PURPOSE: The purpose of this study was to validate physical activity measured with an accelerometer and a diary against doubly labeled water (DLW). METHODS: Of the 260 women enrolled in the validation study; 136 African American (44.1%) and Hispanic (55.9%) women had complete data for these analyses (Mean age = 49.7 ± 7.3 years; Mean BMI = 29.9 ± 6.6 kg/m2). Total energy expenditure (TEE) was measured over 14 days using the DLW methodology. Physical activity energy expenditure (PAEE) was computed as: 0.90(TEE) - resting metabolic rate (measured with indirect calorimetry). During the second week of the DLW protocol, participants wore an accelerometer for 7 consecutive days and completed a 7-day physical activity diary. Incremental coefficients of variation were calculated to assess the association between the accelerometer and diary with DLW TEE and PAEE. Allometric scaling was used to determine the appropriate exponent to adjust the DLW TEE and PAEE for the effect of body weight. Pearson product-moment correlations were used to assess the validity of the two instruments against the weight-adjusted DLW TEE and PAEE. RESULTS: Multiple regression correlations for models that included weight and the instruments showed moderate association with DLW TEE (Multiple R = 0.62, accelerometer; R = .59, diary) and low association with DLW PAEE (Multiple R = .30, accelerometer; R = 0.24, diary). Incremental r-square for the accelerometer and diary were less than 1% (instrument was added after weight was accounted for). Correlations among the instruments and weight-adjusted DLW TEE and PAEE were less than 0.30. CONCLUSIONS: These results suggest that the accelerometer and diary account for a small portion of TEE and PAEE. Most of the variance in DLW-measured energy expenditure was explained by body weight. Supported by the Women's Health Initiative through CDC U48/CC609653

PURPOSE: The purpose of this study was to assess the effect of mediating variables on the association between physical activity of parent and child. METHODS: In September COctober 1996, computer-assisted telephone interviews were conducted among 1,504 parent-child pairs residing in the U.S. Trained interviewers asked each parent and a randomly selected child from the same household a series of questions about physical activity, for the parent and the child, and five mediating variables: parental encouragement, role modeling, financial support, and attitudes toward physical activity of the parent and of the child. At least four attempts were made to reach the telephone number before it was retired from the sample. Of households contacted, fourteen percent had eligible parent-child pairs, among whom 72% completed the survey. The mean interview length for both parent and child was 34 minutes. Structural equation modeling was used to test the hypotheses that parental physical activity was: (1) directly associated with physical activity of the child, or (2) was indirectly associated via mediating psychosocial variables. RESULTS: The direct association between parental and child physical activity was not statistically significant (p > 0.05). However, a test of the aggregate model that included all five mediating variables revealed a significant indirect association (r = 0.20) with parental physical activity and physical activity of the child. For individual mediating variables that were significant, we declared as important those associations whose absolute value was > = 0.30. Important positive associations were observed between parental physical activity and parental mediating variables (values in parentheses): financial support (0.43), attitude (0.32), and role modeling (0.43). Parental role modeling was also associated with the child's attitude toward physical activity (0.30). CONCLUSION: These cross-sectional data reveal that parental role modeling of physical activity behaviors may positively influence children's attitudes toward physical activity and suggest an important focus for media as part of health promotion efforts.

The primary aim of this study was to determine the reliability of usual physical activity data using the CSA accelerometer. Participants were 117 African American and 127 Hispanic women (mean age = 49.2 ± 7.0 years; mean BMI = 30.0 ± 6.4 kg/m2) in the Women On The Move physical activity validation study. Participants wore the CSA accelerometer during waking hours, for seven consecutive days. Data from the accelerometer were classified into all CSA counts, counts ≥ 3 METs, and minutes ≥ 3 METs. Seven-day reliability was estimated by calculating intraclass correlation coefficients, which ranged from 0.83 to 0.86 (see Table). Since high reliability was demonstrated over 7 days of recording, the Spearman-Brown prophecy formula was used to estimate the number of days of monitoring needed to obtain lower reliabilities of 0.70, 0.75, and 0.80. In this sample of women, physical activity monitored over seven consecutive days was shown to be highly reliable. Reliable estimates of usual physical activity may be obtained with three to five days of monitoring.TableSupported by the Women's Health Initiative through CDC U48/CC609653

RATIONALE:Recent physical activity recommendations suggest that comparable amounts of prescribed physical activity, done as a single continuous bout or as a set of intermittent bouts, will produce equal amounts of energy expenditure (EE) during the prescribed activity as well as throughout the day.HYPOTHESES:In a field setting, we tested two hypotheses: (1) continuous and intermittent walking conditions will result in significantly greater total daily EE than a control condition, and (2) continuous and intermittent walking conditions will result in similar total dailyMETHODS:Thirty women (mean age [yr] = 43.7+/-5.8; mean body mass index [kg x m(-2)]= 24.7+/-4.0) participated in a repeated-measures design so that each woman participated in three walking conditions on successive days of the week: a single 30-min brisk walk (continuous): three 10-min brisk walks (intermittent); and no activity (control). Throughout the study protocol, women wore a TRITRAC-R3D accelerometer programmed to estimate EE in 2-min intervals.RESULTS:Mean total EE estimates (kcal) for the three walking conditions were as follows: continuous: 2181+/-308; intermittent: 2121+/-305; and control: 1948+/-270. A repeated-measures analysis of variance omnibus test indicated that EE differed significantly by experimental condition [F(2,58) = 40.2, P < 0.001). To test the first hypothesis, contrasts were examined revealing that EE in the continuous and intermittent conditions was significantly greater than EE in the control condition [F(1,29) = 58.2, P < 0.001]. To test the second hypothesis, contrasts revealed that EE in the continuous condition was significantly greater than EE in the intermittent condition [F(1,29) = 7.0, P = 0.013].CONCLUSION:For the purposes of total EE, selecting a continuous mode of walking may offer additional benefit over an intermittent mode, given the same total prescribed duration.

The purpose of this study was to determine the association between intensity of physical activity and indices of obesity. One hundred eleven African American and 116 Hispanic women (mean age = 49 ± 7 years) from the Women On The Move project participated in the study. Women completed physical activity diaries for seven consecutive days, in which they recorded the type, intensity, and duration of activities lasting 10 or more minutes. Because activity intensity is relative to age, we used established sources to create three MET groupings for two groups of women: Ages 40–59, light (2.0–3.4), moderate (3.5–4.9), hard (5.0+); Ages 60–71, light (2.0–2.9), moderate (3.0–3.9), hard (4.0+). Trained observers measured height, weight, waist, and hip circumferences using a standardized protocol. We estimated percent body fat via the doubly-labeled water technique. Participation in light and moderate intensity physical activity was not associated with indices of obesity (p > 0.05). Without adjustment for confounding variables, we noted the following regression coefficients (β) for associations between hard intensity activity and: body mass index, −0.017, percent body fat −0.00017, and waist circumference −0.034, which changed to - 0.015,-0.00019, and −0.029, respectively, with adjustment for age and race. For this sample of women, only relatively hard intensity activity was associated with indices of obesity. Supported by the Women's Health Initiative through CDC U48/CC609653.

1249 Recent surveys suggest that a large proportion of African-American and Hispanic women are sedentary. One explanation for these findings is that many surveys do not assess time spent in non-leisure modes of physical activity. Hence, the Women on the Move Study sought to determine: (1) the time spent in leisure and non-leisure physical activity among African-American and Hispanic women, and (2) if leisure and non-leisure physical activity varied by age, education, or household income. 58 African-American and 56 Hispanic women([horizontal bar over]Xage=48.0±10.5) completed physical activity diaries for seven consecutive days during the Fall of 1996. We calculated median minutes of physical activity of ≥ 3 METs spent in four categories (decided a priori): household, occupation, sport and exercise, and walking. Among African-American women, median minutes per week spent in physical activity for household, occupation, sport and exercise, and walking were: 130.0, 22.5, 0.0, and 50.0, respectively. Among Hispanic women, corresponding median minutes spent were: 170.0, 0.0, 7.5, and 105.0. Time spent in leisure and non-leisure physical activity did not vary by race/ethnicity, age, education, or household income (p>0.05). Among these women, most time was spent in physical activity of non-leisure mode, with household activities being most common. Comprehensive assessments of physical activity among minority women should include household activity.

Recent physical activity promotion recommendations suggest that a daily accumulation of 30-min periods of moderate-intensity physical activity is sufficient to promote health and prevent disease, implicitly suggesting that energy expenditure (EE) in intermittent physical activity will be similar to a single continuous bout of physical activity. This study tested the hypothesis that intermittent (INT) and continuous (C) physical activity conditions will result in similar EE over the course of one day. Thirty-one active women (mean age=43.8 ± 5.7 years; mean weight=67.0 ± 12.2 kg) participated in a repeated design such that each woman participated in three activity conditions on three successive days. Activity conditions were defined as: 3 10-min brisk walks/day (INT), one 30-min brisk walk/day (C), and no activity, inactive (INACT). The Tritrac-R3D® accelerometer was programmed to estimate kilocalorie EE in two-min intervals. Mean EE estimates (Kcals) for the three conditions (INACT, INT, C) were: 1961 ± 278; 2146 ± 328; 2182 ± 339. Repeated measures analysis of variance omnibus test indicated that EE was significantly different by condition, F(2,60)=30.1, p<0.001. A-priori contrasts showed that EE in INT and C was significantly higher than EE in the INACT condition, F(1,30)=53.1, p<0.001. EE in INT and C conditions were not significantly different, F(1,30)=1.5, p=0.23. These results support the hypothesis that EE is similar in intermittent and continuous forms of physical activity among these women.