Steve Martino, Ph.D., Yale University School of Medicine and Veterans Administration Connecticut Healthcare
As a proponent of integrating addiction treatment into medical care, I find myself in the unique position of serving as the devil’s advocate in this edition of The Bridge. The question I have been asked to address is: How could collaboration and coordination of health and addiction care among providers carry potential risks or complications?
An integrated care model has been the recent darling of our field, with the Institute of Medicine calling for the normalization of this model in the United States (National Academy of Sciences, 2006). To date, placing SBIRT (Screening, Brief Intervention, and Referral to Treatment) in medical settings is the most well-known and prevalent effort to promote integrated care. Indeed, the US Preventive Services Task Force (2013) the National Institute on Alcohol Abuse and Alcoholism (NIAAA, 2005) and American College of Obstetricians and Gynecologists (ACOG, 2004) now recommend universal screening for alcohol and drug use and brief intervention in primary care.
Reasons for integrating addiction treatment and medical care have been widely noted (Gordon et al., 2013; Walley et al., 2012):
1. Patients presenting to medical settings have plenty of risky tobacco, drinking and drug use;
2. Patients with substance use disorders often do not perceive a need for treatment and therefore do not seek care in specialty addiction treatment settings;
3. Improvement in one area (substance use) might mediate improvements in the other (medical);
4. Integrated care could create “one-stop shopping” for patients and spur more continuous relationships with treatment team providers;
5. Providers who work within an integrated model have more opportunity for direct communication and collaboration about patient care issues.
Further bolstering the argument for integrated care is the Affordable Care Act. It will give more people access to health care and to services provided under mental health and addiction treatment parity rules. Therefore, more patients with unhealthy alcohol or other drug use will likely be seen in primary care and other medical settings. An integrated care model might be able to more efficiently manage the interconnected problems of patients and improve their treatment outcomes than the bifurcated system of care that still dominates the United States.
With all this promise, l will now transform me into the devil’s advocate. I present potential unforeseen or unintended consequences of integrating addiction treatment into medical care. My aim is to point out the challenges of integration that, if addressed, could improve the integrated care model.
First, integrated care is a promising idea that has limited empirical support (Nilsen, 2010; Saitz, 2010; Saitz et al., 2010). For example, SBIs show consistent efficacy among primary care patients who drink alcohol in risky or unhealthy ways (Kaner et al., 2009). However, SBIs have unproven efficacy with alcohol dependent patients, illicit drug users or those who misuse prescription medications, and within different patient populations (e.g., adolescents) and medical settings (e.g., inpatient). A brief alcohol intervention designed to fit the demand characteristics of busy primary care settings might be inadequate when exported to new and often more complex applications. Moreover, the effect of SBIs on critical indices of health service utilization (e.g., emergency department admissions, re-hospitalizations) remains unclear. Despite this absence of evidence, federal and state authorities continue to encourage medical providers to use SBIs widely (Knopf, 2007; Madras et al., 2009). The same enthusiasm has been demonstrated for chronic care management to address medical and mental health issues simultaneously within patient-centered medical homes (Bindman et al., 2013; Katon et al., 2010). However, when applied to patients with alcohol and other drug dependence, a chronic care management approach compared with a primary care appointment only condition did not increase self-reported abstinence over one year (Saitz et al., 2013). The field is pushing approaches that might not work as intended and needs more research demonstrating the effectiveness and cost-effectiveness of integrating addiction treatment into medical care.
Second, it is unrealistic for medical providers to address all the preventive medical care mandates. Some estimate it would require between 3.5 – 10.6 hours per day for primary care clinicians to follow all recommended guidelines for screening and behavioral management of the top 10 chronic diseases (Ostbye et al., 2010). Medical providers consistently report lack of time, lack of training, insufficient knowledge of area referral and follow-up resources, and low confidence when addressing addiction issues (Delgado et al., 2011; Rose et al., 2008). These busy providers might simply vote with their feet and not integrate addiction treatment into their medical practices, as has often occurred with SBIs (Nilsen, 2010; Kaner, 2010). Integrative strategies that decompress the medical provider workload, rather than squeeze it further, could have a much greater likelihood of utilization. For example, “warm hand-offs”, in which the health care providers literally walks the patient to a behavioral care provider co-located in the medical setting, might be a feasible model of integration (Cummings et al., 2009), though warm hand-offs as an integrative strategy has seldom been studied for its effectiveness in treatment engagement or outcome improvement.
Another way to address provider burden is to incorporate technology-based approaches to addiction treatment (e.g., computer- or web-delivered SBI) in medical settings. When used as brief alcohol interventions, these approaches are effective (Rooke et al., 2010). However, as with provider-delivered SBIRT overall, it is unknown if computer-based SBIs for alcohol use generalize to other substances and a range of patient populations and settings, and they may not be preferred by all patients even if they reduce provider burden. Recently, Cucciare and colleagues (2013) found that a web-based brief alcohol intervention using normative feedback delivered to veterans in a primary care clinic conferred no additional benefit to treatment-as-usual. While the authors discussed several possible reasons for the absence of differential effects, they speculated that brief alcohol interventions were originally developed to address college student drinking. The veterans in this study included older individuals with chronic alcohol problems/dependence. Older individuals may be less comfortable using technology-based approaches and prefer to talk with a provider. Offering patients choice in how they receive addiction treatment might maximize the benefits of integration. Research efforts should aim to identify specific integrated care approaches and models that work best for specific populations.
Even if medical settings could offer brief interventions delivered by person or computer, these interventions generally are not geared toward those with the most severe addiction problems, and health care providers often feel ill-equipped to handle alcohol and drug dependent patients. It remains to be seen if specialty addiction services (e.g., methadone maintenance, ambulatory detoxification) can be embedded within real-world medical settings or if in some cases it might be better to embed medical services in specialty addiction programs. In either case, both treatment systems would need the resources to deliver the absent element of care on site. A two-way street of integration may be necessary to accommodate the broad range of medical and addiction problems patients bring to both settings.
A question also exists about the degree to which the integration of addiction treatment into medical care might reach those most in need of it. Mulia, Schmidt, and Greenfield (2011) acknowledge efforts to bring more evidence-based behavioral practices, such as SBI, into primary care. However, they note that racial/ethnic minorities and low-SES members of these groups are less likely to be seen by primary care providers. Increasing the availability of addiction treatment in primary care, while generally good, could have the unintended consequences increasing disparities in access to health care for substance use problems. They and others (Saitz et al., 2013; O’Connor, 2013) underscore the need to broaden the venues for delivering integrated care in places where people with co-occurring addiction and medical problems are likely to go. These include schools, work settings, and community centers.
In addition, integrating addiction treatment into medical care might reveal stigmatization of patients by health care professionals. Bitarello et al. (2012) showed that having drug dependence, with or without alcohol dependence, was associated with receiving significantly worse patient-rated quality of primary care across most domains involving physician-patient relationship. Specifically, injection drug use was consistently associated with lower quality of primary care, raising the possibility that methadone or buprenorphine maintained patients could be most prone to stigmatization when seeking care in medical settings. Current practices of withholding antiviral Hepatitis C treatment or dropping people from transplant lists because of relapse may be manifestations of the broader issue of stigmatization. Health care providers need more education and training about addiction and treatments as integrated care moves forward. Fortunately, the American Board of Addiction Medicine launched an effort in 2011 to train more physicians in addiction medicine, accrediting residencies and fellowships at 10 sites in the United States (Frisch, 2013).
Integrated care also might not result in expected improvements in inter-professional communication. For example, Kim et al. (2013) examined the degree to which non-physician primary care clinicians documented the results of substance use screening and brief interventions. They found that most patients with unhealthy substance use had no electronic medical record documentation of SBI even though all patients had received one. They state “that beyond the division of labor, team-based care requires effective forms of communication between team members (p. 207)”. Further, they call for active efforts to coordinate team activities and develop effective communication processes, lest the physician’s warm hand off might become an occasion to wash one’s hands of the patient’s addiction issues. Requirements for documenting behavioral interventions, such as SBI, in the medical record must be clear. Moreover, physicians must value the substance use screening results and brief interventions provided by nurses, medical assistants, or health educators and use them to guide the patient’s care.
Finally, integrating addiction treatment into medical care involves an implementation process. Effective implementation strategies that best support integrated care in medical settings are unclear, though many (e.g., academic detailing, information technologies, prompting, and performance feedback) show promise.18 As one example, the American Medical Association and the Centers on Medicare and Medicaid have adopted billing codes that directly reimburse physicians for providing SBI services (Knopf, 2007). The extent to which the magnitude of this incentive is sufficient for changing practice is unknown. Furthermore, the complex interplay of political policy, organizational, provider, and patient characteristics will influence the implementation of integrated care. The field needs to better understand these factors for integrated care to be sustainable in medical settings. In this regard, NIDA has funded six projects via RFA-DA-12-008, “Integration of Drug Abuse Prevention and Treatment in Primary Care Settings (R01)” to help illuminate this area.
The field must address these devilish issues to avoid undermining the success of integrating addiction treatment into medical care. Forging ahead with the best intentions does not always yield hoped-for-results. Widespread implementation of debriefing after traumatic events is a case in point (Rose et al., 2003; Emmerik et al., 2002). Likewise, increased attention on pain treatment during the past decade may have contributed to recent increases in opioid analgesic misuse and overdose, and reducing the supply of prescription opioids to pain patients in reaction to this miscalculation could result in increasing heroin use (Gordon et al., 2013). If done poorly, integration of addiction treatment into medical care might yield poor outcomes or costly null effects, overburdened and poorly communicating health care providers, dissatisfied, stigmatized, and excluded patients, and abandonment by the addiction treatment and medical fields.