Practical and Relevant Self-Report Measures of Patient Health Behaviors for Primary Care Research

Dr. Marvel has highlighted a key area for further consideration in assessing the recommended behavioral measures in our article. Given that a large and growing proportion of the US population reads English as a second language and many others are at or below the 5th grade reading level, we readily agree that literacy and cultural meaning of any practice -based measure is a critical consideration and challenge to inclusive delivery of care. Our purpose for this project was to expeditiously identify existing measures and so we were reliant on the availability of brief assessments that had already considered low-literacy forms or had been translated into languages other than English. Several instruments such as the adult physical activity assessment gave special consideration to these matters and have used pictorial prompts, as well as offering translations in several languages. Other measures are arguably more problematic in addressing cultural meaning and literacy issues; especially problematic is addressing these needs in children’s self-reported behaviors. In general, published information on reading level was unavailable, in part we believe because there are complexities in making these determinations. We certainly agree that much more information is needed in this area and we encourage our colleagues to take up the challenge of designing methods and measures that are sensitive to cultural meaning and literacy needs of patients.

Competing interests:   None declared

We thank Dr. Fisher for his thoughtful and perceptive comments, which discuss the next steps in the use of brief practical measures of behavior change in primary care. We concur with essentially all the directions and recommendations that he makes, and below respond to and provide slight elaborations on each of his three main points.

1. We concur that a small set of additional measures should be added to round out these measures of behavior change (which was the specific task we were given). We also agree that the PHQ-9 would be one of our top choices in expanding the measures to include mental health issues. We would also suggest inclusion of a brief health related quality of life measure, such as the CDC Health Days measure (www.cdc.gov/hrqol) as another important addition.

Extending this line of reasoning a bit further, it could also be beneficial to have available parallel brief practical measures of clinician and staff behaviors related to delivering evidence-based behavior change interventions. In the same way that HEDIS measures are used in many settings, brief, validated measures of staff actions congruent with, for example, the Chronic Care Model or delivery of the 5 A’s of behavior change endorsed by the USPSTF could be useful performance measures, and the focus of quality improvement interventions.

2. We also concur that the way in which information from these brief behavioral assessments is formatted and reported back to providers and patients is important, so that it informs and enhances patient- clinician communication and shared decision-making. Methods to provide print outs or displays that are simple to interpret and understandable by low literacy (and low numeracy) patients, is an important endeavor to improve linkage of these assessments to prompt behavior change.

3. Finally, we agree that there is great potential for automated administration, scoring, and feedback reports tailored to patient ‘profiles’ of results. Such automation, if offered in a patient-centered collaborative manner, holds great potential to extend what the busy clinical team can do in the office. There is a growing and generally encouraging literature on such ‘tailored behavior change interventions’. In the article cited below, we discuss some of our specific thoughts about how interactive technology can be used to facilitate measurement and intervention to address behavioral risk factors.

We share Dr. Fisher’s enthusiasm regarding the future uptake and opportunity of implementing these health behavior assessments in the reality of primary care practice.

Glasgow RE, Bull SS, Piette JD, Steiner JF. Interactive behavior change technology: A partial solution to the competing demands of primary care. Amer J Prev Med, 2004, 27, S80-87.

Competing interests:   None declared

The article by Glasgow, et al, makes an important contribution for assessing behavior change. There is a clear need for practical self- report measures to briefly assess health behaviors. In the article, the authors outline the criteria used to select items. A question that occurred to me is whether literacy received adequate attention in the selection criteria. The authors list "Cultural and Literacy Issues" as a secondary criterion, citing the need for measures to have been translated into other languages. Consideration of the reading level of items should also be included. Limited literacy skills have a significant affect on patients' ability to accurately complete self-report measures. As I scanned the items listed in the article, I questioned whether all items could be understood by patients with limited literacy, especially the items on risky drinking. The article would be strengthened by providing information about the reading levels of the selected items.

Competing interests:   None declared

I thank Dr. Montori for his comments on our article above and essentially agree with all of them. I would like to elaborate on two points raised. The first is that many clincians do not feel equipped to intervene on these health behaviors, and especially not on mulitple behaviors. This is certainly true, and I would hope that assessment and tracking of health behavior status is only the first step in implementing a practical, primary care based system to also assist patients and their families in improving their health behaviors. Detailed discussion of such approaches is beyond the scope of this discussion, but I refer readers to the article below on the potential of applying the evidence- based 5 As model of behavior change and interactive technology as one way to address this need.

Second, I could not agree more that one of, if not the key issue, in identifying practical and useful measuresfor primary care is sensitivity to change due to intervention. Such responsiveness is a different, and I feel more important, criterion than the more typically emphasized criterion of internal consistency , and I refer readers to the article referenced below by Kristal and colleagues on this topic. My colleagues and I did include senstivity to change as one of our primary criteria, but more research on and more consistent reporting of sensitivity to change is certainly needed.

Glasgow RE, Bull SS, Piette J, Steineer J. Interactive behavior change technology: A partial solution to thhe competing demands of primary care. Amer J Prev Med, 2004, 27, S80-87.

Kristal AR, Beresfoird SA, Lazovich D. Assessing change in diet- intervention research. Amer J Clin Nutrit, 1994, 59, 1855 ff.

Competing interests:   None declared

The paper by Glasgow, et al. makes a number of very important contributions to the primary care literature: it highlights the need to include health behaviors as an important target of care for patients with or without major acute or chronic disease, it emphasizes the need to create measures that fit within the primary care context, it provides criteria for evaluating health behavior measures, and, lastly, it identifies an initial set of patient behaviors and measures to begin the process. Several issues come to mind in this paper about how behaviorally related lifestyle issues are addressed in primary care, both for patients with defined diseases and for patients without defined diseases, how they are included as an essential aspect of a patient’s health, and how they are assessed over time, if at all.

Hopefully, an essential contribution of this paper, as it is digested by the primary care community, will be a continuation of what I like to call the ‘PHQ-9 effect.’ In the late 90’s several publications revealed that depression was infrequently recognized in primary care, despite its prevalence, and that it was poorly treated even when recognized. In addition to the professional concern expressed about this problem and a number of continuing education programs, patient registries and other important and effective mechanisms to increase physician awareness, patient treatment, patient tracking, and patient follow-up, there was the introduction of a simple, easy to use and easy to interpret paper and pencil scale to assess depression in the office that meshed well with the primary care context. It arrived carrying with it only nine items, it had well-defined cut points for diagnosis, and it led to a choice of specific treatments based on patient and physician preference, e.g., medication, psychotherapy, a combination of both. To put it simply, the scale worked and physicians felt comfortable using it. In this sense, the PHQ-9 became both a means to an end and an end in itself. As a means to an end, it improves care by accurately assessing depression, perhaps as a screen, but certainly as a solid first step in addressing this major clinical problem. But in my view, the presence of the PHQ-9 in the primary care office also became an end in itself. The scale made a major contribution simply by making depression and its care visible within primary care. It is like having the blood pressure gauge on the exam room wall. It is hard to ignore, impossible to forget, and difficult to avoid. In this sense, the presence of this easy to use scale itself highlighted depression care and seated it squarely into the general stream of primary care practice. Assessing depression no longer was something one considered only in special cases.

My hope is that this same sequence of events emerges with the introduction of these consolidated measures of health behaviors within primary care. In many ways, health behavior assessment and intervention is to primary care now what depression care was prior to the late 90’s. Although primary care practices need to gear up to address behavioral lifestyle and disease risk issues, like diet, physical activity, etc., on a practical but comprehensive basis, having a useful set of readily available measures that are easily interpretable (e.g., recognizable cut- off scores) and that lead to clear courses of action brings this realm of intervention into the physician’s purview in ways that are sensible and do -able in a busy practice.

It seems to me, however, that we need to take three additional steps before these scales will have widespread utility, again following the PHQ- 9 metaphor. First, we need to add to these scales, refine them and integrate them in ways that broaden their usefulness without burdening patients. This might include adding to the battery the PHQ-9 or other mood scale, and perhaps adding a scale for non health-generated stress (family, work, economic, etc.). These two domains of behavioral health are missing from the current battery, and their inclusion makes the scales more useful for the vast majority of patients. Second, we need to format the scales in ways that make them easy to complete, score and interpret, perhaps using a simple, single page graphic that compares each score to a set of relevant norms or to the patient’s last reported behavior to observe change. Then the scales can be used in a doctor-patient dialogue about behavior and health to identify where changes need to occur and what barriers to change may exist, following traditional behavioral principles. The consolidation of a comprehensive set of measures also opens the door to developing computer-assisted administration and scoring systems that are inexpensive and efficient, and that allow both the patient and the physician to keep the current record and to compare it to a later one to assess change. Third, we need to link specific treatment or action options to specific score configurations. The options are available, but they have not as yet been organized in ways that facilitate their use in the office, nor have primary care physicians been trained well in their application. As in depression care, we have learned that assessment without good treatment options leads to no assessment at all.

In these ways, the Glasgow, et al. paper introduces health behavior assessment and intervention into the real world of primary care. It will be very interesting to watch its up-take over time.

Competing interests:   None declared