Perspectives in Implementing a Primary Care–Based Intervention to Reduce Alcohol Misuse

Abstract

In 2013, the U.S. Preventive Services Task Force (USPSTF) recommended screening followed by brief behavioral counseling to reduce alcohol misuse. Our study, Options Regarding Consumption of Alcohol (ORCA), was one of the studies included in an evidence review that comprised 23 RCTs. ORCA was designed to test whether a primary care–based intervention would reduce alcohol misuse among patients who screened positive for risky or hazardous drinking. Data collection occurred between 1995 and 1999; data analysis was conducted in 2000–2002. Study design and implementation built from a behavioral counseling research paradigm with four components: (1) population-based screening; (2) centralized delivery of intervention components; (3) involvement of primary care practitioners to motivate and reinforce behavior change; and (4) personalization of intervention components. In this paper, we assess the study features using the Pragmatic–Explanatory Continuum Summary Model domains. As a randomized trial, the study included some explanatory features (e.g., standardized follow-up surveys administered by study personnel); however, several aspects of the study were highly pragmatic. Practicable recruitment and training of providers, embedding population-based screening in pre-visit surveys, and keeping the delivery of the primary care intervention components consistent with the tempo and competing priorities of practice are three key features that contributed to the study's success and relevance to the USPSTF.

Introduction

The selection of alcohol misuse as a prevention priority by the U.S. Preventive Services Task Force (USPSTF) reflects the robust body of evidence showing that alcohol misuse contributes substantially to premature morbidity and mortality.¹ Negative health, social, and economic consequences of alcohol misuse occur across a range of behaviors, from risky or hazardous drinking to severe alcohol use disorder, a chronic condition resulting from loss of control over alcohol use.^2,3 Over the lifespan, heavy drinkers have mortality ratios of two or more in comparison with moderate drinkers and abstainers.⁴ The risk of death, injury, and negative social consequences increases even through occasionally engaging in risky drinking patterns of binge drinking (consuming five or more drinks on a single occasion) or of driving after consuming three or more alcoholic beverages.^5,6

The majority of individuals across sociodemographic groups make at least one healthcare visit per year.⁷ Thus, integrating screening and intervention for risky or hazardous drinking into the healthcare setting enables broad reach for prevention strategies that could disrupt progression to an alcohol use disorder. Discussing drinking patterns in the context of a primary care visit can focus on overall health as a motivation to change drinking patterns, and routine medical care can provide ongoing support and encouragement for behavior change. In addition, the prestige of health professionals can enhance the social influence of their advice.

Following an extensive review of evidence regarding screening and intervention for alcohol use, the USPSTF recommended that “clinicians screen adults aged 18 years or older for alcohol misuse and provide persons engaged in risky or hazardous drinking with brief behavioral counseling interventions to reduce alcohol misuse. (B recommendation).”⁸ The evidence review included 23 RCTs, of which 11 were conducted in the U.S. All but four studies were published prior to 2003.

This paper describes the development and implementation of one of the primary care–based studies that informed the USPSTF recommendation. Options Regarding Consumption of Alcohol (ORCA) was conducted in 23 primary care practices at Group Health Cooperative, a nonprofit, consumer-governed HMO in western Washington in the mid-to-late 1990s. Data collection occurred between 1995 and 1999; data analysis was conducted in 2000–2002. Results from the study have been published previously and the main outcome paper was published in 2003.^9,10 The intervention achieved significant reductions in risky drinking practices that included chronic drinking, binge drinking, and drinking and driving.

Here, we describe the intervention research paradigm that informed the ORCA study and the strategies used to recruit and retain primary care physician practices. Note that the intervention was delivered using formats appropriate to the primary care practice setting in the mid-1990s. Since that time, advances in health information technology and changes in the organization of primary care practice provide opportunities to modify the delivery of the intervention. However, the active ingredients of behavioral counseling interventions (e.g., motivational feedback, self-monitoring, problem solving, social support) remain state of the art.

The primary focus of this paper is on the recruitment and implementation of a primary care–based behavioral counseling intervention. We use the Pragmatic–Explanatory Continuum Indicator Summary (PRECIS) model¹¹ to evaluate the components of the study design, implementation, and analysis on a continuum of pragmatic to explanatory research and conclude with a few key “lessons learned” that could inform future primary care behavioral counseling research studies.

Intervention Research Paradigm

The considerable appeal of the primary care setting for addressing alcohol misuse as part of routine preventive care is tempered by challenges related to lack of time, competing priorities, and a lack of infrastructure and staffing for delivery or referral to behavioral counseling. Trends in healthcare toward organizing for chronic disease management rather than acute care, endorsement of evidence-based guidelines by professional organizations, increasing investment in clinical information systems and electronic medical records, and prioritization of quality of care and patient satisfaction can help address these challenges. The Chronic Care Model, which has been widely adopted as a standard for organization of healthcare delivery, is an overarching framework that informed the development of the ORCA study's protocol.^12,13 The model highlights the importance of clinical information systems (e.g., electronic health records and patient registries), decision support (e.g., evidence-based practice guidelines), delivery system design (e.g., specified roles for all practice personnel), and self-management support (e.g., behavioral counseling interventions) that may be supported by community resources.

At the time this study was being designed and implemented, the research team had been involved in several primary care–based behavioral counseling studies.^14–20 Although the behavioral targets and study protocols varied, all of the studies benefited from the Chronic Care Model as a guiding framework as well as from a foundational commitment to “get practice into research.” We operationalized this commitment by including frontline medical staff as full partners in our research, selecting research questions that would inform innovation and evidence-based guidelines, and using research designs that were compatible with emerging care models. Over time, these goals coalesced into a behavioral counseling research paradigm with four main components:

1. population-based identification of the target population;

2. centralized delivery of intervention components through written materials and trained behavioral interventionists;

3. involvement of the primary care practice team as adjunctive motivators and reinforcers; and

4. personalization of intervention components through modalities including computerized feedback and outreach telephone counseling.

Implementing Options Regarding Consumption of Alcohol

The ORCA study was funded by the National Institute on Alcohol Abuse and Alcoholism with the primary aim of conducting a randomized trial of a primary care–based at-risk drinker intervention compared with usual care. Secondary aims were to describe the alcohol use patterns of patients making routine healthcare visits and to evaluate treatment mediators and moderators. Table 1 summarizes the main components of the ORCA study with reference to the ten dimensions outlined in the PRECIS model.

Table 1.

Summary of ORCA Trial and PRECIS Model Domains

PRECIS domain	Assessment of domain for ORCA trial
Participants
Participant eligibility criteria	The trial enrolled primary care patients with advance appointments who completed a telephone interview and scored 15 or below on the Alcohol Use Disorders Identification Test (AUDIT)1 and had at least one of the following drinking patterns: (1) consuming an average of two or more alcoholic drinks per day in the past month (chronic drinking); (2) two or more episodes of binge drinking (consuming five or more drinks on a single occasion in the past month); (3) one or more episodes of driving after consuming three or more drinks in the past month. Providers reviewed lists of scheduled patients and ruled ineligible those who were known to be pregnant, terminally ill, cognitively impaired or to have an alcohol use disorder (AUD).
	Explanatory/Pragmatic: The study screened all patients in advance of non-urgent appointments and provided the intervention to those with the targeted self-reported risky drinking patterns, rather than all primary care patients who drink alcohol. However patients were not required to make any special office visits for alcohol counseling.
Interventions and expertise
Experimental intervention flexibility	Provider-delivered intervention A flow chart to guide a 1-5 minute motivational discussion was clipped to the front of a patient's chart on the day of their visit. The flow chart followed the “ask, advise, assist” format and included a simple algorithm based on the patient's current intention to change their alcohol consumption to guide the provider's discussion. Providers reconfirmed the patient's self-reported drinking patterns, provided supportive advice about potential risks associated with those drinking patterns, asked the patient if they had thought about changing their drinking habits, and gave a motivational message that acknowledged the patient's current intentions. Pragmatic: Providers were given a flow chart outlining steps to take in a discussion; but the specific discussion was up to the provider.
	Self-management support booklet Providers gave patients a booklet, “Drinking Alcohol: A Guide for Evaluating and Changing Drinking Patterns”.2 The booklet included safe drinking limits on the inside cover and had five sections “Take Stock of Your Drinking,” “Decide to Change Your Drinking Habits,” “Set Limits,” “Stay Within Limits,” and “Keep a Healthy Balance.” Pragmatic: A standard booklet was given to all patients.
	Printed Personalized Feedback Providers handed the patient a sheet with personalized feedback that: (1) provided normative information about the prevalence of the patient's reported drinking patterns and associated risks; (2) highlighted the patient's reported intrinsic motivators for changing drinking patterns and compared them to others who have successfully changed; and (3) highlighted the “cons” of at-risk drinking patterns that they endorsed on the screening survey. Explanatory: Patients’ survey answers and theory-based computerized algorithms determined individualized printed feedback.
	Telephone Counseling A telephone counselor made three outreach calls approximately 1-2 weeks after the patient's clinic appointment, four weeks after the first call and four weeks after the second call to encourage the patient to use the self-management support booklet and reinforced the motivational messages they received in the personalized feedback. Explanatory/Pragmatic: The telephone counselor was provided a manual that included goal-driven protocols for each call that depended on the patient's readiness to change. However the format for the approximately 15 minute calls was open-ended and flexible.
Experimental intervention practitioner expertise	Provider-delivered intervention All providers delivering primary care in each practice group were involved. Providers were trained individually via academic detailing in which a study staff member got on each provider's office visit schedule for a brief education and demonstration session lasting 15 minutes to an hour. Pragmatic: All primary care providers were involved with minimal training.
	Telephone Counseling The telephone counselor was a graduate-level clinical psychology student with prior training in behavioral management of alcohol misuse. The counselor received training over a series of study team meetings and received ongoing supervision. Explanatory: The telephone counselor had a background in behavior change counseling and received additional training and supervision.
Comparison intervention	The intervention was compared to usual care. For patients in the usual care group, their medical charts were not flagged and providers were given no materials to provide to patients. Pragmatic: The comparison intervention was usual practice.
Comparison intervention practitioner expertise	The same primary care providers delivered care to both intervention and usual care patients. Pragmatic: All primary care providers were involved.
Follow-up and outcomes
Follow-up intensity	Two follow-up telephone surveys were conducted at 3 and 12 months post-randomization by interviewers masked to the patient's intervention status. Explanatory: Study personnel conducted the follow-up surveys for both intervention and control patients.
Primary trial outcome	The primary outcomes were self-reported prevalence of two at-risk drinking practices and self-reported weekly alcohol consumption. Explanatory/Pragmatic: Although the primary trial outcomes were the outcomes that the intervention was expected to have direct effect on (explanatory), the outcome status did not require research team adjudication and relied on self-report (pragmatic).
Compliance/adherence
Participant compliance with “prescribed” intervention	This was an intent-to-treat trial and not all participants received the “full dose” of intervention components. Receipt of the intervention was tracked for descriptive purposes. The telephone counselor kept records of each call's duration and content, and patients were asked about whether they recalled receiving telephone counseling and used the self-management support booklet at follow-up. Pragmatic: The telephone counselor made repeated attempts to contact patients but there were no other adherence-boosting measures.
Practitioner adherence to study protocol	Primary care providers Delivery of the intervention to patients was traced through completion of a checkbox on the flow chart attached to each patient's chart. If providers did not deliver the intervention they were asked to note the reason on the flow chart. Study personnel retrieved the flow charts daily. Pragmatic: Provider adherence was measured mostly for descriptive purposes although providers knew the flow charts would be collected and monitored.
	Telephone counselor The telephone counselor completed call summary sheets after every call. Calls were discussed during supervision with study investigators, focusing mostly on patients who were difficult to reach or difficult to engage. Explanatory/Pragmatic: Telephone calls were not monitored or taped, however call summary sheets were reviewed and coaching was given for the purpose of improving the counselor's expertise.
Analysis
Analysis of the primary outcome	An intention-to-treat analysis was conducted including all participants regardless of dose of intervention received and patient or provider compliance with intervention protocols. Pragmatic: All randomized patients were included in the primary analysis. No patients were excluded post randomization.

¹Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro, MG (2001). AUDIT: The Alcohol Use Disorders Identification Test. Guidelines for Use in Primary Care. Second Edition. Geneva: WHO.

²Spivak K, Sanchez-Craig M, Davila R. (1994). Assisting problem drinkers to change on their own: Effect of specific and non-specific advice. Addiction, 89: 1135-1142.

The ORCA study had components that were pragmatic, some that were explanatory, and many that fit squarely at the nexus of each continuum. Components that were particularly pragmatic involved the selection, training, and brief intervention delivery by the primary care providers. Rather than selected providers, all providers at the study clinics received minimal training via educational outreach visits or academic detailing.²¹ This involves the trainer arranging a brief visit to the provider's office. Training included demonstrations and role plays. The intervention was delivered during patient encounters scheduled for other concerns. Providers received a chart cue and conducted a brief flexible discussion based on the familiar “Ask, Advise, Assist” format. Their adherence was not strictly monitored. The comparison intervention was usual care delivered by the same providers. Another pragmatic element was the intent to treat analysis; all randomized patients were included in the trial's primary analysis regardless of whether or how much of the intervention they received.

Characterized as more explanatory were components delivered to participants in the intervention condition by the telephone counselor. The telephone counselor had some prior expertise in behavioral alcohol misuse interventions and received additional training, support, and monitoring from study personnel over the course of the trial. Study personnel were also used to conduct baseline and follow-up telephone surveys. Baseline survey responses were used to screen patients for eligibility and produce printed personalized feedback to intervention patients. The study screened all patients who had advance appointments and enrolled only those with the targeted self-reported risky drinking patterns rather than all primary care patients. More pragmatically, patients were not required to make any special office visits for alcohol counseling. Follow-up surveys conducted outside of regular clinical care at 3 and 12 months assessed the primary trial outcomes.

It is important to note that even some of these explanatory/non-pragmatic aspects of the study protocol could be more pragmatic in practice. The increased use of pre-visit assessments and electronic health records would make screening and flagging of patients and even the creation of a personalized feedback document possible without research infrastructure. In addition, as primary care teams evolve to include allied professionals with behavioral counseling expertise, follow-up phone contacts (currently used more for chronic disease management) could be easily integrated with routine practice.

We identify three key components of the study's success and relevance to the USPSTF recommendations that can inform the design of future studies of primary care–based counseling. First, recruiting and training of providers must be practicable. In this study, all provider contact occurred in the course of their regularly scheduled workday and was personalized based on individual provider's knowledge, attitudes, and confidence. We recruited providers during their regularly scheduled staff meetings. For training, we met providers in their offices at their preferred times including before, during, or after clinic hours and adapted the training for the time allotted by each. Providers expressed strong feelings and varied widely in their beliefs about safe drinking limits and we personalized the training to respond respectfully to their attitudes and beliefs. For example, some providers thought the nationally recommended weekly limits were actually much higher and some thought they were lower. In either case, the trainer nonjudgmentally presented the correct limits for men and women.

A second key aspect of the intervention was embedding population-based screening and cohort identification in pre-visit surveys. Providers must be able to know immediately who to counsel and the specifics of the patients’ at-risk behavior patterns in order to customize brief advice and discussion. We identified potential participants through pre-visit telephone surveys that used validated measures of quantity–frequency²² and Behavioral Risk Factor Surveillance System survey questions²³ to assess binge drinking and drinking and driving. Chart flags and flow charts outlining the patients’ at-risk behaviors (e.g., chronic drinking, binge drinking, or driving after drinking) and discussion guidelines were clipped to the top of patients’ medical records. As mentioned above, prescreening and chart flagging are now greatly facilitated by the widespread use of electronic health records and web- or mobile-based pre-visit screenings.

Finally, intervention delivery must also be efficient, team-based, and in line with the tempo and competing priorities of primary care. In the ORCA study, primary care providers gave brief motivational advice and had a short discussion with their patients about the health risks associated with at-risk drinking practices. They assessed patients’ readiness to change their drinking practices and then customized how they introduced the self-management support materials based on readiness. Providers informed patients that a telephone counselor would be calling to follow up. In the study, the behavioral counselor was a graduate student, but an increasing focus on team-based care and promotion of non-physician staff practicing to the top of their licenses has carved out a role for others to provide much of the ongoing motivation and self-management support to patients.

Conclusions

Building from a commitment to “get practice into research,” the ORCA study provided a rigorous evaluation of a primary care feasible brief intervention to reduce alcohol misuse. Assessing study components against the PRECIS domains illustrates that informative primary care–based studies can include a mix of pragmatic and explanatory features. The four-component paradigm of population-based identification of the target population, centralized intervention delivery, involvement of primary care providers and motivators, and personalization through tailored written feedback and phone counseling can guide future primary care–based behavioral counseling research.