Varying the types of goals participants received did not substantially change patterns of walking among sedentary individuals with type 2 diabetes using an automated Internet-mediated walking program. In particular, SG did not result in more bout steps than LG, primarily because most of the successful LG participants chose to reach their total step goals by increasing their bout steps. The bout step increases of almost 2000 steps indicate that participants walked an average of one mile more every day at the end of the program than they were walking at the beginning of the program. At a pace of 3 miles per hour, that would be an additional 20 minutes of moderate intensity activity each day. Because many of the successful participants in our LG group did increase their bouts of moderate intensity walking, presumably they gained the same health benefits that would have resulted from success in a program using structured goals targeting moderate intensity bouts of walking.
It is perhaps not surprising that LG induced increases in bout steps for many participants. With low goals, it may be easy to add in a few extra steps here and there throughout the day, rather than going on a sustained moderate-intensity walk. However, as goals increase, it becomes more and more difficult to achieve success without incorporating sustained moderate-intensity bouts of walking into the day.
Participants who received structured goals were less satisfied with and less adherent to the intervention. This difference was demonstrated in participant responses to closed-end satisfaction questions, in coded themes from answers to open-ended questions, and in the data on adherence to wearing the pedometer. These differences in participant satisfaction are likely to impact negatively long-term adherence to a pedometer-based walking program. While our pilot testing suggested that SG might not be as well received as LG, we were somewhat surprised by the magnitude of the difference in satisfaction between the two groups, particularly considering that many components of the two interventions were identical. Themes identified in response to open-ended questions suggest that the lack of continuous reinforcement for SG participants was part of what caused dissatisfaction.
In another pedometer-based intervention study, Le Masurier (2003) found that about 49% of individuals who reached a 10,000 step target did not meet the minimum criteria of 30 minutes activity in bouts lasting for at least 10 minutes . Because Le Masurier did not report baseline bout activity and because bouts of activity were measured using an accelerometer, the increase in the proportion of steps that occurred during bouts could not be determined. Goals assigned to the Stepping Up to Health participants were based on actual steps taken during the previous week, and were for the most part significantly lower than the 10,000 step goals often used for thinner, fitter, and healthier individuals. In the LG group, most participants who increased their total steps also increased their bout steps even with these relatively low LG.
This pilot study was designed to compare two versions of the Stepping Up to Health intervention, and included no true control group for evaluating the impact of Stepping Up to Health on patients' walking. Although pre-post increases in walking may reflect a Hawthorne effect, the change in step counts was substantial, and supports other evidence that pedometer-based walking programs are effective interventions for increasing physical activity among people with diabetes .
Despite the fact that all participants in this pilot study had type 2 diabetes, and thus were at high risk for adverse cardiovascular events, none of the participants experienced any serious or cardiovascular-related adverse events while participating in this unsupervised, unmonitored home-based walking program. This supports previous research demonstrating that moderate intensity physical activity programs are safe even in high-risk groups [30, 31].
The results of this pilot study must be interpreted in light of its limitations. The most important limitations of this study are the short duration of the intervention and the small number of participants. Increasing walking for only six weeks may produce transient improvements in some risk factors, but unless it is sustained over a significantly longer time, the benefits are not likely to persist. The number of participants was relatively small yet provided a sufficient sample size to detect statistically significant differences. Nonetheless, it would be reassuring to see these results replicated in a larger sample. Recruitment for this trial was limited to those who had computer and Internet access. Because of this, our results may not be generalizable to those without Internet access or to those with lower socioeconomic status. Additionally, those randomized to the structured goals group were slightly more active at baseline than those randomized to the lifestyle goals group. By calculating individually tailored step count goals based on objectively measured walking history, we minimized the effect of this baseline difference between group comparisons of step count increases. Finally, the definition of a bout step – steps taken during a walk lasting for at least 10 minutes at a pace of at least 60 steps per minute – is a relatively low-intensity threshold, approximately 2 miles per hour, and it might be too low of an intensity threshold to induce a cardio-respiratory training effect on the more fit participants in the trial.
Currently available and affordable information technology has made it possible to automate personally tailored physical activity interventions that can be delivered at low marginal cost over the Internet. However, we know little about how to design these interventions to promote sustained adherence and to optimize health outcomes. There are many more unanswered questions in the design of such complex systems. One critical question is how to help those 57% who were non-responders in this study – those who did not significantly increase their total steps to start walking. Larger and longer studies are needed to provide further evidence-based optimization of automated internet-mediated walking programs.