Introduction

Alcohol use disorder (AUD) is characterized by compulsive alcohol use and loss of control over alcohol intake, and it is associated with substantial morbidity and mortality in North American populations [1–3]. In Canada, more than 3000 deaths/year are attributable to alcohol consumption [3]. Globally, 5.3% of all deaths and 5.0% of the global burden of disease and injury (measured in Disability-Adjusted Life Years) in 2016 have been attributed to alcohol, translating into approximately 3 million alcohol-attributable deaths after taking into account the potential benefits of low-risk drinking [4]. Alcohol use is directly related to many other adverse physical and psychosocial health outcomes, including cancer, infectious disease, cardiovascular disease, neuropsychiatric disease, and intentional and unintentional self-injury [4–6]. Moreover, alcohol drinking affects not only individuals, but has a negative impact on society as a whole, as it is often associated with increased rates of child abuse and/or neglect [7–9], intimate partner violence [10, 11], and hospitalization and mortality rates [12–14].

Despite the substantial individual, societal, and global public health burden of AUD, it appears to be largely untreated [1, 15]. Even though effective pharmacotherapies for AUD have been available for decades and are supported by numerous systematic reviews and meta-analyses establishing both their efficacy and safety [16–19], drugs such as naltrexone and acamprosate have been found to be severely underutilized among AUD patients in select populations [20–22]. For example, in studies of US military veterans with AUD, fewer than 5% received any kind of pharmacotherapy [21, 22]. In another study enrolling low-income AUD patients in Ontario, Canada, the rate of AUD medications dispensed was similarly low [23]. While these low rates are concerning, questions as to the generalizability of previous findings remain, since many of the existing studies examining AUD pharmacotherapy enrolled non-representative samples. Thus, the rate of treatment uptake in the general population is still largely unknown.

Our study aims to address this gap in the literature, leveraging comprehensive whole-population data to determine the proportion of individuals receiving pharmacotherapy for AUD. We used administrative data from the general population of Manitoba, Canada to examine the use of pharmacotherapy (naltrexone, acamprosate, and/or disulfiram) among individuals with AUD. We describe their sociodemographic and clinical characteristics and their healthcare use, determine the medical specialties most likely to prescribe them AUD medications, and identify variables that predict whether or not individuals with AUD are likely to fill prescriptions for AUD pharmacotherapy. The study’s findings shed light on prescribing patterns among physicians caring for AUD patients, and inform clinical practice for treating AUD in general populations.

Methods

Results

The proportion of Manitobans with AUD using pharmacotherapy is presented in Table 1. Between 1996 and 2015, 37,388 individuals were diagnosed with AUD, representing 2.9% of the Manitoba population. Among these, 493 individuals received a prescription medication for AUD and 36,895 individuals did not receive any AUD-specific pharmacotherapy. Among individuals who received pharmacotherapy for AUD, 58.0% were prescribed naltrexone, 36.3% were prescribed acamprosate, and 5.7% were prescribed disulfiram. A higher proportion of males vs females filled a prescription for each type of AUD medication. Table 1 also presents the sociodemographic characteristics of the study cohort. The mean age at index date for both groups was 40 years. More than half of individuals who filled a prescription for an AUD medication were first diagnosed between the ages of 30–50, while non-prescription users were more evenly distributed among age groups from 12–79 years. The majority of those who filled a prescription lived in urban vs rural settings, and they were fairly evenly represented across income quintiles, whereas those who did not fill a prescription demonstrated a clear income gradient.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociodemographic characteristics of Manitobans with alcohol use disorder.

https://doi.org/10.1371/journal.pone.0257025.t001

Table 2 presents the clinical characteristics and healthcare use of the study cohort. The majority of AUD medication prescription users fell into Band 3 or 4 on the RUB scale (moderate health resource use), while many non-prescription users were in Band 0 (no significant health resource use). Prescription users were more likely than non-prescription users to have a mood- or anxiety-related diagnosis leading up to their index date. Those who filled prescriptions for AUD medication had more hospitalizations and ambulatory visits in the year prior to index date, and were also more likely to have mental health-related ambulatory care visits than those who did not fill a prescription. Neither the number of emergency department visits nor the number of prescription medications for chronic conditions in the year prior to index date was different between the two groups. Prescription users were more likely than non-users to receive SSRI antidepressants, but not TCA antidepressants. They were also more likely to receive sedatives, anxiolytics and the anti-epileptic medication lamotrigine.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Clinical characteristics and healthcare use of Manitobans with alcohol use disorder.

https://doi.org/10.1371/journal.pone.0257025.t002

Fig 1 shows that the majority of prescriptions for AUD medications came from family physicians practicing in urban settings, followed by psychiatrists and family physicians in rural settings. Fig 2 presents the results of the hierarchical regression model, showing which variables were important predictors of individuals with AUD receiving pharmacotherapy. These findings indicate that individuals with AUD were more likely to fill an AUD prescription medication if they lived in an urban setting or had a comorbid mood or anxiety disorder diagnosis. Older individuals had reduced odds of receiving an AUD medication. Sex was not a significant predictor of receiving a prescription medication for AUD.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Physician specialties responsible for prescribing individuals’ first alcohol use disorder medication.

Index dates 1996/97 to 2014/15.

https://doi.org/10.1371/journal.pone.0257025.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Variables predicting pharmacotherapy use among individuals with alcohol use disorder.

Odds Ratio with 95% CI.

https://doi.org/10.1371/journal.pone.0257025.g002

Discussion

In the general population of Manitoba, only 1.3% of individuals diagnosed with AUD filled a prescription for an AUD medication over the course of the 20-year study period; naltrexone was the medication most commonly prescribed. This finding clearly highlights the profound underutilization of evidence-based medications for AUD, and is consistent with other research on low prescribing rates of AUD medications among US veterans [20, 21], low income Canadians [23], and Australians [32]. Low prescribing rates, uptake and utilization of AUD medications is a multi-faceted and complex matter influenced by patient, physician, and health system factors, including perceived low patient demand for the treatment, lack of provider knowledge about pharmacotherapy for alcohol dependence, and lack of confidence in treatment efficacy [33, 34].

In our study, the majority of individuals who filled a prescription spanned ages 30 to 50 years, a lengthy span which reflects the lag time typically observed between AUD diagnosis and onset of treatment [32]. While AUD often begins in young adulthood, the wide range of ages among individuals being treated for AUD points to a considerable gap between diagnosis and treatment, highlighting a need for improved AUD recognition in primary care settings and more aggressive outreach before individuals with AUD begin to present in emergent care. Older individuals in our study were less likely to receive pharmacotherapy for AUD than the 30–50 year old group. A similar finding has been found in other studies of health service use and treatment for substance use among older adults, despite positive treatment outcomes [35, 36]. These findings point to an opportunity to improve access to care and broader treatment options for older patients with AUD.

The vast majority of patients with AUD who received pharmacotherapy lived in urban settings, perhaps reflecting better access to prescription medications for AUD in urban vs rural areas in Manitoba. Individuals with AUD living in urban settings were also evenly distributed across income quintiles. This is an interesting finding as previous research has shown that lower income individuals are more likely to receive treatment for AUD [37]. Our finding can perhaps be explained by the availability of full coverage for eligible medications for all Manitoba residents after an income-based deductible is met. Finally, those who received AUD medications were also more likely to have overall higher health resource utilization and, in the year before diagnosis of AUD, have higher rates of hospitalizations. These findings likely reflect the impact of AUD on an individual’s physical health, contributing to increased resource use and hospitalization prior to pharmacotherapy being initiated, and are corroborated by a number of other studies [12, 30, 38].

In Manitoba, the majority of AUD medication prescriptions were provided by family physicians practicing in urban settings, with psychiatrists being the second most frequent prescribers. These results are in contrast to other studies of AUD medication prescribing, which found that psychiatrists are more likely than other medical specialties to use pharmacotherapy in the general population of individuals with AUD [34, 39]. Research from the US Veteran’s Health Administration has also shown that clinicians with dedicated training in addictions or physicians working at specialized addiction treatment centres tend to use pharmacotherapy at a higher rate than family physicians. However, in Manitoba, a large portion of addiction care is provided by family physicians, which may partially explain why they were the specialty most likely to prescribe AUD medications in our study. Overall, prescription rates for AUD medications were quite low, warranting improved education for physicians about the pharmacological options available to individuals with AUD. Family medicine, with its high prevalence of collaborative care models, is an ideal setting in which to initiate pharmacological treatment for these patients [40].

Our study found that individuals with AUD were more likely to be prescribed an AUD medication if they had a co-morbid mood or anxiety disorder diagnosis. This finding aligns with studies of co-morbidities among US veterans [20, 21]. Patients with a co-occurring disorder are more likely to present to specialty psychiatric care, and thus may benefit from psychiatrists’ increased knowledge of and/or experience with AUD medications. Furthermore, the efficacy of naltrexone in combination with the antidepressant sertraline has been demonstrated in a population with co-occurring AUD and depression [41]. Sex was not associated with receiving an AUD medication prescription, which is in keeping with research from Ontario, Canada [23].

Conclusion

This large population-based longitudinal cohort study examined receipt of naltrexone, acamprosate, and disulfiram treatment among individuals with an AUD diagnosis in Manitoba, Canada. Despite established evidence for pharmacotherapeutic treatment of AUD, these medications were profoundly underutilized, potentially contributing to considerable public health harms for individuals, families and society at large. Future research in this area should focus on medication adherence and time to prescriptions following diagnosis in the general population, as well as specific barriers preventing pharmacotherapy treatment for individuals with AUD.

Supporting information

Acknowledgments

We acknowledge the Manitoba Centre for Health Policy for use of data contained in the Population Health Research Data Repository (HIPC Project Number 2014/2015–30). Dr. Bolton is supported by a Canadian Institutes of Health Research operating grant and a NARSAD Young Investigator Award.