(continued)
We developed two interrelated generic analytic frameworks to guide the systematic review of behavioral topics (Figures 1 and 2). These analytic frameworks were derived from those developed for screening topics.2 They separately frame the two main questions to consider when systematically reviewing relevant clinical behavioral intervention research, namely:
More in-depth Key Questions (KQs) for each main question are detailed in the text notes on each analytic framework diagram, and the relevant sections of the diagram are numbered to correspond to these key questions.
Clinical interventions are predicated on a foundation of epidemiological research that adequately substantiates the link between particular behaviors and health outcomes,2 as depicted in Figure 1 (Analytic Framework 1, KQs 1, 2, 5). For instance, there is strong consistent evidence that tobacco use, sedentary lifestyle, and improper diet lead to negative clinical and functional health outcomes,1 and, conversely, that smoking cessation, exercise improvement, and dietary improvement lead to positive clinical and functional health outcomes. However, few behavior change intervention studies actually document long-term health outcomes (KQ 8). Therefore, we usually must rely on linking separate bodies of evidence (represented here by the two interrelated but separate analytic frameworks) to demonstrate whether clinical interventions improve health behaviors and lead to better health outcomes.
The USPSTF may elect to summarize, but not systemically review, the evidence supporting the link between health-behavior change and outcomes (shown here in Analytic Framework 1) when either: (1) the evidence has been reviewed in a previous USPSTF report and addresses all issues of current concern, or (2) a good-quality systematic review conducted by another reputable body is available that meets USPSTF standards for grading evidence and addresses the behaviors and outcomes that the USPSTF is interested in. In such instances, Analytic Framework 1 may be dispensed with altogether and attention focused on the literature addressing interventions to effect the desired behavior change (discussed under Analytic Framework 2).
However, even when an evidence review does not formally undertake the key questions in Analytic Framework 1, the epidemiologic evidence linking health behavior change to health benefits illustrated in this diagram can help define appropriate behavior-change outcome measures for the systematic review of behavioral counseling interventions represented by Analytic Framework 2 (Figure 2). Ideally, behavior-change outcome measures of interest in a particular behavioral review are defined as those related epidemiologically to reductions in morbidity and mortality directly (KQ 6) or through intermediate outcomes (KQs 2 and 5 linked together). For behaviors such as improper diet and insufficient physical activity, intermediate outcomes may include physiological risk factors, such as blood pressure, weight, and cholesterol level, through which reductions in morbidity and /or mortality are mediated. In reality, the preferred outcome measures may not be widely available in the literature, because behavioral outcome definitions often vary widely among studies. Sustained behavior changes potentially affect other outcomes of importance to the patient (changes in other behaviors or quality of life) or to the health care system (utilization or patient satisfaction) (KQ 4), and may also induce adverse effects, such as increased injury rates in those increasing physical activity (KQ 3).
As new epidemiologic evidence becomes available, the behavioral outcomes of interest to reviewers may also shift. For tobacco, illicit drug, and alcohol misuse, abstinence has been the primary treatment goal and the most important behavioral outcome. Recently, increased attention has been paid to the health benefits from reducing smoking,91 increasing safe needle use in intravenous drug users,92 and stressing moderation in alcohol use.93 Thus, future reviews may include interventions addressing such behavioral outcomes.
Once a behavior change has been clearly related epidemiologically to improved health outcomes, the most critical issue for clinicians is knowing whether interventions in the clinical setting help patients change their behavior and, if so, how to deliver them effectively and practically. Analytic Framework 2 (Figure 2) contains the logic and critical questions to systematically evaluate the evidence for recommending specific strategies in clinical care to promote healthy behaviors.
Earlier USPSTF experience suggested the need for studies that develop and validate risk-screening and intervention-assessment tools and that examine the efficacy or effectiveness of interventions based on these assessments.94 Assessment (KQ 3) specifies how best to identify patients in need of behavioral intervention and to measure quickly any key characteristics by which the intervention should be individualized (KQ 1, 2). Assessment itself may have adverse effects, such as anxiety, misdiagnosis, or distraction from appropriate care, which would detract from any overall benefit (KQ 4).
|
Example: Physical Activity Interventions Using physical activity as an example, the majority of adults may be sedentary, but not all who visit the clinician need an exercise intervention, and there is no way to determine the need for activity counseling without a specific assessment. In a recent study, fully half of older adults in community-based medical practices who were willing to receive exercise counseling were already active enough not to need further encouragement.95 Activity assessments include standard questions about the frequency, duration, and intensity of physical activity, as well as medical factors that would dictate the exercise type or regimen to prevent harms or complications. Exercise assessment is often individualized further to address motives, barriers, and supports for increasing activity levels. The efficacy of exercise interventions appears to be enhanced when varied according to factors such as the patient's readiness to change, exercise preferences, or past experiences.96 |
The next arrow or link in Analytic Framework 2 examines whether clinical setting interventions are effective in changing behavior (KQ 5) and specifying for whom (KQ 7). For behavioral counseling interventions, no less than for other primary care treatment regimens, it is critical to know intervention details97 (KQ 6): What were the key elements of the intervention, and to whom were they delivered? How were they delivered—when, where, and by whom? What were the time and intensity of the intervention contact? How often and over what time period was the intervention delivered? What was the total intervention "dosage" in terms of frequency, intensity and duration? What were the extent and the duration of the treatment effect (KQ 8)?
Many successful interventions provide repeated contacts and supports that can be modified to fit the individual path of change undertaken by the patient (KQ 9). The USPSTF also considers other benefits (KQ 10) or potential harms (KQ 11) associated with the behavior change. Evaluation of intervention processes as well as content determines the extent, fidelity, and quality of intervention implementation.47
Finally, the review can consider how characteristics of the health care setting influence the likelihood that appropriate individuals will be identified and will receive behavioral interventions (KQ 12a-c), and how larger sociocultural environmental forces influence individuals' ability to change their behavior (KQ 13).98-101 Since individuals are embedded within social, political, and economic systems that shape their behaviors and constrain their access to resources for change, it is important to incorporate these broader factors into our evaluation of interventions.47
To gain the maximum benefit from interventions in clinical settings, we need to extend our perspective beyond efficacy (i.e., it works in research settings) or even effectiveness (it works in real-world clinical settings) to consider the degree to which tested interventions are feasible for adoption into those real-world clinical settings and sustainable over long periods of time.31,63 These perspectives are critical to realizing the public health benefits of modest clinical interventions.
Evidence-based analyses help define the most effective and efficient interventions for specific risk behaviors. Unfortunately, the state of the evidence for behavioral counseling interventions precludes a simple, consistent approach to conducting and reporting the results of these evidence reviews, particularly across a variety of behaviors.
Lack of detail and inconsistency of terms describing behavioral interventions in published reports seriously hamper rigorous reviews and limit the potential for research replication. Similarly, methodologic approaches to these topics are evolving as we consider whether and how special methodologic considerations apply regarding adequacy of research design or unique threats to internal and external validity when evaluating behavioral counseling interventions.
These issues are important to understand, particularly given the gap between available behavioral research and current standards of high-level evidence developed for other fields of medicine.47 However, under the best of circumstances, it remains to be seen how far we can go in specifying standardized approaches for clinicians to the variety of patients for a variety of behaviors. There may be a limit as to how well we will ultimately be able to define any standardized approach, given the multiplicity of factors (patient, family, community, clinician, and health care setting) influencing behavioral change, and the range of states within each factor. This is an important area for ongoing research.
Thus, the current literature, while much improved over the past, may still be insufficient to unequivocally define for the clinician what does and does not work across all primary care behavioral counseling interventions. However, given the prevalence and health impact of unhealthy behaviors, clinicians may still use the time and resources readily available to them to reinforce the importance of healthy behaviors with their patients. For detailed evidence-based consideration of behavioral counseling interventions for specific behaviors, readers are referred to the USPSTF recommendations (and associated systematic reviews).102
Given the inconsistencies in terms and intervention descriptions in the current behavioral counseling intervention literature, the USPSTF decided to use a unifying construct to describe these interventions more consistently across a range of approaches and behaviors. The USPSTF also recognized the need to contribute to the development of a new conceptual and linguistic synthesis for health behavioral counseling interventions in clinical care. Given that no single empirically validated model captures the broad range of interventions across risk behaviors, the USPSTF chose to adopt the Five A's construct because it was judged to have the highest degree of empirical support for each of its elements and because of its use in the existing literature. We describe and then illustrate the use of this construct in the next section, which also updates the 1996 USPSTF summary of the range of research-supported strategies for clinicians interested in delivering behavioral counseling intervention in clinical care.1
The Four A's construct (ask, advise, assist, arrange) was originally developed by the National Cancer Institute to guide physician intervention in smoking cessation.103 Recently, the Canadian Task Force on Preventive Health Care proposed that clinicians use a Five A's construct (adding an agree step) to organize their general approach to assisting patients with behavioral counseling issues (W. Elford, Canadian Task Force on Preventive Health Care, personal communication, December 2000). The U.S. Public Health Service12 used the A's construct to report on high-quality, controlled clinical trials in tobacco cessation, many conducted in primary care settings to test brief, feasible population-level interventions. The A's construct has also been applied to brief primary care interventions for a variety of other behaviors.70,75,95
To be congruent with the U.S. Public Health Service and Canadian Task Force concepts of the A's construct, we adopted the following terminology to describe minimal contact interventions that are provided by a variety of clinical staff in primary care settings:
The content of each step in the Five A's construct necessarily varies from behavior to behavior, but clinical intervention targeting any behavior change can be described with reference to these five intervention components. While we promote the idea of a unifying construct to describe behavioral counseling interventions across behaviors, we acknowledge that the type and intensity of behavior change strategies needed may vary by the complexity of the change, whether the behavior is being added or deleted, and by factors individual to the patient, as described in Theories and Models of Behavior Change. Our brief description of each "A" of this unifying construct uses selected examples from recent research to detail current options and challenges in providing behavioral counseling interventions in clinical care.
Assess: Because behavioral risks are largely invisible and are rarely the main reason for seeking clinical care, explicit systems for behavioral risk-factor assessment in clinical populations serve two purposes. First, they identify all those in need of some intervention for a given behavior (e.g., sedentary or underactive individuals vs already active).96 Second, they gather data needed to target (group) those needing different interventions and, if warranted, to individualize (tailor) brief interventions for maximum effectiveness or health benefit.104
Depending on the behavior, groups are targeted for intervention by factors such as current practices (e.g., current tobacco users vs recent quitters),12 intention (e.g., intending to breast-feed vs not),105 readiness to change the behavior (e.g., soon vs not),106 and presence of medical/physiological factors defining treatment options (e.g., pregnant vs not). Within target groups, moderating factors such as age,107 gender,108 ethnicity,109 comorbidity, or health literacy110 can help clinicians individualize (tailor) intervention emphasis104 once such tailoring has been proven beneficial. Such assessment for intervention individualization may be delayed to a later point in the A's process12 (see Agree section). Assessment can also identify contraindications to intervention, such as general promotion of physical activity in the presence of recent morbidity96 or the safety and appropriateness of nicotine replacement therapy as a behavioral treatment adjunct.12
Systematic, routine assessment is the foundation for proactive behavioral counseling interventions, particularly to realize their public health potential. For instance, having a system in place to identify and document tobacco-use status triples the odds of clinician intervention.12 Adequate assessment can help the clinician consider patient priorities and medical risks, particularly among those with multiple behavioral risks.111 Little research currently exists in effective methods for prioritizing among competing behavioral risks, but ongoing work by the Behavior Change Consortium, sponsored by the National Institutes of Health, may help address these issues.112
|
Assessment Strategies Ideal assessment strategies for clinical practice settings are feasible, brief, and able to be interpreted or scored easily and accurately, and they enhance intervention appropriateness and effectiveness.113,114 Assessment ranges from a few focused questions added before the clinician visit ("Have you used tobacco products at all in the last seven days? If yes, are you seriously thinking about quitting in the next 6 months? If no, have you used them in the last 6 months?"115) to more comprehensive tools, such as health risk appraisal (HRA). An HRA is a multi-page questionnaire that collects patient information to identify risk factors and is typically used to produce an individuated feedback report to promote health, sustain function, and prevent disease. HRA feedback, alone or in combination with single-session counseling by a clinician, is generally ineffective in producing behavior change,111 but the HRA can be a low-cost, easy method to gather data systematically about a variety of modifiable health behaviors and related factors. Challenges for behavioral assessments include the tension between accuracy and feasibility.116 To be practical, many tools are abbreviated to require as little patient and clinician time as possible; thus, good evaluations consider both accuracy and applicability for any assessment approach. Most behaviors besides tobacco use—such as poor diet, physical inactivity, or risky sex—are complex to assess because clinicians need some details of usual practices, such as the frequency, intensity, and duration of various physical activities96 or "usual" intake of specific food items, both to identify individual candidates for intervention and to measure their progress.116,117 One approach to the demands of a more lengthy assessment is to obtain brief assessment by telephone in advance of the clinic visit.95,117 This has been shown to produce reasonably accurate results, at least for physical activity.118 Assessments rely on self-report and recall of customary behavior, and these can suffer from lapses in individual memory, errors in estimation, and the imprecise mapping of self-reported activities to meaningful, physiologically related measures.116 Overall, when reliable biological or biomechanical markers are available for comparison, self-reported health behaviors and risk factors tend to underestimate the proportion of general-population individuals considered "at risk."113 Accuracy and self-disclosure are enhanced by selecting assessment tools designed to maximize the accuracy of self-report information.113 |
Advise: As discussed above, clinician advice establishes behavioral issues as an important part of health care and enhances the patient's motivation to change. Such advice is most powerful when personalized by specifically linking the behavior change to the patient's health concerns, past experiences, family, or social situation,119 and tempering it with the individual's level of health literacy.120 Clinician advice primarily gives the cue to action, while other health professionals and media provide the details.29,56 In this scenario, the clinician is a uniquely influential catalyst for patient behavior change69 and is best supported by a coordinated system to accomplish and maintain that change.
Feedback from current or previous assessments can help personalize health risks and health benefits as well as enhance motivation for change.59 Well-delivered advice supports the patient's self-determination.121 Using minor qualifications such as, "As your physician, I feel I should tell you," for an advice message, rather than "You should," is a subtle but powerful way to convey respect for, and avoid undermining, patient autonomy.
|
Advice Strategies Effective clinician advice has several important elements. Personalized feedback can be biological (laboratory or physiological test results), normative (compared with results for others of the same age, race, and gender), or ipsative (compared with one's previous scores). How the clinician's advice is delivered matters—a warm, empathetic, and non-judgmental style elicits greater cooperation and less resistance, particularly for patients not currently interested in change.59,119 A respectful, individualized approach first considers patient interest in change before warning about health risks or trying to convince the patient to take action.122 Helpful clinician advice also emphasizes the clinician's confidence in the patient's ability to change the behavior (building self-efficacy), and reassures the patient that there are multiple ways to approach successful change and sources to support the behavior change once it is undertaken.119 Acknowledging a patient's previous success in making changes can also boost the patient's confidence. Even considering all these elements, advice messages can be compactly constructed and short (30 to 60 seconds), particularly when coupled with additional assistance. Some clinicians are reluctant to advise patients because people seeking clinical care are not consciously seeking medical advice about their behavior. However, well-delivered advice is actually associated with improved satisfaction among smokers54 and other patients with behavioral risk factors.30 Experts recommend providing anticipatory advice for preventing risky sexual activity or tobacco, alcohol, and illicit drug use to all members of special populations, such as adolescents, even before risky behaviors are evident.123 |
Agree: Here the patient and clinician "come to common ground"51 on area(s) where behavior change is to be considered or undertaken. When both agree that change is warranted, they then collaborate to define behavior-change goals or methods. The importance of collaborative care and patient agreement in a course of action was not explicit in the original Four A's model, but medical thinking has shifted over recent decades to greater patient participation in many aspects of medical care.124 Increasingly, treatment decisions are based on clinician-patient agreement after considering treatment options, consequences, and patient preferences.125 Shared decisionmaking is specifically recommended by the USPSTF for preventive services that involve conflicting or highly individualized risk-benefit trade-offs.126 Similarly, a collaborative approach that emphasizes patient choice and autonomy is critical in behavioral counseling intervention, where the patient retains ultimate control.
Patient involvement in decisionmaking about behavior change offers important benefits, even when decisions involving competing risks and benefits are not the overriding concern. Patients who are actively involved in health care decisions have a greater sense of personal control,127 an important factor for successful behavior change. Also, patient involvement in decisions promotes choices based on realistic expectations and patient values,128 which are important determinants of patient adherence or compliance.129 Patient-centered approaches in which the patient and clinician mutually agree on specific changes may require less visit time than provider-centered ones.130
|
Agreement Strategies Additional questions will help frame the rest of the intervention. For example, current tobacco intervention guidelines recommend assessing whether the patient is willing to make a quit attempt within the next 30 days.12 If not, subsequent behavior-change assistance will consist of a motivational intervention to bolster confidence and readiness and address environmental and other barriers to change. If the patient is ready to take action, then further behavioral counseling is provided, along with adjunctive medication or medical devices, if appropriate. For many behaviors, a few brief questions such as "How important is it for you to..." or "How confident are you that you can..." easily assess a person's motivation and confidence to change a particular behavior, and quickly identify the most promising avenues for further assistance.121 This type of open-ended exchange can engage even the minimally interested patient in a nonthreatening way that may also increase knowledge, self-confidence, and motivation. Actively engaging a patient's agreement before proceeding with further behavioral counseling can also prevent resistance.121 Agreement considers the multiple treatment or intervention options available to help the patient achieve selected behavior change goals. For instance, patients can select home-based or fitness center options to increase their activity levels, nicotine fading or "cold turkey" approaches to smoking cessation, the use of varied contraceptive methods and/or abstinence to prevent pregnancy, and the choice of a wide variety of approaches to improving diet. Moreover, for each of these changes, patients can often choose between reliance on self-help and more intensive clinic methods, based on preference and perceived need for the more intensive skill training and higher levels of social support that clinic-based and face-to-face counseling provide. For people with multiple behavioral risks, agreement is needed about which behavior change(s) to tackle first. |
Assist: In providing assistance, the primary care clinician or other health care staff offers additional treatment to address barriers to changes, increase the patient's motivation and self-help skills, and/or help the patient secure the needed supports for successful behavior change. Effective primary care interventions seek to teach self-management and engage problem-solving/coping skills, thereby enabling the patient to undertake the next immediate step(s) in the targeted behavior change.70
Those not ready to commit to making a specific behavior change in the near future often benefit from assistance strategies that explore ambivalence and enhance motivation.59 As emphasized earlier, additional assistance through effective behavior-change techniques need not be provided directly by the primary clinician solely within the context of a primary care visit. Clinicians may provide assistance through referral to other health care staff within the clinic or outside in the larger health care system or community. Importantly, such approaches typically involve multiple communication channels and intervention methods, which also improve intervention outcomes.12,19
Additional assistance within or outside the patient visit is likely to produce better outcomes than minimal-contact, advice-only treatment. For example, even though 1-3 minutes of advice and counseling have been found to double smokers' 6-month quit rates, time-intensive interventions and more numerous contacts produce even better effects.12 Increasing the total contact time in an intervention (time per intervention X number of contacts) from the minimal 1 to 3 minutes to more than 30 minutes doubles the long-term quit rates yet again. Similarly, a recent analysis at the U.S. population level estimated the expected ex-smoker yields of increasing the proportion of physicians who provide systematic advice (1-3 minutes) to their smoking patients from 60 percent to 90 percent. That estimate was compared with also providing additional counseling assistance (10 minutes) by the clinician or other staff for the 50 percent of advised smokers interested in quitting.66 The results showed that increasing rates of physician advice alone would yield an additional 63,000 quitters per year. Coupling the higher advice level with brief counseling assistance would increase annual quitters by a factor of 10 (630,000).
|
Assistance Strategies Assistance techniques vary according to the behavior and the individual patient's needs but include practical counseling (problem-solving skills training) to replace the problem behavior with new behaviors and to tackle environmental and physiological barriers to change. Assistance also can include direct support from the health care provider/team, guidance in obtaining social support from friends and family, the provision of self-help materials to support self-change efforts, and the provision or prescription of appropriate medication or medical devices (e.g., pharmacotherapy for tobacco dependence, contraceptives for prevention of unplanned pregnancy, and dietary supplements for certain weight loss regimens). Other effective behavior-change techniques include modeling and behavioral rehearsal, contingency contracting, stimulus control, stress-management training, and the use of self-monitoring and self-reward.131 Involving a variety of staff and using diverse, complementary intervention methods improve the feasibility and the effectiveness of providing further behavior change assistance. Interactive videos can deliver standardized portions of behavioral counseling interventions.24 Telephone counseling and well-developed self-help materials provide additional channels for efficiently delivering effective interventions.23,27 If proven effective, computer-driven interventions will someday offer direct, interactive personalized contact through computer kiosks or the Internet that bypasses use of office staff and resources.26 Within certain health care environments, such as managed care and health maintenance organizations, staff outside the clinical setting undertake written and telephone counseling that can result in feedback to the provider or medical chart.68 For settings with few of these options, the delivery of appropriate behavior change assistance is more feasible if intervention activities are spread across clinical staff (e.g., clinician, nurse, medical assistant, and receptionist).29,132,133 |
Arrange: Arranging followup challenges us to reconceptualize behavioral risk factors as chronic problems that change over time.67 No matter how intensive the initial assistance, some form of routine followup assessment and support through repeat visits, telephone calls, or other contact is generally deemed necessary in behavior change interventions. For one thing, followup contacts provide the opportunity to evaluate and adjust the behavior-change plan. Usually, this is accomplished by briefly repeating the first four A's (assess, advise, agree, assist) to update the behavior-change plan, taking into account the patient's intervening efforts, experience, and current perspective. Followup allows for support of behavior-change maintenance134 and relapse prevention for those who have already made some significant behavior change.93,135 In general, followup is best scheduled within a relatively short time period (e.g., one month), although the timing can be geared to provide support for a specific event (such as calling a few days after a set quit-smoking date). After initial intervention followup, future contacts are often spaced at successively longer intervals to provide needed support and continuity in a gradually reduced manner.
|
Arranging Implementation and Followup Behavioral interventions can involve "stepped-care" approaches, similar to those used for hypertension management, with the need for referral to more intensive treatment or outside resources determined after evaluating response to briefer, less-intensive interventions during followup.48 Simply notifying patients that followup will occur seems to be a powerful motivating factor,136 communicating that the behavior change is important and that followup assistance will be available if needed. Clinical staff can systematically arrange followup assessment and support through repeat clinical visits, telephone calls, or other methods of contact between the patient and the health care system. Completion rates for followup and outside referral are important implementation process measures. Recent advances in health communications can assist both clinicians and patients as they engage in appropriate adjustment of the behavior-change plan. For example, interactive computer programs coupled with the capacity for individually tailored output can track individual progress and adjust health promotion strategies to respond to the individual's preferences, rate of progress, and changing environments.137 The diversity of populations that clinicians serve increases the importance of adjusting behavior change plans to the culture, social circumstances, and economic status of clients; such adjustment of health behavior change plans over time and across changing circumstances is an area where many health professionals need increased preparation and expertise.138 |
Behavioral counseling interventions in clinical settings are an important means of addressing prevalent health-related behaviors, such as lack of physical activity, poor diet, substance (tobacco, alcohol, and illicit drug) use and dependence, and risky sexual behavior that underlie a substantial proportion of preventable morbidity and mortality in the United States. Important advances in the ways primary care interventions have been packaged have resulted from the past 2 decades of research. Most importantly, brief interventions designed to fit into everyday practice have been found to produce clinically meaningful changes in the population for a growing number of behavioral risk factors.
Future progress will depend on further refinement of the science supporting behavioral counseling interventions in clinical care through ongoing behavioral research and further development of standards and methods for the reporting and systematic review of behavioral counseling interventions. These advances will facilitate subsequent recommendation development for behavioral counseling topics. They will also facilitate the identification of common, as well as unique, key elements of behavioral counseling interventions across behaviors and populations and, thus, enhance their practical implementation by real clinicians and real patients in everyday clinical settings.