adequate family caregivers or other support systems. (Weiss, 2007, p. 375)

A series of papers on “Managing Complexity in Chronic Care” based on this definition appears as a special issue of the Journal of General Internal Medicine (Kupersmith, 2007; Safford, Allison & Kiefe, 2007; Weiss, 2007). The Dutch authors cited above also think of patient complexity as interference with standard care and offer an important distinction with a set of domains and a tool that were first designed for use in inpatient settings.

It is appealing to distinguish between complexity that arises from characteristics of a patient—such has having multiple interacting diseases that may complicate each other … and complexity of care delivery, such as involvement of multiple systems and specialties that require interdisciplinary communication to be effective …. (de Jonge, Huyse, & Stiefel, 2006, p. 680)

This foundational Dutch work is the basis for a US outpatient adaptation that employs similar but modified concepts and domains:

These domains, each with two areas of inquiry and definitions of increasing levels of complexity, are illustrated in Fig. 14.1 as the Minnesota Complexity Assessment Method, an outpatient adaptation of the original INTERMED domains and questions (Stiefel et al., 2006). Each item in Fig. 14.1 represents a separate source of potential complexity, which if high, flags an area to potentially address as an interference with standard care.

The goal of these assessment models is to more usefully understand the nature of the obstacles that are keeping usual care from yielding the expected positive outcomes. By noting these barriers and sorting them into specific categories or domains, the team can more accurately understand the specific context of patients who are not making the expected progress toward health. The goal then becomes helping the patient deal more successfully with those specific barriers. If the barriers are not at a high level, then no action is needed beyond routine medical care. But if the specific area of challenge or “domain” is noted to be a large barrier for a specific patient, the primary care clinician and/or care manager on the primary care team can enlist other team members or community services to more constructively address the barrier. This enlarged team is only fully engaged when the patient assessment suggests it is necessary. However, more detailed understanding of the targeted domain needing attention can determine who else should be involved in the team-based care and which community resources and/or patient resources may be needed. For example, if the patient has never trusted a health care professional, the first step can be to arrange for more focused contact and trust-building time with the primary care ­provider or someone else the patient recognizes as part of the team. In other instances, the patient may need someone from community social services to facilitate a connection to a nearby community group such as YWCA or other social group to overcome social isolation.

A patient may have a list of medical diagnoses suggesting medical complexity. That same patient may also face interfering factors (social or care complexity) in the domains of social and health systems. The combination of high medical and high social/health system complexity makes usual care and care planning much less likely to yield the desired outcome. This patient faces the dilemma of dealing with both medical and social/care system complexity or must manage the “area under the line” as noted in Fig. 14.2 (Peek, 2010).

A hypothesis (or way of thinking) is that the net challenge for care of a particular patient may be a product of both axes (medical complexity and care complexity), not just the more familiar medical complexity. To address both, the clinic should have both these axes in mind. Providers and patients getting together not only to plan care but to create strategies to reduce predictable interferences with that care is shown in Fig. 14.1. The investment in medical care plans is not protected unless patients and providers together ask themselves what person-specific factors will predictably interfere with those care plans and with usual or customized care for the patient’s conditions. On the patient side, patients can often be confused by their own illnesses, complicating life factors, and stresses. Whether articulated or not, these undermine confidence in their ability to do carry out their own self-care and patient engagement behaviors. Patients can begin to feel like a failure, “difficult,” or that no one wants to see them anymore. But this is often not due to lack of motivation or “noncompliance,” but to very real interferences with care from social or care delivery sources (Peek, Baird, & Coleman, 2009; Peek, 2010).

High levels of complexity on the social/care delivery axis can point clinicians and care teams toward broader or supplementary interventions or connections than their medical care plans might do alone. For example, here are some aspects of care planning that might be included for high levels of social or care complexity:

In Minnesota (and possibly other states) medical practices with state “health care home” certification are now paid a small care management fee based upon the medical complexity of each patient. This fee has been augmented modestly to encourage care teams to address two domains of social complexity: having language barriers and a mental health diagnosis. The last factor is a known risk for increased cost for patients needing medical care but can be a blended factor with despair, isolation, and an overwhelmed emotional state for patients. Being overwhelmed is not a medical diagnosis, but the despairing symptoms can result in a diagnosis of depression, not otherwise specified (NOS). This intellectual dilemma was discussed in the first part of this chapter. In fact, that diagnosis, Depression NOS, is common in primary care and can yield increased attention to help the patient. Unfortunately, if that assistance consists only of medication, it does not assist the patient in dealing with their domains of social and care system complexity that are overwhelming them such as lack of trust in health professionals, a disengaged stance toward their illness, lack of social support, lack of resources/insurance, and uncoordinated care from a variety of providers.

However, increased payment for some dimensions of social complexity is an early sign that some reimbursement systems are seeking ways to fund care systems that seek to assist patients in the reality of their complex lives. These increased payments for social complexity are a beginning of efforts to find sustainable payment models to support primary care practices as they try to improve outcomes for patients when usual care does not succeed.

The gradual emergence of payment models to address some dimensions of social complexity serves as a reminder that not all “complexity” is from a mental health condition. This reminds us to go beyond the familiar mental health taxonomy into a new vocabulary for what makes patients complex—including the predictable interferences with care. The DSM-based model rests upon a descriptive taxonomy rather than underlying etiology (Althoff & Waterman, 2011). We diminish our positive clinical impact on patients when our assessment methods are descriptive of symptoms but less sophisticated about clarifying the nature of their distress. Psychiatric training changes are needed, as recommended by Althoff & Waterman, to improve the understanding of patients’ entire situations beyond psychiatric diagnosis and medication tolerance. In primary care, we also need to gain more insight into patient’s total dilemma rather than base our clinical insights primarily upon the diagnosis as defined in the DSM. We need to go beyond treating with medications alone or medications supported by mere encouragement, “pep talk,” or other positive messages that do not necessarily get at the specific factors that are distressing people (Salazar-Fraile, Sempere-Verdu, Mossakowski, & Bryan, 2010). Our approach to mental health diagnosis has been compared to the “field guide” approach of a bird-watching book—identifying mental disorders with checklists of characteristics similar to those used by birdwatchers (Lane, 2007). This indeed describes what we see, but leaves us without tools that improve our understanding of why we see different symptoms or patterns of distress in our patients. A new model of care that deals with patient complexity is possible, does not depend upon the DSM diagnostic model, and can address very real issues that confound patients and providers and interferes with usual care based upon diagnosis alone.