Individualized, tailored measures of physical activity are gaining in popularity, as they are more accurate in estimating the amount of time spent in different intensity levels than a group based measure [19, 20]. Unlike previous studies conducted in laboratory settings with protocols unrealizable in community settings, the method presented herein is (1) relatively simple to conduct, (2) useful for those with limited laboratory techniques training, and (3) suitable for community based research settings as a simple but ecologically valid way to measure the amount of time spent in different physical activity intensities.
Previous studies required the use of expensive, indirect calorimetry equipment, the use of a treadmill, and required participants to spend a large amount of time in a laboratory. The purpose of this study was to develop an ecologically valid approach to measure physical activity levels in community dwelling, women of color and document the amount of physical activity done during a typical week. Such an approach, and the data derived from it, could be used by other investigators involved in community based interventions, to design personalized physical activity programs. The methodology presented here is novel in that the documented amount of increased physical activity is calculated relative to each participant's self-selected and self-identified recreational walking pace. The ecological validity of the measure provides assurances to those implementing community based physical activity intervention programs that each participant is doing a level of physical activity that exceeds their typical physical activity level but is unlikely to be so taxing as to discourage the participants and lead to drop out. Other than the cost associated with accelerometers and weight and height measuring devices, no other equipment was needed for the described method. This method can also be completed in a short amount of time and in a community based setting, placing little burden on research or intervention program participants.
The decision to label the mean plus one SD of the activity counts obtained during the orientation walk as 'increased' physical activity is intended to illustrate how the methodology can be used to personalize the development of individualized physical activity thresholds for ecologically accurate measurement and intervention milestones. Alternatives to using standard deviations to establish intervention goals could include the selection of an absolute number of activity counts or a simple percentage increase beyond a given baseline level. Regardless, a baseline level of physical activity should be tied to a self-selected walking pace that corresponds to a Borg scale exertion level of 'very light'. Given that our participants spent very little time in what we defined as increased physical activity, establishing mean baseline values upon data obtained during more vigorous activities than our orientation walk pace and then establishing physical activity goals much above that baseline are unlikely to be met by the populations included in this study.
Although it would have been interesting to incorporate a broader range of physical activities into the development of our measure, we felt comfortable anchoring our measure to that of a 'recreational' walking pace, since walking is the most commonly reported mode of physical activity in the US and is typically cited by public health recommendations as appropriate for the general population . Despite using our relatively low threshold for increased physical activity tied to individual participant's comfortable walking pace, we identified low levels of physical activity done by our sample. Although our inclusion criteria required that the participants exercise fewer than 90 minutes per week, our data confirmed previously reported data  on the dearth of physical activity done by these segments of the US population. Identifying the mean activity counts associated with the orientation walk plus one standard deviation resulted in a relatively low threshold of physical activity particularly for the AA women. This fact points to the need to develop physical activity programs whose starting points are based upon the individual's typical baseline physical activity level, which are likely to be below the minimum recommended published guidelines. Prescribing physical activity levels well beyond an individual's typical physical activity level presents the possibility of a host of acute negative physical health consequences, loss of motivation, and failure to adopt and maintain physical activity. The negative relationship between the intensity of physical activity and intervention program adherence suggests that it is important to initially identify personalized levels of physical activity that the participants are comfortable performing so as to increase the likelihood of adherence during the intervention and beyond [9, 10].
Although not the focus of this paper, we did identify that, on average, AA women spent significantly more time in increased physical activity compared to HL women. Given that the AA women walked at a slower pace during the orientation walk (i.e. fewer counts per minutes) than the HL women, AA women had, on average, lower thresholds for identification of increased physical activity, based upon our methodology. Thus, AA women would need to do lower levels of physical activity during the data collection period to be identified as involving in increased physical activity. Conversely, HL women, on average, walked at an orientation walk pace that was consistent with 'moderate' exercise meaning they required a greater level of physical activity to be labeled as engaged in increased activity . As noted, they reached this threshold of physical activity during less than one percent of the data collection period. There are a variety of potential reasons for these related findings, including differences between AA and HL women regarding family and socioeconomic status differences, cultural norms associated with physical activity, nutrition and other health habits, personal motivations, and built environment issues. These findings reinforce the contention that ecologically valid measures of physical activity are of critical importance. Regardless of the reasons for the identified differences, it remains important to focus on the fact that neither of our ethnic groups approached the recommended weekly amount of physical activity that would contribute to a healthy lifestyle, let alone weight loss or weight loss maintenance . Although our inclusion criterion of fewer than 90 minutes per week of exercise ensured our sample was 'sedentary', our data indicate than none of the participants even approached the 90 minutes threshold.