Chronic, non-cancer pain, defined as pain of greater than 3 months duration, is a common, important health issue. The prevalence of chronic pain ranges from 20% to 60% [1]. The prevalance of low back pain is greater than 30%[2], and the prevalance of migraine is approximately 15%[3, 4]. Pain disorders, including headache, back pain, arthritis and other musculoskeletal pain, are estimated to cost the United States economy $61 billion per year in lost productive time [5]. It is frequently asserted that chronic non-cancer pain is undertreated [6–8].

The optimal approach to treating chronic pain is controversial [9–13]. The realization that acute pain and cancer pain were often undertreated led to a liberalization of opioid use in these populations in the late 1980's and throughout the 1990's [6, 14–16]. Subsequently, many advocates and practitioners called for an increase in the use of opioids for patients with chronic non-cancer pain as well [14, 17, 18]. Other experts assert that use of opioids for chronic pain should be limited [10, 19]). used with caution [20], or eschewed altogether due to potential for substance misuse and lack of proven efficacy [13].

In large part, uncertainty about how to best manage patients with chronic pain stems from a lack of research in this area [21]. The limited available research comes from studies performed in specialty practice settings [11, 22–25]; few studies have examined the care of chronic pain in primary care [26, 27]. Unlike other common chronic diseases such as congestive heart failure or diabetes, there are few clinical trials to inform practice guidelines.

Chronic pain management is complex. Like other chronic diseases, chronic pain is multi-faceted, and it is attended by its own set of comorbidities. It is complicated by substantial psychological and functional impairment in the forms of depression, disability, and loss of livelihood [28–30]. It is also costly. Patients with depression and back pain or migraine incur 3 to 4 times higher medical costs than patients with these pain conditions alone [31].

Generalists, even if well-versed in the biopsychosocial model of disease, often feel unprepared to manage chronic pain. They may lack the training in using certain pharmacological regimens, such as combining chronic opioids with psychiatric and pain modulating agents [27, 32]. They may fear regulatory scrutiny and sanction when prescribing opioids, and they may be wary of fostering opioid dependence, misuse or addiction [33, 34]. These barriers result in variations and inconsistencies in care that leave both patients and providers frustrated.

We developed a disease management program to improve the management of chronic pain in opioid-treated patients attending an academic general medicine practice. Traditional models of office-based care focus on diagnonsis and acute management of medical problems in the context of a single provider-patient relationship. In contrast, disease management emphasizes: (1.) The use of multi-disciplinary teams providing integrated care; (2.) Evidence-based algorithms; (3.) Interval visits to monitor response to therapy; and (4.) Information systems that permit tracking of outcomes and adjustment of therapy [35–39]. We applied these principles in developing and implementing an intensive, structured, and coordinated program to improve the management of patient with chronic pain. We focused our program on patients treated with opioid medications and attempted to provide an environment that would allow the safe and effective use of these drugs. We sought to determine if this program could improve pain, functional, and psychiatric outcomes in a 3 month uncontrolled trial.

We found that a multi-disciplinary, primary care-based, disease management program can improve pain, depression and disability scores in opioid-treated patients with chronic pain in a 3 month uncontrolled trial. The improvements across all outcomes support an improved quality of life resulting from the intervention. These improvements were obtained using an approach that balanced the potential benefits and adverse effects of opioids.

We hyothesize that these improvements resulted from the combined effects of systematizing pain management and treating depression. Improvements appear independent of opioid dosing. The clinical significance of the 12% to 15% improvement in pain scores is unclear. Uncontrolled trials in specialty pain clinics have reported a 20% to 25% reduction in pain scores [57]. Some research suggests that a 30% decrease in pain scores (i.e. about 2 points) represents clinically significant relief of pain [58], but the issue of how to interpret pain scales clinically is not resolved.

The improvement in depression scores was clinically important and may reflect combined effects of intensification of pharmacological management for depression and pain and the systematization of care. Although the reciprocal relationship between pain and depression has been established in previous studies, the effect that treating one condition has on the other has not been well-assessed [29]. One recent study demonstrated that the presence of severe pain predicted a poor response to antidepressant therapy, and thus it is plausible that intensifying pain management would have a beneficial effect on depression [59]. Clearly, the improvements in depression scores seen in this study cannot be attributed to increasing anti-depression pharmacological therapy alone because the proportion of patients treated with anti-depressants increased from 44% to 52% only. We have since added a structured treatment algorithm to increase the use of anti-depressants. It is important to note that there was a statistically signficant trend toward greater depression among substance misusers who did not complete the trial; thus, it is possible that the trial overestimates the effect of multi-disciplinary management on depression outcomes.

To our knowledge, this is the first study to prospectively examine the effects of multi-discplinary pain management on the outcomes of pain, disability, depression, and substance misuse in an academic primary care practice caring for a wide range of patients. Previous studies conducted in a military clinic[60] and a health maintenance organization [26] demonstrated improved pain and functional scores with systematic, multidisciplinary intervention. The military trial was uncontrolled and enrolled referred patients into a specialty clinic. The HMO trial was conducted in a primary care setting. It was controlled, and did evaluate pain, function and mood outcomes. Neither study systematically monitored substance misuse.

We documented a high prevalence of substance misuse (32%). We did not assess addiction per se. The prevalence of substance misuse and addiction in patients receiving chronic opioids is unclear and depends on the populaton under study. Some authors have asserted that addiction and substance misuse are uncommon consequences of opioid use for pain. One widely cited reference estimated the prevalence of addiction at approximately 4 in 10,000 treated patients [61]. Others have reported prevalences of addiction ranging from 3% to 17% [62, 63]. A recent retrospective study in a primary care setting documented a high prevalence of opioid misuse: 24% in a resident physician clinic and 31% in a Veterans Administration outpatient clinic[64], but the criteria used to determine the prevalence of opioid misuse were limited by chart review. Some behaviors defined as indicators of opioid misuse (e.g. lost or stolen medications, requests for early refills) could be construed as indicators of inadequately treated pain (i.e. "pseudoaddiction"[65]) and not substance misuse or addiction.

Opioid misuse not only complicates the management of pain in the individual patient, but has negative societal consequences as well, especially when opioids are diverted from their intended use [66–68]. Several states have documented increases in unintentional deaths from opioids, especially diverted methadone [69–71]. National surveys demonstrate dramatic increases in the non-medical use of OxyContin^® and other prescription drugs among teens and young adults [72, 73]. The trauma literature has documented recent increases in opioid use among patients admitted to trauma centers [74]. In response, there have been state and national initiatives to reduce prescription drug misuse [75, 76]. Though high rates of substance misuse are a source of concern, our program may serve as an example for how care can be organized to reduce misuse without eschewing the benefits of opioid medications.

Although the prevalence of substance misuse in our study population is higher than reported in clinical trials of opioids, these trials have occurred in specialty settings with selected populations. They excluded patients with a history of substance misuse, or have not systematically monitored patients for substance misuse [11, 12, 23, 66, 77]. In addition, they commonly excluded patients with psychiatric illness (including depression) which is a strong predictor of substance misuse [23, 66, 78–80]. Patients in our program had a high baseline prevalence of depression, previous substance and alcohol abuse, and other psychiatric disorders (Table 1.). The strong relationship between mental illness and substance abuse disorders is well known and thus the high prevalence of substance misuse is not entirely unexpected [81]. Previous studies of mental illness have documented a high coexisting prevalence of substance and alcohol misuse: 32% with unipolar depression; 61% with bipolar depression; 47% with schizophrenia; 84% with personality disorders: and 24% with anxiety disorders [78, 82].

We specifically sought out patients whose pain management was difficult for providers or in whom substance misuse was suspected. Many had established or suspected psychiatric diagnoses. How to identify chronic pain patients at risk for drug misuse and to treat their pain remains a challenge [83–89]. The pattern of substance misuse in our population often suggested polysubstance abuse; this places patients at especially high risk of morbidity and mortality [90]. Although patients committing substance misuse were offered substance abuse treatment referral, only two followed through and most did not return to our program. The option of pain management without the use of opioid analgesics was offered to all patients who committed substance misuse.

The difficulty in obtaining mental health and substance abuse treatment services is a pressing public health issue and a topic of national debate in the United States [81]. In our sample there was a clear trend toward increased comorbid depression among substance misusers who did not complete three month follow up (Table 2.). Despite the availability of on-site psychiatric consultation, our program was not successful in retaining and managing a challenging subset of patients with substance misuse and depression.

We are aware that some of our substance misusing patients migrated to other practices in order to obtain opioids and other controlled substances. Our program implemented policies to prevent migration of patients within our practice (Appendix B.) and the University of North Carolina Health Care System. The cornerstone of these policies was meticulous documentation in an electronic medical record that is accessible to all physicians at our medical center and to hospitals and physicians affiliated with our system in the surrounding communities. In general, though, we have no direct control over migration that occurs outside of our practice and our health care system. In order to curtail migration and "doctor shopping," some states have implemented centralized monitoring systems for opioids and other controlled substances. North Carolina is currently exploring the feasibility of such a system. A description of the operational state monitoring programs is available online through the United States Drug Enforcement Agency Diversion Control Program website [91].

We believe that our results may be generalizable to other academic primary care practices that serve diverse patient populations with a high burden of medical and psychiatric comorbidity. The etiologies and sites of pain were similar to those reported in population-based epidemiological surveys and clinical trials in primary care, except that headache was under-represented in our population [1, 12, 92]. The results replicate epidemiological research that demonstrates a strong interaction between pain and the psychiatric comorbidities of depression and substance misuse [87, 88].

Our study may be more applicable to the general medical setting than previous trials examining the effects of opioids on chronic pain because we did not exclude patients with serious psychiatric comorbidity or those suspected of substance abuse. To be effective, pain management should encompass more than pharmacological management directed at pain scores; it should address a variety of behavioral and psychosocial factors that contribute to suffering [86, 93].

Our study has several limitations. It was uncontrolled and of relatively short duration. The research assistants were not blinded to pre- and post-treatment assessments. The improvements could reflect secular trends, although the chronic nature of our patients' symptoms and disability makes this less likely, and improvements were noted across all of the pre-specified outcomes. Not all patients receiving chronic opioids in our practice were referred. Thus, we may have over-estimated the prevalence of substance misuse because providers were more likely to refer "problem" patients in whom they suspected opioid misuse.

Another limitation of our study is its individualized nature. We did not adhere to strict algorithms for diagnosis and treatment and did not test a single intervention. The evidence-base for managing chronic pain in the general medicine setting is limited and the multi-modal nature of our intervention was by necessity empirical and exploratory. As such, we decided to allow more latitude and individualization in treatment choice. We have used our experience and the data collected to develop more robust algorithms to guide the management of pain and depression and to make psychiatric referral when appropriate.

As a corollary to our multi-modal approach, it is difficult to ascertain if the improvements derived from pain medications, intensification of depression therapy, or simply participation in an organized program. Improvements and changes in behavior that occur as a result of becoming a target of special interest in a program are often referred to as a Hawthorne effect [94, 95].

In a 3 month trial conducted in an academic primary care setting, a systematic, multi-disciplinary approach to chronic pain management that included the use of opioids and tools to prevent misuse was effective in improving pain, depression, and function scores. Future efforts will be directed at examining their durability and promoting their sustainability. A randomized control trial would determine whether these are real effects or represent a secular trend.

Comorbid depression and substance misuse were common. Efforts will also be made to further characterize the interaction of these and other comorbid psychiatric conditions with chronic pain. Chronic pain patients with substance abuse are a challenging subset of patients who could benefit from new research and initiatives to mitigate the risk of abuse while ameliorating pain control.

Patient Name ____________________________ Diagnosis ______________________________

Physician Name __________________________ Telephone Number _______________________

UNC General Internal Medicine Practice

Pain Review Committee:

Policy on Serious Opioid or Controlled Substance Misuse & Misconduct

Definition

Procedure

Serious misuse or misconduct is a special category of misuse. It results in the immediate discontinuation of opioids in the Internal Medicine Clinic. The Pain Review Committee will address instances of serious misuse on an expedited basis. Individual cases will not require the review of the entire committee. The following procedure applies: