Clinical Scholarship
Physical Activity in Low-Income Postpartum Women Susan Wilkinson, Chiu-Mieh Huang, Lorraine O. Walker, Bobbie Sue Sterling, Minseong Kim Purpose: To validate the 7-day physical activity recall (PAR), including alternative PAR scoring algorithms, using pedometer readings with low-income postpartum women, and to describe physical activity patterns of a low-income population of postpartum women. Methods: Forty-four women (13 African American, 19 Hispanic, and 12 White) from the Austin New Mothers Study (ANMS) were interviewed at 3 months postpartum. Data were scored alternatively according to the Blair (sitting treated as light activity) and Welk (sitting excluded from light activity and treated as rest) algorithms. Step counts based on 3 days of wearing pedometers served as the validation measure. Findings: Using the Welk algorithm, PAR components significantly correlated with step counts were: minutes spent in light activity, total activity (sum of light to very hard activity), and energy expenditure. Minutes spent in sitting were negatively correlated with step counts. No PAR component activities derived with the Blair algorithm were significantly related to step counts. The largest amount of active time was spent in light activity: 384.4 minutes with the Welk algorithm. Mothers averaged fewer than 16 minutes per day in moderate or high intensity activity. Step counts measured by pedometers averaged 6,262 (SD=2,712) per day. Conclusions: The findings indicate support for the validity of the PAR as a measure of physical activity with low-income postpartum mothers when scored according to the Welk algorithm. On average, low-income postpartum women in this study did not meet recommendations for amount of moderate or high intensity physical activity. C 2004 SIGMA THETA TAU INTERNATIONAL. JOURNAL OF NURSING SCHOLARSHIP, 2004; 36:2, 109-114.
[Key words: physical activity, pedometer, low income, postpartum]
*
P
hysical activity is a leading health indicator in Healthy People 2010 (U.S. Department of Health and Human Services, 2000). The Centers for Disease Control and Prevention and the American College of Sports Medicine have advocated that to receive health benefits: “Every U.S. adult accumulate 30 minutes or more of moderate-intensity physical activity on most, preferably all, days of the week” (Pate et al., 1995, p. 404). Thus, having valid measures to assess physical activity in diverse populations is important. This study was focused on physical activity in a population of low-income postpartum women in an attempt to clarify conflicting evidence of sedentary lifestyles in low-income women (Ford et al., 1991; Hinton & Olson, 2001; Lewis, Raczynski, Heath, Levinson, & Cutter, 1993; Ransdell & Wells, 1998) and to document special considerations that might affect validity of physical activity measures when applied to women in caregiving roles. For example, many activity questionnaires have been designed and validated with men and they do not include household or childcare activities. Consequently, they may misclassify activity patterns in women and underestimate total daily energy expenditure by up to 57% (Ainsworth, 2000; Ainsworth,
*
* Richardson, Jacobs, & Leon, 1993). This problem might be especially present for postpartum women for whom nonleisure, household, and infant care activities are the major source of daily physical activity. To our knowledge, no physical activity measures for women after childbirth have been validated against independent measures of physical activity and only a few studies have examined physical activity patterns of low-income postpartum women (Gennaro Susan Wilkinson, RN, PhD, Epsilon Theta, Research Associate, School of Nursing, University of Texas at Austin, TX; Chiu-Mieh Huang, RN, PhD, Assistant Professor, School of Nursing, Chang-Gung University, Taiwan, ROC; Lorraine O. Walker, RN, EdD, Epsilon Theta, Luci B. Johnson Centennial Professor in Nursing, School of Nursing, Bobbie Sue Sterling, RN, PhD, Epsilon Theta, Project Director, Austin New Mothers Study, School of Nursing, Minseong Kim, MED, Graduate Research Assistant, School of Nursing, all at University of Texas at Austin, TX. This work was supported by grant NR04679 from the National Institute of Nursing Research. We acknowledge Sally Wilging for her contribution to this study. Correspondence to Dr Walker, University of Texas at Austin, School of Nursing, 1700 Red River Street, Austin, TX 78701-1499. E-mail:
[email protected] Accepted for publication September 24, 2003.
Journal of Nursing Scholarship
Second Quarter 2004
109
Physical Activity in Postpartum Women
& Fehder, 2000; Olson, Strawderman, Hinton, & Pearson, 2003). In this study we examined the 7-day physical activity recall (PAR; Blair et al., 1985; Sallis et al., 1985), a wellestablished interview protocol of physical activity that may be well suited to new mothers. Unlike activity inventories with lists of physical activities, the PAR interview involves asking an interviewee to recall activities from the past 7 days that were of “moderate,” “hard,” and “very hard” intensity. Welk, Thompson, and Galper (2001) proposed an alternative algorithm for scoring the 7-day recall to increase its precision. Thus, the first aim of this study was to validate the 7-day physical activity recall (PAR), including alternative PAR scoring algorithms, against pedometer readings with low-income postpartum women. The second aim was to describe physical activity patterns of a low-income population of postpartum women. In this research, as in most research with childbearing women, low income is defined as meeting eligibility for programs such as Medicaid and women-infantchildren programs (WIC).
Background Childbearing generally results in altered activity patterns for women (Sternfeld, Sidney, Jacobs, Anderssen, & Bild, 1992) and potentially leads to an extended time of reduced physical activity. The American College of Obstetricians and Gynecologists (2002) has recommended that prepregnancy exercise routines can be resumed gradually after delivery by increasing time and intensity depending on individual capability. Because traditional postpartum medical care generally ceases around 6 weeks after delivery, many women might not be aware of the importance and value of physical activity after childbearing, or not know how best to resume regular activity or increase the intensity of their daily activities. Thus, valid measures of physical activity for women after pregnancy are important, especially for low-income women with limited resources and more barriers to a physically active lifestyle than those of higher-income women. Brownson and colleagues (2001) reported availability of places for physical activity was higher among men than women. Among women, those with higher incomes reported greater access. Barriers to physical activity reported by women with low incomes included heavy traffic, unattended dogs, foul air from cars and factories, and high neighborhood crime rates. Existing studies of postpartum women have included varied measures to assess physical activity, such as single or multiple-item questions (Harris, Ellison, & Clement, 1999; Ohlin & Rossner 1994; Sampselle, Seng, Yeo, Killion, & Oakley, 1999), activity inventories (Boardley, Sargent, Coker, Hussey, & Sharpe, 1995; Gennaro & Fehder, 2000), rating scales (Olson et al., 2003; Walker, 1996; Walker & Freeland-Graves, 1998), and the PAR (Janney, Zhang, & Sowers, 1997). In only two studies was the validity of the physical activity measure assessed for use with postpartum women (Olson et al., 2003; Walker, 1996), and those validations were based on correspondence with other 110
Second Quarter 2004
Journal of Nursing Scholarship
questionnaire data. Thus, none of these measures has been specifically validated for use with postpartum women using independent estimates of activity, such as electronic pedometers or accelerometers. Testing the validity of the PAR (Blair et al., 1985; Sallis et al., 1985) as a measure of physical activity in postpartum women would advance research with this population. The PAR is a measure of activity that may be especially useful with women during postpartum because it is based on activities elicited from interviewees. Thus, it may potentially avoid biases inherent in physical activity measures designed for men or for other phases of the life span. It has the advantage also of indicating duration and intensity of physical activity as well as energy expenditure. Research reports (Richardson, Ainsworth, Jacobs, & Leon, 2001) have shown that the PAR may be more sensitive to vigorous activity (hard and very hard) than to low intensity activities when using the standard PAR scoring algorithm described (Blair et al., 1985) and revised by Blair. An alternative algorithm designed to be more sensitive to differences in sitting time and light activity has been proposed and tested by Welk and colleagues (2001). The Welk algorithm may be especially useful in measuring physical activity in populations, such as postpartum women, who may not reach recommended moderate intensity physical activity levels.
Methods Design and Sample As part of the Austin New Mothers Study (Walker, Timmerman, Sterling, Kim, & Dickson, 2004), a longitudinal study of postpartum women, the 7-day PAR (Blair et al., 1985) was administered and pedometer readings were gathered at the 3-month postpartum assessment. Eligibility criteria for the Austin New Mothers Study included: Medicaid coverage, age 18 years or more, absence of medical risk factors, full-term delivery, and ability to read and speak in English. Protocols for the study had institutional review board approval. The sample for this report consisted of 44 women (13 African American, 19 Hispanic/Latina, and 12 White) with a mean age of 22.4 years (SD=3.9). The mean body weight was 70.0 kg (154 lb; SD=14.58 kg) with a mean BMI of 27.47 (SD=6.35). The median number of children in the household was two (range 1-4). Most participants (68%) were either high school graduates or had college or technical education. Eighteen participants (41%) reported being employed at the 3-month observation. Of these, six (33%) were employed less than 20 hours, seven (39%) 2039 hours, and five (28%) 40 or more hours per week. Instruments The 7-day PAR is an interview-based, general-purpose measure of habitual physical activity. The PAR is based on several premises including: (a) sampling across 7 days of the week provides a representative level of usual physical activity; and (b) time spent in moderate, hard, or very hard
Physical Activity in Postpartum Women
intensity activities is easier for participants to recall because these activities are performed less often (Blair et al., 1985). Administering the PAR interview requires approximately 20 minutes. The interviewer guides the participant through the recall process to determine frequency, duration, and intensity of physical activities (including moderate, hard, and very hard activity) for the previous 7 days. Sleeping time (and sitting time in the Welk algorithm) is elicited to determine inactivity in each day. The residual time comprised duration of light activity. The duration and intensity of various activities was evaluated for inclusion based on PAR administration criteria. For example, to be included, an activity must be done continuously for 10 minutes; the exact number of minutes was recorded when the activity exceeded 10 minutes. Sleep time was rounded to the closest quarter hour. Classifying the intensity of reported activities was based on the Compendium of Physical Activities (Ainsworth et al., 1993). From the PAR interview, the number of hours spent in sleep, light, moderate, and vigorous (hard and very hard) intensity activities was determined. Because duration of sitting was elicited, it also could be quantified. A second type of data derived from the PAR was an estimate of standardized energy expenditure (kcal/kg/d). PAR energy expenditure estimates are based on metabolic equivalents (MET): 1 MET represents the metabolic rate of an individual at rest, and is set at 3.5 ml of oxygen consumed per kilogram of body weight per minute, or 1 kcal/kg/h (Kriska & Caspersen, 1997). In the standard Blair algorithm for scoring energy expenditure, the duration of activities in the categories of light (includes sitting time, if elicited), moderate, hard, and very hard activity is multiplied by the values of 1.5, 4.0, 6.0, and 10.0 kcal/kg/h, respectively. Sleep is multiplied by 1.0. In the alternative algorithm proposed by Welk and colleagues (2001), the duration of activities in the categories of light (excludes sitting time), moderate, hard, and very hard activity is multiplied by the values of 1.5, 4.0, 6.0, and 10.0 kcal/kg/h, respectively, with sleep and sitting multiplied by 1.0. Thus, the Welk algorithm specifically excludes sitting from light activity and weights it comparably to sleep. Comparative analysis showed that correlations between energy expenditure estimates from PAR and accelerometer data were higher for those based on the Welk algorithm than on the Blair algorithm (Welk et al., 2001). Reliability of the PAR has been supported by significant test-retest correlations (Jacobs, Ainsworth, Hartman, & Leon, 1993; Richardson et al., 2001; Williams, Klesges, Hanson, & Eck, 1989). Validity of the PAR in adult populations has been established through relationships between the PAR and other measures of physical activity, fitness, or energy expenditure, such as doubly labeled water (Conway, Seale, Jacobs, Irwin, & Ainsworth, 2002; Leenders, Sherman, Nagaraja, & Kien, 2001), maximum oxygen consumption (Dishman & Steinhardt, 1988; Jacobs et al., 1993), motion sensors (Jacobs et al., 1993; Welk et al., 2001), and self-report measures of physical activity (Blair et al., 1985; Dishman & Steinhardt, 1988). Despite its
widespread validity testing, no reports have indicated study of the validity of the PAR with low-income new mothers. The Yamax Digi-Walker® (Tokyo, Japan) was chosen as the measure by which to validate the PAR with low-income postpartum women because it is a cost-effective (about $20) and reliable pedometer. It has the advantage of measuring body movement associated with walking, whether done for recreation or in the course of household and other activities of daily living, and has been used in several field studies (Bassett et al., 1996; Bassett, Cureton, & Ainsworth, 2000; Welk et al., 2000). Functioning as a step counter, the Yamax avoids errors of miscalculations or changes in stride length that may corrupt measures of distance traveled. In a comparison of five different brands of electronic pedometers for measuring distance walked, the Yamax pedometer was the most accurate in recording all steps taken (Bassett et al., 1996). Using the Yamax Digi-Walker® to assess daily physical activity patterns, Welk and colleagues (2000) showed that the mean step counts were within 3%-5% of the recorded values for participants. It is highly consistent between two units worn on left and right sides simultaneously of a participant for each trial; accuracy of the Yamax model is not affected by type of walking surface (Bassett et al., 1996). Because the original Yamax Digi-Walker® (DW500) was no longer available, the Yamax SW series 200, comparable to the original Yamax (D. Bassett, personal communication, May 3, 1999; Le Masurier & Tudor-Locke, 2003), was used in this study.
Procedures At 3 months after delivery, an interviewer administered the PAR to mothers using the semi-structured interview protocol. After completing the PAR interview, each participant received the Yamax Digi-Walker® pedometer and a printed instruction sheet for home review. The mothers were instructed to wear the pedometer according to manufacturer’s guidelines (attached to a belt or waistband and 5-7 cm from the umbilicus) from the time they got up until the time they went to bed except for time spent in bathing or swimming for three separate days. They recorded the step counts on forms that also included the time worn and related information such as side of body worn on. The completed recording sheets were returned by mail by all participants included in this analysis. Cases included in this analysis were selected according to the following: (a) the PAR interview was administered by an interviewer trained at a national training center for the PAR and who had an academic background in the field of exercise; and (b) the pedometer readings monitored walking for at least 30 hours over a 3-day period and within 31 days of the PAR interview. Although we were able to get adjacent, not contemporaneous, assessments of physical activity on the PAR interviews and pedometers, this approach is consistent with the purpose of the PAR as a measure of habitual physical activity. Our approach also accommodated the reality of demands faced by mothers of young infants. Journal of Nursing Scholarship
Second Quarter 2004
111
Physical Activity in Postpartum Women
Data Analysis Data from the PAR interviews included: (a) hours spent per day in sleep, sitting, light, moderate, hard, and very hard activity; and (b) estimated daily energy expenditure (kcal/kg/d). Step counts were averaged to yield a measure of mean steps per day. Statistical analysis was conducted using the Statistical Package for the Social Sciences 11.0 (SPSS Inc., Chicago, IL). Descriptive analysis was performed to reveal the mean daily duration of various physical activity intensities performed by new mothers. Data were examined for skewness. Pearson correlations were computed for normally distributed PAR variables, and Kendall’s tau-b correlations (Pett, 1997) were computed for skewed variables (e. g., total kcal/kg/d).
Results PAR Validation Findings Table 1 shows correlation coefficients between minutes spent in component activities based on the Welk and Blair algorithms and step counts. Using the Welk (2001) algorithm, minutes spent in light intensity activity, total activity (sum of light, moderate, hard, and very hard activity), and energy expenditure were positively and significantly correlated with step counts. Minutes in total active time approximated the period during which mothers could potentially be walking and thus were considered as a key measure for validity estimation. In turn, minutes spent in sitting were negatively and significantly correlated with step counts. No PAR component activities derived using the Blair algorithm were significantly correlated with step counts. Descriptive Findings on Activity for Postpartum Women Table 2 shows mean minutes spent per day in component activities of the PAR. Women averaged 502.7 minutes Table 1. Pearson Correlations Between Step Counts and Physical Activity Recall Components in Two Scoring Algorithms Correlation with step counts
Physical inactivity and activity Sleep (min) Sitting (min) Total inactivity (min) (Sleep + sitting)
Welk algorithm
Blair algorithm
.07 −.45∗∗ −.40∗∗
—a —a
.07
Light activity (min) Moderate activity (min) Hard activity (min)
.35∗ .14b .12b
−.21 .14b
Very hard activity (min)
−.10b
−.10b
Total moderate & higher activity (min) (M+H+V) Total activity (min) (L+M+H+V) Total kcal/kg/d
.10b .40∗∗ .22∗b
.12b .10b −.07 .01b
Note. L=light, M=moderate, H=hard, and V=very hard intensity activity. a Not a component of the Blair algorithm. b Kendall’s tau was used because of skewness of the variable distribution. ∗ p
