Douglas Berger

Factors Underlying Quality Problems with Alcohol Screening Prompted by a Clinical Reminder in Primary Care: A Multi-site Qualitative Study

ABSTRACTBACKGROUNDPopulation-based alcohol screening is recommended in primary care, and increasingly incentivized by policies, yet is challenging to implement. The U.S. Veterans Health Administration (VA) achieved high rates of screening using a national performance measure and associated electronic clinical reminder to prompt and facilitate screening and document results. However, the sensitivity of alcohol screening for identifying unhealthy alcohol use is low in VA clinics.OBJECTIVEWe aimed to understand factors that might contribute to low sensitivity of alcohol screening.DESIGNThis was an observational, qualitative study.PARTICIPANTSParticipants included clinical staff responsible for conducting alcohol screening and nine independently managed primary care clinics of a single VA medical center in the Northwestern U.S.APPROACHFour researchers observed clinical staff as they conducted alcohol screening. Observers took handwritten notes, which were transcribed and coded iteratively. Template analysis identified a priori and emergent themes.KEY RESULTSWe observed 72 instances of alcohol screening conducted by 31 participating staff. Observations confirmed known challenges to implementation of care using clinical reminders, including workflow and flexibility limitations. Three themes specific to alcohol screening emerged. First, most observed screening was conducted verbally, guided by the clinical reminder, although some variability in approaches to screening (e.g., paper-based or laminate-based screening) was observed. Second, specific verbal screening practices that might contribute to low sensitivity of clinical screening were identified, including conducting non-verbatim screening and making inferences, assumptions, and/or suggestions to input responses. Third, staff introduced and adapted screening questions to enhance patient comfort.CONCLUSIONSThis qualitative study in nine clinics found that implementation of alcohol screening facilitated by a clinical reminder resulted primarily in verbal screening in which questions were not asked vertbatim and were otherwise adapted. Non-verbal approaches to screening, or patient self-administration, may enhance validity and standardization of screening while simultaneously addressing limitations of the clinical reminder and issues related to perceived discomfort.

Comparison of provider-documented and patient-reported brief intervention for unhealthy alcohol use in VA outpatients

BACKGROUND Performance measures for brief alcohol interventions (BIs) are currently based on provider documentation of BI. However, provider documentation may not be a reliable measure of whether or not patients are offered clinically meaningful BIs. In particular, BI documented with clinical decision support in an electronic medical record (EMR) could appear identical irrespective of the quality of BI provided. We hypothesized that differences in how BI was implemented across health systems could lead to differences in the proportion of documented BI recalled and reported by patients across health systems. METHODS Male outpatients with unhealthy alcohol use identified by confidential satisfaction surveys (2009-2012) were assessed for whether they reported receiving BI in the past year (patient-reported BI) and whether they had BI documented in the EMR during the same period (documented BI). We evaluated and compared the prevalence of documented BI to patient-reported BI across 21 VA networks to determine whether documented BI had a variable association with patient-reported BI across the networks. RESULTS Of 9896 eligible male outpatients with unhealthy alcohol use, 59.0% (95% CI 57.4-60.5%) reported BI (50.4-64.9% across networks) and 37.4% (95% CI 36.0-38.9%) had BI documented in the EMR (28.0-44.2% across networks). Overall, 72.9% (95% CI 70.8-75.5%) of patients with documented BI also reported BI. The association between documented BI and patient-reported BI did not vary across VA networks in adjusted logistic regression models. CONCLUSIONS Performance measures of BI that rely on provider documentation in EMRs appear comparable to patient report for comparing care across VA networks.

Screening for Unhealthy Alcohol Use

Primary medical care should include routine alcohol screening to identify patients with unhealthy alcohol use, whether an alcohol use disorder or risky drinking. Screening should include all adult primary care patients as well as adolescents. Although it is important that screening be done with a validated instrument, there are a number of similarly accurate alcohol screening questionnaires for identifying the entire spectrum of unhealthy drinking, and some of them are very brief. We believe that primary care clinicians should commit two validated questions to memory: one to assess any alcohol use and one single-item alcohol screening question (SASQ). However, in settings with EMRs and systems for universal screening, we favor use of the AUDIT-C because of the information it provides on reported typical drinking and on the severity of unhealthy drinking, as well as the utility of a scaled marker for measuring the severity of unhealthy alcohol use and changes over time. Finally, primary care clinicians and others conducting screening must all be carefully trained about recommended drinking limits and risks associated with drinking above them, the purpose of alcohol screening and the efficacy of brief alcohol interventions, including practicing asking validated screening questions in a comfortable manner.

Barriers to and Facilitators of Alcohol Use Disorder Pharmacotherapy in Primary Care: A Qualitative Study in Five VA Clinics

BackgroundThree medications are FDA-approved and recommended for treating alcohol use disorders (AUD) but they are not offered to most patients with AUD. Primary care (PC) may be an optimal setting in which to offer and prescribe AUD medications, but multiple barriers are likely.Objective This qualitative study used social marketing theory, a behavior change approach that employs business marketing techniques including “segmenting the market,” to describe (1) barriers and facilitators to prescribing AUD medications in PC, and (2) beliefs of PC providers after they were segmented into groups more and less willing to prescribe AUD medications.DesignQualitative, interview-based study.ParticipantsTwenty-four providers from five VA PC clinics.ApproachProviders completed in-person semi-structured interviews, which were recorded, transcribed, and analyzed using social marketing theory and thematic analysis. Providers were divided into two groups based on consensus review.Key ResultsBarriers included lack of knowledge and experience, beliefs that medications cannot replace specialty addiction treatment, and alcohol-related stigma. Facilitators included training, support for prescribing, and behavioral staff to support follow-up. Providers more willing to prescribe viewed prescribing for AUD as part of their role as a PC provider, framed medications as a potentially effective “tool” or “foot in the door” for treating AUD, and believed that providing AUD medications in PC might catalyze change while reducing stigma and addressing other barriers to specialty treatment. Those less willing believed that medications could not effectively treat AUD, and that treating AUD was the role of specialty addiction treatment providers, not PC providers, and would require time and expertise they do not have.ConclusionsWe identified barriers to and facilitators of prescribing AUD medications in PC, which, if addressed and/or capitalized on, may increase provision of AUD medications. Providers more willing to prescribe may be the optimal target of a customized implementation intervention to promote changes in prescribing.

Influence of a targeted performance measure for brief intervention on gender differences in receipt of brief intervention among patients with unhealthy alcohol use in the Veterans Health Administration

AIMS Women are less likely than men to receive brief intervention (BI) for unhealthy alcohol use. In 2007, the U.S. Veterans Health Administration (VA) used a national performance measure to implement BI. Although AUDIT-C scores ≥3 for women and ≥4 for men optimize sensitivity and specificity for identifying unhealthy alcohol use, VAs performance measure required BI only among a targeted subgroup defined by a non-gender-specific score threshold (AUDIT-C ≥5). This may have influenced gender differences in receipt of BI among those optimally eligible for BI. Therefore, we evaluate differences in proportions of women and men offered BI before and after BI implementation. METHODS National secondary chart review data (7/06-6/10) identified all outpatients with unhealthy alcohol use for whom BI would be indicated (AUDIT-C ≥3 women, ≥4 men). Logistic regression, including a time-by-gender interaction, estimated the prevalence and 95% confidence interval (CI) of BI for women and men pre- and post-implementation. FINDINGS Among patients optimally eligible for BI (n=51,272, 8206 women and 43,066 men), the prevalence of BI increased more steeply for men than women after implementation (interaction p-value <0.0001). Pre-implementation rates of BI were 21% (95% CI, 18-24) for women and 26% (95% CI, 24-29) for men, and post-implementation rates were 32% (95% CI, 30-34) for women and 47% (95% CI, 45-49) for men. CONCLUSIONS Healthcare systems implementing BI with performance measures may wish to consider that specifying a single alcohol screening threshold for men and women may increase gender differences in receipt of BI among patients likely to benefit.

Factors underlying quality problems with alcohol screening in routine care

Since 2004 >90% of outpatients in the US Veterans Health Administration (VA) have been screened for unhealthy alcohol use with the AUDIT-C. However, research suggests variability in the quality of screening. To understand factors underlying variable quality, we conducted two qualitative studies: 1) an ethnographic study where we observed clinical staff performing screening, and 2) a key-informant study where we conducted 1:1 interviews with clinical staff. For Study 1, four researchers observed alcohol screening at 9 primary care clinics and took handwritten notes, which were transcribed. For Study 2, snowball sampling was used to recruit key informants (n=29) at 5 additional clinics who completed 20-30 minute semi-structured interviews, which were recorded and transcribed. Both qualitative datasets were analyzed using an a priori coding template. In Study 1, we observed 58 clinical staff caring for 166 patients. Alcohol screening was observed 74 times. Clinical staff appeared uncomfortable conducting verbal alcohol screening, and most screening was not verbatim. Study 2 interviews found that clinical staff and providers believed that addressing unhealthy alcohol use is an important part of care but had not received standard training regarding how or why to conduct alcohol screening. Information on alcohol screening was provided to clinicians via email announcement of the availability of electronic clinical decision support and ad-hoc peer-to-peer demonstration of its use. Participants perceived the screening questions to be sensitive and reported modifying questions to increase patient comfort. Participants were largely focused on identifying patients with the most severe condition—alcohol dependence—for which brief intervention does not have confirmed efficacy. Lack of training and discomfort are barriers contributing to variability in screening quality. Addressing the spectrum of unhealthy alcohol use is not yet viewed as part of a preventive agenda. Additional strategies are likely needed to improve screening quality.

Primary Care Management of Alcohol Misuse

More than one in four American adults consume alcohol in quantities exceeding recommended limits. One in 12 have an alcohol use disorder marked by harmful consequences. Both types of alcohol misuse contribute to acute injury and chronic disease, making alcohol the third largest cause of preventable death in the United States. Alcohol misuse alters the management of common conditions from insomnia to anemia. Primary care providers should screen adult patients to identify the full spectrum of alcohol misuse. A range of effective treatments are available - from brief counselling interventions and mutual help groups to medications and behavioral therapies.

Electronic Medical Record Alert Associated With Reduced Opioid and Benzodiazepine Coprescribing in High-risk Veteran Patients

Background: Over the past decade, overdoses involving opioids and benzodiazepines have risen at alarming rates, making reductions in coprescribing of these medications a priority, particularly among patients who may be susceptible to adverse events due to high-risk conditions. Objectives: This quality improvement project evaluated the effectiveness of a medication alert designed to reduce opioid and benzodiazepine coprescribing among Veterans with known high-risk conditions (substance use, sleep apnea, suicide-risk, age 65 and above) at 1 Veterans Affairs (VA) health care system. Methods: Prescribers were exposed to the point-of-prescribing alert for 12 months. For each high-risk cohort we used interrupted time series design to examine population trends in coprescribing 12 months after alert launch adjusting for coprescribing 12 months before launch, demographics and clinical covariates. Trends at the alert site were compared with those of a similar VA health care system without the alert. Secondary analyses examined population trends in opioid and benzodiazepine prescribing separately. Results: Over 12 months, the alert activated for 1332 patients. Proportions of patients with concurrent prescriptions decreased significantly postalert launch among substance use [adjusted odds ratio (AOR)=0.97; 95% confidence interval (CI)=0.96–0.99; 12-month decrease=25.0%], sleep apnea (AOR=0.97, 95% CI=0.95–0.98, 12-month decrease=38.5%), and suicide-risk (AOR=0.94, 95% CI=0.91–0.98, 12-month decrease=61.5%) cohorts at the alert site. Decreases in coprescribing were significantly different from the comparison site among suicide-risk (AOR=0.92, 95% CI=0.86–0.97) and sleep apnea (AOR=0.98, 95% CI=0.96–1.00) cohorts. Significant decreases in benzodiazepine prescribing trends were observed at the alert site only. Conclusions: Medication alerts hold promise as a means of reducing opioid and benzodiazepine coprescribing among certain high-risk groups.

Increased Rates of Documented Alcohol Counseling in Primary Care: More Counseling or Just More Documentation?

BackgroundClinical performance measures often require documentation of patient counseling by healthcare providers. Little is known about whether such measures encourage delivery of counseling or merely its documentation.ObjectiveTo assess changes in provider documentation of alcohol counseling and patient report of receiving alcohol counseling in the Veterans Administration (VA) from 2009 to 2012.DesignRetrospective time-series analysis.ParticipantsA total of 5413 men who screened positive for unhealthy alcohol use at an outpatient visit and responded to a confidential mailed survey regarding alcohol counseling from a VA provider in the prior year.Main MeasuresRates of provider documentation of alcohol counseling in the electronic health record and patient report of such counseling on the survey were assessed over 4 fiscal years. Annual rates were calculated overall and with patients categorized into four mutually exclusive groups based on their own reports of alcohol counseling (yes/no) and whether alcohol counseling was documented by a provider (yes/no).Key ResultsProvider documentation of alcohol counseling increased 23.6% (95% CI: 17.0, 30.2), from 59.4% to 83.0%, while patient report of alcohol counseling showed no significant change (4.0%, 95% CI: −2.3, 10.3), increasing from 66.1% to 70.1%. An 18.7% (95% CI: 11.7, 25.7) increase in the proportion of patients who reported counseling that was documented by a provider largely reflected a 14.7% decline (95% CI: 8.5, 20.8) in the proportion of patients who reported alcohol counseling that was not documented by a provider. The proportion of patients who did not report counseling but whose providers documented it did not show a significant change (4.9%, 95%CI: 0.0, 9.9).ConclusionsIf patient report is accurate, increased rates of documented alcohol counseling in the VA from 2009 to 2012 predominantly reflected improved documentation of previously undocumented counseling rather than delivery of additional counseling or increased documentation of counseling that did not meaningfully occur.

Comparison of DSM-IV and DSM-5 criteria for alcohol use disorders in VA primary care patients with frequent heavy drinking enrolled in a trial

AbstractBackgroundCriteria for alcohol use disorders (AUD) in the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) were intended to result in a similar prevalence of AUD as DSM-IV. We evaluated the prevalence of AUD using DSM-5 and DSM-IV criteria, and compared characteristics of patients who met criteria for: neither DSM-5 nor DSM-IV AUD, DSM-5 alone, DSM-IV alone, or both, among Veterans Administration (VA) outpatients in the Considering Healthier drinking Options In primary CarE (CHOICE) trial.MethodsVA primary care patients who reported frequent heavy drinking and enrolled in the CHOICE trial were interviewed at baseline using the DSM-IV Mini International Neuropsychiatric Interview for AUD, as well as questions about socio-demographics, mental health, alcohol craving, and substance use. We compared characteristics across 4 mutually exclusive groups based on DSM-5 and DSM-IV criteria.ResultsOf 304 participants, 13.8% met criteria for neither DSM-5 nor DSM-IV AUD; 12.8% met criteria for DSM-5 alone, and 73.0% met criteria for both DSM-IV and DSM-5. Only 1 patient (0.3%) met criteria for DSM-IV AUD alone. Patients meeting both DSM-5 and DSM-IV criteria had more negative drinking consequences, mental health symptoms and self-reported readiness to change compared with those meeting DSM-5 criteria alone or neither DSM-5 nor DSM-IV criteria.ConclusionsIn this sample of primary care patients with frequent heavy drinking, DSM-5 identified 13% more patients with AUD than DSM-IV. This group had a lower mental health symptom burden and less self-reported readiness to change compared to those meeting criteria for both DSM-IV and DSM-5 AUD. Trial Registration ClinicalTrials.gov NCT01400581. 2011 February 17