|Home | About | Journals | Submit | Contact Us | Français|
Clinical decision making can be described as answering one question: “What is the best next thing for this patient at this time?” In addition to incorporating clinical information, research evidence, and patient preferences, the process requires considering contextual factors that are unique to each patient and relevant to their care. The failure to do so, thereby compromising that care, can be called a “contextual error.” Although proponents of evidence-based clinical decision making and many scholars of the medical interview emphasize the importance of individualizing care, no operational definition is provided for the concept, nor is any methodology proposed for the interpretation of clinically relevant patient-specific variables. By conceptualizing the physician–patient encounter as a participant–observer case study with an N of 1, this essay describes how existing approaches to studying social systems can provide clinicians with a systematic approach to individualizing their clinical decision making.
Most physicians probably recognize the importance of assessing patient context in making clinical recommendations, but may not approach the task in a systematic and coherent way that considers the full breadth of the patient's life circumstances. While physicians are trained to assess a patient's clinical state (from the history of present illness, physical exam, laboratory, and other tests), their preferences (values), and the research evidence (through a critical appraisal of the literature), integrating patient context into health care decisions remains an element of what the evidence-based medicine literature calls “clinical expertise.”1 Although a recently published prescriptive model of evidence-based decision making mentions patient “circumstances” (see Fig. 1), the term is narrowly qualified as “physical” and “clinical circumstances.”2 There has been little methodological consideration of how one identifies, from the complexity of each patient's life, that which is pertinent to their care.
The cognitive challenge of individualizing clinical decisions at a discrete moment and place with limitations of resources and possibilities is the point of departure for this paper. While a number of models of the medical interview outline the overarching goals and objectives of the physician-patient encounter, few conceptualize the reasoning applied in answering the ever-present question: “Under the circumstances, what is the best next thing for this patient at this time?” In his seminal writing on the biopsychosocial model, George Engel introduces general systems theory as a framework for broadening conceptions of illness beyond the biomedical model to include social, psychological, and behavioral dimensions, noting that “Systems theory holds that all levels of organizations are linked to each other in a hierarchical relationship so that change in one affects change in the others.”3 In a subsequent essay, “The clinical application of the biopsychosocial model,” he illustrates how perturbations in biomedical and psychosocial systems affect one another.4 What he does not offer, however, is a methodology for tracking those perturbations amid the infinite complexity of a patient's life, given the constraints of time and resources during a medical encounter.
Engel's model has stimulated many projects to define and describe the medical interview in a manner that incorporates psychosocial and biomedical elements into patient care.5–10 Stewart et al. have developed “patient-centered communication,”12 aptly described as a “cluster of physician behaviors” that promote “knowing the patient as a person, in addition to accurately diagnosing their disease.”11 Greater emphasis has been placed on the functions of the medical interview (collecting information, responding to the patient's emotions, influencing behavior)13 rather than on just its structure (e.g., greeting, followed by chief complaint, then history of present illness, etc.). This “three-function” model, introduced by Bird and Cohen-Cole, has been elaborated by Lazare, Putnam, and Lipkin as: determining and monitoring the patient's problem; developing, maintaining, and concluding the therapeutic relationship; and carrying out patient education and implementation of treatment plans. For each function, there are tasks such as “acquiring the knowledge base of psychosocial issues” and “eliciting data for the biomedical, psychological, and conceptual domain,” for determining and monitoring a patient's problem, for instance.14 Functional analysis thus describes what successful clinicians do but not how they think. We shift here from a focus on the objectives for the clinician providing patient-centered, individualized care, to consideration of the cognitive processes for accomplishing those objectives.
This essay draws on the literature on qualitative methods theory to describe an approach to individualizing clinical decisions, which is called “contextualization.” Contextualization involves identifying what is relevant to the immediate clinical problem from across the spectrum of a patient's life, including their cognitive abilities, emotional state, cultural background, spiritual beliefs, economic situation, access to care, social support, caretaker responsibilities, attitude to their illness, and relationship with health care providers (Table 1). In the following case, consider how the failure to contextualize caretaker responsibilities results, initially, in a failure to determine what is the best next thing for this patient at this time:
Ms. Gloria Dawson (a pseudonym) presented in the preoperative testing clinic for evaluation for gastric bypass surgery to treat morbid obesity. She was 44 years old, 5′ 6″ tall, and weighed 270 lbs. She had been referred to a surgeon experienced in the procedure from the weight loss clinic after unsuccessful nonsurgical attempts to lose weight. He scheduled her for an open rather than laparoscopic procedure because of concerns about adhesions from a prior cholecystectomy. The internist who examined Ms. Dawson noted well-controlled diabetes and hypertension. Despite her weight she could walk up two flights of stairs, and her exam was unremarkable. Ms. Dawson commented that she looked forward to losing weight so that she could get around more easily and better assist her son who is disabled.
Concerned about the passing reference to her caretaking responsibilities, the physician asked her about her disabled son. He was aware that abdominal surgery for obesity can be fraught with complications, including wound dehiscence and a prolonged course. Average recovery time for the procedure recommended to Ms. Dawson is 46 days.15 In the discussion that followed, the patient described how she was the sole functioning adult member of her household at a critical time. Her son was dying of end-stage muscular dystrophy and her husband, who was an alcoholic, was abusive and unhelpful. She also had an 8-year-old daughter. She had not “thrown out” her husband, she said, because she needed money that came from his disability income and social security to support the family.
As they discussed her home situation and her plans for surgery, Ms. Dawson expressed dismay at the thought of being unable to care for her children for weeks or months. “They need me now more than ever,” she said. The physician also pointed out that assisting her son with toileting and bathing, which required heavy lifting, could complicate her recovery. She noted that she had been informed about the possibility of a lengthy postoperative course, but had been focused on the positive aspects of surgery, particularly greater mobility. She concluded, and the physician concurred, that this was the wrong time for the procedure.
Although Ms. Dawson met the criteria for bariatric surgery, sending her to the operating room at that time would have been an error for several reasons: first, it would have been inconsistent with her immediate priority to care for her children, and protect them from her husband's abuse; second, her need to assist her son might have led to unwise physical activities during recovery; and, finally, any serious complications may have wreaked havoc on her family's fragile homeostasis. How were these issues initially missed? After all, her doctor had discussed the surgery with her and she had indicated her preferences. The problem was that nobody had identified the incongruity between her personal situation (her context) and the proposed plan of care. One might call this a “contextual error.”
How can clinicians prevent contextual errors? Need they just do a better job at eliciting patients’ preferences, as in a utility analysis, so that patients can independently determine what is best for them?16,17 Or should they exhaustively explore with each individual how every aspect of their life situation relates to potential outcomes? While the former approach does not take into account the wide spectrum of patient comprehension of the implications of context, the latter is impractical. In Ms. Dawson's case, it is the physician who has the greater potential to recognize that premature weight bearing, precipitated by the need to care for a disabled child, might compromise her recovery. Patients need physicians not only for their knowledge and procedural skills, but for their skills as contextual thinkers.
As with the other components of clinical decision making, contextualization requires a systematic approach that prevents physicians from succumbing to personal biases and preconceptions, or relying exclusively on clinical experience. The following section introduces such an approach using three concepts from the qualitative sciences to integrate patient context into clinical decision making: The first is explanatory theory building,18,19 the second is reflexivity,20 and the third is triangulation.21
How is research evidence applied to individual patients? In the evaluation of Ms. Dawson at the time when she was referred for bariatric surgery, it would presumably have involved the following hypothesis-driven, deductive process:
Hypothesis: “Ms. Dawson fits the criteria used in studies that demonstrate a benefit of gastric bypass surgery to certain patients in terms of quality of life and reduced morbidity.” Hypothesis testing: tests would then have been selected and performed to determine whether she did, in fact, fit those criteria, which requires measurements of her body mass index, blood pressure, and fasting serum glucose, among others. Deduction: the results apparently confirmed that she fit the accepted criteria, and therefore belonged to the category of patients who are known to benefit as a group from the procedure.
Concluding at this point, however, that Ms. Dawson would likely benefit from surgery was premature. Potential contextual qualifiers, which could override expected benefit, needed to be explored. Such consideration of context required moving beyond categorization to an inductive, theory-building process in which unique elements of the patient's life are discovered and assessed for clinical relevance.
A contextual assessment often begins with the clinician picking up on possible “red flags,” from the patient's narrative or appearance:
Query: “Of all the potential concerns related to having surgery, why did Ms. Dawson interject that reference to her disabled child?” the clinician would wonder, recognizing that herein lies an issue that warrants further exploration. A dialogue follows and from that dialogue come new data: her obesity was not a central issue for her until her primary care physician brought it to her attention and referred her to a weight loss clinic; her initial decision to consent to surgery was based principally on faith in her physician's judgment; her chief concern is not her weight but the well being of children whom she believes need her full resources now; state services and private foundations are already maximized; and her weight problem, while disabling, remains stable. A theory emerges of a woman who, on the one hand, is maintaining a precarious homeostasis in a fragile ecosystem and, on the other, has entered a health care system encouraging her down a clinical pathway that may disrupt that balance.
In contrast to “hidden agendas,” in which the patient hides or hesitates to reveal critical information,22 contextual red flags are often alluded to by the patient without an awareness of their significance to clinical care. As such, pursuing them is not only a matter of communication and rapport in the doctor–patient relationship, but requires specific cognitive skills on the part of the clinician.
Theory building extends beyond the hypothesis-driven, deductive approach to interpreting clinical and laboratory data that is essential to diagnosis. While the hypothetico–deductive approach serves a necessary preliminary function of labeling through a process of categorization, contextualization requires identifying what is unique about each situation through a process of discovery. Determining that Ms. Dawson meets criteria for bariatric surgery, for instance, is a categorization. Determining that she is not at an appropriate time in her life for the surgery is a discovery. Theory as used in this approach is explanatory rather than predictive. Facilitating the emergence of explanatory theory from data has been elaborated in the writings of Glaser and Strauss on grounded theory.18,19
The question to ask before acting on any theory is whether the theory is good. “Good” to the empiricist versed in the scientific method means valid, reliable, and objective. Determination that a patient is anemic or hypertensive is based on tests that are assessed according to these criteria. While the measurement properties of the hemocytometer or sphygmomanometer are easily demonstrable, how does one evaluate the quality of the contextual assessment a physician provides in trying to help a patient like Ms. Dawson? Can one measure validity, reliability, and objectivity when N = 1 and the data processing instrument is a human being?
Qualitative methodologists, wrestling with these issues, have adopted a corresponding set of criteria: validity—a measure of the degree to which findings correctly map the phenomenon in question—becomes credibility, which is gauged by “the ‘truth’ of the findings, as viewed through the eyes of the observed and within the context” in which the inquiry is being pursued.23 We can determine whether our theory is credible, for instance, based on whether Ms. Dawson agrees with our assessment and whether it fits with what we know about her and our health care system. Reliability—a measure of reproducibility—becomes dependability, “the extent to which the inquiry would produce similar or consistent findings if carried out as described, including taking into account factors that may have affected the research results.”23 Dependability may be assessed by asking whether our colleagues would have reached similar conclusions if given the same information. Finally, objectivity—the extent to which findings are free from bias—becomes confirmability, the extent to which evidence coming directly from subjects and research context can corroborate those findings. We confirm our findings by talking with family members, social workers, or others knowledgeable about our patient's situation. While these are not novel approaches to most clinicians, the language of the qualitative sciences makes explicit how they are valuable and require systematic attention.
In contrast to quantitative analysis, which is rooted in the use of computers, laboratory equipment, and other technologies, qualitative methodologies depend on the researcher as the key instrument of analysis. This leads invariably to bias, which unless systematically addressed leads to erroneous inference.20 It is equally evident in the clinical realm: “At present, clinicians’ individual preferences (as distinct from clinical expertise) often play a large role in their actions, leading to large practice variations in managing similar cases.”1
To the extent possible, physicians’ preferences and assumptions must be confronted explicitly and accounted for. Ms. Dawson appeared to her physicians as a soft-spoken but articulate white woman of middle class background. Assumptions of a stable family environment could account for why the plight of her children and her dysfunctional marriage were initially overlooked. Also, clinicians caring for her may not have been attentive to the possible effect of their demeanor or social position on her willingness to raise concerns that should have been explored.
Qualitative methodologists have coined the term reflexivity, an “attitude of attending systematically to the context of knowledge construction…at every step of the research process.”20“Reflexivity” not only addresses bias and the skewed perceptions it may generate, but also recognizes the effect that the researcher (or clinician) may have on the data-gathering process as well. In clinical care, as in fieldwork, one's comportment is an important variable in what gets revealed. The issue was addressed as early as the 1950s by the psychoanalyst Michael Balint who, in the setting of physician support groups, explored how doctors unconsciously impose constraints on what is “allowable” for discussion in the physician–patient relationship.24 Clinicians who are perceived as detached or intimidating may not elicit clinically relevant information. Reflexive physicians are sensitive to the effect of their personal presence and style of interaction on the patient's level of comfort during the clinical encounter. They are also alert to their own response for what it might indicate about the patient. Reflexivity parallels recent conceptions of “mindfulness.”25
A key objective of the physician–patient encounter is getting the story “right,” and making sure it is complete. Triangulation, one of the central principles of case analysis in the social sciences, illustrates how multiple methods of data collection are integrated to provide new insights or enable resolution when there are competing explanations.26 (Its usage here is distinctly different than in family systems theory, where it refers to the introduction of an uninvolved third party as a coping mechanism in situations of interpersonal stress.27) Consider the elderly hypothyroid patient who denies missing medication dosages but continues to have an elevated thyroid-stimulating hormone level despite escalating dosages of thyroid replacement. Potential explanations are wide ranging but include, among others, laboratory error, pharmacy error, patient error, and diagnostic error. Sorting these out may involve examining pills, contacting the pharmacist, repeating laboratory tests, talking with family members or other health care professionals familiar with the patient, and assessing the patient's understanding of the illness and the treatment involved. Identification of poor medication adherence and deteriorating cognitive function as a likely explanation might in turn lead to neurocognitive testing and an assessment of living situation.
Triangulation, a term borrowed from land surveying, involves combining various data sources to locate a point in space.21 Qualitative researchers and clinicians alike draw on diverse sources of information to find explanations and solutions for problems. Yet the process of triangulation goes beyond using multiple data sources simply to confirm a result. In case analysis, inconsistencies are not necessarily regarded as evidence of poor data quality, but as windows to new insight. This has been described as “focusing on what is learned by the degree of convergence.”21 The decision to pursue a diagnosis of dementia when an elderly patient's story and laboratory values are not consistent reflects such an approach. Thus, triangulation both builds credibility for explanations and guides clinicians’ thinking toward reconciling discrepancies in data to establish a correct diagnosis and optimal treatment plan.
It is not sufficient to base health care decisions on current evidence and patient preferences alone. Individuals may be well informed about the risks and benefits of a proven intervention but still overlook its implications for their specific life circumstances. This essay explored the case of an obese woman who agreed to elective weight reduction surgery at a time when she had conflicting priorities. The term “contextual error” may be applied to such instances in which medical decisions fail to incorporate factors unique to an individual patient and relevant to their care.
Because of their knowledge and experience, physicians are often better equipped than patients to make connections between a proposed intervention and some aspect of their life situation. Making these connections consistently and accurately, without distorting information, is akin to the challenges qualitative researchers face when studying complex social systems. While physicians have successfully embraced the systematic approaches inherent in quantitative methodologies for interpreting clinical and laboratory data, they have yet to adopt systematic approaches to qualitative inquiry available from other disciplines. Considering the doctor–patient encounter as a participant–observer case study with an N of 1, clinicians can contextualize health care recommendations by adopting specific methodologies from qualitative field research. First, they must be skillful in picking up on cues to contextual factors relevant to the case. Second, they must be attuned to personal biases that color perceptions about what is relevant. Third, they must adopt a systematic approach to corroborate findings involving multiple sources. These three concepts of theory building, reflexivity, and triangulation may guide clinicians in individualizing care.
Dr. Weiner received support from the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program.
I would like to acknowledge Simon Auster, MD, JD for many shared insights, particularly the formulation of the question, “What is the best next thing for this patient at this time?” I would also like to thank Arthur Elstein, PhD, Alan Schwartz, PhD, Kelly J. Devers, PhD, Kevin Weiss, MD, Stefan Kertesz, MD, Avery Hart, MD, and two anonymous reviewers for comments on drafts of this manuscript that led to substantive changes.