Primary care engagement is associated with increased pharmacotherapy prescribing for alcohol use disorder (AUD)

Alcohol use disorder (AUD) is a significant cause of morbidity and mortality but remains under-diagnosed and undertreated. In the United States, 68.5 million adults develop an AUD during their lifetime [1] and 88,000 alcohol-related deaths occur annually [2]. In 2010, high risk alcohol use cost the United States $249 billion [3]. Despite the heavy human and economic toll of AUD, only 71.1% of adults receiving outpatient medical care reported receiving any assessment of alcohol use, and only 2.9% of those with alcohol abuse and 7.0% of those with alcohol dependence were offered information about treatment at the time of diagnosis [4]. Furthermore, only 8.3% of adults with AUD actually receive treatment for unhealthy alcohol use at a specialized treatment center [5].

AUD pharmacotherapy (acamprosate, naltrexone, disulfiram, and topiramate) is efficacious in improving consumption outcomes (prevention of return to any drinking or return to heavy drinking) and broadly recommended by clinical guidelines [6, 7]. Historically, AUD pharmacotherapy prescribing occurred in specialty settings; however integration of AUD care into primary care is increasingly recommended [8‐10]. Primary care settings provide an opportunity for AUD treatment because of primary care providers’ ability to diagnose AUD both through screening and through detecting medical conditions or symptoms caused or exacerbated by alcohol use [11, 12]. Further, primary care provider skills, such as longitudinal monitoring and medication management, and organizational strategies can improve chronic disease outcomes and are readily generalized to AUD pharmacotherapy [13, 14].

Chronic diseases like AUD require assessment and longitudinal follow-up from a regular source of care. Primary care engagement represents a series of visits with a primary care provider [15] and is associated with improved health outcomes for several chronic diseases [16]. Greater patient engagement in primary care may make this setting superior for delivering treatment than specialty AUD treatment settings [17]. Given the difficulty of connecting AUD patients to specialty care, improving primary care engagement among AUD patients and increasing pharmacotherapy prescribing in primary care may be a promising strategy for improving AUD outcomes [18]. Despite this potential, the link between primary care engagement and prescribing of AUD pharmacotherapy remains unclear. To address this gap in knowledge, we conducted a retrospective cohort study of patients in a large urban health care delivery system. During the study period, this health care delivery system did not have organizational guidelines or programs to support the prescribing of AUD pharmacotherapy.

Of 21,159 adult AUD patients in our cohort, almost half were between the ages of 45–64 (45.4%), and the majority were male (71.1%; Table 1). Non-Hispanic Black (35.7%) was the most common race/ethnicity. Only 2.1% of patients were prescribed AUD pharmacotherapy. Of those individuals prescribed AUD pharmacotherapy, topiramate (0.9%, n = 196) and naltrexone (0.8%, n = 161) were the most prescribed medications followed by acamprosate (0.4%, n = 82) and disulfiram (0.2%, n = 42).

During the 2 years after AUD diagnosis, 57.1% of AUD patients had no primary care, 16.9% had limited primary care, and 26.0% were engaged with primary care. Compared to patients with no primary care, patients with limited primary care and patients engaged with primary care were significantly different in all demographic, clinical, and health care utilization characteristics.

In the no primary care, limited primary care, and engaged with primary care groups, the unadjusted percentages of patients receiving an AUD pharmacotherapy prescription were 1.6%, 2.7%, and 3.0%, respectively. In adjusted analyses, the probability of patients with no primary care receiving an AUD pharmacotherapy prescription was 1.61% (95% CI 1.39, 1.84). The probability of patients with limited primary care receiving an AUD pharmacotherapy prescription was 2.56% (95% CI 2.06, 3.06) and 2.89% (95% CI 2.44, 3.34%) for patients engaged with primary care (Table 2).

In our sensitivity analysis, compared with patients with no primary care, the probability of receiving an AUD pharmacotherapy prescription remained higher for patients with limited primary care (adjusted difference: 0.71 percentage points, 95% CI 0.18, 1.24) and for patients engaged with primary care (adjusted difference: 1.21 percentage points, 95% CI 0.69, 1.72). In our original analysis, 63 study subjects entered the cohort with a problem list entry “alcohol consumption binge drinking” or “alcohol consumption heavy,” representing less than 0.01% of the total study sample. Removing these individuals from the study did not change the overall portion (2.1%) of patients who received AUD pharmacotherapy.

In our study examining the relationship between primary care engagement and AUD pharmacotherapy prescribing, primary care engagement was associated with a higher, but modest, probability of receiving an AUD pharmacotherapy prescription. However, the overall percentage of AUD patients prescribed AUD pharmacotherapy was less than 3% even among those engaged with primary care, suggesting broad underutilization.

Our findings are consistent with previous work showing that most adults with an AUD do not receive evidence-based treatment [1, 4, 5]. Further, the low numbers of patients prescribed AUD pharmacotherapy in our study are similar to Veterans Health Administration facilities data and studies of national prescription rates [22‐27]. Our study extends these low prescribing numbers to patients engaged with primary care services. Barriers to AUD pharmacotherapy prescribing include a lack of provider and patient knowledge about pharmacotherapy, formulary restrictions, provider comfort and perceptions of pharmacotherapy effectiveness [28, 29]. Our findings represent prescribing within a health care delivery system without guidelines or programs to address these barriers and improve prescribing. Evidenced-based primary care implementation strategies addressing these barriers are needed to achieve higher rates of AUD pharmacotherapy prescribing in primary care.

Our results show an association between primary care engagement and AUD pharmacotherapy prescribing. Further research should determine if improving AUD patient primary care engagement increases pharmacotherapy prescribing. Patients prescribed AUD pharmacotherapy may be more likely to engage with primary care services; therefore, changing the level of engagement among additional AUD patients may not necessarily increase prescribing. However, strategies to improve primary care engagement in other chronic diseases have improved medication initiation and adherence [30, 31]. The management of AUD as a chronic disease may similarly benefit from improved primary care engagement.

While increased efforts to engage AUD patients in primary care could translate into higher rates of prescribing, nearly 60% of our cohort had no contact with primary care services at our institution. Strategies to link patients from emergency or inpatient units to AUD pharmacotherapy and primary care services for longitudinal management are necessary to reach a broad population of AUD patients. In a recent study, a standard protocol for treatment of hospitalized AUD patients resulted in increased AUD pharmacotherapy prescribing at discharge, which supports the feasibility and efficacy of inpatient AUD pharmacotherapy strategies, but the subsequent level of primary care engagement among such patients remains unclear [32].

This study has several limitations. First, while our absolute AUD pharmacotherapy rates are similar to other studies, our results may not be generalizable to other settings which differ from a large urban academic medical center. Second, because the determination of level of primary care engagement required following patients for the entire duration of the follow up period (2 years), we cannot determine if the prescribing of pharmacotherapy preceded or followed primary care engagement and can only determine an association between primary care engagement and prescribing of AUD pharmacotherapy. Third, we could not account for health care utilization and prescribing of AUD pharmacotherapy at outside institutions. Fourth, because we used ICD-9-CM diagnosis codes and problem list entries, some patients may be misclassified as having AUD or not having AUD and we could not precisely measure AUD severity. Most guidelines recommend consideration of AUD pharmacotherapy for patients with moderate or severe disease; therefore, our estimates of AUD pharmacotherapy may be conservative. Finally, we focused on a measure of initial prescription of AUD pharmacotherapy and did not assess adherence (prescriptions dispensed) or duration of therapy.

Overall, while primary care engagement was associated with a higher, but modest, probability of receiving a prescription for AUD pharmacotherapy, few patients were prescribed AUD pharmacotherapy. Strategies to greatly increase overall rates of AUD pharmacotherapy in health care settings are needed to improve the proportion of patients receiving evidence-based care. Primary care engagement may be one potential strategy to increase prescribing of AUD pharmacotherapy.