[Skip to Content]
[Skip to Content Landing]
Invited Commentary
May 28, 2012

Leveraging Technology for Multiple Risk Factor Interventions Comment on “Multiple Behavior Changes in Diet and Activity”

Author Affiliations

Author Affiliation: Clinical Applications and Prevention Branch, Division of Cardiovascular Sciences, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland.

Arch Intern Med. 2012;172(10):796-798. doi:10.1001/archinternmed.2012.1633

Health risk behavior change research has focused predominantly on a single risk factor, but most of the general population (58%) has 2 or more chronic disease risk factors.1 Intuitively, interventions that target multiple risk factors should improve the prevention of disease better than single risk factor interventions, but systematic reviews of multiple risk factor interventions have produced disappointing results.2 In this issue of the Archives, Spring et al3 provide examples of 2 innovative research directions that have the potential to improve outcomes in multiple risk factor intervention research.

One factor that contributes to the disappointing results of multiple risk factor intervention research is the diverse and heterogeneous array of risk factor outcomes in the literature that do not lend themselves to aggregation for pooled meta-analyses.2 Based on recent recommendations from Prochaska et al4 to address this problem, Spring et al3 calculated a composite score of the 4 health behaviors targeted by the interventions studied: saturated fat intake, fruit and vegetable consumption, physical activity, and sedentary behavior. This “Composite Diet-Activity Improvement Score” weighted the changes in each risk factor equally by normalizing and standardizing a z score change averaged across the 4 targets. The adoption of a composite primary outcome in multiple risk factor interventions is a critical first step toward developing a common outcome metric that can be compared across studies.

Using this composite index, Spring et al3 show that an intervention targeting increased fruit and vegetable consumption and reduced sedentary activity produces more of an effect than does the more traditional diet and activity intervention targets (decreasing saturated fats and increasing physical activity). But what do these findings mean in the context of a composite index? Spring et al note a modest correlation between decreased sedentary time and reduced saturated fat intake in the fruits and vegetables/sedentary activity target condition as evidence of a potential untargeted effect of reduced saturated fat resulting from decreased sedentary time. Alternatively, the improved outcomes from the fruits and vegetables/sedentary behavior intervention could have resulted from these targets being more easily changed by the intervention than the other targets, thus contributing more to the composite index change. Given the choice of earning $175 for engaging in physical activity 60 minutes or more per day and consuming less than 8% of calories from saturated fat per day vs eating 5 fruits and vegetables per day and limiting sedentary activity to 90 minutes or less per day, I suspect that most people would choose the latter because these targets are potentially easier to achieve. Selecting easier targets first is a logical strategy when staggering multiple risk factor interventions, but these differences in behavior change difficulty between risk factors highlight the need to consider differential weighting in any composite index.

One method of differential weighting to consider in future studies is to weight each risk behavior by its disease risk. For example, the INTERHEART study found comparable risk ratios for acute myocardial infarction from fruit and vegetable consumption and regular physical activity, so equal weighting in a composite index may be reasonable, but abnormal lipid profiles conveyed a much greater risk of myocardial infarction, so perhaps saturated fat should be weighted more heavily.5 Although determining appropriate disease risk weights, especially across diseases, will be challenging, estimating the long-term health benefits from these short-term health behavior changes could provide the field with a composite index that reflects the relative contribution of each risk factor to disease prevention.

The second innovative example to improve outcomes in this study is the application of mobile technologies. Spring et al3 use mobile technology for delivery of the intervention and study procedures. The ubiquity of mobile phone use, domestically and globally,6 now allows researchers and physicians to deliver study procedures and interventions in ways previously not possible. Various components of many clinical studies can now be conducted remotely via technology,7 reducing study costs and participant burden while also obtaining more intensive longitudinal data in more representative samples. For this article, Spring et al tested 2 comparisons during the intervention phase and 3 during the follow-up phase, but the wealth of daily longitudinal data available could provide important findings on variability and patterns of change from these multiple risk factor interventions.

Perhaps most promising is the potential of mobile technologies to deliver these multiple risk factor interventions more intensively but at less cost. In their review of multiple risk factor interventions, Ebrahim et al noted that “more intensive interventions might be expected to produce better effects although those used in many of the trials would far exceed what is feasible in routine practice.”2(p14) By leveraging mobile technologies for the self-monitoring and feedback components of the interventions, Spring et al3 limited human resource needs to a few remote coaching sessions. Since these coaching sessions were manualized to ensure treatment fidelity, even the coaching could be off-loaded to web or mobile computer algorithms. Incentive provision also could be automated given the increasing commerce functionality of Internet and mobile technologies.

Although all therapeutic aspects of the human encounter may not lend themselves to computerization, there are numerous advantages to such automated interventions. Treatment fidelity is fully maintained, not only between participants in a study but also from research to practice. Once developed, such interventions are fully scalable, with nearly limitless reach at minimal additional cost. In contrast to Internet-delivered interventions, mobile interventions also can be pushed to the patient throughout the day, in the context of the behavior, and adapted to current status, environmental context, and previous intervention effects.8

Via technology, we will soon be able to deliver fully automated and configurable multiple risk factor interventions that monitor progress continuously and can be delivered throughout the day every day. It remains an empirical question, however, whether these technological advances improve outcomes, reduce costs, or both. Spring et al3 have contributed to the empirical evidence of the value of these technologies, but many more research contributions such as this are needed to establish that technologically delivered multiple risk factor interventions improve outcomes.

Back to top
Article Information

Correspondence: Dr Riley, Division of Cardiovascular Sciences, National Heart, Lung, and Blood Institute, 6701 Rockledge Dr, MSC 7936, Bethesda, MD 20892 (wiriley@mail.nih.gov).

Financial Disclosure: None reported.

Disclaimer: This commentary represents the views of the author and not those of the National Institutes of Health.

References
1.
Fine LJ, Philogene GS, Gramling R, Coups EJ, Sinha S. Prevalence of multiple chronic disease risk factors: 2001 National Health Interview Survey.  Am J Prev Med. 2004;27(2):(suppl)  18-24PubMedArticle
2.
Ebrahim S, Taylor F, Ward K, Beswick A, Burke M, Davey Smith G. Multiple risk factor interventions for primary prevention of coronary heart disease.  Cochrane Database Syst Rev. 2011;(1):CD001561PubMed
3.
Spring B, Schneider K, McFadden HG,  et al.  Multiple behavior changes in diet and activity: a randomized controlled trial using mobile technology.  Arch Intern Med. 2012;172(10):ioi120016789-796
4.
Prochaska JJ, Velicer WF, Nigg CR, Prochaska JO. Methods of quantifying change in multiple risk factor interventions.  Prev Med. 2008;46(3):260-265PubMedArticle
5.
Yusuf S, Hawken S, Ounpuu S,  et al; INTERHEART Study Investigators.  Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study.  Lancet. 2004;364(9438):937-952PubMedArticle
6.
Pew Internet and American Life Project.  Americans and their cell phones. Pew Research Center. http://pewinternet.org/Reports/2011/Cell-Phones.aspx. Published August 15, 2011. Accessed March 12, 2012
7.
Lipset C. Virtual clinical trials to improve patient participation and access. ThinkScienceNow website. http://www.thinksciencenow.com/blog-post/virtual-clinical-trials-to-improve-patient-participation-and-access/. Published June 7, 2011. Accessed March 12, 2012
8.
Riley WT, Rivera DE, Atienza AA, Nilsen W, Allison SM, Mermelstein R. Health behavior models in the age of mobile interventions: are our theories up to the task?  Transl Behav Med. 2011;1(1):53-71PubMedArticle
×