The source of the innovation, the strength and quality of the evidence supporting it, the relative advantages the innovation provides, and its adaptability, trial-ability, complexity, design quality and packaging, and cost.
The CFIR identifies eight characteristics of an intervention that facilitate or hinder implementation health innovations: (1) source; (2) evidence; (3) advantage; (4) adaptability; (5) trial-ability; (6) complexity; (7) design quality; and (8) cost.109
PCPs positively regard strong policy statements about innovations from government health authorities,110, 111 their professional medical societies,71 and easily understood evidence-based practice guides as the most credible sources of clinical information.71, 111 SBI for risky alcohol use has been supported by many of the most highly regarded sources of medical information: the USPSTF,39 the Robert Wood Johnson Foundation's Cutting Back program,100 and federal health agencies including SAMHSA, the Office of National Drug Control Policy, HRSA, the VA, and the U.S. Department of Defense. A widely circulated Clinician's Guide from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) instructs PCPs on alcohol SBI,112 and in July 2014, the HHS Centers for Disease Control (CDC) released an operational guide to help PCPs integrate alcohol SBI into their practices.113 A recent CDC study and related public relations initiative underscored the need for PCPs to better address their patients' unhealthy drinking.114 In 2008, the American Medical Association determined that alcohol SBI is a primary care medical procedure and created Current Procedural Terminology codes that are now used by Medicaid, Medicare, and commercial insurers to pay for SBI. Although federal health authorities support SBI for drug and prescription medication misuse, decisions by the USPSTF not to endorse SBI for drugs due to insufficient controlled research and recent, well-controlled trials115 that showed no clinical effects of drug SBI in primary care may blunt support for generalizing alcohol SBI to include other drugs.
Strength of Evidence
Despite the endorsement of alcohol SBI by credible authorities and dozens of RCTs, many PCPs are unconvinced that alcohol SBI and treatments for substance dependence are effective.91, 116, 117, 118, 119 Many disagree with government definitions of "risky use" that warrant their clinical intervention,12 and many express "therapeutic nihilism,"120 that there is nothing they as primary care clinicians can do to help a person with an addiction. Surveys of PCPs in several studies found widespread disbelief that SBI could be effective with their patients or that heavy drinkers or drug users in their practices could change.117, 119 The inconsistent results from SBI research likely contribute to the skepticism about the strength of the evidence for adopting it. McCambridge and Rollnick121 observe that although researchers generally show that SBI reduces drinking, research does not usually find an impact of SBI on health service use or on other non-drinking, health-compromising behaviors. Little research demonstrates a clinical effect of screening and brief counseling for primary care patients who use illicit drugs, misuse psychoactive prescription drugs, or combine alcohol and other drugs, who have comorbid psychiatric conditions, and who are under 18 years of age or over 65.4, 43, 47, 49, 54, 59, 63, 64 Neither this disbelief in the strength of the evidence or ability to help patients nor disagreement with definitions of risky use was found among key informants or FQHC site staff. Moreover, many were aware of the gaps in the literature on effectiveness with patients who used illicit drugs and prescription medications for non-medical reasons and of the lack of a sufficient body of evidence to support routine use with adolescents.
A PCP's expertise is in managing a wide breadth of problems;17 PCPs are "specialist generalists." Although some SBI studies have shown reductions in risky alcohol use, reductions in alcohol-related injuries, economic benefits for employers and insurers, and fewer arrests and vehicle accidents,4, 6, 8, 39, 43 these advantages are usually not obvious to PCPs. For the most part, SBI exemplifies the prevention paradox:122 Preventive interventions may benefit public health at a population level but produce little clinically observable benefit to most individuals. In a 2013 JAMA editorial, Fineberg noted: "Prevention of disease is often difficult to put into practice. Among the obstacles: the success of prevention is invisible, lacks drama, often requires persistent behavior change, and may be long delayed; statistical lives have little emotional effect, and benefits often do not accrue to the payer; avoidable harm is accepted as normal, preventive advice may be inconsistent, and bias against errors of commission may deter action; prevention is expected to produce a net financial return, whereas treatment is expected only to be worth its cost; and commercial interests as well as personal, religious, or cultural beliefs may conflict with disease prevention."123 Thus, SBI is a "hard sell." Providers' and patients' time, attention, and resources are limited.111, 121, 124, 125, 126, 127
Key informants and FQHC staff echoed that selling SBI to providers and staff can be challenging and a slow process during the early stages of adoption but that as providers see patients' health improving as a result of SBI services, the value of SBI becomes more evident and the need to "sell" SBI dissipates. Some informants and FQHC staff, however, did not indicate that patients' time and attention were common barriers to conducting SBI. Most (but not all) patients seem willing to engage with the provider about substance use issues, although in some communities, the stigma of discussing it was a factor. Patients rarely present with specific substance-related symptoms or substance use service requests.111, 121, 124 FQHC staff indicated that most patients present with a range of health concerns but that rarely is the need specific to substance use and nor do patients request substance use services. FQHC staff recognized the important role they play in making the connection between substance use and health during a primary care visit that is focused on a completely different health concern. Most patients who benefit from SBI exhibit only slight reductions in substance abuse and do not exhibit dramatic improvements in health outcomes or functioning.120 When comparing the relative importance of tobacco and alcohol as health risks that they should attend to in brief primary care visits, PCPs choose tobacco.118 Key informants and health center staff indicated that addressing tobacco may be less difficult because it is a less stigmatized substance than alcohol or drug use, and tobacco cessation methods and products are more accessible than are other substance use treatments. Moreover, requirements to document tobacco screening and intervention make those screenings more likely.
For specific at-risk populations, SBI does produce observable benefits for patients in timeframes that clinicians can appreciate.128 For example, in a study of prenatal SBI in primary care, Kennedy et al.129 reported that an emphasis on assisting pregnant women to have healthy babies fostered buy-in in that staff valued helping women abstain from alcohol and/or drug use while they were pregnant so that their babies would be healthy. Goler128 demonstrated that the perinatal SBI program Early Start improves maternal-infant health outcomes and leads to lower overall costs by an amount that is significantly greater than the costs of the program. In a series of systematic reviews, Turner and McLellan41 link substance use and such common clinical conditions seen in primary care practice as diabetes, sleep disorders, hypertension, depression, bone disease, back pain (opioids), and lung disease (marijuana). But statistical relationships between substance use and medical conditions are difficult to act on in busy fee-for-service primary care practices, although newer models of patient-centered medical homes and accountable care organizations may elevate prevention efforts.39, 71, 127, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155
Which components of SBI are core and which can be modified or omitted have not been settled.116 Screening ranges from a single question, urine or blood toxicology screens, standardized questionnaires, and clinical impressions to structured research diagnostic assessments. Screening can be administered face-to-face (by physicians, nurses, community health workers, peer counselors, and others), self-administered with pen-and-paper, administered online and/or via computer, and using interactive voice response modes. Brief interventions range in length from a single session of less than five minutes to more than 15 minutes and may be conducted over multiple sessions. Brief intervention has immediately followed screening or lagged screening by a week or more. Computer, telephone, and face-to-face brief intervention have been tested. Participating study sites indicated that face-to-face brief interventions immediately following screening were most common; however, some sites indicated the need to build capacity to provide SBIRT services via telephone or other telehealth/telepsychiatry mechanisms, particularly in rural areas.
Health center staff indicated that adaptation and tailoring is happening at the clinic level even when sites have multiple clinics. Clinics are figuring out what works best within their operations and patient flows. This impacts the selection of screening tools (most commonly, a validated tool such as the AUDIT-C, AUDIT, NIAAA, or National Institute on Drug Abuse [NIDA] single-item screens), which provider (nurse, social worker, health educator, MA, or physician) will conduct the SBIRT services, and how many sessions will be offered. In some cases, this meant that more than one provider would deliver different components of SBIRT and conduct hand-offs within the office visit to the next provider (e.g., the MA or nurse, following the screening, hands off to the clinical social worker). Sites had varying perspectives on which staff member was best equipped to provide SBIRT services. As one site noted, doctors believe that BHPs are best equipped to offer brief intervention, whereas BHPs report patients are more likely to engage in brief intervention and follow recommendations when they come from the PCP. Other sites encountered similar discussions among staff surrounding the use of MAs and nursing and social work staff. In some cases, which staff performed each part of the SBIRT service delivery process was determined by what disrupted business operations and workflow and in other cases was influenced by perceptions of which staff would be most effective.
Each FQHC site utilized an EHR and indicated that integrating the substance use screening into the EHR was key to facilitating the implementation of services regardless of the type of provider. All sites agreed that electronic and integrated patient health records facilitated communication between providers and improved continuity of care, allowing for tracking service delivery and the need to follow up or address the issue of substance use at the next visit. The challenges, however, included training staff and transitioning from paper records to electronic record keeping, which were problematic and time-consuming. Overall, sites expressed that adapting to changes in patient record keeping, new workflows, integrated service delivery models, and adopting new practices were initial challenges that improved over time as staff became more familiar and processes were modified based on experiences on the ground.
Fourteen studies found adaptability challenges to implementing SBI. Requiring or expecting that SBI be administered to every patient was perceived by many clinicians as too burdensome in busy primary care practices.116, 117, 156 Key informants and site discussions indicated that when components of SBI are delivered by more than one provider and do not rely solely on the physician, conducting SBI may be viewed as less burdensome and as enhancing services. Some preferred delivering SBI at wellness check-ups, at specific times of the year, or in response to patient symptoms.156 Site visit informants suggested that decisions about how often to screen may be driven by a leadership perspective. In some sites, leadership felt it was critical to screen everyone and at every visit. At one site, however, SBIRT was perceived as a specialty service rather than a universal standard of care--its preventive and early intervention nature was not widely recognized or understood. The extensive intake and follow-up substance use questions required by federal grant programs were particularly seen as inflexible and disruptive of clinical flow.157 Beich et al.158 reported that PCPs resisted implementing SBI out of concern that brief counseling would impede patient flow and add stress on overburdened PCPs but would also be insufficient to deal with patients' complex substance use problems.
Many studies of SBI implementation describe the substantial tailoring of interventions to fit local settings.99, 120, 129, 132, 135, 136, 159, 160, 161 Such adaptations, described both in the literature and by key informants and health center staff, have included shortening alcohol and drug screening and assessment tools, embedding them in other questionnaires, administering them electronically, modifying them for cultural sensitivity, tailoring EHR systems, placing posters and brochures in exam rooms, reassuring patients with information on privacy and confidentiality, establishing and refining clinic workflow and documentation flow, delivering services by telemedicine, devising financial incentives for PCP participation, and establishing regular meetings to review program progress.162 A study that tracked the course of the first cohort of SAMHSA SBIRT grantees157 found that the programs had adapted: (1) from full-length screening questionnaires to shorter screeners; (2) from screening for substance use risk only toward screening for multiple risk factors; (3) from in-house generalist providers toward contracted specialist models; (4) from PCP settings toward high-volume emergency settings; and (5) from external referral for SUD treatment toward PCP management of care. Some of these adaptations were also made by FQHCs in this project. Adaptations including a short screen (often referred to as a "prescreen") such as the three-item AUDIT-C for alcohol use and the two-item PHQ-2 for depression were conducted to determine if use of a longer screen such as the full ten-item AUDIT or nine-item PHQ-9 were warranted. In addition, it was not uncommon to start with tobacco and then phase in alcohol screening followed by screening for multiple risk factors and expanding to other substances and emotional health concerns such as depression.
Trying out an innovation is easier if full commitment is not necessary from the outset. In the extensive review of SBI implementation studies, Williams identified two that specifically addressed trial-ability via pilot work. The WHO employed a four-stage process: development, training, piloting, and implementation. The VA program pilot-tested screening for unhealthy alcohol use,98 a clinical reminder for brief intervention,163 and several methods of measuring brief intervention performance before large-scale implementation.97 Pilot testing gave the VA implementation team opportunities to test the intervention on a small scale, to revise or reverse course if warranted, and to build experience and expertise within the organization, in addition to time to reflect upon and test the intervention without having to fully commit to a specific strategy. Pilot testing also gave implementers in the New Mexico FQHC trial opportunities to adapt to local culture.111 If PCPs must expand their teams to include SBI clinicians, the up-front costs may create trial-ability hurdles. Similarly, key informants and FQHC staff indicated that piloting the SBIRT workflow and modifying it as needed was important to integrating it into routine practice. Flexibility to adapt mid-course facilitated implementation and sustainability.
The more complicated the innovation, the more challenging its implementation. SBI programs may require staff to take on new roles and responsibilities. For example, receptionists may be required to ask patients to complete questionnaires. Nurses or MAs may be tasked to make "warm hand-offs" of patients with positive screens to brief intervention counselors. PCPs may need to shift roles toward reinforcing patient participation in SBI counseling and managing SUD pharmacotherapy.164 Others have implemented online screening, administered screens as part of the "rooming" (that is, social history and vital sign collection) process,110, 134, 165, 166 incorporated screening into brief nurse visits,167 instituted prompts and screens by EHRs,56, 91, 134, 168 had MAs administer validated assessment questionnaires,134 and incorporated SBI into similar services for other behavioral risks and disorders.157
The specific content, timing, and intensity of brief counseling that is necessary to achieve clinical change is not settled. A review of 24 systematic reviews of SBI in primary care that covered 56 RCTs45 suggests that screening alone or screening with as little as five minutes of feedback and advice may be as effective as lengthy or repeated brief interventions. Another review found that brief intervention must include at least two of three elements--feedback, advice, and goal setting--to be effective.169 Other studies find that more intensive motivational interviewing techniques involving engagement, focusing, evoking "change talk", and planning must be provided to achieve positive results.121, 170 A systematic analysis of 47 SBI studies did show that initiating brief intervention immediately after screening produced better outcomes than did delaying it.17 Primary care patients who are not seeking substance abuse treatment rarely return for second or subsequent brief intervention services, reinforcing the value of immediately linking brief interventions and screening.169
SBI, a seemingly simple service, can be quite complex to implement in busy primary care settings. Shifts in workflow, staff responsibilities, and communication patterns to integrate SBI into primary care can be difficult.101, 158, 171, 172, 173 Key informants and health centers indicated that a strong internal champion, open lines of communication to leadership and among staff, and an environment that was willing to adapt and shift course facilitate adoption and sustainability overtime, although this may not happen quickly.
The literature is generally silent about how SBI programs are bundled, presented, or assembled. In their systematic review of the effectiveness of brief alcohol interventions in primary care, Nilsen et al.16 found only one study that had a well-developed communications strategy.12, 174, 175
One design challenge is that substance use screening often relies on pen-and-paper or verbally administered questionnaires that require scoring to assess risk.120 Screening questionnaires are generally not part of usual PCP practice, in which diagnosis relies more on empirical observation, ad hoc verbal questioning, and lab tests.120, 176 Another design challenge involves integrating prompts to administer SBI into EHRs in ways that do not irk PCPs or hinder their other clinical work.16, 133, 168 PCPs may simply turn off, ignore, or dismiss EHR prompts if they are not perceived as useful. Several SAMHSA SBIRT programs have invested significant resources into simple, elegant implementation designs. SBIRT Colorado177 and SBIRT Oregon178 make it very easy for PCPs to adopt SBI in their practices, providing screening tools, talking points, billing guides, information on workflow, scripts for various staff roles, and patient and clinician testimonials, as requested by PCPs.71, 111, 179
FQHC sites indicated that integrating all patient notes and treatments into one shared electronic record improves record access, continuity of care, patient tracking, provider response time, and communication between providers and speciality care, reduces health care costs, improves record confidentiality, and leads to staff efficiency and a more effective referral process. During site visits, staff offered a number of recommendations for enhancing EHRs to facilitate SBIRT implemention including building in clinical decision support tools and developing methods and incorporating tools to measure fidelity and quality and assess patient outcomes. Other suggestions included requiring companies that sell EHRs to FQHCs to include appropriate SBIRT data fields and reports in order to automatically generate quality measure data for internal improvement efforts and for Uniform Data System reporting; flagging systems for clinical quality measures to indicate when patient screens are due; and implementing patient confidentiality measures surrounding substance use. Some expressed that FQHCs should not have to adapt their own EHRs for SBIRT; the process is time-intensive and can take years; it is costly; and it is a difficult process to navigate. Some were aware of EHR vendors who had developed this capacity but did not know how to obtain the add-on modules and thought that more education and resources for FQHCs were needed to purchase existing modules and avoid "re-creating the wheel."
Costs to initiate and sustain SBI are substantial. Substantial staff time and economic costs are incurred in hiring and training new staff, setting up new workflows, training current staff in new responsibilities, exploring billing opportunities, and adapting medical record and billing systems. Opportunity costs are substantial, as well, because SBI must also compete with other demands on providers' and patients' time, attention, and resources.111, 124, 125, 126, 127, 156, 158, 180, 181 Many feel that the workload to implement SBI is not balanced by observable clinical or financial benefits.46, 101, 158 Neushotz and Fitzpatrick172 suggest that SBI must show benefits that outweigh its costs to implement and sustain: benefits to patients in improved health and functioning, to clinicians in improved outcomes for their patients and improved clinical interactions, and to the institutions in improved reputation, efficiency, and resources. SBI models that are dependent upon researchers or external funding for their sustainability or are too complex and costly for continued use in practice will not be sustained even if they are found to be clinically effective.119
Facilitators: Recommendations from Key Informants and FQHCs
Interviews with key informants and health center staff revealed a number of recommendations for facilitating the implementation of SBIRT interventions in health centers. Below is a summary of those recommendations:
Shift SBIRT and other behavioral health interventions from a diagnosis to a problem focus, with primary attention to improving functioning and quality of life.
Use an integrated behavioral health consultant model; that is, integrate a behavioral health generalist with substance use training into primary care teams.
Develop a standard model for SBIRT implementation across states and organizations, providing clarity and the opportunity for states to learn from one another's best practices.
Consider paired visits in which primary care and BHPs work in the same area. Pairing behavioral health assessment of risky behavior during a primary care, pediatric, or perinatal visit highlights the areas of key importance and intervention for the PCP or behavioral health counselor.
Use primary care appropriate language. Avoid referring to substance use and depression screening as "social history"; instead use "health history" or "health information". Avoid utilizing the SBIRT acronym because it creates confusion about the process and the target population. Avoid using stigmatizing language such as "addict" or "alcoholic."
Ease the rigidity of SBIRT paperwork and protocols to enable more flexibility to match the timing and flows of primary care.