Study Details
The purpose of On the Move, funded by the National Cancer Institute, was to understand the sources of measurement error in self-reported measures of physical activity (PA) and to improve the accuracy of physical activity measurement in epidemiological studies. The specific aims included:
Determination of the reliability and validity of two self-reported physical activity questionnaires used in on-going cohort studies; Evaluation of whether reliability and validity of the two PA questionnaires varies by factors such as age, gender and health status.
Correction of estimated measures of association between physical activity (as measured by the two study PA questionnaires) and health outcomes for observed measurement error.
The two PA questionnaires that On the Move was designed to evaluate are from the LACE (Life After Cancer Epidemiology) Study, funded by NCI, to examine behavioral factors and breast cancer recurrence and survival among approximately 2,000 women diagnosed with breast cancer, and the California Men’s Health (CMH) Study, funded by the California Cancer Research Program to investigate etiologic factors related to prostrate cancer in a large multi-ethnic cohort. Both questionnaires inquire about duration, frequency and intensity of specific activities and produce summary measures of PA in Met-minutes/week.
Study Methods
Pilot Study: The purpose of the pilot study was to a) develop an electronic physical activity record; b) compare the electronic PA record to a traditional paper and pen record; and c) document the cognitive difficulties respondents encounter when completing the two study PA questionnaires.
Development of Electronic PA Record: To develop the electronic PA record, we collaborated with a private telecommunications company, BeWell Mobile (www.bewellmobile.com). The application runs on a Cingular platform on Motorola phones (V557, RAZR V3) and operates through soft keys and arrow keys. The basic structure is an activity list organized into 15 different domains (e.g. quiet activities, job activities, home, sports), extracted from the Compendium of Physical Activities (Ainsworth et al, 2000). The stem question is: “Since [time of last recording], did you do any [domain] activities for at least 10 minutes at a time?” A “Yes” response prompts a menu (and often submenu) of specific activities in that domain. Selection of a specific activity prompts questions about duration (hours and minutes) and intensity (based on Borg rating of perceived exertion).
The application requires respondents to move through all 15 domains before completing their entries and allows backward movement for review, additions and corrections. The final step is wireless transmission to the study database located on the study server. The database is structured with a row for each recorded activity and columns for study id, time, date, activity name, activity code, MET value, duration and intensity. All data are saved on the cell phone until successful transmission, and automatic pre-set text messages are sent if expected entry is not received.
Comparison of Electronic PA Diary with Paper Diary: Approximately 25 men and 25 women between the ages of 45 and 65 were asked to record three times a day all of the activities they did for at least 10 minutes at a time on four consecutive days (Friday-Monday) in three consecutive weeks. Half of the group used electronic records in weeks 1 and 3 and a paper record in week 2, while the other half used paper records in weeks 1 and 3 and the electronic record in week 2. A comparison of the electronic and paper diaries demonstrated generally similar estimates of physical activity from both types of records, higher data quality (e.g.. fewer missing data) with the electronic record and generally greater user preference for the electronic record.
A manuscript reporting on these findings is in preparation.
Understanding Cognitive Processes: 19 men and 21 women participated in cognitive interviews regarding their responses to the two PA questionnaires (LACE or CMH). Qualitative analyses of the interviews indicated that both questionnaires presented significant cognitive challenges. The manuscript reporting on these observations, Physical Activity Questionnaire Comprehension: Lessons from Cognitive Interviews, is available on this website Based on these findings, both physical activity questionnaires were modified for use in the validation component of the study.
Validation Study: For the validation study, we enrolled 296 men and 327 women members of the Kaiser Foundation Health Plan, ages 45 to 65, 102 of whom were prostrate cancer survivors and 123 breast cancer survivors. Participants had three in-person visits to the research clinic over a six month period. At the first visit they completed the two physical activity questionnaires, performed a submaximal graded exercise treadmill test, and had body size measured and body composition assessed with bioelectrical impedance. Approximately one month later, participants returned to the clinic to complete the two questionnaires a second time and receive instructions on how to record their physical activities using the cell phone and how to wear and use the Actigraph accelerometer. Their activity was monitored with these two instruments for seven days following that visit. About five months later, participants returned for their third study visit to complete the two questionnaires for a final time. Following that visit, their activity was monitored again for a seven days period, with the cell phone and accelerometer.
Data from the validation component of On the Move are now being analyzed and will be reported in the next year.