Over a 4-year project period (2018-2022), 4 multidisciplinary project teams led by researchers experienced in alcohol SBI training and implementation, partnered with a multi-clinic healthcare system within their respective regions to implement alcohol SBI and AEP prevention protocols. Key strategies for these projects, with their associated outputs and outcomes, are outlined in Figure 1.

Throughout implementation, planned and unplanned adaptations were required to respond to dynamic circumstances. Planned adaptations that all projects used included modifying training materials to reflect patient demographics, adding educational content on FASDs for providers, and promoting a universal approach to alcohol SBI that included strategies for preventing AEP (ie, asking about any alcohol use). For example, staff training and patient education materials were adapted to include “case” scenarios for patients who were drinking at lower risk levels, but due to their alcohol use and current contraception practices, were at risk for AEP. Planned systems-level strategies to adapt infrastructure, such as modifying the EHR to include the screening instrument and automated scoring, scripted statements for introducing screening questions to patients, provider alerts for patients that screen positive, fields for documenting alcohol SBI encounters, and codes that support alcohol SBI billing, were accomplished in the health systems where authority to make system-level modifications was centralized. In the health systems with decentralized management structures, modifying the EHR depended on local authority, capacity, and willingness to make the required changes. Strategies for monitoring alcohol SBI quality also varied by health setting, with some healthcare systems establishing performance goals (eg, 80% of eligible patients/visits screened) and then monitoring reach and fidelity using EHR-generated data. Performance data were communicated via standardized dashboards, newsletters, or reports, and used to identify systems glitches, booster training needs, or necessary refinements.

The health systems with decentralized decision-making structures require customization of implementation strategies at the clinic level. For example, in a system like Boston HealthNet, where sites are autonomous, pre-implementation strategies that targeted the network and administrator levels were developed to increase awareness and receptivity to enhancing alcohol SBI protocols. At the clinic level, implementation strategies were informed proactively by conducting environmental scans to identify site-level needs, such as adapting systems and infrastructure to support universal alcohol SBI. Accomplishing such systems-level adaptations was complicated by the variety of EHRs in use by clinics. While most of these systems already contained some universal screening protocols, and all systems used moderate drinking guidelines to define recommended drinking limits,³³ screening practices varied widely. In some cases, screening prompts were added to the EHR for staff and decision support with recommended education, intervention, or referral being added for point-of-care providers.

Evaluations of e-SBI for alcohol use in the perinatal period have shown preliminary but encouraging results.^31,32,34-36 The complexity of integrating e-SBI into visit systems and workflows required the SBI-Tech Michigan team to react with some unplanned adaptations. For example, a planned implementation strategy was to use e-learning to reduce the burden and facilitate consistency of training clinic teams, a method that has repeatedly demonstrated its efficacy in building knowledge and skills, and in reviews has shown better results than face-to-face learning.³⁷ However, providers expressed a lack of interest in e-learning models that required dedicated time to complete. In response, HFH used a multi-pronged training approach with providers—email, grand rounds, and clinic huddles—that did not require separate scheduling, and also offered one-on-one training, consistent with preceptor training models. Also, the initial BPA was designed to be minimally disruptive to providers’ workflow but was not attended to by providers. Following provider feedback, the BPA was adapted to be more “interruptive” in the Epic EHR workflow.

An unanticipated cross-cutting change in implementation circumstances was the COVID-19 pandemic, which disrupted clinical services across all implementation sites due to paused delivery of prevention services for varying periods of time, increased use of telehealth, and staff shortages and turnover. Adaptations made in reaction to this unforeseen change varied by program site. At the reproductive healthcare sites, the continued uncertainty of the pandemic and limited access to sites was addressed by converting live or hybrid training strategies to all-virtual platforms, such as digital modules delivered asynchronously through the health systems’ LMS, synchronous online modules delivered during clinical staff meetings, and facilitated practice/role play sessions delivered via remote videoconferencing.

Since all training ceased for the Boston HealthNet sites during the spring and summer of 2020, B SMART produced a publicly available three-part podcast to continue their education strategy for healthcare teams, providing CME/CNE credits to listeners.³⁸ Once training began again virtually, additional topics of interest and interactive online formats were added.

The pandemic also prompted unplanned adaptations for HFH’s e-SBI. The original design was to have patients complete e-SBI on clinic-owned tablets while in the waiting area. Distancing precautions, along with increased use of telehealth, resulted in fewer in-person appointments and limited the time patients spent in the waiting area. In response, the e-SBI software was adapted for patients to use on personal devices, allowing them to complete the e-SBI from home using a link sent via patient portal message prior to their appointment. Table 2 provides a cross-case summary of planned and unplanned adaptations to strategies.

The 4 project teams identified key questions associated with successful implementation at each SBIRT Program Matrix level, drawing on constructs from the implementation science literature.^23,24 These questions can be used to guide adaptations made to content or implementation methods, whether planned or unplanned, to identify supports and address challenges identified at each level. Table 3 summarizes key questions, challenges, and suggested strategies to pre-empt or address challenges related to alcohol SBI implementation.

For example, all project teams expected to adapt alcohol SBI for patients who may be at risk for AEPs, which required modifying the interpretation of screening tools as well as expanded intervention protocols and training content. Although this planned adaptation provided an opportunity to educate healthcare systems and teams about FASD and their role in prevention, it also required providers to address any alcohol use (ie, change the definition of lower-risk use) for the subset of patients who are pregnant or might be pregnant. Thus, addressing the elevated risk of AEP required additional changes to screener interpretation and inclusion of other relevant information, such as sexual history and contraception practices. These adaptations were challenging from a systems perspective, given the need to combine data from different places within the EHR to alert providers. Such complex programming was not easily executed within the participating health systems’ EHRs and workflows and will continue to require future adaptation.

The ability to proactively adapt alcohol SBI so that it is delivered consistently and universally within a setting was facilitated by health system management structures that authorized system-level changes such as modifying the EHR to embed screening tools and procedures, and documentation fields. While centralized management structures promoted the efficiency of planning and adapting alcohol SBI for uniform dissemination across multiple sites, we observed that decentralized management structures required greater site-level participation, local champions, and customization of approaches to enhance acceptance among providers and staff. Regardless of structure, the involvement of key personnel in logistical planning for implementation, including systems administration and quality improvement staff, facilitates the integration of alcohol SBI into routine practice.

Although all 4 of our cases collected data for evaluation purposes, only 3 of the health systems were able to adapt the EHR to document delivery of alcohol SBI key components during the patient visit. The adapted EHR provided these systems with an integrated method for monitoring the reach and consistency of alcohol SBI delivery that could be used to inform continuous quality improvement.

While receptivity from health system leadership and providers is fundamental to practice change, the willingness or ability of health systems to invest administrator, provider, and staff time to participate in pre-implementation planning and start-up activities, including adapting the EHR and procedures, requires an upfront commitment. Our study found that the number and type of adaptations made to alcohol SBI implementation strategies varied by management structure. Centralized structures using more planned adaptations than unplanned adaptations, dropped relatively few planned strategies, likely due to an emphasis on pre-implementation planning and structural-level changes that promoted consistency of delivery and were maintained when the pandemic struck. Planned adaptations to enhance cultural acceptability or support fidelity of intervention delivery, whether they involve tailoring training or customizing systems, are considered enhancements provided they do not dilute the reach or effectiveness of alcohol SBI.¹⁸ However, unplanned adaptations, sometimes dismissed as improvisation, can still promote the goals of implementing an evidence-based intervention if they relate to the core implementation or pre-implementation strategies. These include responding to unanticipated patient visit protocol changes such as those implemented during the pandemic, which required unplanned adaptations to allow patients to access screening from their personal devices; and responding to unanticipated site-level barriers in adopting alcohol SBI that required pre-implementation FASD education to increase receptivity to implementing or expanding delivery of alcohol SBI.^18-20

The COVID-19 pandemic presented unanticipated challenges across all 4 cases, including competing clinical priorities related to patient care, shutdowns, staff shortages, the shift to telehealth, and changes to the location and length of patient visits. The importance of rapidly reacting and adapting strategies to accommodate the diverse needs of the healthcare systems⁴² provided an opportunity to innovate while still advancing the goals of alcohol SBI implementation and AEP/FASD prevention.

While we used a theoretically informed logic model and implementation frameworks to guide the 4 cases, themes, and lessons learned, our data sources were the observations and reflections of the authors, as opposed to standardized measures. Also, we are unable to compare the relative success of the implementation strategies used across projects using a common metric, such as screening and brief intervention rates, given differences in EHR documentation and access to data across the 4 cases. In addition, since implementation approaches were targeted to different levels of the SBIRT Program Matrix depending on each health system’s receptivity, capacity, priorities, and unique challenges, other indicators, such as variation in the use of different types of strategies, whether planned or unplanned, were more aligned with this paper’s focus. Finally, while the 4 cases illustrate implementation approaches used within different types of healthcare settings and management structures, healthcare delivery is context-specific; thus, the lessons from our 4 cases may not be generalizable to all types of settings.