Abstract
The aim of this study was to update and validate the Pregnancy Physical Activity Questionnaire (PPAQ), using novel and innovative accelerometer and wearable camera measures in a free-living setting, to improve the measurement performance of this method for self-reporting physical activity. A prospective cohort of 50 eligible pregnant women were enrolled in early pregnancy (mean = 14.9 weeks’ gestation). In early, middle, and late pregnancy, participants completed the updated PPAQ and, for 7 days, wore an accelerometer (GT3X-BT; ActiGraph, Pensacola, Florida) on the nondominant wrist and a wearable camera (Autographer; OMG Life (defunct)). At the end of the 7-day period, participants repeated the PPAQ. Spearman correlations between the PPAQ and accelerometer data ranged from 0.37 to 0.44 for total activity, 0.17 to 0.53 for moderate- to vigorous-intensity activity, 0.19 to 0.42 for light-intensity activity, and 0.23 to 0.45 for sedentary behavior. Spearman correlations between the PPAQ and wearable camera data ranged from 0.52 to 0.70 for sports/exercise and from 0.26 to 0.30 for transportation activity. Reproducibility scores ranged from 0.70 to 0.92 for moderate- to vigorous-intensity activity and from 0.79 to 0.91 for sports/exercise, and were comparable across other domains of physical activity. The PPAQ is a reliable instrument and a valid measure of a broad range of physical activities during pregnancy.
Keywords: activity assessment, epidemiologic methods, exercise, reproducibility of results, women
Abbreviations
- MET
metabolic equivalent of task
- PA
physical activity
- PPAQ
Pregnancy Physical Activity Questionnaire
- SD
standard deviation
The American College of Obstetricians and Gynecologists (ACOG) recommends regular physical activity (PA) during pregnancy due to its positive impact on physical fitness, weight management, and psychological well-being. Physical inactivity during pregnancy is implicated in excessive gestational weight gain, gestational diabetes mellitus (GDM), pre-eclampsia, and preterm birth (1) and is consequently an urgent public health concern. In light of these observations, ACOG has called for additional research to clarify the optimal type, frequency, and intensity of exercise during pregnancy (1). Similarly, the 2018 Physical Activity Guidelines Advisory Committee Scientific Report (2) concluded that there are limited data concerning the dose-response relationships between type of PA and important pregnancy outcomes, such as GDM and pre-eclampsia (3).
Self-reported assessments of PA continue to be the most common method for surveillance, epidemiologic, and intervention studies of pregnancy PA, given lower compliance with activity monitor wear during pregnancy, the need for a measure of long-term or habitual activity patterns, and large sample sizes, as well as the participant burden and high cost of assessing PA with wearable activity monitors (4).
However, a critical knowledge gap remains in how best to quantify PA during pregnancy using commonly administered self-report tools. According to a recent review, evidence concerning the measurement properties of self-administered pregnancy PA questionnaires is limited and predominantly of lower quality (5). Many of these tools do not measure light intensity and sedentary behavior, and do not include contemporary sedentary activities (e.g., “screen-time” via smart phones, tablets; time texting/on social media) (5).
The Pregnancy Physical Activity Questionnaire (PPAQ) (6) was developed in 2004 as the first validated PA questionnaire for pregnancy and has since become one of the most widely used instruments for assessing pregnancy PA (5). The PPAQ has been translated and validated for use in over 13 languages and used by researchers in approximately 70 countries. In the United States, the PPAQ is being used in the Environmental Influences on Child Health Outcomes (ECHO) study, an NIH-funded population-based longitudinal birth cohort across 35 centers (7).
Over the past 2 decades, advances have been made in accelerometer and direct observation (e.g., wearable cameras) calibration and validation methods that can be used to determine the validity of self-reported PA and sedentary behavior in free-living settings. Therefore, our first goal was to update the PPAQ to include contemporary sedentary behaviors and the most recent metabolic equivalent of task (MET) values. Our second goal was to evaluate the reproducibility and validity of the updated PPAQ in a free-living setting using a novel validation system of a wrist-worn accelerometer and a wearable camera. We hypothesized that the updated PPAQ measures of PA intensity (sedentary, light, and moderate to vigorous) and type (i.e., household/caregiving, occupational, sports/exercise, and transportation) would be comparable to the objective measures.
METHODS
Study population
Participants (n = 50) were recruited in early pregnancy via flyers at prenatal care centers, health clinics, and advertisements in local papers and using paid Facebook (Menlo Park, California) advertisements. Enrollment commenced in March 2019 and follow-up continued until January 2021. Participants were screened at recruitment with a single-item rating scale for level of PA (8) to ensure they represented a range of people who regularly undertake a diverse set of activities of various intensities (e.g., low-active to high-active participants). Women were considered ineligible for the study if they had any of the following characteristics: 1) >20 weeks’ gestation, 2) under 16 or over 40 years of age, 3) pregnant with twins or triplets as this is often accompanied by clinically prescribed PA restrictions, 4) musculoskeletal issues that limited ambulation, 5) chronic disease (e.g., diabetes requiring insulin, hypertension or heart disease requiring medication, chronic renal disease, emphysema) or life-threatening illnesses, or 6) lack of a telephone.
Each participant read and signed a written informed consent approved by the Institutional Review Board of the University of Massachusetts, Amherst.
Study design
A total of 3 7-day assessments were conducted in early, middle, and late pregnancy (Figure 1). At the beginning of each assessment period, the updated PPAQ (Web Appendix 1, available at https://doi.org/10.1093/aje/kwad130) was interviewer-administered in person (n = 40) or virtually (n = 10), and participants were instructed how to wear the accelerometer and wearable camera over the following 7 days in a free-living setting. Participants were provided with a wear log to record times that the monitors were not worn (e.g., during sleep or water-based activities). Video demonstrations of how to correctly wear the monitors and links to paper-based materials were provided on a study website. At the end of each assessment period, the updated PPAQ was repeated.
Figure 1.
Updated Pregnancy Physical Activity Questionnaire (PPAQ) study timeline, PPAQ novel validation study, 2019–2021. Participants wore an accelerometer (GT3X-BT; ActiGraph, Pensacola, Florida) on the nondominant wrist and a wearable camera (Autographer; OMG Life (defunct)) for 7 days.
Physical activity assessment
Pregnancy Physical Activity Questionnaire.
The PPAQ is a semiquantitative instrument that measures duration and intensity of time spent in household/caregiving, occupational, transportation, and sports/exercise activities (6). Prior leisure-time sedentary behavior questions on the original PPAQ (questions 11–13) had been limited to sitting and using a computer, writing/reading/talking on the telephone while not at work, or watching TV or a video. For the purposes of the current study, we updated the PPAQ to capture the full range of contemporary sedentary behavior (e.g., “screen-time” via smart phones and tablets; time spent texting and on social media). The new wording changes were consistent with recent questions used to measure sedentary behavior in nonpregnant populations (9–11) (Web Table 1).
To classify time spent in categories of PA intensity (i.e., sedentary, light, and moderate to vigorous), we used MET values from pregnancy studies (12, 13) (n = 8 activities), and, when not available (n = 25 activities), the 2011 Compendium of Physical Activity (14) MET values (Web Table 2). The selection of MET values was also guided by Activities Collected over Time over 24-hours (ACT24), a PA assessment tool made available by the National Cancer Institute (15) and widely used for research purposes (16). For questions addressing walking to go places (question 20) and walking for fun/exercise (question 23), MET values were also informed by the recent findings of Marshal et al. (17) that pregnant women accumulate most of their walking time at cadences <50 steps/minute, which according to McAvoy et al. (18) qualifies as light-intensity ambulation. Overall, these MET value updates led to a change in MET value for 22 of the 31 closed-ended questions and a change in intensity category for 4 of these activities (questions 9, 11, 14, 27).
We multiplied the number of hours spent in each reported activity by its MET level using standard cutpoints to define activity by intensity—sedentary (at least 1 and less than 1.5 METs), light (at least 1.5 and less than 3 METs), moderate to vigorous (>3 METs) (19)—and by type (sports/exercise, household/caregiving, occupational, and transportation). Coding instructions for the updated PPAQ are presented in Web Appendix 2.
Accelerometer.
Participants were instructed to wear the ActiGraph GT3X-BT (Pensacola, Florida) on the nondominant wrist for 3 7-day periods (early, middle, and late pregnancy) except during sleep or water-based activities. Accelerometer devices were initialized to collect data at 80 Hz using ActiLife 6 software (version 6.13.1; ActiGraph). After each assessment period, data were exported as raw and 1-second epoch files.
Wear time was based on self-report or, if missing, using the Choi wear time algorithm (20). Assessment periods with at least 10 hours of wear time for at least 4 days were included (21). Time and frequency-domain features were computed from raw accelerometer data in 15-second nonoverlapping epochs and used in previously trained random forest models to classify accelerometer data according to intensity (22). In secondary analyses, we used methods developed by Hildebrand et al. (23) and Montoye et al. (24) for raw acceleration and count data. We calculated MET-hours/day and % wear-time hours/day spent in each intensity category. Last, we calculated average counts per minute (CPM) for each assessment period.
Wearable camera.
Participants were instructed to wear an Autographer (OMG Life (defunct)) on a lanyard, clip, or chest harness during the 7-day period between PPAQ administrations. The Autographer is a wearable camera used to categorize the type and context of accelerometer-identified episodes of PA across the entire PA spectrum, including sedentary and nonsedentary activities, within the full range of household/caregiving, occupational, sports/exercise, transportation activities.
Participants were instructed that they were free to remove the camera or turn the camera off at any time to protect their privacy and those of others. They were also instructed to deactivate the camera in a prespecified list of places (e.g., restrooms, changing rooms, doctor’s office, automated teller machine or bank, and wherever requested by others). Participants were encouraged to seek verbal permission from family members, cohabitants, workplace managers, and others when wearing the camera and were provided with a prepared statement to read to anyone with concerns. After each assessment time period, participants were given the opportunity to delete any images prior to review by research staff.
The wearable camera recorded up to 4,000 “wearer’s perspective” images per day (at approximately 15-second intervals; median capture rate is approximately 3 images/minute). A minimum of 600 images/day was used as the criterion for inclusion in analyses. Episode images, in addition to the surrounding contextual images before and after the PA bout, provide visual clues as to the activity type. We aggregated images into counts of activity durations and sequences using established procedures (25, 26). PA episodes were annotated according to type and subtype using the 21 categories and 821 subcategories from the 2011 Compendium (14). We normalized camera data by classified recognizable hours of wear time. We then calculated % wear-time hours/day spent in each activity type (i.e., household/caregiving, occupational, sports/exercise, and transportation).
Covariates
At the baseline assessment, information on age, first day of last menstrual period, expected delivery date, parity, and race/ethnicity were collected via self-report, and height and weight were measured using a stadiometer and a calibrated scale. For virtual assessments, height and weight information was collected via self-report. Early pregnancy body mass index was calculated as weight (kg)/height (m)2. Gestational age (weeks) was calculated using first day of last menstrual period and/or expected delivery date.
Statistical analyses
We calculated the mean and standard deviation for variables derived from the PPAQ, accelerometer, and wearable camera. To evaluate the validity of the PPAQ, we calculated Spearman correlation coefficients between the PPAQ (the mean of the 2 assessments) compared with weekly summaries of the PA measurements from the accelerometer and wearable camera (27), within each measurement period. Based on the Qualitative Attributes and Measurement Properties of Physical Activity Questionnaires (QAPAQ) (28) and the systematic review of pregnancy PA questionnaires by Sattler et al. (5), validity scores for correlations with accelerometers of >0.4 were deemed reasonable. We inspected scatterplots as graphical displays of these relationships. We also calculated limits of agreement (29), defined as the mean change in scores of repeated measurements (−1.96 × standard deviation (SD) of this change; SD change) to assess absolute validity in addition to relative validity.
To describe the reproducibility of the PPAQ measures, intraclass correlation coefficients were calculated as the proportion of total variance explained by between-subject variance within each measurement period. Based on QAPAQ (28) and the systematic review by Sattler et al. (5), reproducibility scores for intraclass correlations of >0.70 were deemed reasonable. Between- and within-subject variance components were estimated using log-transformed activity data assuming a compound symmetric covariance structure (30). Last, we conducted several sensitivity analyses evaluating whether findings differed according to hours of wearable camera use and according to season.
RESULTS
We excluded PPAQ assessments in which participants reported >20 hours/day of total activity or >85% hours/day of nonsedentary activities (n = 4) and accelerometer assessments not meeting the minimum wear time criteria (n = 1). A total of 6 participants did not contribute to the middle (n = 3) or late pregnancy (n = 3) assessments; however, due to our use of repeated measures (31), these participants contributed to the analyses for which they had complete data, and therefore the final sample size throughout the analysis was 50. Throughout the study, complete data for the accelerometer ranged from 90% to 98% and for the wearable camera from 82% to 94%. Average wear time was 12.80 (SD, 1.63) hours/day for the accelerometer and 3.48 (SD, 2.17) hours/day for the wearable camera; mean MET-hours/day and correlations with the PPAQ did not meaningfully differ between those with those with ≤4 hours vs. >4 hours of camera wear time (F = 2.12). Similarly, there was no meaningful difference in mean MET-hours/day between participants who attended their baseline assessment visit in person vs. virtually (F = 0.1508).
In the total sample, participants were on average 32.9 (SD, 4.2) years old and were predominantly non-Hispanic White (80%), and parous (1.1 (SD, 1.4) prior pregnancies), with an early pregnancy body mass index of 25.1 (SD, 3.6) at a mean of 14.9 weeks’ gestation (Table 1). At baseline (early pregnancy), the mean for MET-hours/day of total PA was 20.73 (SD, 4.96) according to the PPAQ and 21.52 (SD, 9.15) according to the accelerometer (Table 2). Mean MET-hours/day did not meaningfully differ across seasons (F = 0.8776, P = 0.48).
Table 1.
Participant Characteristics (n = 50), Novel Validation Study for the Pregnancy Physical Activity Questionnaire, 2019–2021
| Characteristic | No. | % | Mean (SD) |
|---|---|---|---|
| Age, years | 32.9 (4.2) | ||
| Body mass indexa | 25.1 (3.6) | ||
| Parity | |||
| Nulliparous | 22 | 44 | |
| 1–2 | 20 | 40 | |
| ≥3 | 8 | 16 | |
| Gestational age, weeks | |||
| Early pregnancy assessment | 14.9 (3.8) | ||
| Midpregnancy assessment | 26.0 (3.0) | ||
| Late pregnancy assessment | 35.9 (1.6) | ||
| Race/ethnicity | |||
| Asian | 3 | 6 | |
| Black | 1 | 2 | |
| Hispanic | 3 | 6 | |
| Multirace | 3 | 6 | |
| Non-Hispanic White | 40 | 80 |
Abbreviation: SD, standard deviation.
a Weight (kg)/height (m)2.
Table 2.
Physical Activity as Measured Using the Pregnancy Physical Activity Questionnaire, Accelerometer, and Wearable Camera, According to Pregnancy Time Period, Novel Validation Study for the Pregnancy Physical Activity Questionnaire, 2019–2021
| Measure | Total No. | PPAQ, Mean (SD) | ActiGraph a , Mean (SD) | Wearable Camera a , Mean (SD) | |||
|---|---|---|---|---|---|---|---|
| MET-hours/day | % Waking Day | MET-hours/day | % Wear-Time, hours/day | Counts Per Minute | % Wear-Time hours/day | ||
| Overall pregnancy | 50 | ||||||
| Total PA | 20.17 (5.69) | 21.88 (8.63) | |||||
| Counts per minute | 2,651.59 (701.58) | ||||||
| PA type | |||||||
| Household/caregiving | 7.46 (5.00) | 30 (18) | 80 (25) | ||||
| Occupational | 6.41 (4.36) | 35 (20) | 9 (20) | ||||
| Sports/exercise | 1.34 (1.21) | 4 (3) | 6 (13) | ||||
| Transportation | 1.76 (1.41) | 8 (6) | 5 (11) | ||||
| PA intensity | |||||||
| Sedentary | 8.19 (3.21) | 56 (19) | 5.31 (1.68) | 64 (9) | |||
| Light | 8.14 (4.38) | 34 (17) | 11.09 (4.11) | 29 (7) | |||
| Moderate to vigorous | 3.84 (2.85) | 9 (7) | 5.48 (2.84) | 8 (4) | |||
| Early pregnancy | 50 | ||||||
| Total PA | 20.73 (4.96) | 21.52 (9.15) | |||||
| Counts per minute | 2,621.13 (703.42) | ||||||
| PA type | |||||||
| Household/caregiving | 7.30 (4.73) | 28 (17) | 78 (26) | ||||
| Occupational | 6.98 (3.76) | 38 (19) | 11 (21) | ||||
| Sports/exercise | 1.40 (1.18) | 4 (3) | 6 (13) | ||||
| Transportation | 2.11 (1.54) | 9 (6) | 5 (10) | ||||
| PA intensity | |||||||
| Sedentary | 8.46 (2.93) | 57 (19) | 5.66 (1.8) | 64 (9) | |||
| Light | 8.18 (4.50) | 33 (16) | 9.93 (4.31) | 27 (7) | |||
| Moderate to vigorous | 4.09 (2.34) | 10 (7) | 5.93 (3.04) | 8 (4) | |||
| Midpregnancy | 47 | ||||||
| Total PA | 20.7 (5.98) | 22.86 (8.74) | |||||
| Counts per minute | 2,710.92 (741.95) | ||||||
| PA type | |||||||
| Household/caregiving | 7.58 (5.01) | 30 (19) | 80 (27) | ||||
| Occupational | 6.67 (5.29) | 34 (21) | 9 (20) | ||||
| Sports/exercise | 1.39 (1.28) | 3 (3) | 6 (16) | ||||
| Transportation | 1.79 (1.21) | 8 (7) | 6 (16) | ||||
| PA intensity | |||||||
| Sedentary | 8.27 (3.43) | 56 (19) | 5.10 (1.63) | 63 (9) | |||
| Light | 7.98 (3.95) | 34 (17) | 12.19 (4.26) | 30 (8) | |||
| Moderate to vigorous | 4.46 (3.61) | 10 (7) | 5.57 (2.84) | 7 (4) | |||
| Late pregnancy | 44 | ||||||
| Total PA | 18.96 (6.06) | 21.27 (7.47) | |||||
| Counts per minute | 2,623.49 (667.89) | ||||||
| PA type | |||||||
| Household/caregiving | 7.51 (5.39) | 31 (19) | 82 (23) | ||||
| Occupational | 5.48 (3.82) | 32 (21) | 8 (20) | ||||
| Sports/exercise | 1.22 (1.18) | 4 (4) | 5 (11) | ||||
| Transportation | 1.32 (1.34) | 6 (5) | 5 (8) | ||||
| PA intensity | |||||||
| Sedentary | 7.80 (3.29) | 56 (19) | 5.16 (1.58) | 63 (9) | |||
| Light | 8.27 (4.75) | 36 (18) | 11.24 (3.41) | 30 (7) | |||
| Moderate to vigorous | 2.89 (2.19) | 7 (5) | 4.88 (2.48) | 7 (4) | |||
Abbreviations: MET, metabolic equivalent of task; PA, physical activity; PPAQ, Pregnancy Physical Activity Questionnaire; SD, standard deviation.
a In early, middle, and late pregnancy, participants wore an accelerometer (GT3X-BT; ActiGraph, Pensacola, Florida) on the nondominant wrist and a wearable camera (Autographer; OMG Life (defunct)) for 7 days.
Mean MET-hours/day of total PA were comparable between the PPAQ and the accelerometer for overall pregnancy (20.17 vs. 21.88 MET hours/day for PPAQ and accelerometer, respectively) and within each pregnancy time period (Table 2). The majority of MET-hours/day were spent in household/caregiving activities, followed by occupational activities, transportation, and sports/exercise, according to both PPAQ and wearable camera across all time periods. However, according to the PPAQ, a greater percentage of time was spent in occupational activity than classified by the wearable camera, which is likely due to participant instructions to remove the wearable camera during sensitive work settings. In terms of intensity of PA, sedentary MET-hours/day were consistently higher, while light and moderate- to vigorous-intensity MET-hours/day were lower, on the PPAQ vs. the accelerometer across all time periods although both methods were consistent in indicating that the largest percent time was spent in sedentary behavior. Findings were comparable when Hildebrand et al. (23) and Montoye et al. (24) accelerometer cutpoints were used (Web Table 3).
Validity
For total activity MET-hours/day, correlations between the PPAQ and the accelerometer ranged from 0.37 to 0.44 across pregnancy time periods (Table 3). In terms of the type of PA, correlations between the PPAQ and wearable camera were highest for sports/exercise PA (0.70 in early pregnancy to 0.52 in late pregnancy) followed by occupational activity (0.30 in early pregnancy to 0.26 in late pregnancy). While correlations were low for household/caregiving activity in early pregnancy (r = 0.03), they increased to 0.29 in late pregnancy. Correlations for transportation activity were lower, ranging from −0.01 to 0.20.
Table 3.
Spearman Validity Coefficients Between the Pregnancy Physical Activity Questionnaire, Accelerometer, and Wearable Camera According to Pregnancy Time Period (n = 50), Novel Validation Study for the Pregnancy Physical Activity Questionnaire, 2019–2021
| Early Pregnancy | Midpregnancy | Late Pregnancy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ActiGraph a | Wearable Camera a | ActiGraph a | Wearable Camera a | ActiGraph a | Wearable Camera a | |||||||
| Measure | Correlation Coefficient | 95% CI | Correlation Coefficient | 95% CI | Correlation Coefficient | 95% CI | Correlation Coefficient | 95% CI | Correlation Coefficient | 95% CI | Correlation Coefficient | 95% CI |
| Total PA (MET-hours/day) | 0.37 | 0.10, 0.59 | N/Aa | N/Aa | 0.44 | 0.17, 0.65 | N/Aa | N/Aa | 0.41 | 0.12, 0.63 | N/Aa | N/Aa |
| PA type (% wear-time hours/day) | ||||||||||||
| Household/caregiving | N/Aa | N/Aa | 0.03 | −0.26, 0.31 | N/Aa | N/Aa | 0.17 | −0.13, 0.45 | N/Aa | N/A | 0.29 | −0.02, 0.55 |
| Occupational | N/Aa | N/Aa | 0.30 | 0.02, 0.54 | N/Aa | N/Aa | 0.26 | −0.04, 0.52 | N/Aa | N/A | 0.26 | −0.05, 0.52 |
| Sports/exercise | N/Aa | N/Aa | 0.70 | 0.52, 0.82 | N/Aa | N/Aa | 0.63 | 0.41, 0.78 | N/Aa | N/A | 0.52 | 0.25, 0.71 |
| Transportation | N/Aa | N/Aa | 0.20 | −0.09, 0.46 | N/Aa | N/Aa | −0.01 | −0.31, 0.30 | N/Aa | N/A | 0.11 | −0.20, 0.41 |
| PA intensity (MET-hours/day) | ||||||||||||
| Sedentary | 0.31 | 0.02, 0.54 | N/Aa | N/A | 0.23 | −0.07, 0.49 | N/Aa | N/Aa | 0.35 | 0.06, 0.59 | N/Aa | N/Aa |
| Light | 0.25 | −0.04, 0.50 | N/Aa | N/A | 0.19 | −0.11, 0.46 | N/Aa | N/Aa | 0.27 | −0.03, 0.53 | N/Aa | N/Aa |
| Moderate to vigorous | 0.22 | −0.07, 0.47 | N/Aa | N/A | 0.53 | 0.28, 0.71 | N/Aa | N/Aa | 0.30 | −0.01, 0.55 | N/Aa | N/A |
| PA intensity (% wear-time hours/day) | ||||||||||||
| Sedentary | 0.45 | 0.18, 0.65 | N/Aa | N/A | 0.35 | 0.06, 0.58 | N/Aa | N/Aa | 0.38 | 0.09, 0.61 | N/Aa | N/Aa |
| Light | 0.42 | 0.16, 0.63 | N/Aa | N/A | 0.31 | 0.01, 0.55 | N/Aa | N/Aa | 0.40 | 0.11, 0.62 | N/Aa | N/Aa |
| Moderate to vigorous | 0.17 | −0.12, 0.44 | N/Aa | N/A | 0.42 | 0.14, 0.63 | N/Aa | N/Aa | 0.17 | −0.13, 0.45 | N/Aa | N/Aa |
Abbreviations: CI, confidence interval; MET, metabolic equivalent of task; N/A, not applicable; PA, physical activity.
a In early, middle, and late pregnancy, participants wore an accelerometer (GT3X-BT; ActiGraph, Pensacola, Florida) on the nondominant wrist and a wearable camera (Autographer; OMG Life (defunct)) for 7 days. The accelerometer does not measure PA type (context). The wearable camera does not measure PA intensity.
In terms of the intensity of PA, across the MET-hours/day and % wear-time hours/day categories, correlations between the PPAQ and accelerometer were highest for sedentary behavior (ranging from 0.23 to 0.45) and light-intensity activity (ranging from 0.19 to 0.42) across all pregnancy time periods. Correlations for moderate- to vigorous-intensity activity ranged from 0.17 in early and late pregnancy to 0.53 in middle pregnancy. Overall, correlations tended to be highest in early pregnancy. Findings were comparable when Hildebrand et al. (23) and Montoye et al. (24) accelerometer cutpoints were used (Web Table 4). Correlations with the accelerometer did not meaningfully differ across seasons. Finally, correlations with the wearable camera did not meaningfully differ between participants with ≤4 hours vs. >4 hours of camera wear time. Limits-of-agreement scores, as a measure of absolute validity (Web Table 5), were relatively wide.
Reproducibility
Reproducibility between the 2 administrations of the PPAQ for total PA MET-hours/day ranged from 0.55 to 0.83 (Table 4). In terms of activity type, reproducibility was highest for sports/exercise (0.79–0.91), household/caregiving (0.71–0.92), and occupational activity (0.70–0.91), followed by transportation (0.42–0.67). In terms of activity intensity, reproducibility was highest for moderate to vigorous activity (0.70–0.92), followed by light-intensity activity (0.75–0.87) and sedentary behavior (0.63–0.82). The reproducibility of the ActiGraph and Autographer, respectively, have been previously published (32, 33).
Table 4.
Reproducibility (Intraclass Correlation Coefficients) of the Pregnancy Physical Activity Questionnaire According to Pregnancy Time Period, Novel Validation Study for the Pregnancy Physical Activity Questionnaire, 2019–2021
| Measure | Early Pregnancy (n = 50) | Midpregnancy (n = 47) | Late Pregnancy (n = 44) | |||
|---|---|---|---|---|---|---|
| ICC | 95% CI | ICC | 95% CI | ICC | 95% CI | |
| Total PA (MET-hours/day) | 0.55 | 0.31, 0.71 | 0.81 | 0.68, 0.89 | 0.83 | 0.70, 0.90 |
| PA type | ||||||
| Household/caregiving | 0.71 | 0.53, 0.82 | 0.83 | 0.72, 0.90 | 0.92 | 0.85, 0.95 |
| Occupational | 0.70 | 0.53, 0.82 | 0.91 | 0.85, 0.95 | 0.80 | 0.66, 0.89 |
| Sports/exercise | 0.79 | 0.66, 0.88 | 0.79 | 0.65, 0.88 | 0.91 | 0.84, 0.95 |
| Transportation | 0.67 | 0.50, 0.80 | 0.42 | 0.14, 0.63 | 0.64 | 0.42, 0.79 |
| PA intensity | ||||||
| Sedentary | 0.63 | 0.43, 0.78 | 0.82 | 0.70, 0.89 | 0.76 | 0.60, 0.86 |
| Light | 0.75 | 0.59, 0.85 | 0.77 | 0.65, 0.87 | 0.87 | 0.77, 0.92 |
| Moderate to vigorous | 0.70 | 0.53, 0.82 | 0.81 | 0.69, 0.90 | 0.92 | 0.86, 0.96 |
Abbreviations: CI, confidence interval; ICC, intraclass correlation coefficient; MET, metabolic equivalent of task; PA, physical activity.
Variance components
To assess the ability of the updated PPAQ to detect clinically significant change in PA, we calculated the within- and between-individual variance components for total MET hours/day and total moderate to vigorous hours/day from the PPAQ. For total MET hours/day, the between-person SD was 4.59 MET hours/day and the within-person SD was 3.29 MET hours/day. For total moderate to vigorous hours/week, the between-person SD was 3.29 hours/week and the within-person SD was 3.84 hours/week.
DISCUSSION
In this study of the validity and reproducibility of the updated PPAQ in a free-living setting using novel validation techniques, we found that the PPAQ had reasonable validity in detecting total PA, sedentary, light, and moderate- to vigorous-intensity activities compared with the accelerometer. Compared with the wearable camera, the PPAQ had reasonable validity in detecting sports/exercise PA. The overlapping confidence intervals for household/caregiving and occupational activities indicate that there were no substantive differences in performance between these domains. We observed sufficient reproducibility for all intensities and types of physical activities, with moderate to high reproducibility for total activity and intensity and type of activities.
Our validity scores for moderate- to vigorous-intensity activity (ranging from 0.17 to 0.53) were slightly lower than the corresponding scores for sports/exercise activity (ranging from 0.52 to 0.70). This may be due to the fact that structured exercise is recalled more reliably than other activities, such as household/caregiving activities (34). Indeed, sports/exercise makes up only a portion of moderate to vigorous PA on the PPAQ, which also queries moderate-intensity household/caregiving and occupational activities. This finding may also be due to differences in the validation tools for these domains (i.e., accelerometer for moderate to vigorous PA vs. wearable camera for sports/exercise).
Because context of PA was validated via the wearable camera, lower validity correlations may be due to the fact that participants did not feel comfortable wearing the camera or were asked to turn the camera off in public spaces (35). We observed the lowest validity scores for transportation activities (ranging from −0.01 to 0.20). In the Physical Activity and Pregnancy Questionnaire (PAPQ), the only other pregnancy PA questionnaire that included transportation activity, the authors observed a weak correlation between walking for transportation and pregnancy weight gain (r = −0.117, P = 0.015) (36). For populations in which transportation activities make up a larger proportion of total PA, the Transport and Physical Activity Questionnaire has observed fair reproducibility (ranging from 0.59 to 0.61) and validity (r = 0.27 with moderate to vigorous PA) (37).
Our validity findings for the updated PPAQ (ranging from 0.37 to 0.44 for total activity) are, in general, higher than those observed for the original PPAQ (ranging from 0.08 to 0.43). Validity scores were also higher for light-intensity activity (updated PPAQ: 0.19–0.42 vs. original PPAQ −0.08 to 0.22) and for sedentary behavior (updated PPAQ 0.23–0.45 vs. original −0.34 to 0.12) (6). Validity scores for moderate to vigorous PA for the updated PPAQ (0.17–0.53) were largely overlapping with validity scores for the original (0.20–0.49). The original PPAQ was validated using hip-worn accelerometers, which are limited in their capacity to detect movements of the upper extremities. Given that many of the PPAQ questions are designed to capture activities such as taking care of children or household chores, which involve upper-extremity movements, the overall observed improvements in validity scores may in part be due to the use of the wrist-wear location and accelerometer methods calibrated from diverse activity patterns.
Our reproducibility findings for the updated PPAQ (ranging from 0.55 to 0.83 for total activity) were comparable to the original PPAQ (0.78) (6). Scores for moderate- to vigorous-intensity activity (updated PPAQ 0.70–0.92 vs. original 0.81–0.82), sedentary behavior (updated PPAQ 0.63–0.82 vs. original 0.79), sports/exercise (updated PPAQ 0.79–0.91 vs. original 0.83), and occupational activity (updated PPAQ 0.70–0.91 vs. original 0.93) were also comparable.
It is also important to compare the updated PPAQ with existing pregnancy PA questionnaires. Only the Leisure-Time Physical Activity Questionnaire (LTPAQ) (38), the Physical Activity and Pregnancy Questionnaire (39) and Q1 of MoBa (40) were developed specifically for pregnant women. However, the LTPAQ focuses on leisure-time PA only and Q1 of the MoBa focuses on exercise only. For the LTPAQ, Spearman correlations ranged from 0.27 to 0.47 for total leisure-time PA and Q1 of the MoBa showed validity scores of 0.32 for exercise. In comparison, we observed higher validity scores for sports/exercise, ranging from 0.52 to 0.70. In terms of activity intensity, our scores were comparable to the scores for the Physical Activity and Pregnancy Questionnaire, which observed 0.20 for light-intensity activity (vs. 0.19–0.42 for the updated PPAQ), 0.15 for moderate activity, and 0.59 for vigorous-intensity activity (vs. 0.17–0.53 for moderate- to vigorous-intensity activity for the updated PPAQ).
In terms of questionnaires not developed for the pregnant population but that have been evaluated in pregnant women, poor validity has been observed for the International Physical Activity Questionnaire (IPAQ) (41), the IPAQ-short form (IPAQ-SF) (42), and the Global Physical Activity Questionnaire (GPAQ) (43).
There are several advantages of the PPAQ as compared with existing pregnancy PA questionnaires. First, the PPAQ remains the only questionnaire that uses pregnancy-specific METs when available. Second, the PPAQ is the only tool to be concurrently validated against the accelerometer and wearable camera to assess both activity intensity and context. Prior pregnancy PA questionnaires relied upon accelerometers alone or consumer-grade devices such as pedometers, logbooks, or other PA questionnaires (4, 5). Those studies that used accelerometers placed the accelerometer on the waist and hip and therefore were limited in detecting movements involving the upper extremities (6).
Strengths and limitations
Strengths of the updated PPAQ include the assessment of total PA, sedentary behavior, type of PA (including household/caregiving activity), and the fact that it was developed using a novel, data-driven approach among pregnant women (6). Other strengths include the prospective design and use of a free-living setting for validation.
Our study also had several limitations. The validity results are affected by limitations in the objective measures and in the PPAQ measures. For example, cutpoints used to categorize accelerometer data into intensity levels vary according to subject characteristics and the activities performed during the generation of calibration equations. In addition, because wearable cameras are a novel validation tool, there are no references for a minimal acceptable number of wear-time hours. Participants had permission to remove the wearable camera at work to protect coworker and client privacy, limiting the number of hours. In spite of these limitations, the wearable camera, unlike the accelerometer, is unique in its ability to provide information on the context of PA. Last, because errors associated with the accelerometer, wearable camera, and PPAQ are largely independent, our correlation coefficients are likely not overstated (27).
Reactivity (i.e., change in behavioral pattern due to the knowledge that PA is being monitored) may have occurred. However, a national sample of US adolescents and adults did not observe sufficient evidence of accelerometer reactivity (44). In addition, because both objective measures and the PPAQ could lead to such changes in behavior, and as the analysis is within-subject, the impact of reactivity on the validity findings is reduced.
While we used MET values for pregnant women when available, the majority of MET values in the PA Compendium were not developed in pregnant women. However, because all 3 instruments (i.e., PPAQ, wearable camera, and accelerometer) rely upon the same intensity categories, the accuracy of Compendium values will not affect comparisons between them (i.e., the validity scores). Future studies designed to measure the metabolic intensity of physical activities among pregnant women in the field are warranted.
Because recruitment and follow-up spanned a 2-year period, each trimester of pregnancy spanned different seasons. This reduces the concern that validity scores for sports/exercise may be affected by weather. In addition, occupational and household/caregiving activities tend to be consistent across seasons. In our sensitivity analyses, results did not meaningfully vary across seasons.
Activities were selected for inclusion in the original PPAQ based upon their ability to discriminate between subjects, consistent with the primary objective of a questionnaire to classify subjects into activity rankings and to avoid unnecessarily lengthy questionnaires (6). For the same reason, the PPAQ should not be used to calculate absolute energy expenditure. Our findings demonstrate that the updated PPAQ is sensitive to differences in PA levels and patterns between women and therefore will help investigators avoid misclassifying active pregnant women as sedentary when the opposite may be true.
Conclusion
The updated PPAQ provides valid and reliable estimates of PA and sedentary behavior in free-living pregnant women. The updated PPAQ can be used for: 1) surveillance studies assessing compliance with PA guidelines, 2) etiologic studies of PA and maternal-fetal outcomes designed to identify the optimal dose of pregnancy PA for the prevention of these disorders, and 3) to determine the impact of exercise intervention studies.
Supplementary Material
ACKNOWLEDGMENTS
Author affiliations: Department of Biostatistics and Epidemiology, School of Public Health and Health Sciences, University of Massachusetts, Amherst, Amherst, Massachusetts, United States (Lisa Chasan-Taber, Susan Park); Department of Kinesiology, School of Public Health and Health Sciences, University of Massachusetts, Amherst, Amherst, Massachusetts, United States (Robert Marcotte, Patty Freedson); Department of Mathematics and Statistics, College of Natural Sciences, University of Massachusetts, Amherst, Amherst, Massachusetts, United States (John Staudenmayer); and Zilber College of Public Health, University of Wisconsin Milwaukee, Milwaukee, Wisconsin, United States (Scott Strath).
Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health, under award number R21HD094565.
The data set is available from the corresponding author.
The views expressed in this article are those of the authors and do not reflect those of the National Institutes of Health.
Conflict of interest: none declared.
REFERENCES
- 1. Mota P, Bø K. ACOG Committee Opinion No. 804: physical activity and exercise during pregnancy and the postpartum period. Obstet Gynecol. 2021;137(2):376. [DOI] [PubMed] [Google Scholar]
- 2. 2018 Physical Activity Guidelines Advisory Committee . 2018 Physical Activity Guidelines Advisory Committee Scientific Report. Washington, DC: Department of Health and Human Services; 2018. https://health.gov/sites/default/files/2019-09/PAG_Advisory_Committee_Report.pdf. Accessed May, 11, 2023. [Google Scholar]
- 3. Dipietro L, Evenson KR, Bloodgood B, et al. Benefits of physical activity during pregnancy and postpartum: an umbrella review. Med Sci Sports Exerc. 2019;51(6):1292–1302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Evenson KR, Chasan-Taber L, Symons Downs D, et al. Review of self-reported physical activity assessments for pregnancy: summary of the evidence for validity and reliability. Paediatr Perinat Epidemiol. 2012;26(5):479–494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Sattler MC, Jaunig J, Watson ED, et al. Physical activity questionnaires for pregnancy: a systematic review of measurement properties. Sport Med. 2018;48(10):2317–2346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Chasan-Taber L, Schmidt MD, Roberts DE, et al. Development and validation of a pregnancy physical activity questionnaire. Med Sci Sports Exerc. 2004;36(10):1750–1760. [DOI] [PubMed] [Google Scholar]
- 7. National Institutes of Health . Environmental Influences on Child Health Outcomes (ECHO) Program. https://www.nih.gov/research-training/environmental-influences-child-health-outcomes-echo-program. Accessed December 7, 2022.
- 8. Jackson AW, Morrow JR, Bowles HR, et al. Construct validity evidence for single-response items to estimate physical activity levels in large sample studies. Res Q Exerc Sport. 2007;78(2):24–31. [DOI] [PubMed] [Google Scholar]
- 9. Mackay LM, Schofield GM, Oliver M. Measuring physical activity and sedentary behaviors in women with young children: a systematic review. Women Heal. 2011;51(4):400–421. [DOI] [PubMed] [Google Scholar]
- 10. Healy GN, Clark BK, Winkler EAH, et al. Measurement of adults’ sedentary time in population-based studies. Am J Prev Med. 2011;41(2):216–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Saint-Maurice PF, Welk GJ. Web-based assessments of physical activity in youth: considerations for design and scale calibration. J Med Internet Res. 2014;16(12):e269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Campbell CG, Foster RC, Lanningham-Foster LM, et al. The modified obstetric metabolic equivalent (MET): finding a MET that fits in pregnancy. J Dev Orig Health Dis. 2012;3(3):159–165. [DOI] [PubMed] [Google Scholar]
- 13. Chasan-Taber L, Freedson PS, Roberts DE, et al. Energy expenditure of selected household activities during pregnancy. Res Q Exerc Sport. 2007;78(2):133–137. [DOI] [PubMed] [Google Scholar]
- 14. Ainsworth BE, Haskell WL, Herrmann SD, et al. 2011 compendium of physical activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011;43(8):1575–1581. [DOI] [PubMed] [Google Scholar]
- 15. Health NCINI of . ACT24 Activities Collected over Time over 24-hours instrument. https://act24.cancer.gov/act24/researcher/login. Accessed November 7, 2022.
- 16. Matthews CE, Berrigan D, Fischer B, et al. Use of previous-day recalls of physical activity and sedentary behavior in epidemiologic studies: results from four instruments. BMC Public Health. 2019;19(suppl 2):1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Marshall MR, Montoye AHK, George AJ. Pregnancy walking cadence does not vary by trimester. Gait Posture. 2018;65:81–85. [DOI] [PubMed] [Google Scholar]
- 18. McAvoy CR, Moore CC, Aguiar EJ, et al. Cadence (steps/min) and relative intensity in 21 to 60-year-olds: the CADENCE-adults study. Int J Behav Nutr Phys Act. 2021;18(1):1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tremblay MS, Aubert S, Barnes JD, et al. Sedentary Behavior Research Network (SBRN)—terminology consensus project process and outcome. Int J Behav Nutr Phys Act. 2017;14(1):1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Choi L, Liu Z, Matthews CEBM. Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc. 2011;43(2):357–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Migueles JH, Cadenas-Sanchez C, Ekelund U, et al. Accelerometer data collection and processing criteria to assess physical activity and other outcomes: a systematic review and practical considerations. Sports Med. 2017;47(9):1821–1845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Staudenmayer J, He S, Hickey A, et al. Methods to estimate aspects of physical activity and sedentary behavior from high-frequency wrist accelerometer measurements. J Appl Physiol. 2015;119(4):396–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Hildebrand M, Van Hees VT, Hansen BH, et al. Age group comparability of raw accelerometer output from wrist-and hip-worn monitors. Med Sci Sports Exerc. 2014;46(9):1816–1824. [DOI] [PubMed] [Google Scholar]
- 24. Montoye AHK, Clevenger KA, Pfeiffer KA, et al. Development of cut-points for determining activity intensity from a wrist-worn ActiGraph accelerometer in free-living adults. J Sports Sci. 2020;38(22):2569–2578. [DOI] [PubMed] [Google Scholar]
- 25. Doherty AR, Kelly P, Kerr J, et al. Using wearable cameras to categorise type and context of accelerometer-identified episodes of physical activity. Int J Behav Nutr Phys Act. 2013;10:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Martinez J, Decker AE, Cho CC, et al. Validation of wearable camera still images to assess posture in free-living conditions. J Meas Phys Behav. 2021;4(1):47–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kendall MG. Rank Correlation Methods. 4th ed. London, UK: Griffin; 1970. [Google Scholar]
- 28. Terwee CB, Mokkink LB, Van Poppel MNM, et al. Qualitative attributes and measurement properties of physical activity questionnaires: a checklist. Sport. Med. 2010;40(7):525–537. [DOI] [PubMed] [Google Scholar]
- 29. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Int J Nurs Stud. 2010;47(8):931–936. [PubMed] [Google Scholar]
- 30. Stoffel MA, Nakagawa S, Schielzeth H. partR2: partitioning R2 in generalized linear mixed models. PeerJ. 2021;9:1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Fitzmaurice GM, Laird N, Ware J. Applied Longitudinal Analysis. Hoboken, NJ: John Wiley & Sons; 2004. [Google Scholar]
- 32. Aadland E, Ylvisåker E. Reliability of objectively measured sedentary time and physical activity in adults. PloS One. 2015;10(7):1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Davies A, Allman-Farinelli M, Owen K, et al. Feasibility study comparing physical activity classifications from accelerometers with wearable camera data. Int J Environ Res Public Health. 2020;17(24):1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Matthews CE, Moore SC, George SM, et al. Improving self-reports of active and sedentary behaviors in large epidemiologic studies. Exerc Sport Sci Rev. 2012;40(3):118–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Doherty AR, Hodges SE, King AC, et al. Wearable cameras in health: the state of the art and future possibilities. Am J Prev Med. 2013;44(3):320–323. [DOI] [PubMed] [Google Scholar]
- 36. Haakstad LAH, Voldner N, Henriksen T, et al. Physical activity level and weight gain in a cohort of pregnant Norwegian women. Acta Obstet Gynecol Scand. 2007;86(5):559–564. [DOI] [PubMed] [Google Scholar]
- 37. Adams EJ, Goad M, Sahlqvist S, et al. Reliability and validity of the Transport and Physical Activity Questionnaire (TPAQ) for assessing physical activity behaviour. PloS One. 2014;9(9):e107039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Aittasalo M, Pasanen M, Fogelholm M, et al. Validity and repeatability of a short pregnancy leisure time physical activity questionnaire. J Phys Act Heal. 2010;7(1):109–118. [DOI] [PubMed] [Google Scholar]
- 39. Haakstad LAH, Gundersen I, Bø K. Self-reporting compared to motion monitor in the measurement of physical activity during pregnancy. Acta Obstet Gynecol Scand. 2010;89(6):749–756. [DOI] [PubMed] [Google Scholar]
- 40. Brantsæter AL, Owe KM, Haugen M, et al. Validation of self-reported recreational exercise in pregnant women in the Norwegian Mother and Child Cohort Study. Scand J Med Sci Sport. 2010;20(1):1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Harrison CL, Thompson RG, Teede HJ, et al. Measuring physical activity during pregnancy. Int J Behav Nutr Phys Act. 2011;8(1):19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Sanda B, Vistad I, Haakstad LAH, et al. Reliability and concurrent validity of the International Physical Activity Questionnaire short form among pregnant women. BMC Sports Sci Med Rehabil. 2017;9(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Watson ED, Micklesfield LK, Van Poppel MNM, et al. Validity and responsiveness of the Global Physical Activity Questionnaire (GPAQ) in assessing physical activity during pregnancy. PloS One. 2017;12(5):1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Davis RE, Loprinzi PD. Examination of accelerometer reactivity among a population sample of children, adolescents, and adults. J Phys Act Heal. 2016;13(12):1325–1332. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.