|Home | About | Journals | Submit | Contact Us | Français|
There are no proven, comprehensive treatments in primary care for patients with medically unexplained symptoms (MUS) even though these patients have high levels of psychosocial distress, medical disability, costs, and utilization. Despite extensive care, these common patients often become worse.
We sought to identify an effective, research-based treatment that can be conducted by primary care personnel.
We used our own experiences and files, consulted with experts, and conducted an extensive review of the literature to identify two things: 1) effective treatments from randomized controlled trials for MUS patients in primary care and in specialty settings; and 2) any type of treatment study in a related area that might inform primary care treatment, for example, depression, provider-patient relationship.
We developed a multidimensional treatment plan by integrating several areas of the literature: collaborative/stepped care, cognitive-behavioral treatment, and the provider-patient relationship. The treatment is designed for primary care personnel (physicians, physician assistants, nurse practitioners) and deployed intensively at the outset; visit intervals are progressively increased as stability and improvement occur.
Providing a comprehensive treatment plan for chronic, high-utilizing MUS patients removes one barrier to treating this common problem effectively in primary care by primary care personnel.
The medical system fares poorly with patients seeking care for medically unexplained symptoms (MUS).1 We prefer the more encompassing and descriptive term MUS, but many use terms like somatization1 and somatoform disorders2 for these patients. Unfortunately, many MUS patients avidly seek care to find an organic disease they fear but do not have. Doctors then may test for and even treat (nonexistent) organic disease. This produces high utilization of services, unnecessary laboratory testing and consultation, increased costs, and high iatrogenic complication rates (e.g., ill-advised tests, drug addiction, and trial treatments for presumed but absent organic diseases).3–9
Further problematic is that we overlook patients' basic problem of psychosocial distress.10–13 Rates of mental and physical dysfunction are high; disability and poor work records are common; relationships are poor; and personal distress prevails. Magnifying the impact, MUS patients constitute a large proportion of all outpatients.14–18 The entire system suffers not only from a harmful effect upon patients, but also from excessive costs and utilization.12,13,19–25 With some urgency, evidence-based treatment is required if we are to reverse these problems. This article describes a comprehensive treatment plan for MUS patients designed for primary care personnel, a well-recognized need.26
In a literature review, we sought not only to understand work in primary care and specialty settings, but also to go beyond present approaches to find additional research that might inform treatment. Some urge that simply doing better with what we now have is insufficient, that synthetic new models designed for primary care are needed, and that producing evidence for the models is the essential next step.27,28
We synthesized the treatment from our own experiences, consultation with experts, and from an extensive review of the literature. We searched the medline database from April 2002 back to 1966 using the Silverplatter WinSPIRS interface. Terms such as “somatoform disorders,”“chronic fatigue syndrome,”“functional colonic diseases” (e.g., irritable bowel syndrome, colitis), “fibromyalgia,”“premenstrual syndrome,”“myofascial pain syndromes,” unexplained “chronic pain,” functional “dyspepsia,” and psychological components of allergies (“hypersensitivity”) were used. Wherever possible, these terms were searched as Medical Subject Headings (MeSH) selected from the database Thesaurus, and other subjects not adequately represented in the Thesaurus were searched as keywords in the title and/or abstract fields. The combined set of citations for all subjects of interest was then limited to randomized controlled trials and articles in English and in human populations. Separate searches also were run on “nurse practitioners” and/or “primary care” insofar as either subject relates to the above terms, limited to English. Using the same interface, all of the above topics were searched again in the American Psychological Association's PsychINFO database from April 2002 back to 1967.
We evaluated over a thousand abstracts produced by our searches, and reviewed several hundred full articles derived from the search (in addition to our own files). We looked for two things: 1) clinical trials in primary care and specialty settings; and 2) any type of study that might inform a new treatment. We included reviews. There were many unidimensional treatment studies of high quality, and we used these data to inform which components we considered for what proved to be a multidimensional treatment. We selectively reference our findings.
The vast majority of randomized controlled trials (RCTs) have been conducted by mental health specialists in mental health settings,26,29 whereas treatments in primary care have been limited and brief, and the topic rarely studied.26–28,30–33 A single consultation letter to primary care physicians decreased costs (via decreased hospitalization) and, secondarily, it minimally improved functional status, symptoms, and quality of life,31 and had a mild effect on physical functioning33 in severe MUS patients.34,35 A subset of patients underwent additional group therapy experiences conducted by a specialist, and they improved psychologically and physically.36 Because of more stringent rules for hospitalization in the present environment, it is unlikely that the cost benefits would be obtained today. Also, this approach gained little following because it lacks the comprehensive, multidimensional type of care most believe will be needed for complex problems in mental health care like MUS.32,33,37–39 Other brief, limited interventions also have not been effective.40 We found only 1 study of a multidimensional treatment for primary care. That intervention entailed eight 2-hour sessions for groups of 17 in cognitive-behavioral treatment (CBT) techniques, relaxation, meditation, and stress management, and showed no impact compared to controls.41
We found much valuable experience in multidisciplinary chronic pain and other somatization management programs,33,35,36,42–47 and in consultation-liaison psychiatry efforts with difficult, high-utilizing patients.24,48–51 All employed CBT.44,45,52–54 Many have shown that CBT interventions delivered in specialty care settings have improved short-term outcomes in MUS patients.46,47 Kroenke and Swindell provide a nice review of CBT in MUS, including its effectiveness in just 5 to 6 sessions.55
However, health care systems have not incorporated these models in routine care, because of cost and because MUS patients often refuse to seek services outside of the primary care setting.56–59 Some report as few as 10% completing referral60; in contrast, 81% of MUS patients were willing to have psychosocial treatment in primary care by their physician.61 For those who agree to referral, treatment for MUS has not proven effective over the long-term,56–59 largely because of the time-limited exposure to MUS patients, in the range of once weekly for 8 to 10 weeks (e.g., counseling, pain clinic). Although there are occasional reports of effects lasting upwards of 1 year and 1 report of incomplete improvement over 5 years,62 most MUS patients revert to their baseline state upon discontinuing specialty care.63
We concluded that CBT was potentially useful in primary care and we configured its basic skills for use by primary care personnel. While we eschewed the more complex, specific techniques that might require special training, such as desensitization and immersion therapy, we included active cognitive reorientation at many visits based upon an understanding of the patient's explanatory model,64 often facilitated by using symptom diaries.65 Further, behavioral approaches based upon operant mechanisms52 were employed65 (e.g., regular visits and medication schedules independent of symptoms). We also were guided by an earlier RCT, in which we showed that the CBT model presented later was easily learned and effectively deployed by primary care residents, and that it had an impact on MUS patients.66 Although data were lacking, we further concluded that deploying cognitive behavioral treatment over longer periods of time merited exploration.
The difficulty physicians experience in treating depression led experts to recommend a multidimensional approach, as used often in pain clinics, rather than focusing on physician education.49 Collaborative care leads to interactions among various parts of complex health problems and, continuing to involve the physician, multiple other domains are integrated with medical care, e.g., social services, physical therapy, pain clinic, specialty medical referral, psychological referral.67–69 These activities typically are not coordinated by the physician but, rather, by a case manager.70,71 Moreover, following stepped-care principles, these services are used only as needed on an individualized basis.67,72 In our view, collaborative, stepped-care approaches appeared promising for meeting the mental health needs of other primary care patients with mental health problems like MUS.67,72
We adapted these principles for use in primary care by primary care personnel. Rather than initially delegating the key coordinating role to someone other than the primary provider, however, we believed that primary care providers could not only handle the coordinating role but also could themselves conduct some of the work rather than triage it.70 Depending upon their skills and interests, physicians, physician assistants (PAs), and nurse practitioners (NPs) could conduct first-line treatment in the following areas: social worker, physical/occupational therapist, exercise/relaxation instructor, and diet counselor, as well as conducting the intervention (to be described). We believed this approach would be less costly, more efficient, and more effective, in addition to greatly enhancing the provider-patient relationship (PPR) by actively involving the provider in most dimensions of care. Providers would need to recognize failures in each area and, at that point, to make appropriate referrals. We anticipated also that referral could occur with successes, such as the patient who improves (gains insight) and wishes formal counseling. Certain problems also would initially be outside the provider's capability and necessitate referral: psychological/psychiatric treatment, osteopathic manipulative therapy, and specialty medical treatment. In addition, whereas PAs or NPs are primarily responsible for treatment, the usual care physician would remain actively involved in a collaborative way as a back-up and informal consultant for ongoing issues, thus remaining integral to treatment.
The literature is just as replete with reports of the centrality of the PPR to the care of MUS patients as it is with concerns about typically poor relationships with MUS patients.30,73,74 Although the centrality of communication and the PPR has been recognized,75 we found only 1 intervention that explicitly addressed the PPR,76 but it did not report a systematic procedure to establish it. We concluded that this major barrier to care should be addressed.8,39,77–80
We used a patient-centered method described by Smith81,82 and tested by our group.66,83,84 This method describes in behavioral terms how to communicate and establish the PPR. Our group showed in an RCT that it was learned easily by first-year primary care residents and, in turn, that there was evidence of a positive impact upon patients. We incorporated this PPR method as the centerpiece of the intervention for MUS patients, where we integrated it with the CBT model.
Consistent with this orientation of viewing the person and their relationship with others as paramount,85 we went beyond isolated reliance upon the PPR (while maintaining it central) to consider social issues and social relationships,86 and more recent interpretations of sociosomatics that focus on how everyday processes relate to cultural expectations and values.87 In going beyond the dyadic relationship, the treatment was consistent with relationship-centered care88 and with sociopsychosomatic conceptualizations.89 Treatment reflected incorporation of the broader psychosocial dimension by, for example, incorporating significant others, actively addressing role constriction and social isolation, and generally working to decrease both perceived and real social marginalization from healthy states.87
The 5 steps and 21 substeps of the evidence-based patient-centered interviewing method are relied upon as the mainstay throughout treatment.81,82 The method is a provider's main data-gathering, relational, and therapeutic activity. Table 1 shows the steps and substeps providers use, always integrating them with doctor-centered interviewing for disease and other relevant details.
Providers focus upon actively eliciting and responding to emotions. Distress, upset, anger, and sadness are among myriad emotions that MUS patients will express, and these are explored, including those surrounding troublesome medical experiences.8 Emotion is addressed in a way that enhances the relationship. Specifically, when patients express emotion, the provider Names it, Understands it, Respects it, and Supports it; that is, one NURSs the emotion.81,82 Past research suggests that these skills are important tools that physicians learn in improving relationships with and satisfaction of patients.83,84 In addition to patient-centered communication, there are additional ways providers can improve their relationships with chronic MUS patients:
Each of 4 provider behaviors occurs at every visit, initially introducing the material and later reaffirming it, and introducing evolving new issues. These behaviors are designed to alter thinking patterns and associated behaviors around symptoms and other aspects of patients' lives.92,93 The PPR principles above are liberally employed and integrated with CBT, especially NURS, in discussions of the frequent difficult problems and attendant emotions that these patients typically have.
The provider, in collaboration with the patient, first identifies achievable long-term goals with the patient,8,44 e.g., decreased symptoms, improved functioning and well-being, less disruptive behavior, improved work/school record, and improved interpersonal relationships.78 Patients are told not to plan on cure,8,94,95 and instead the provider negotiates specific treatment goals with the patient.44,59Short-term goals are facilitated by the provider at each visit in a way that the patient has been responsible for identifying goals rather than having them prescribed. Short-term goals are designed to actualize long-term goals, e.g., a patient with chronic pain and fatigue negotiates to walk one-half block 3 times daily until the next visit. Providers identify achievable goals and limit their number to no more than 2 to 3 for the next visit; these are often labeled homework.
Initially, the provider determines the patient's explanatory model,64 starting with an open-ended, patient-centered approach81,82 and later pinning down the following specifics if they do not develop open-endedly: learning what the patient believes is wrong, the type of treatment expected, the hoped for outcome of the treatment, the cause of illness, the reasons for symptoms at any one time, the mechanism of the illness and symptoms, and the patient's understanding of the course of the illness and of its treatment. These ideas of the patient must be understood, because reorientation and reattribution are often needed as part of treatment.44 To achieve this, patients must be understood in an empowering way.96 Patients typically are worried about the causes of their symptoms and have definite expectations for care.97 Understanding these enhances the PPR and provides guidance for treatment. Explanations that do not blame the patient are important.96 Differences in doctors' and patients' expectations seem to be a factor contributing to the frequent poor PPR.98
The provider gives an overview of the treatment options for the patient,8 e.g., antidepressants, tapering of addicting medications, exercise, and regular visits. The next step is crucial: obtaining an explicit commitment to work on the treatment program.44,92,93 This involves the patient equally in their program, another part of the negotiation process, and establishes the patient's responsibility for their health. The provider also will often obtain a key family member's commitment to work as comanager.
Providers conduct multiple elements of the treatment plan (summarized in Table 3). The specific order and initial emphasis varies according to unique patient circumstances and needs, e.g., exercise might receive more initial emphasis in a de-conditioned patient. Further, the specific treatment elements are negotiated and developed over time rather than being unilaterally prescribed. The patient's progress and response to initial work, as well as her/his interests, determine the course of treatment. Providers are careful not to overload patients and to match introduction of new aspects of treatment with the patient's capacities and assimilation of previous material. The provider encourages patients to identify and formulate pertinent questions to encourage responsibility and assertiveness.100 Patients also are asked to keep diaries of their symptoms and the associated activities, thoughts, and emotions which, with some, will be a focus of treatment.44
The specifics are negotiated on a time-contingent schedule, e.g., visits, medications, and exercise are regularly scheduled.52 This schedule dissociates the patient's symptoms from the treatment and offsets treatment acting as a reinforcer to the symptoms. Another guideline is to use treatments that have worked before and avoid those that have not.95 Nonnarcotic analgesics and other more benign symptomatic medications can be effective and inexpensive, e.g., for irritable bowel syndrome.101 Their use signals the provider's concern and that she/he is doing something.
Uniquely designed special programs for relaxation, physical exercise, and physical therapy can be helpful, occupy the patient's time with healthy behaviors, can be done without consultation, and also show that the provider is involved and helpful. Prescriptions must be individualized. Simple relaxation exercises102–104 or progressive muscular relaxation105 can be taught in addition to simple physical therapy procedures.54 Usual prescriptions are for 20 to 30 minutes once or twice daily. These procedures are taught to patients by providers and integrated into the remainder of the program. To further offset the effects of often-severe deconditioning, exercise programs are demonstrated, discussed, and individualized rather than just giving the patient a print-out or costly referral.54,106,107 It almost always is necessary to warn patients to adhere to the negotiated agreement even when they are having the most pain (or other difficulty, such as fatigue or palpitations) ever and, conversely, not to exceed the agreement even if they are having no problem whatsoever. Any change in treatment schedule is negotiated at the next visit. Such pacing and adherence to graded, negotiated programs is essential. These types of CBT work require repetition over time and remain prominent throughout treatment.44–47,52,53,65,108
Providers do not initiate narcotics and tranquilizers. For patients already taking these agents, a program of gradual tapering is negotiated, understanding that compromise around speed of reduction and, occasionally, complete discontinuation may be necessary. The key points for tapering and discontinuing addicting medications are summarized in Table 4. This slow reduction precludes physiological withdrawal and provides for comfortable psychological withdrawal in a safe, supportive circumstance. By following this approach, providers can have many patients completely off all addicting medications in 2 to 4 months. Some patients, however, will be unable to completely discontinue the medications, and it is acceptable to negotiate just a reduction in the dosage. The remainder of the program provides alternatives to narcotics. Sometimes, it can be helpful to negotiate a specific contract with the patient about amount of narcotics to be used.90
Some MUS patients have dysthymia or major depression, varying from mild to severe; when combined they have “double depression.”2 Sometimes, just vegetative manifestations are present.109,110 Although data are limited on the treatment of depression in patients who somatize, these data suggest that treatment is successful111,112; other data suggest improvement in pain at low doses and independent of an antidepressant effect.113,114 We briefly summarize an approach derived from recently validated115,116 Agency for Health Care Policy and Research guidelines117,118 in Table 5 and refer the reader to more detailed reviews.119 Treatment can involve both counseling and antidepressants in full doses, preferably selective serotonin reuptake inhibitors (SSRIs) because of their lower side-effect profile.120,121 Of note, SSRIs have not been shown to be different from tricyclic antidepressants in rates of falls among nursing home residents.122
Many chronic MUS patients have panic disorder and other types of anxiety, often in combination with depression.123 As in the above review, providers use SSRIs,119,124 but we do not recommend the use of either monoamine oxidase (MAO) inhibitors or tranquilizing agents because of their severe side-effect profiles; patients already taking tranquilizing agents will have them weaned and fully discontinued in many instances.
It is explicitly negotiated with the patient that they neither see other providers nor visit the emergency room without prior consultation with the provider; similarly, patients are asked not to get medications refilled from prior sources.78 In addition, patients must sometimes be educated to distinguish emergencies from those problems that can be handled during working hours.
Except with a change in pattern of complaints or the development of objective manifestations of organic disease, providers avoid additional investigation or hospitalization.31 During regularly scheduled office visits, while continuing to integrate patient-centered approaches, including the patient's agenda,44,79,81 it is important that the provider determine that there is no change in symptoms suggesting organic disease and, on the other hand, recognize that not every complaint needs investigation. Recalling that change is slow and treatment occurs over a long period, the provider often must shift the conversation away from the somatic complaints by showing a preferential interest in the psychosocial aspects of the patient's story,8,78 but she/he always performs a brief physical examination pertinent to the patient's complaint.79
The provider also reviews the patient's diaries, helps the patient understand the relationship of symptoms to stress and, when necessary, reviews previous material (e.g., no cancer and no need for further evaluation). New plans are negotiated for activity levels or other actions to be conducted before the next visit. When patients have comorbid organic disease, treatment of the disease is integrated with the 4-point program and may become a major focus.
The provider asks patients to have a significant-other person in their life accompany them for a visit. During the visits, providers are open-ended and empathic and keep the conversation focused upon the patient's health and its impact upon the family member. Providers carefully observe the interaction, how supportive the family member is, and assess if and how the family member can become an ally in care. One must determine the family member's perception of the need for and willingness to seek change; when this is absent, progress is unlikely.125 Because the patient's basic illness behaviors must change, the best allies often are less supportive of the patient's present efforts, less understanding of failure to improve, and less facilitative of continuing the patient's current situation.125 It also is important, oftentimes, to reiterate much of the same material the patient needed to understand.125
These already high-utilizing patients require regular, frequent visits,1 particularly at the outset when most need to be seen weekly; as stability occurs and progress begins, the visit intervals are progressively lengthened as long as the patient remains stable.31,48,78 Usual intervals with stable patients are 4 to 12 weeks. With crises, frequent visits may again be needed. Visits should be time-contingent; this conveys that the provider is more interested in the patient than in the symptom. Providers also can reduce the number of visits with 3-to 5-minute phone calls.126 Intake visits take longer than usual, and 60 to 90 minutes are allotted, sometimes spreading the time over the first 2 visits. Thereafter, visits are 15 to 20 minutes.31,48,78 This recommended method can be continued indefinitely in managing MUS patients.
This literature review, while systematic, was a screening procedure to ensure comprehensiveness, in contrast to a formal literature review comparing treatments and characterized by more stringent control of bias. There were few treatments in primary care to compare, and our task was, rather, to synthesize existing work from all sources.
Although not yet published, our experiences using NPs to deploy this treatment with over 100 MUS patients have been that most patients can be diagnosed and treated without referral for medical or psychological assistance. NPs successfully treated the majority of depressed patients, performed all the drug tapering and discontinuation, and managed most comorbid organic diseases. Psychological and medical consultation were sought if failure in parts of the treatment occurred, e.g., unresponsive depression, worsening headache. Providing back-up for NPs was key, because none had managed a panel of patients independently; they needed to ask questions about medical management, especially at the outset. These issues around treating comorbid organic diseases would not be a problem for a physician provider. Indeed, NPs were in contact daily with the primary care physician and problems usually were first addressed at this level; sometimes, the physician saw the patient but generally offered advice about a patient they already knew quite well in a fully collaborative arrangement.
Preliminary results suggest that patients have been pleased with the treatment, as have physicians and administrators. NPs understandably have been more measured in their responses although they overwhelmingly favor the approach, would do it again, and, most telling, continue to see these patients in their practices after the research treatment concludes.
Even an established treatment, however, will not solve the other problems that also prevent effective management of MUS patients. We will need to address these in parallel. Encouragingly, the field has moved away from blaming the physician and has recognized the many realistic impediments to effective treatment that need to be addressed: 1) through no fault of their own, physicians' education does not prepare them for dealing with mental health and psychosocial problems127–129; 2) it is not yet known whether these so-called “difficult” patients130–132 will respond to treatment in primary care,14,133,134 nor does anyone know how to handle their high rates of resistance and noncompliance135,136; 3) high rates of refusal by patients of referral for mental health consultation also can impair care56–59,129; 4) compensation also has been an issue with these patients, who may require more and longer visits, as well as telephone contacts, and they also may be impecunious129,137; 5) a satisfactory classification/diagnosis system has not been established, rendering it difficult to identify a treatment when we cannot make a diagnosis129,138–143; 6) the role of medical comorbidity in MUS patients has not been understood129,144,145; and 7) competing demands upon the provider compound the problem by, in the absence of enough time to do everything, forcing him/her to make an active choice from the multiple priorities of patients, the practice ecosystem, the policy environment, and themselves.129,137,146,147
Therefore, we must walk soberly, wittingly, and fully aware as we face one of the great problems of this new century in primary care and, indeed, all of medicine—at least if we want a medicine rooted in scientific and humanistic principles.127,128
We have identified an intervention for managing MUS patients that: 1) is conducted by primary care personnel in primary care; 2) utilizes CBT methods similar to those previously reported effective (in specialty care) but deploys them more intensively over a much longer time period; and 3) employs a strong emphasis on communication and provider-patient relationships. We thus have identified a longer, more comprehensive, and more intensive MUS management program for primary care. We presently are testing the method to determine if it will lead to improved mental and physical health outcomes.
This work was supported by National Institute of Mental Health grant MH57099.