Patient-Reported Offers of Alcohol Treatment for Primary Care Patients at High-Risk for an Alcohol Use Disorder

Abstract

Background

This study assessed patient-reported alcohol treatment offers by healthcare providers following routine annual screening for alcohol use in primary care.

Methods

A telephone interview within 30 days of the annual screen assessed demographics, alcohol and other drug use, mental health symptoms, and offers of formal treatment for alcohol by a VA healthcare provider. We included male patients (n = 349) at high-risk for an alcohol use disorder (AUD) who had not received alcohol treatment in the past three months. We assessed self-reported receipt of any offers of formal treatment for alcohol, and associations of offer of formal treatment for alcohol with demographic and clinical variables.

Results

145 (41.5%) patients reported an offer of at least one type of formal treatment for alcohol. More severe alcohol misuse (OR 1.07, 95% CI 1.03 to 1.11) and younger age (OR 0.97, 95% CI 0.95 to 0.99) were associated with reporting an offer of formal treatment.

Discussion

Most primary care patients at high-risk for an AUD were not offered treatment following an annual screening. Our results highlight the importance of training primary care providers in what constitutes appropriate medical treatment for this population and the most effective ways of making a treatment offer.

Introduction

Alcohol use disorders (AUDs) are prevalent, disabling, and have significant negative consequences;¹ nonetheless, fewer than 20% of those with lifetime AUDs in the United States ever receive treatment.² The availability of evidence-based treatments has become a health policy priority as a result of reforms focused on better access to and quality of behavioral health services.³ Alcohol treatment offers made by health professionals during a primary care visit provide an opportunity to make evidence-based treatments more available to those with AUDs,^4-6 and may be a precursor to primary care patients accessing alcohol treatment when needed.^7,8

Health professionals in primary care settings are likely to encounter patients who demonstrate high risk for an AUD, such as those scoring eight or more on the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) or have an AUD diagnosis yet have not been receiving alcohol treatment.^9-12 Clinical practice guidelines recommend offering these patients alcohol treatment, such as referring them to specialty care, monitoring their drinking, and offering addiction-focused pharmacotherapy.¹³ Because patients with an AUD generally do not visit primary care to seek alcohol treatment,¹⁴ it is important to examine whether clinicians offer treatment when high risk for an AUD is identified during screening, even if patients do not mention or ask for it.¹⁵ Health professionals in primary care settings may not be aware of the various evidence-based pharmacological or behavioral treatment options for patients with AUDs.¹⁶ Consequently, the type of treatment offered should also be examined to ensure that appropriate care is recommended.^17,18

Patient-reported measures are useful for examining alcohol treatment offers in primary care settings.¹⁹ Information on alcohol treatment offers during clinical encounters is typically not captured in administrative data, may not be documented accurately by providers,²⁰ and is expensive to examine via medical record review.²¹

Objectives

This study examined patient-reported receipt of alcohol treatment offers by a healthcare provider to patients identified as high risk for an AUD following routine annual screening for alcohol misuse in primary care. We assessed the (1) proportion of patients who reported receipt of an offer of formal treatment for alcohol, (2) types of treatment offer reported, and (3) predictors of patient-reported treatment offers.

Methods

Research Procedures

Between February 2013 and February 2014, we conducted a telephone survey within 30 days of all patients screening positive for alcohol misuse (five or more on the AUDIT-C in VA administrative data) during routine annual screening.⁹ Patients received alcohol screening in one of several outpatient clinics at the VA Greater Los Angeles Healthcare System (GLA). The RAND Human Subjects Protection Committee and GLA Institutional Review Board approved this study.

Eligibility Criteria

We restricted these analyses to the subset of male patients at high risk for an AUD: i.e., patients receiving an eight or more on the AUDIT-C during the annual screen or with an AUD diagnosis on the screening date (clinician-recorded administrative claims data using ICD-9 codes) who had not received alcohol treatment in the past three months.^10-12 The VA Clinical Practice Guideline recommends these criteria to identify individuals who should be referred to specialty SUD care.¹³ Further eligibility criteria included: aged 18 years or older, no cognitive impairment (measured via administrative claims data using ICD-9 codes), engaged in care at GLA, a telephone number recorded in the VA system, and completing the interview within 30 days of the annual screen.

Measures

Demographics

We collected self-reported race/ethnicity, marital status, education level, employment status, income, insurance coverage outside VA, and whether participants received all or most of their care through VA or an outside provider. We obtained information on age and sex from VA administrative records and all other information from the survey.

Clinical Measures

The full AUDIT²² was used to assess severity of alcohol misuse. Participants were also asked if they had used any illegal drug in the past 30 days. Depression symptoms were assessed using the Patient Health Questionnaire (PHQ-9),²³ and anxiety symptoms were assessed using the Generalized Anxiety Disorder Assessment (GAD-7).²⁴ Items from the SF-12 Health Survey were used to assess overall physical and mental health.²⁵ Stage of change was assessed using the Readiness to Change Questionnaire.²⁶

Treatment Offers

Participants reported whether a VA doctor or other healthcare provider, in the last 30 days, offered the following specifically for their alcohol use: (1) therapy or counseling, (2) medication (e.g., acamprosate, disulfiram, naltrexone), (3) referral to intensive outpatient treatment or a residential program, or (4) unspecified medical treatment. We then derived a binary indicator for receiving an offer of at least one type of formal treatment.

Statistical Analyses

We calculated descriptive statistics for all measures. We then conducted a multivariable logistic regression to assess associations of any formal treatment offer with demographic and clinical variables;²⁷ we chose model adjustment variables based on scientific grounds, prior to estimation, with the intent to control for potential confounding. We did not make use of any automated variable selection procedures.

Results

Of 1,922 approached patients, 112 (5.8%) were ineligible, 435 (22.6%) could not be reached, 324 (16.9%) declined participating, and 19 (1.0%) could not participate due to a health condition; 973 (50.6%) participated in our survey. Logistic regressions did not indicate nonresponse bias by age, sex, income, marital status, or period of military service. Of all participants in our survey, 349 (35.9%) were eligible for the analysis: 146 (41.8%) met only the AUDIT-C criterion, 116 (33.2%) met only the AUD diagnosis criterion, and 87 (24.9%) met both criteria. Of this sample, 342 (98.0%) participants were seen in primary care, and the remaining participants were seen in other outpatient settings. Average age was 55 years (SD = 15), with a majority being white (52.7%), not married or living as married (62.8%), having completed at least some college education (60.8%), getting all or most of their medical care through the VA (84.3%), and not using illegal drugs in the last month (89.1%; see Table 1).

Table 1.

Participant Demographics

Variable	Total Sample (n = 349)
Age
Years (sd)	54.9 (15.2)
Married or living as married
n (%)	129 (37.2%)
Highest education completed
Not complete high school (%)	18 (5.2%)
High school or GED (%)	119 (34.1%)
Some College (%)	159 (45.6%)
College grad or beyond (%)	53 (15.2%)
Ethnicity/race
Hispanic (%)	92 (26.4%)
Non-Hispanic White (%)	184 (52.7%)
Other (%)	73 (20.9%)
Insurance coverage and medical care
Insurance coverage outside the VA (%)	133 (38.3%)
All/most medical care through VA (%)	291 (84.3%)
High-risk for alcohol use disorder
AUDIT-C score ≥ 8 only (%)	146 (41.8%)
AUD diagnosis code on screening visit only (%)	116 (33.2%)
Both AUDIT-C ≥ 8 and AUD diagnosis on index date (%)	87 (24.9%)
AUDIT score
Mean (sd)	14.4 (7.3)
Mental health symptoms
Mean PHQ-9 (sd)	5.9 (6.0)
Mean GAD-7 (sd)	4.8 (5.5)
SF-12:
Mean physical health score (sd)	46.5 (11.0)
Mean mental health score (sd)	50.9 (11.9)
Readiness to Change
Precontemplation (%)	40 (11.7%)
Contemplation (%)	182 (53.2%)
Action (%)	120 (35.1%)
Past Month Drug Use
Yes (%)	38 (10.9%)

Overall, 145 (41.5%) patients reported receiving an offer of at least one type of formal treatment for alcohol from a VA healthcare provider in the prior 30 days. Patients were offered one or more of the following: therapy/counseling (n = 121; 34.2%), medication (n = 18; 5.1%), referral to intensive outpatient treatment or a residential program (n = 19; 5.4%), or unspecified medical treatment (n = 20; 5.7%). Only 17 patients (5%) reported both an offer of therapy or counseling and either referral to either a residential program or intensive outpatient treatment program. Only more severe alcohol misuse (i.e., higher full AUDIT scores; OR 1.07, 95% CI 1.03 to 1.11) and younger age (OR 0.97, 95% CI 0.95 to 0.99) were associated with reporting an offer of at least one type of formal treatment (see Table 2).

Table 2.

Multivariate Logistic Regression of Formal Treatment Offer Receipt on Demographic and Clinical Variables

Variables	Any Formal Treatment Offer
	OR (95% CI)
AUDIT score	1.07 (1.03, 1.11)*
Age at screening visit	0.97 (0.95, 0.99)*
Married or living as married	1.61 (0.96, 2.71)
Non-Hispanic White	0.98 (0.60, 1.59)
At least some college	0.72 (0.44, 1.17)
Employment Category
Employed	1 (referent)
Unemployed	1.30 (0.63, 2.67)
Out of labor force	1.08 (0.58, 2.04)
All/most care received at VA	1.63 (0.78, 3.37)
Readiness to Change category
Pre-contemplation	1 (referent)
Contemplation	2.34 (0.93, 5.88)
Action	2.33 (0.89, 6.10)
SF-12 Measures
PCS	0.98 (0.96, 1.00)
MCS	1.00 (0.98, 1.02)
Illegal drug use or prescription misuse	1.66 (0.79, 3.49)

Notes: OR = Odds Ratio.

AUDIT = Alcohol Use Disorders Identification Test, full version.

^*p < 0.05.

We used continuous measures for age and AUDIT score.

Discussion

In our sample of patients at high risk for an AUD, over half (58.5%) did not report being offered formal treatment for alcohol following routine annual screening for alcohol misuse in primary care. This result conforms with research demonstrating low rates of offering patients information about formal treatment following alcohol screening.¹⁹ When offered, therapy/counseling was most prevalent, while offers of medication and referral to intensive outpatient treatment or a residential program were rare. Moreover, while, as expected, patients’ severity of alcohol misuse was significantly associated with patient-reported offers of medical treatment, we also found that older populations—the least likely age group to seek alcohol treatment^28,29—were less likely to report being offered treatment.

In another manuscript, we found that some patients reported receiving advice to drink less or abstain from drinking (n = 262)³⁰ or a referral to a self-help group (n = 70) (not reported). Although such interventions are likely not sufficient for individuals at high risk for an AUD, these findings suggest that, even when providers did not offer formal treatment, most providers were aware that the patient was drinking at unhealthy levels and provided some type of intervention. Our results highlight the importance of training primary care providers in what constitutes appropriate medical treatment for this population, and then in the most effective ways of making a treatment offer.^7,31

Patient-reported measures can be used in conjunction with medical records and administrative data for a more complete assessment of treatment offers.³² A limitation of these measures, however, is retrospective recall bias. It is possible that patient inability to remember recent treatment offers indicates the need for providers to better engage patients about their alcohol use in treatment encounters. A potential limitation regarding the generalizability of this study involves the number of approached patients that could not be reached via phone (22.6%) or declined participating in the phone survey (16.9%). While analyses indicated that nonresponse did not seem related to key demographic variables, future research may wish to involve sampling and survey methods other than by phone. Reimbursement for formal treatment will be important to consider in future research as well, given that reimbursement varies by payer. For instance, reimbursement limitations may be an issue, particularly for patients with unhealthy alcohol use that does not meet formal diagnostic criteria and therefore may not qualify for reimbursement depending on the health insurance provider. Future research could also examine additional predictors of patient-reported treatment offers (e.g., provider- and setting-level variables), the degree to which patient-reported measures differ from other data sources (e.g. medical records),^27,33,34 and the impact of different criteria for a new treatment episode than the criterion used in this study (i.e., no AUD treatment in the previous three months). Such research would inform efforts to provide timely, quality treatment for AUDs that is in accord with patients’ treatment preferences.^16,35