Do electronic health records used by primary care practices support recommended alcohol-related care? Authors: Katharine Bradley, James McCormack, Megan Addis, Leah K Hamilton, Gwen T Lapham, Daniel Jonas, Dawn Bishop, Darla Parsons, Cheryl Budimir, Victoria Sanchez, Jennifer Bannon, Gabriela Villalobos, Alex H Krist, Theresa Walunas, Anya Day Published: 2024 ABSTRACT AbstractObjectiveThe quality of alcohol-related prevention and treatment in US primary care is poor. The purpose of this study was to describe the extent to which Electronic Health Records (EHRs) used by 167 primary care practices across 7 states currently include the necessary prompts, clinical support, and performance reporting essential for improving alcohol-related prevention and treatment in primary care.Materials and MethodsExperts from five regional quality improvement programs identified basic EHR features needed to support evidence-based alcohol-related prevention (ie, screening and brief intervention) and treatment of alcohol use disorders (AUD). Data were collected regarding whether EHRs included these features.ResultsEHRs from 21 vendors were used by the primary care practices. For prevention, 62% of the 167 practices’ EHRs included a validated screening questionnaire, 46% automatically scored the screening instrument, 62% could report the percent screened, and 37% could report the percent screening positive. Only 7% could report the percent offered brief intervention. For alcohol treatment, 49% of practices could report the percent diagnosed with AUD, 58% and 91% allowed documentation of referral and treatment with AUD medication, respectively. Only 3% could report the percent of patients diagnosed with AUD who received treatment.DiscussionMost EHRs observed across 167 primary care practices across 7 US states lacked basic functionality necessary to support evidence-based alcohol-related prevention and AUD treatment. Only 3% and 7% of EHRs, respectively, included the ability to report widely recommended quality measures needed to improve the quality of recommended alcohol-related prevention and treatment in primary care.ConclusionImproving EHR functionality is likely necessary before alcohol-related primary care can be improved.Keywords: alcohol drinking, alcohol use disorders, electronic health records, quality improvement, performance measurement Objective Background and significance Alcohol accounts for extensive disability, and over 140 000 U.S. deaths annually, exceeding deaths due to opioid overdose.1,2 Experts agree that alcohol-related prevention and treatment in primary care are essential to decrease morbidity and mortality due to alcohol,3–6 and US performance measures have been developed for monitoring clinical care and insurance plans (Table 1).18,19,27–29 Nevertheless, less than 1% of US adults receive recommended alcohol-related prevention,31 and less than 10% receive alcohol treatment.32 Table 1.Evidence-based prevention and treatment of unhealthy alcohol use in primary care and associated national performance measures.Unhealthy alcohol use—which includes the full spectrum from drinking at levels that often impact health to severe AUD—is common in primary care, with a prevalence often between 20% and 30%.7 Over the past 25 years, research has defined evidence-based care for both alcohol-related prevention and treatment of AUD. AUD was previously referred to as alcoholism or alcohol abuse or dependence, but research has shown AUD to be a single spectrum with the count of symptoms the optimal measure of severity (eg, 2-3 symptoms, mild; 4-5 moderate; and 6-11 severe AUD based on the Diagnostic and Statistical Manual 5th edition [DSM-5] criteria).8 Evidence-based alcohol-related prevention and treatment are outlined below, along with associated national performance measures for each. Preventive alcohol screening and brief counseling to decrease alcohol use. Primary care alcohol-related prevention, recommended by the US Preventive Services Task Force (USPSTF) consists of screening and behavioral counseling in primary care.3 Prevention is focused on decreasing alcohol use by patients who do not have an AUD. The USPSTF recommends screening with one of two brief validated instruments (a single item alcohol screen or the 3-item AUDIT-C)3,9 optimally administered on paper or electronically (eg, tablet, kiosk, or online EHR portal) to maximize sensitivity,10–12 followed by patient-centered brief counseling for patients who screen positive (hereafter, “brief intervention”). Primary care brief interventions are patient-centered and typically provide information on alcohol-related health risks and recommended drinking limits and encourage patients to decrease alcohol consumption.3,13 Successful implementation14–17 has used iterative quality improvement with performance measurement and plan-do-check-adjust (PDCA) cycles.16,17National Performance Measures: EHR-based performance measures for alcohol screening and brief interventions were initially used in the US Veterans Administration health system.14,15 National Quality Forum (NQF# 2152), the NCQA HEDIS18 and the Centers for Medicare and Medicaid Services (CMS) Merit-based Incentive Payment System (MIPS #431) all now use similar measures.19 These measures require EHRs to be used to identify when patients are screened, when they screen positive for unhealthy alcohol use, and if so, when they are offered brief intervention.18,19 Initiation and engagement in AUD treatment. Experts recommend that patients with unhealthy alcohol use be assessed for AUD with one of several approaches.20–22 Once AUD is diagnosed, diverse alcohol treatments can improve outcomes for patients with moderate-severe AUD. Evidence-based, patient-centered management of AUD in primary care includes offering patients treatment options—including medications for AUD,4 behavioral counseling,23,24 specialty addiction treatment,6 and peer support.25 No medication or behavioral treatment has been shown superior to others,26 so evidence-based treatment focuses on helping patients find a treatment acceptable to them.26National Performance Measures: The NCQA has used HEDIS measures of AUD treatment initiation (eg, within 14 days of a new diagnosis) and engagement (eg, within 34 days of initiation) for several decades, and the Centers for Medicare and Medicaid Services (CMS) now uses similar measures.27,28 The CMS MIPS has an electronic clinical quality measure (eCQM 137v10)29 and EHRs can be certified (MIPS #305) if they can report this measure.30 Evidence-based prevention and treatment of unhealthy alcohol use in primary care and associated national performance measures. Preventive alcohol screening and brief counseling to decrease alcohol use. Primary care alcohol-related prevention, recommended by the US Preventive Services Task Force (USPSTF) consists of screening and behavioral counseling in primary care.3 Prevention is focused on decreasing alcohol use by patients who do not have an AUD. The USPSTF recommends screening with one of two brief validated instruments (a single item alcohol screen or the 3-item AUDIT-C)3,9 optimally administered on paper or electronically (eg, tablet, kiosk, or online EHR portal) to maximize sensitivity,10–12 followed by patient-centered brief counseling for patients who screen positive (hereafter, “brief intervention”). Primary care brief interventions are patient-centered and typically provide information on alcohol-related health risks and recommended drinking limits and encourage patients to decrease alcohol consumption.3,13 Successful implementation14–17 has used iterative quality improvement with performance measurement and plan-do-check-adjust (PDCA) cycles.16,17National Performance Measures: EHR-based performance measures for alcohol screening and brief interventions were initially used in the US Veterans Administration health system.14,15 National Quality Forum (NQF# 2152), the NCQA HEDIS18 and the Centers for Medicare and Medicaid Services (CMS) Merit-based Incentive Payment System (MIPS #431) all now use similar measures.19 These measures require EHRs to be used to identify when patients are screened, when they screen positive for unhealthy alcohol use, and if so, when they are offered brief intervention.18,19 Preventive alcohol screening and brief counseling to decrease alcohol use. Primary care alcohol-related prevention, recommended by the US Preventive Services Task Force (USPSTF) consists of screening and behavioral counseling in primary care.3 Prevention is focused on decreasing alcohol use by patients who do not have an AUD. The USPSTF recommends screening with one of two brief validated instruments (a single item alcohol screen or the 3-item AUDIT-C)3,9 optimally administered on paper or electronically (eg, tablet, kiosk, or online EHR portal) to maximize sensitivity,10–12 followed by patient-centered brief counseling for patients who screen positive (hereafter, “brief intervention”). Primary care brief interventions are patient-centered and typically provide information on alcohol-related health risks and recommended drinking limits and encourage patients to decrease alcohol consumption.3,13 Successful implementation14–17 has used iterative quality improvement with performance measurement and plan-do-check-adjust (PDCA) cycles.16,17National Performance Measures: EHR-based performance measures for alcohol screening and brief interventions were initially used in the US Veterans Administration health system.14,15 National Quality Forum (NQF# 2152), the NCQA HEDIS18 and the Centers for Medicare and Medicaid Services (CMS) Merit-based Incentive Payment System (MIPS #431) all now use similar measures.19 These measures require EHRs to be used to identify when patients are screened, when they screen positive for unhealthy alcohol use, and if so, when they are offered brief intervention.18,19 Initiation and engagement in AUD treatment. Experts recommend that patients with unhealthy alcohol use be assessed for AUD with one of several approaches.20–22 Once AUD is diagnosed, diverse alcohol treatments can improve outcomes for patients with moderate-severe AUD. Evidence-based, patient-centered management of AUD in primary care includes offering patients treatment options—including medications for AUD,4 behavioral counseling,23,24 specialty addiction treatment,6 and peer support.25 No medication or behavioral treatment has been shown superior to others,26 so evidence-based treatment focuses on helping patients find a treatment acceptable to them.26National Performance Measures: The NCQA has used HEDIS measures of AUD treatment initiation (eg, within 14 days of a new diagnosis) and engagement (eg, within 34 days of initiation) for several decades, and the Centers for Medicare and Medicaid Services (CMS) now uses similar measures.27,28 The CMS MIPS has an electronic clinical quality measure (eCQM 137v10)29 and EHRs can be certified (MIPS #305) if they can report this measure.30 Initiation and engagement in AUD treatment. Experts recommend that patients with unhealthy alcohol use be assessed for AUD with one of several approaches.20–22 Once AUD is diagnosed, diverse alcohol treatments can improve outcomes for patients with moderate-severe AUD. Evidence-based, patient-centered management of AUD in primary care includes offering patients treatment options—including medications for AUD,4 behavioral counseling,23,24 specialty addiction treatment,6 and peer support.25 No medication or behavioral treatment has been shown superior to others,26 so evidence-based treatment focuses on helping patients find a treatment acceptable to them.26National Performance Measures: The NCQA has used HEDIS measures of AUD treatment initiation (eg, within 14 days of a new diagnosis) and engagement (eg, within 34 days of initiation) for several decades, and the Centers for Medicare and Medicaid Services (CMS) now uses similar measures.27,28 The CMS MIPS has an electronic clinical quality measure (eCQM 137v10)29 and EHRs can be certified (MIPS #305) if they can report this measure.30 Electronic health records (EHRs) are essential for improving the quality of primary care33 and have been used to strengthen alcohol-related prevention and treatment.11,14–17,34,35 EHR prompts, clinical decision support, and reporting for quality improvement are standard tools used by clinicians for improving management of chronic conditions, and have been used in successful efforts to improve the quality of care for unhealthy alcohol use. However, successful efforts to improve alcohol-related primary care have often occurred in integrated health systems using custom-built EHRs to support alcohol-related care; it is unclear whether EHRs used by other U.S. primary care practices include the features necessary to improve alcohol-related prevention and treatment. The purpose of this study was to describe the extent to which EHRs used by primary care practices currently include the necessary prompts, clinical support, and performance reporting essential for improving alcohol-related prevention and treatment in primary care. Methods Study design and overview This clinical research informatics study, integrating knowledge integration across clinical and research information systems,36 emerged from a large federally funded research program designed to implement evidence-based alcohol-related prevention and treatment in real-world primary care practices in US regions with a high prevalence of unhealthy alcohol use.37 Program teams were experienced in using EHRs for quality improvement but discovered that alcohol-related quality improvement was challenging due to deficiencies in EHRs (Table 2). The Alcohol EHR Study Workgroup was formed to describe gaps in the EHR functionality needed for improving the quality of alcohol-related care based on prior research and existing clinical informatics infrastructure available in primary care. The workgroup, which included experts in primary care practice facilitation, EHRs, quality improvement, and alcohol-related prevention and treatment, conducted work in 3 phases. In Phase 1, the workgroup defined basic EHR features needed to support alcohol-related clinical care and quality improvement in primary care based on prior research. In Phase 2, each of the 5 regional programs collected data on the presence or absence of each feature, defined in Phase 1, in each EHR used by primary care practices they worked with. In Phase 3, the workgroup summarized qualitative barriers related to use of EHRs for alcohol-related quality improvement identified throughout workgroup discussions. Below we summarize methods of each of the three phases. Table 2.Case study.A solo, independent primary care practice in Illinois decided to participate in a quality improvement program for alcohol-related care. The practice was experienced in using EHRs to support quality improvement, having participated in a previous quality improvement program sponsored by the AHRQ for cardiovascular conditions using national quality measures. Prior to the alcohol-related quality improvement program, the practice documented alcohol use under social history, specifying whether the patient drank alcohol (yes/no) without using a validated screening instrument, and the provider noted they often did not document AUD diagnoses in the EHR because of the potential stigma patients might experience when viewing their records via the EHR portal. As a result, performance measures for alcohol screening and AUD management produced by the EHR at the start of the project were zero. By the end of the 6-months of practice facilitation, the practice routinely conducted alcohol screening on paper using a validated tool. However, the provider preferred to use the AUDIT-C38 instead of the single-item alcohol screen supported by their EHR, and it was not possible to document AUDIT-C results in the EHR. Further the practice wanted to use the online EHR portal but the portal could not support alcohol screening. In addition, the practice continued to avoid documenting brief intervention or AUD diagnosis codes due to fears of upsetting patients. As a result of these factors, important alcohol-related care was still often not documented in the EHR. In addition, the built-in EHR reports on performance for alcohol screening only allowed practices to generate annual reports and could not be used to extract data for shorter periods to assess performance and support iterative quality improvement. As a result, no reports could be generated at 3 and 6 months regarding the proportion of patients with a documented AUD diagnosis who were prescribed medications for AUD.The 5 regional quality improvement programs and their participating primary care practices encountered significant barriers to using practices’ EHRs to support alcohol-related quality improvement, even when some basic necessary features were present. This case study demonstrates additional gaps in EHR functionality and how it affected practices’ ability to improve alcohol-related care. Case study. The 5 regional quality improvement programs and their participating primary care practices encountered significant barriers to using practices’ EHRs to support alcohol-related quality improvement, even when some basic necessary features were present. This case study demonstrates additional gaps in EHR functionality and how it affected practices’ ability to improve alcohol-related care. Phase 1. Identification of basic EHR features necessary for alcohol-related quality improvement In Phase 1, the workgroup identified three domains of basic EHR functionality as central to improving alcohol-related prevention and treatment in primary care based on prior successful implementation11,14–17,34,35: (1) point-of care prompts to identify patients due for screening, brief intervention, assessment, or treatment of alcohol use disorder (AUD); (2) clinical support, such as validated instruments for screening or AUD assessment, and documentation of referral, peer support or medication treatment for AUD; and (3) reports on performance to support iterative quality improvement, as well as reporting for national performance measures or incentive programs (Table 3). The lead author proposed basic EHR features in each domain necessary to support the implementation of alcohol-related prevention and treatment based on prior research,11,14–17,39,40 which the workgroup iteratively edited until consensus was achieved. Table 3.Basic EHR features necessary to support alcohol-related care and the proportion of the 21 EHRs evaluated that had each feature.Alcohol-related prevention: alcohol screening and brief intervention% EHRs with feature … Prompts for … Alcohol screening. Is the EHR able to prompt primary care teams to conduct alcohol screening when “due” (eg, annually)?10% Brief intervention. Is the EHR able to prompt brief intervention for patients who have a “positive” screen for unhealthy alcohol use?10%Clinical Support … Validated screens. Does the EHR include questions and data fields to document responses to a validated brief screen (single-item Alcohol Screening Question and AUDIT-C)?3,933% Patient can complete screen in EHR. Does the EHR support e-screening via online portal, smartphone, kiosk or tablet that directly enters patient responses into the EHR?29% Automatic screen scoring. Does the EHR calculate the score based on responses (eg, AUDIT-C questions each 0-4; total 0-12 points)?24% Screen result in note. Are the screening instrument and score, and date included in EHR progress notes (automatically or as option for providers to import into progress notes)?33% Suggested brief intervention actions. Are any elements of brief intervention outlined (eg, express concern, recommended limits, connect alcohol to health, elicit response)?14%Reporting Performance … Alcohol screening. Does the EHR retain discrete data on the date a validated screen is documented?62% Positive screen. Does the EHR retain discrete data on the date a positive screen is documented?29% Brief intervention. Does the EHR retain discrete data on the date a brief intervention is documented?14% Alcohol use disorders (AUD) diagnosis and treatment% EHRs with feature … Prompts for … Assessing risk of AUD. Is the EHR able to prompt for assessing risk of AUD in patients with unhealthy alcohol use, high-risk screens (eg, AUDIT-C 7-12) or prior AUD?19% Initiating AUD treatment. Is the EHR able to prompt providers to initiate AUD treatment when patients have a new AUD diagnosis?10% Treatment engagement. Is the EHR able to prompt providers to encourage continued AUD treatment engagement in patients who have initiated treatment?10%Clinical Support … Questionnaires to assess risk of AUD. Does the EHR include the questions and data fields to document measures of AUD risk or AUD symptoms?20–2238% Automatic scoring of AUD risk. Does the EHR score and interpret measures of risk of AUD (eg, full AUDIT, 2-item AUD screen, Alcohol Symptom Checklist)?20–2238% Referral for AUD treatment. Does the EHR support documentation of referral for AUD treatment?62% Referral for peer support. Does the EHR support documentation of referral or recommendation for peer support?14% AUD medication orders. Does the EHR include documentation of medication orders for AUD treatment?95%Reporting Performance … AUD diagnosis. Does the EHR identify the date of a documented code for diagnosis of active AUD by a clinician?43% AUD treatment initiation. Does the EHR identify patients who initiate treatment (eg, have a visit with AUD code documented within 14 days of the initial AUD diagnosis etc.)?10% AUD treatment engagement. Does the EHR identify patients who engage in treatment (eg, 2 more visits with AUD codes or AUD medications in the 34 days after initiating)?5%Abbreviations: AUD, alcohol use disorder; AUDIT, World Health Organization’s 10-item Alcohol Use Disorders Identification Test; AUDIT-C, Alcohol Use Disorders Identification Test-Consumption screening questionnaire; EHR, electronic health record; BI, brief intervention. Basic EHR features necessary to support alcohol-related care and the proportion of the 21 EHRs evaluated that had each feature. Abbreviations: AUD, alcohol use disorder; AUDIT, World Health Organization’s 10-item Alcohol Use Disorders Identification Test; AUDIT-C, Alcohol Use Disorders Identification Test-Consumption screening questionnaire; EHR, electronic health record; BI, brief intervention. Phase 2. The availability of EHR features needed for alcohol-related care and quality improvement In Phase 2, the workgroup evaluated EHRs used by primary care practices in each region for the presence of EHR features defined in Phase 1 as necessary for alcohol-related quality improvement. Participants A total of 167 practices across seven states—Illinois, Indiana, Michigan, North Carolina, Oregon, Virginia, and Wisconsin—participated in this phase. Some practices that had participated in the programs before the necessary data were collected were not included. Data collection Between August 2022 and June 2023, practice facilitators from each of the five programs collected data on the EHR used by each primary care practice participating in their programs via EHR audits, primary care team self-report and resulting program records on EHR functionality. A template was developed that included all the necessary EHR features identified in Phase 1 (Table 3). For each practice, practice facilitators used the template to collect the following data: EHR vendors, whether the EHR was practice- or externally owned and controlled, and the presence or absence of each of the previously agreed upon EHR features needed for alcohol-related quality improvement (Table 3). More primary care practices provided data for prevention (n = 167) than for AUD treatment (n = 158) because Phase 1 specifications for EHR features for alcohol-related prevention were developed earlier, and nine primary care practices had completed work with the program before data collection for AUD treatment began. Each practice collected data separately because products from the same EHR vendor differed across practices, even when practices participated in the same network (eg, an EHR vendor might have multiple products, practices could have installed different modules and add-ons, and some practices had local software configurations of EHRs). Although some practice facilitators had developed custom EHR features for primary care practices, these features that were “custom-built” for the program were not counted because they reflected the expertise of the external practice facilitator, not the features of the EHR. Analyses All sites transferred their collected data to the lead author’s site, where data were combined for analyses. Initial analyses summarized findings at the level of the EHR vendor to indicate whether any product or version of the EHR vendor (across all versions observed) included a specific necessary feature. In addition, the proportion of all primary care practices whose EHR provided each needed EHR feature for alcohol-related prevention or treatment was summarized graphically. Phase 3. Additional barriers related to use of EHRs for alcohol-related quality improvement in primary care During Phases 1-2, the workgroup identified important additional EHR-related barriers to alcohol-related quality improvement illuminated by the case study (Table 2). These issues are described qualitatively. Results Table 3 shows basic EHR features defined in Phase 1 as necessary to support primary care quality improvement of alcohol-related prevention and treatment. Ten EHR basic functions needed for prevention included: point-of-care prompts for when screening or brief intervention were due; built-in clinical support such as validated screening instruments with automated scoring, and guidance on key elements of brief intervention; and EHR data elements needed to allow reporting on performance including the prevalence of screening, positive screens, and brief intervention (Table 3). Eleven EHR features needed to increase AUD treatment in primary care included: point-of-care prompts to assess AUD risk in patients with high-risk drinking, and prompts to support initiation and engagement in treatment when AUD is diagnosed; EHR clinical support such as validated questionnaires to assess risk of AUD (eg, full AUDIT, 2-item screen for AUD, DSM-5 Alcohol Symptom Checklist),20–22 and documentation of specific treatment offered; and EHR data elements needed to allow reporting on performance, including AUD diagnosis, and treatment initiation and engagement (Table 3). EHRs used by primary care: availability of EHR features needed for alcohol-related care The 167 participating practices used a total of 21 EHR vendors (“EHRs” hereafter). The most widely used EHRs were Epic (89 practices); eClinicalWorks (18 practices), and Athena (15 practices), with many practices (n = 127) using EHRs that were externally owned and controlled (Table S1). Table 3 summarizes the proportion of the 21 EHRs that were ever observed to include each basic feature required for evidence-based alcohol-related care (ie, in any of the 167 primary care practices). While 62% of the EHRs documented discrete data on alcohol screening and 95% documented orders for AUD medications in at least one primary care practice, they were often lacking important functionality. Only 1 of the 10 basic EHR functions needed to improve alcohol-related prevention in primary care, and 2 of 11 needed to improve AUD treatment, were observed in over half of the EHRs (Table 3). Moreover, considerable variability was observed across EHRs, and across versions of the same EHR observed in different practices. EHRs observed in multiple practices and states tended to have more of the basic features needed to support alcohol-related care than those observed in one primary care practice and one state (Tables S2 and S3). Figure 1A shows the percent of the 167 practices that had EHRs with each feature needed to improve alcohol-related prevention. Only 17% and 8% of the primary care practices, respectively, had EHRs that allowed for a prompt for alcohol screening and—if the alcohol screen was “positive”—brief intervention. While 62% of practices’ EHRs included a validated screen and 46% scored the screen, only 13% allowed screening via an online portal and 19% displayed the result of screening in a progress note. Finally, despite an ability to report the prevalence of screening in over half of the practices, only 7% of practices had an EHR with the ability to report the prevalence of brief intervention consistent with performance measures. The proportion of primary care practices with EHRs that included each basic feature needed to support alcohol-related care. (A) EHR features needed for alcohol-related prevention (n = 167). (B) EHR features needed for AUDs diagnosis and treatment (n = 158). Source: Practice facilitators across 5 regional quality improvement programs observed and documented EHR functions available in 21 EHRs used by primary care practices. Figure 1B shows that of the 158 practices whose EHRs were evaluated for features needed to support AUD treatment, few had EHRs with the necessary clinical prompts: 15% had prompts to assess AUD risk, 8% had prompts to initiate AUD treatment among those with a new AUD diagnosis, and 6% had prompts for clinicians to support AUD treatment engagement among patients who had initiated treatment. While 91% of the 158 primary care practices’ EHRs captured data on AUD medications ordered by a primary care provider, and 58% of practices used an EHR that allowed documentation of referral to AUD treatment, only 28% of practices had an EHR that provided a questionnaire to assess risk of AUD and only 6% had EHRs that documented referral for treatment or recommendation of peer support. Finally, despite almost half of practices having EHRs that allowed reports of the prevalence of AUD diagnoses, only 6% and 3% of practices, respectively, had an EHR that supported reporting performance measures for AUD treatment initiation and engagement. Other barriers to use of EHRs for alcohol-related quality improvement in primary care Limited choice of alcohol-related measures Several brief screening questionnaires are recommended by the US Preventive Services Task Force (USPSTF) for primary care practices to choose from for alcohol screening—consisting of 1-3 items, each with strengths and limitations.3 Similarly, there are several alternative validated approaches that have been used or recommended by experts for assessment of the risk of AUD when patients report unhealthy alcohol use.20–22 Some EHRs only included one questionnaire for screening and one for AUD assessment. As shown in the case study (Table 2) that sometimes resulted in lack of documentation of important alcohol-related clinical care provided, if a clinic used another screening questionnaire. Codes used for documenting alcohol-related prevention or treatment Billing or other codes are widely recommended for measuring preventive brief intervention41 and AUD treatment.27–29 However, some providers felt uncomfortable using these codes. Some providers worried the codes could end up on insurance claims or bills, upsetting patients (Table 2); others expressed concern that patients would be billed for alcohol-related counseling they may not have wanted or appreciated. Lack of certification of EHRs for alcohol-related reporting The Centers for Medicare and Medicaid (CMS) Merit-based Incentive Payment System (MIPS) and National Committee for Quality Assurance (NCQA) Healthcare Effectiveness Data and Information Set (HEDIS) have similar alcohol-related measures for both prevention and treatment (Table 1). EHRs often seek certification for MIPS measures. Therefore, to help interpret our study’s findings that most EHRs did not support reporting performance on these measures, we reviewed the public list of certifications maintained by the Office of the National Coordinator of Health Information Technology (ONC), in August 2023.30 At the time of review, there was no evidence on the site that ONC certification was available for the prevention measure (MIPS #341), and only 9 of the 21 EHRs in this study had any active ONC certification for the AUD treatment measure (MIPS #305).30 Need for flexibility in EHR reports on performance Even when EHRs could generate reports for national performance measures for alcohol-related prevention or treatment, they often only allowed reporting for a single aggregate measure for a 1-year timeframe (eg, annual report on brief intervention, or treatment initiation followed by treatment engagement). To conduct iterative quality improvement, primary care practices need to be able to evaluate the individual components of the workflow (eg, screening, brief intervention, AUD diagnosis, treatment initiation, and treatment engagement), and generate reports on the performance of each for variable time periods, typically the time between local quality improvement meetings. Discussion This study assessed 21 EHRs used by 167 primary care practices across 7 states to understand gaps in EHR functionality needed to improve alcohol-related primary care. Results revealed most EHRs lack the necessary functionality to improve alcohol-related prevention and treatment: only 1 of 10 basic EHR functions needed to support evidence-based alcohol-related prevention and only 2 of 11 basic EHR functions needed to support AUD treatment was observed in more than half the 21 EHRs (Table 3). Moreover, although performance measurement is the cornerstone of quality improvement, few of the 21 EHRs evaluated could provide reports for national performance measures for alcohol screening and brief intervention (14%) and AUD treatment (5%). Consistent with these EHR deficits, when evaluating gaps in basic EHR functions across the 167 primary care practices, only 3% of primary care practices had EHRs that could generate reports of national performance measures for preventive screening and brief interventions, and only 7% of the practices had EHRs that could generate reports of national performance measures for initiation and engagement in AUD treatment. In addition, as our case study illustrates (Table 2), even when the required EHR features were available, they often could not be used for iterative quality improvement because they were difficult to use, inflexible, and only generated annual reports of aggregate measures. These gaps in EHR support for alcohol-related primary care are alarming given that alcohol screening and brief intervention is ranked second among all USPSTF recommendations for adult primary care in terms of clinical burden prevented and cost effectiveness42 and the quality of care for AUD has been shown to be poorer than for any other chronic condition,43 with no improvement over the past 20 years.32 EHRs have been a critical component of successful efforts to improve alcohol-related care,11,14–17,34,35 and multiple national performance measures have been developed to improve alcohol-related care, including CMS measures (Table 1).18,19,27–29 One of the earliest objectives of CMS’s “Meaningful Use” in 2011 was to have certified EHR’s provide CMS-approved quality measures and offer clinical decision support.44 EHR clinical decision support and reporting for population health management are, therefore, widely available in primary care for other common chronic conditions45–47 such as hypertension48 and diabetes,49–51 and system capabilities described in this study for alcohol-related care parallel EHR tools available for other conditions.52–54 Moreover, such EHR tools have been shown to improve outcomes including among racial and ethnic minoritized groups.46,55–58 Clearly, large-scale improvements in EHRs are needed to support widely recommended, evidence-based primary care for addressing unhealthy alcohol use in the United States. Achieving these goals will require a focused effort by EHR vendors. Such EHR improvements should provide the same level of support for alcohol-related care as for other prevention and treatment (eg, hypertension and diabetes management)55,59 and be coordinated with improvements necessary for other recommended mental health and substance use screening.60,61 Moreover, the features necessary for evidence-based alcohol-related care should be included as core EHR components for primary care at no additional cost. To achieve this, the Alcohol EHR study workgroup arrived at the following specific recommendations for EHR vendors: EHR support for alcohol-related care must allow flexibility. EHRs must include a choice of the two validated alcohol screening questionnaires recommended by the USPSTF9,38 with automatic scoring, and also measures used to identify AUD.20–22 EHRs must also allow primary care practices to choose the frequency and delivery modality for screening (eg, via online patient portal) and allow practices to set different screening thresholds for prompts aside from those for brief intervention (eg, higher scores might be used for prompting assessment with a checklist for the 11 diagnostic criteria for AUD,22,62 and any alcohol use might be used to prompt follow-up for women who are pregnant). Finally, EHRs could incorporate links to evidence-based recommendations and guidelines to assist in selecting appropriate interventions, treatment, and medications for AUD as the evidence evolves.3,4,63 EHR support for alcohol-related care must allow flexibility. EHRs must include a choice of the two validated alcohol screening questionnaires recommended by the USPSTF9,38 with automatic scoring, and also measures used to identify AUD.20–22 EHRs must also allow primary care practices to choose the frequency and delivery modality for screening (eg, via online patient portal) and allow practices to set different screening thresholds for prompts aside from those for brief intervention (eg, higher scores might be used for prompting assessment with a checklist for the 11 diagnostic criteria for AUD,22,62 and any alcohol use might be used to prompt follow-up for women who are pregnant). Finally, EHRs could incorporate links to evidence-based recommendations and guidelines to assist in selecting appropriate interventions, treatment, and medications for AUD as the evidence evolves.3,4,63 Clinical tools must be easy to access and use. EHR documentation, prompts, and performance reports must be simple and practically feasible in primary care, as highlighted previously for quality improvement in other areas.33,64 Prompts should be automatic as care is delivered (eg, patient completes screen in the portal with mental health screening or on paper and, once documented, a prompt automatically alerts primary care providers if the score indicates follow-up is needed; once follow-up is documented the prompt would be automatically “turned off”). Clinical tools must be easy to access and use. EHR documentation, prompts, and performance reports must be simple and practically feasible in primary care, as highlighted previously for quality improvement in other areas.33,64 Prompts should be automatic as care is delivered (eg, patient completes screen in the portal with mental health screening or on paper and, once documented, a prompt automatically alerts primary care providers if the score indicates follow-up is needed; once follow-up is documented the prompt would be automatically “turned off”). Key elements of alcohol-related care must be captured as discrete data. Documented screening, brief interventions, AUD diagnosis, counseling, referral to treatment or peer support, prescriptions to treat AUDs, and satisfaction of national performance measures, must be captured in the EHR as discrete data elements, with dates. Screening scores must be captured so they can trigger necessary prompts, but also so that they can be used to monitor changes in risk over time.65 Key elements of alcohol-related care must be captured as discrete data. Documented screening, brief interventions, AUD diagnosis, counseling, referral to treatment or peer support, prescriptions to treat AUDs, and satisfaction of national performance measures, must be captured in the EHR as discrete data elements, with dates. Screening scores must be captured so they can trigger necessary prompts, but also so that they can be used to monitor changes in risk over time.65 Obtaining frequent performance reports must be simple and practical. EHRs must provide easily accessible reports on the percent of patients who receive screening and follow-up, and the percent of patients with AUD who receive treatment. While many EHRs can report some of the needed components of performance they do not currently provide the measures needed or the flexibility of time windows to support quality improvement efforts. For iterative quality improvement, primary care practices need to be able to generate reports on the individual steps in alcohol-related workflow (eg, screening, positive screens and brief intervention, AUD diagnosis and AUD treatment initiation and engagement, separately). Reporting needs to be possible for variable time windows (eg, 1-12 weeks) to allow for iterative quality improvement. One-to-four week intervals can be appropriate for measuring alcohol screening (eg, in all adult primary care patients), but longer time periods may be needed for lower prevalence outcomes (eg, AUD diagnosis, and treatment).17 Obtaining frequent performance reports must be simple and practical. EHRs must provide easily accessible reports on the percent of patients who receive screening and follow-up, and the percent of patients with AUD who receive treatment. While many EHRs can report some of the needed components of performance they do not currently provide the measures needed or the flexibility of time windows to support quality improvement efforts. For iterative quality improvement, primary care practices need to be able to generate reports on the individual steps in alcohol-related workflow (eg, screening, positive screens and brief intervention, AUD diagnosis and AUD treatment initiation and engagement, separately). Reporting needs to be possible for variable time windows (eg, 1-12 weeks) to allow for iterative quality improvement. One-to-four week intervals can be appropriate for measuring alcohol screening (eg, in all adult primary care patients), but longer time periods may be needed for lower prevalence outcomes (eg, AUD diagnosis, and treatment).17 At the same time, as is clear from our case study, gaps in alcohol-related care will require more than EHR improvements. Practice facilitation and training of primary care teams that explicitly address stigma will be critical to overcoming negative stereotypes regarding AUD and the historic exclusion of AUD treatment from mainstream medical care. Given stigma, and fears about stigmatized codes appearing on patients bills or in their EHRs, EHR support that allows documentation of brief interventions and assessment of AUD without use of ICD, procedure, or other codes, may also be important to improve clinical care and performance monitoring. For example, one system routinely assesses and documents DSM-5 AUD symptoms in the EHR allowing identification of patients who should be offered shared decision-making about AUD treatment without requiring a diagnosis.66 Such measures of AUD symptoms should be standardized, and interoperability standards developed to completement measures of alcohol use, so that clinical improvement is not hampered by relying on codes. Finally, incentives may be needed for certification of EHRs for alcohol-related performance reports.67 Limitations This study had important limitations. First, the primary care practices that participated in these 5 regional quality improvement programs for alcohol-related care during the COVID-19 pandemic may be different from other primary care practices. Second, some EHR functions could have been missed due to participating practices being unaware of alcohol-related features, using older EHR versions, choosing not to purchase add-on features, or having not “turned on” features that supported alcohol-related care. It is also possible that the features were present in the clinic’s EHR but that program staff did not observe or document these features while delivering practice facilitation to improve alcohol-related care. This study did not capture detailed information on clinic EHRs including product name, version, and local configuration options that would allow comparison across different primary care practices. Finally, as above, EHRs form a critical foundation for quality improvement, but other support for overcoming stigma, developing practical workflows, and team based primary care for AUD as part of integrated mental health care will require added support. Nevertheless, the sheer magnitude of deficits observed across many EHRs suggests the importance of our findings. Conclusion This study of 21 EHRs used in 167 primary care practices across 7 states revealed critical gaps in basic EHR support for alcohol-related care. EHRs lacked most basic functions needed, and even when features were present, they were often not usable or practical for use by busy primary care providers. The combination of these limitations created a critical barrier to alcohol-related quality improvement in most practices. Given the high prevalence of unhealthy alcohol use and AUD, and the significant burden on morbidity and mortality in our country, improved EHR support is critically needed. Clinicians and staff in primary care must have access to tools that can provide needed prompts, incorporate evidence-based clinical decision support, and improve the quality of documentation and performance reporting for alcohol screening, prevention, and treatment. Change will require EHR vendors to add the basic functionality needed for alcohol-related care. Supplementary Material Contributor Information Katharine Bradley, Kaiser Permanente Health Plan of Washington, Kaiser Permanente Washington Health Research Institute, Seattle, WA 98101, United States. James McCormack, Oregon Health & Science University, Oregon Rural Practice-based Research Network, Portland, OR 97239, United States. Megan Addis, Kaiser Permanente Health Plan of Washington, Kaiser Permanente Washington Health Research Institute, Seattle, WA 98101, United States. Leah K Hamilton, Kaiser Permanente Health Plan of Washington, Kaiser Permanente Washington Health Research Institute, Seattle, WA 98101, United States. Gwen T Lapham, Kaiser Permanente Health Plan of Washington, Kaiser Permanente Washington Health Research Institute, Seattle, WA 98101, United States. Daniel Jonas, Division of General internal Medicine, The Ohio State University, Columbus, OH 43221, United States. Dawn Bishop, Altarum, Ann Arbor, MI 48105, United States. Darla Parsons, Altarum, Ann Arbor, MI 48105, United States. Cheryl Budimir, Altarum, Ann Arbor, MI 48105, United States. Victoria Sanchez, Oregon Health & Science University, Oregon Rural Practice-based Research Network, Portland, OR 97239, United States. Jennifer Bannon, Center for Health Information Partnerships, Northwestern University, Evanston, IL 60611, United States. Gabriela Villalobos, Department of Family Medicine and Population Health, Virginia Commonwealth University, Richmond, VA 23219, United States. Alex H Krist, Department of Family Medicine and Population Health, Virginia Commonwealth University, Richmond, VA 23219, United States. Theresa Walunas, Center for Health Information Partnerships, Northwestern University, Evanston, IL 60611, United States. Anya Day, Altarum, Ann Arbor, MI 48105, United States. Author contributions All authors met the ICMJE recommended authorship criteria. Katharine Bradley and Anya Day conceptualized the project. James McCormack, Megan Addis, Leah K. Hamilton, Gwen T. Lapham, Daniel Jonas, Dawn Bishop, Darla Parsons, Cheryl Budimir, Victoria Sanchez, Jennifer Bannon, Gabriela Villalobos, Alex H. Krist, Theresa Walunas, and Anya Day participated in data curation. Megan Addis, Leah K. Hamilton, and Gwen T. Lapham conducted the formal analysis. Katharine Bradley, Megan Addis, Daniel Jonas, Alex H. Krist, Theresa Walunas, and Anya Day acquired funding. Megan Addis administered the project. Katharine Bradley and Megan Addis supervised the project. Katharine Bradley drafted the original manuscript, and all authors reviewed and edited it. Supplementary material Supplementary material is available at JAMIA Open online. Funding This project was supported by Agency for Healthcare Research and Quality [R18HS027076 to K.B., M.A., G.T.L., L.K.H., D.B., D.P., C.B., and A.D.], Agency for Healthcare Research and Quality [R18 HS027078 to D.J.], Agency for Healthcare Research and Quality [R18 HS027080 to J.M. and V.S.], Agency for Healthcare Research and Quality [R18 HS027077 to G.V. and A.H.K.], and Agency for Healthcare Research and Quality [R18HS027088 to J.B. and T.W.]. Conflicts of interest The authors have no competing interests to declare. Data availability The data underlying this article will be shared on reasonable request to the corresponding author. References Associated Data Supplementary Materials Data Availability Statement The data underlying this article will be shared on reasonable request to the corresponding author.