Project Topics
Get Complete Project Material File(s) Now! »
Mode of sedentary behaviors
Type of sedentary behaviors refers to the mode of sedentary behaviors, such as watching TV, using a computer or driving a car. Often, time spent in TV viewing is used as a proxy measure of sedentary behaviors duration (Dunstan et al., 2005). Time-use surveys have reported that, aside from sleeping, watching TV was the behavior that occupies the most time in the domestic setting (Office for National Statistics, 2005; Australian Bureau of Statistics, 2006; United States Departement Labor, 2007). However, it has been suggested that TV viewing may not always be a robust marker of a sedentary lifestyle (Sugiyama et al., 2008, Owen et al., 2010). Therefore, all types of sedentary behaviors need to be measured.
Conceptual models of physical activity and sedentary behaviors
This section will present three conceptual models used in the field of physical and sedentary behaviors epidemiology that can be used to guide research. The work of LaMonte and Ainsworth (2001), Pettee-Gabriel and Morrow. (2012), and Chastin et al. (2013) will be presented following a chronological order. In addition, an ongoing project (ALPHABET project) will be presented.
Measurement model for physical activity and energy expenditure (LaMonte and Ainsworth, 2001)
In 2001, LaMonte and Ainsworth proposed a framework for measuring physical activity and energy expenditure, collectively referred to as human movement (Figure 4) (LaMonte and Ainsworth, 2001). This framework made the distinction between physical activity, as a behavior, and energy expenditure, as the energy cost of the behavior. The framework provides examples of measurement methods using direct and indirect measures of physical activity and energy expenditure. For physical activity, direct measures include motion sensors, direct observation and global positioning system. Indirect measures include physical activity records, 24-hour recalls and questionnaires. For energy expenditure, direct measures include calorimetry and doubly labeled water. Indirect measures include oxygen uptake, heart rate, body temperature and ventilation. For each measurement method, it is possible to extrapolate each metric to energy expenditure for use in analysis of energy expenditure and health outcomes.
Model for the physical activity domains (Pettee-Gabriel and Morrow, 2010)
In 2010, Pettee-Gabriel and Morrow, proposed a framework for human movement, representing physical activity and sedentary behaviors as two components of human movement (see Figure 5) (Pettee-Gabriel and Morrow, 2010). The framework makes the distinction between the behaviors (physical activity and sedentary behaviors) and the physiological results or consequences of movement (energy expenditure and physical fitness). The framework identifies four domains where physical activity can take place (leisure, occupation, household, and transport), and classifies sedentary behaviors as non-discretionary or discretionary. Examples of discretionary and non-discretionary sedentary behaviors are presented. Discretionary sedentary behaviors include sitting, media use, non-occupational, school and computer use. Non-discretionary sedentary behaviors include sleeping, occupation, school, sitting while driving and sitting while riding.
Measurement of physical activity and sedentary behaviors with questionnaires in surveillance systems
Physical activity and sedentary behaviors are complex behaviors; and their assessment is a challenge implying strategic choices. When selecting a measurement tool, one should determine which characteristics of physical activity or sedentary behaviors are of interest as it is unlikely that a tool measure all facets of a behavior, then one must consider which assessment method is best able to measure the characteristics of interest while minimizing bias. Ideally, the measurement is reproducible, valid, and responsive. Methods to measure physical activity and/or sedentary behaviors include subjective instruments (questionnaires, logs, ecological momentary assessment), and objective methods (motion and posture sensors, physiological sensors, direct observation, and context awareness using cameras and GPS). In the setting of surveillance studies, questionnaires commonly are used.
Classification of self-report tools of sedentary behaviors
Dall et al. developed a framework to help in the development, comparison and evaluation of self-report tools. The framework, named TASST for Taxonomy of Self-reported Sedentary Behavior Tool, consists in four domains: type of assessment, recall period, temporal unit and assessment period (Dall et al., 2017). The framework is represented in Figure 7. The type of assessment includes whether sedentary is measured using a single item or a composite item. For single item instrument sedentary time can be measured directly or using a proxy (such as TV viewing). For composite measures, the instrument can ask about the pattern of sedentary behaviors (i.e. how the behavior is accumulated throughout a given period), or the time spent in sedentary behavior can be estimated by summing the time spent in a range of different activities. When summed, the calculation can be made from questions asking about specific behaviors (for example, reading a book) or time spent in specific domains (for example, at home or at work). The recall period is the time frame over which the respondent is asked to consider his sedentary behaviors, and includes previous day, previous week, longer period, and unanchored (i.e. a general period of time such as a typical week). The temporal unit refers to the time frame, within the recall period, that the respondent is asked to report their sedentary behaviors, including single day, week and longer. The assessment period provides information regarding whether a respondent is asked questions about specific days (for example only weekend day) or specific time of a day (for example only morning). Authors mapped self-report instruments of sedentary behaviors to the TASST framework, and reviewed the psychometric properties (accuracy and sensitivity to change) of included instruments. By doing so, Dall and colleagues observed that tools assessing total sedentary time as a sum of behaviors seemed to provide better accuracy than single-item direct measurement tools (Dall et al., 2017), and tools with a previous day recall period tended to provide better accuracy than those with longer periods. Yet, the overall accuracy remained poor for all instruments reviewed, with both over-and under- estimation reported. As for sensitivity to change, almost no information were available.
Validity and reliability studies of the GPAQ
Because the GPAQ has been developed for use worldwide, the validity and reliability of the GPAQ has been investigated frequently in many languages (Bull et al., 2009; Trinh et al., 2009; Thuy et al., 2010; Hoos et al., 2012; Herrmann et al., 2013; Soo et al., 2015; Mumu et al., 2017; Wanner et al., 2017). Bull and colleagues were the first to undertake a validity study for the GPAQv1 (in its 19 items form) (Bull et al., 2009). Criterion validity was investigated in eight countries, against pedometers or accelerometers worn seven days to cover the same time frame than the recall period of the GPAQv1 (i.e., one week). Reliability was assessed by administering the questionnaire at two occasions, 3 to 7 days apart. The reliability and validity was assessed using Spearman’s rho coefficients due to the skewed distribution of the data. The pooled result from six countries showed a correlation between GPAQv1 and pedometers for total physical activity time of r = 0.31 (n=1507). Two countries used accelerometers to assess criterion validity by comparing measures of minutes of total moderate- and vigorous-intensity physical activity, and total sedentary time from the GPAQv1 with measures of time derived from accelerometer counts. Correlations for sedentary time ranged from r = -0.02 to 0.40, and the correlations for moderate and vigorous physical activity ranged from r = -0.03 to 0.23, and r = 0.26 to 0.23, respectively (Bull et al., 2009). Stronger correlations were reported for 3 to 7 days test-retest reliability, ranging from r = 0.67 for vigorous intensity leisure physical activity to r = 0.73 for sedentary and vigorous intensity physical activity at work (Bull et al., 2009).
The first version of the GPAQ with 19 items evaluated by Bull and colleagues (Bull et al., 2009) has been subsequently modified. The GPAQ in its second version1 has 16 items, as some items were deemed redundant and were removed (Armstrong and Bull., 2006). When tested against accelerometry, the GPAQ showed poor correlations, frequently below the threshold of r = 0.50, for moderate- and vigorous-intensity physical activity (Hoos et al., 2012; Herrmann et al., 2013; Mumu et al., 2017; Wanner et al., 2017). For test-retest reliability, correlations ranged from poor (test-retest recall frame = 3 weeks, r = 0.13 for women, r = 0.32 for men) (Thuy et al., 2010), moderate (tet-retest recall frame = 2 weeks, r = 0.69) (Trinh et al., 2009) and good (test-retest recall frame = 10 days, ICC = 0.83-0.92) (Herrmann et al., 2013) for physical activity and sitting time. Long-term reliability (2-3 months) was lower that short-term reliability over 10 to 14 days in Trinh and colleagues’ study (2 weeks test-retest: r = 0.50 to 0.74; 2 months test- retest: r = 0.32 – 0.68) and in Herrmann and colleagues’ study (10 days test-retest: ICC = 0.83 – 0.96; 3 months test-retest: ICC = 0.53 – 0.83) (Trinh et al., 2009; Herrmann et al., 2013).
Responsiveness is less frequently assessed than reliability and validity. Cleland and colleagues assessed the validity of the GPAQ when estimating changes in physical activity and sedentary behaviors over 3 to 6 months (Cleland et al., 2014). Participants worn an accelerometer (ActiGraph GT3X+) for seven days and completed the GPAQ on day 7, on two occasions, with an interval of 3 to 6 months. The extent of change from the first measurement (T1) to measurement 2 (T2) was assessed as the difference in moderate-to-vigorous physical and total sedentary time between measures at T1 minus T2. Spearman’s rho coefficient was calculated to assess correlation between the change scores derived from the two instruments (i.e. accelerometer and questionnaire). Results for agreement in change over time showed moderate correlation (r = 0.52, p = 0.12) for moderate-to-vigorous physical activity and poor correlation for total sedentary time (r = −0.024, p = 0.916) (Cleland et al., 2014).
Table of contents :
Chapter 1. Introduction
Context
Research aims and questions
Research studies
Thesis outline
Chapter 2. Literature review
1 Physical activity and sedentary behaviors: concepts and definitions
1.1 Definitions of physical activity
1.2 Terms used in the measurement of physical activity
1.2.1 Intensity
1.2.2 Duration
1.2.3 Frequency
1.2.4 Mode
1.3 Definition of sedentary behaviors
1.4 Terms used in the measurement of sedentary behaviors
1.4.1 Sedentary time
1.4.2 Interruptions in sedentary time
1.4.3 Frequency of sedentary bouts
1.4.4 Mode of sedentary behaviors
1.5 Conceptual models of physical activity and sedentary behaviors
1.5.1 Measurement model for physical activity and energy expenditure (LaMonte and Ainsworth, 2001)
1.5.2 Model for the physical activity domains (Pettee Gabriel and Morrow, 2010)
2 1.5.3 Taxonomy of sedentary behaviors
1.5.4 ALPHABET project
1.6 Summary
2 Measurement of physical activity and sedentary behaviors with questionnaires in surveillance systems
2.1 Classification of self-report tools of sedentary behaviors
2.2 Type of questionnaires
2.3 Measurement properties of questionnaires
2.3.1 Validity
2.3.2 Reliability
2.3.3 Responsiveness
2.4 Validity and reliability studies of the GPAQ
2.5 Summary
3 Public health surveillance
3.1 Definition and concepts
3.2 Objectives of public health surveillance
3.3 Types of surveillance systems
3.4 Historical overview of WHO non-communicable diseases surveillance
3.5 Surveillance of physical activity and sedentary behaviors
3.5.1 Worldwide surveillance
3.5.2 National surveillance
3.6 Summary
Chapter 3. Personal contribution
Study 1. Surveillance of physical activity and sedentary behaviors: case-study using French surveillance data
Study 2. Results from the first French Report Card on physical activity for children and adolescents (2016)
Study 3. Reliability and validity of the French version of the global physical activity questionnaire
Study 4. Content comparison of sedentary behaviors questionnaires: a systematic review 108
Chapter 4. General discussion
3 1 Study 1. Surveillance of physical activity and sedentary behaviors: case-study using French surveillance data
1.1 Main results
1.2 Discussion
1.3 Strengths and limitations
2 Study 2. Results from the first French Report Card on physical activity for children and adolescents (2016)
2.1 Main results
2.2 Discussion
2.3 Strengths and limitations
3 Study 3. Reliability and validity of the French version of the global physical activity questionnaire
3.1 Main results
3.2 Discussion
3.3 Strengths and limitations
4 Study 4. Content comparison of sedentary behaviors questionnaires: a systematic review
4.1 Main results
4.2 Discussion
4.3 Strengths and limitations
5 Perspectives
5.1 Public health perspectives
5.2 Research perspectives
6 Conclusion
References
