This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Sedentary behavior has received much attention in the scientific community over the past decade. There is growing evidence that sedentary behavior is negatively associated with physical and mental health. However, an in-depth understanding of the social and environmental context of sedentary behavior is missing. Information about sedentary behavior, such as how everyday sedentary behavior occurs throughout the day (eg, number and length of sedentary bouts), where, when, and with whom it takes place, and what people are doing while being sedentary, is useful to inform the development of interventions aimed at reducing sedentary time. However, examining everyday sedentary behavior requires specific methods.
The purpose of this paper is (1) to introduce sedentary behavior–triggered Ecological Momentary Assessment (EMA) as a methodological advancement in the field of sedentary behavior research and (2) to examine the accuracy of sedentary behavior–triggered EMA in 3 different studies in healthy adults. Moreover, we compare the accuracy of sedentary behavior–triggered EMA to simulations of random-trigger designs.
Sedentary behavior–triggered EMA comprises a continuous assessment of sedentary behavior via accelerometers and repeated contextual assessments via electronic diaries (ie, an application on a smartphone). More specifically, the accelerometer analyzes and transfers data regarding body position (a sitting or lying position, or an upright position) via Bluetooth Low Energy (BLE) to a smartphone in real time and triggers the deployment of questionnaires. Each time a participant spends a specified time (eg, 20 minutes) in a sedentary position, the e-diary triggers contextual assessments. To test the accuracy of this method, we calculated a percentage score for all triggered prompts in relation to the total number of bouts that could trigger a prompt.
Based on the accelerometer recordings, 29.3% (5062/17278) of all sedentary bouts were classified as moderate-to-long (20-40 minutes) and long bouts (≥ 41 minutes). On average, the accuracy by participant was 82.77% (3339/4034; SD 21.01%, range 71.00-88.22%) on the study level. Compared to simulations of random prompts (every 120 minutes), the number of triggered prompts was up to 47.9% (n=704) higher through the sedentary behavior–triggered EMA approach. Nearly 40% (799/2001) of all prolonged sedentary bouts (≥ 20 minutes) occurred during work, and in 57% (1140/2001) of all bouts, the participants were not alone.
Sedentary behavior–triggered EMA is an accurate method for collecting contextual information on sedentary behavior in daily life. Given the growing interest in sedentary behavior research, this sophisticated approach offers a real advancement as it can be used to collect social and environmental contextual information or to unravel dynamic associations. Furthermore, it can be modified to develop sedentary behavior–triggered mHealth interventions.
“Sitting is the new smoking” or “Why a sedentary lifestyle is killing you”—these and similar headlines have received a high level of media attention in recent years. There is growing evidence that sedentary behavior is a behavioral risk factor for human health [
Several different definitions have evolved over the past decade [
Previous studies in the field of sedentary behavior research used self-reported methods such as questionnaires, which have limited validity and are prone to recall biases and social desirability [
The latest findings from Stamatakis and colleagues [
However, the epidemiological evidence of sedentary behavior' effects on health is incomplete [
Currently, thigh-worn accelerometers are the method of choice for measuring sedentary behavior accurately [
To the best of our knowledge, there is a lack of studies addressing the social and environmental contexts of sedentary behavior. Fortunately, with EMA, there exists an established approach to assess social and environmental context information in daily life [
The purpose of this paper is (1) to introduce sedentary behavior–triggeredd EMA as a methodological advancement in the field of sedentary behavior research and (2) to examine the accuracy of sedentary behavior–triggered EMA in 3 different studies among healthy adults. Moreover, we compared the accuracy of sedentary behavior–triggered EMA to simulations of random-trigger designs.
EMA, sometimes also called the Experience Sampling Method (ESM), is currently a state-of-the-art methodology for examining within-subject associations in behavioral relationships [
The idea of triggered EMA (or e-diaries) is not entirely new, as Ebner-Priemer and colleagues [
Examples of sedentary behavior–triggered Ecological Momentary Assessment (EMA) in everyday life.
We used the sedentary behavior–triggered EMA system in 3 different studies, aiming to examine the accuracy of this approach.
We recruited 57 university employees from the Karlsruhe Institute of Technology (KIT) in Germany between May and August 2017. Participants carried a smartphone (Motorola Moto G, Motorola Mobility LLC) and three Move 3 accelerometers for 5 consecutive days. Participants wore accelerometers during the entire measurement period, but not during sleep, swimming, and showering. The thigh-worn monitor and the smartphone were connected via BLE. Sedentary behavior–triggered EMA was used within a mixed sampling scheme. In particular, during the time period from 7:30 am to 9:30 pm, participants received sedentary behavior–triggered prompts (ie, after at least 30 minutes were spent in a sitting or lying position) and randomly triggered prompts at various time points. Since sedentary time is a highly prevalent behavior in daily life, triggered prompts may occur several times per day, which may increase participants' burden. A solution to minimize participants' burden is to implement time-out phases, in which researchers define a time period (eg, of 20, 30, or 40 minutes in duration) when the participants receive no EMA prompts after an answered EMA prompt. During this time-out phase, the study design inhibits EMA prompts, even if the sensor detects an event of uninterrupted sedentary time. In particular, in our first study, EMA prompts occurred no more than every 40 minutes. At each EMA prompt, participants were asked about their social (alone versus not alone) and environmental (home versus work versus leisure activities) contexts (
We recruited 97 individuals from the University of Konstanz in Germany between May and July 2019. Sedentary behavior was assessed for 4 consecutive days (Thursday to Saturday) using Move 3 accelerometers, which were coupled with smartphones (Motorola Moto G, Motorola Mobility LLC) via BLE. During the time period between 6:00 am and 10:00 pm, short questions were asked via the smartphone whenever the person sat for 20 minutes. We implemented a time-out phase of 20 minutes. At each EMA prompt, participants were asked about their social and environmental contexts (
We recruited 72 individuals from the University of Konstanz in Germany between January and March 2019. For 4 consecutive days (Monday to Thursday), participants wore a Move 3 accelerometer on their right thigh from the time they got up in the morning to the time they went to bed in the evening. The accelerometer was connected to a smartphone (Motorola Moto G, Motorola Mobility LLC) via BLE. Prior to the assessment, participants received an extensive briefing on the use of the smartphone and accelerometers and completed a paper-pencil questionnaire that included demographic variables (age, gender, and educational level). During the time period between 6:00 am and 10:00 pm, short questionnaires were asked via the smartphone whenever the person sat for 20 minutes (sedentary trigger). We implemented a time-out phase of 20 minutes. At each EMA prompt, participants were asked about their social and environmental contexts (
Data were collected anonymously, and the study fully conformed to the Declaration of Helsinki and the ethics guidelines of the German Psychological Society. Participants received detailed information regarding voluntary participation, the handling of the questionnaires, and the processing of their data, and they gave written informed consent according to the ethics guidelines of the German Psychological Society [
Study characteristics and Ecological Momentary Assessment (EMA) items.
Study characteristics and EMA items | Study 1 (N=57) | Study 2 (N=97) | Study 3 (N=72) | |||||
|
||||||||
|
Duration (days) | 5 | 4 | 4 | ||||
|
Days of the week | Wednesday-Sunday | Thursday-Sunday | Monday-Thursday | ||||
|
Valid participantsa, n (%) | 46 (81) | 73 (75) | 59 (82) | ||||
|
|
|||||||
|
|
Male | 19 (41) | 36 (49) | 31 (53) | |||
|
|
Female | 27 (59) | 37 (51) | 28 (47) | |||
|
Sedentary trigger (minutes) | 30 | 20 | 20 | ||||
|
Time-out phase (minutes) | Minimum: 40 |
20 | 20 | ||||
|
||||||||
|
Environmental context (response options) | |||||||
|
Social context (response options) |
a ≥ 2 days with ≥ 10 hours wear time.
The same technological system (the Move accelerometer and smartphone with Android operating system) was used in all 3 studies. Thus, from study preparation to data preprocessing, the study procedures (
First, the sampling scheme and forms (eg, questions about social and environmental context) were created by using the online platform movisensXS (moviesens GmbH). This step included all set-up, such as the selection of the study duration, specification of the trigger option (eg, triggering after 20 minutes or 30 minutes of sitting or lying), and implementation of the time-out triggers. Second, immediately before data collection, the study smartphone was connected to the movisensXS online platform by using the movisensXS app to download the sampling scheme and forms via an individual participant code. Third, the chosen trigger option (eg, triggering after 20 minutes of sitting) was calibrated to the selected body position (the lateral aspect of the right thigh) and connected to the smartphone via BLE by using the movisensXS app. Fourth, after data collection, the recorded raw acceleration data were processed in 1-minute intervals by using the manufacturers' software DataAnalyzer (verson1.13.5, movisens GmbH). During this step, a band-pass filter (0.25-11 Hz) automatically eliminated gravitational components or artifacts (eg, vibrations when cycling on a rough road surface or sensor shocks). This resulted in an Excel spreadsheet with a self-selected choice of parameters such as body position, movement acceleration intensity (MAI), or activity class. Fifth, the smartphone entries from the participants were downloaded from the movisensXS online platform. Sixth, all accelerometer and EMA files from different participants were synchronized and combined into a single data file using DataMerger (version1.8.0, movisens GmbH). Seventh, prior to the analyses, we parametrized sedentary-specific variables such as sedentary bouts while calculating the cumulated sum of the dichotomous variable body position (1= sitting/lying; 0= upright). Eighth, we excluded participants from the data set if they did not fulfill the wear-time criteria of at least 2 valid days of 10 hours of wear time per day [
Process of study preparation and data preprocessing.
To test the accuracy of sedentary behavior–triggered EMA, we calculated an accuracy score, which is the percentage of all triggered prompts in relation to the total number of all possible triggered prompts. In particular, we first calculated sedentary bouts based on the cumulative sum of the dichotomous variable body position (1= sitting/lying; 0= upright) that was recorded by the accelerometer, and categorized them into the following categories: short bouts (≤ 5 minutes), short-to-moderate bouts (5-19 minutes), moderate-to-long bouts (20-40 minutes), and long bouts (≥ 41 minutes). Second, since earlier studies [
Participants' characteristics (N=178).
Variable | Study 1a (n=46), |
Study 2a (n=73), |
Study 3a (n =59), |
|
Age, in years | 34.0 (9.6; 25-62) | 28.6 (11.6; 19-66) | 26.3 (8.5; 21-60) | |
|
||||
|
Female, n(%) | 27 (59) | 37 (51) | 28 (48) |
|
Male, n(%) | 19 (41) | 36 (49) | 31 (52) |
BMI (kg/m2) | 22.8 (3.3; 17.7-32.1) | 23.5 (3.0; 17.1-32.4) | —b | |
Total smartphone promptsc | 12.31 (1.86; 8-18) | 20.72 (7.85; 3-45) | 14.83 (5.32; 2-29) | |
Total triggered promptsc | 7.3 (2.99; 2-17) | 20.72 (7.85; 3-45) | 14.83 (5.32; 2-29) | |
Compliance (%)d | 79.3 (17.3; 22.2-100) | 43.41 (22.5; 4.6-100) | 47.49 (23.7; 6.8-93.1) | |
Wear-time accelerometer (hr/day)c | 13.6 (1.1; 10.8-16.1) | 14.39 (1.6; 10.2-18.6) | 13.7 (1.3; 10.7-16.5) | |
Physical activity of complete measurement period (milli g-force)c | 86.87 (22.14; 46-148) | 81.18 (24.78; 32-141) | 79.64 (20.26; 44-155) | |
Body position: sitting/lying (hr/day)c | 10.2 (1.6; 7.4-13.7) | 9.3 (1.9; 5.7-16.6) | 9.16 (1.42; 5.5-12.5) | |
Total number of short sedentary bouts (≤ 5 min)c | 11.1 (6.6; 0-29) | 10 (6.6; 0-48) | 11.9 (6.12; 4-39) | |
Total number of short-to-moderate bouts (6- ≤ 19 min)c | 6.3 (2.6; 0-12) | 7.1 (2.6; 1-13) | 8.6 (3.2; 3-18) | |
Total number of moderate-to-long bouts (20- ≤ 40 min)c | 3.5 (1.5; 0-7) | 4 (1.8; 0-12) | 3.8 (1.2; 2-7) | |
Total number of long sedentary bouts (≥ 41 min)c | 4.0 (1.4; 1-7) | 4 (1.4; 1-8) | 3.6 (1.2; 1-6) |
aNumber of monitoring days per study: Study 1=5 days, Study 2=4 days, Study 3=4 days.
b—not available.
cAggregated within the study day per participant.
dPercentage of answered Ecological Momentary Assessment prompts across each study sample.
Our additional analyses revealed that the sedentary behavior–triggered EMA in the mixed-sampling design of study 1 was 8.97% (n=78) and 20.83% (n=182) higher than that of a simulation of a random-trigger design with prompts every 90 minutes and 120 minutes, respectively. In study 2, the accuracy of the purely sedentary behavior–triggered EMA design was 34.42% (n=587) and 43.46% (n=741) higher than that of a simulation of a random-trigger design with prompts every 90 minutes and 120 minutes, respectively. In study 3, the accuracy of the purely sedentary behavior–triggered EMA design was 34.25% (n=501) and 47.88% (n=699) higher than that of a simulation of a random-trigger design with prompts every 90 minutes and 120 minutes, respectively. These results indicated that the sedentary behavior–triggered EMA system triggered more prompts compared to the simulations of random-trigger designs during moderate-to-long sedentary bouts, and thus, it increases the chance of getting social and environmental context information more often, especially during these kinds of sedentary bouts.
Accuracy of sedentary behavior–triggered Ecological Momentary Assessment (EMA). Left side: the amount of accelerometer-recorded sedentary bouts per study (black dots: sedentary bouts within the study period; grey dots: sedentary bouts outside of the study period). Right side: the amount of triggered EMA diaries (red dots: triggered sedentary bouts; black dots: not-triggered sedentary bouts).
Accuracy per study.
Measures | Study 1 | Study 2 | Study 3 |
Number of all moderate-to-long sedentary bouts (≥ 20 min) |
1450 | 1993 | 1614 |
Sedentary bouts prior to 6 am or 7:30 am | 19 | 70 | 7 |
Sedentary bouts after 9:30 pm or 10 pm | 98 | 218 | 147 |
Sedentary bouts > 20 - < 30 min | 464 | N/Aa | N/A |
Total number of bouts that could be triggered | 869 | 1705 | 1460 |
Triggered sedentary bouts | 617 | 1434 | 1288 |
Accuracy of used study design (%) | 71.00 | 84.11 | 88.22 |
Accuracy of 90 min. random triggered simulation (%) | 62.03 | 49.69 | 53.97 |
Accuracy of 120 min. random triggered simulation (%) | 50.17 | 40.65 | 40.34 |
aN/A: Not Applicable
In addition to the accuracy on the study level, we calculated the accuracy per participant.
Distribution of subject-level accuracy separated by sedentary behavior–triggered Ecological Momentary Assessment (EMA) design and simulations of random triggered designs of every 90 and 120 minutes.
Each time the participants responded to the prompt, they were asked about their current environmental and social context. Across all studies, participants answered 2001 EMA prompts, with an average of 11.57 (SD 7.07) prompts per participant. According to the results about the environmental context, participants reported across all studies that 39.98% (800/2001) of all moderate-to-long and long sedentary bouts occurred during work, 32.93% (659/2001) occurred while at home, 22.44% (449/2001) occurred during leisure activities, and 4.65% (93/2001) occurred during transport. According to the results about the social context, participants reported across all studies that in 56.27% (1126/2001) of all moderate-to-long and long bouts, they were not alone. Specifically, data from studies 2 and 3 revealed that the participants were mostly in the company of friends or family members.
In our additional analyses, we found a significant positive correlation (
Results of social and environmental context for each study sample.
Participant responses | Study 1, mean (SD; range) | Study 2, mean (SD; range) | Study 3, mean (SD; range) | ||||
Number of answered prompts | 11.02 (5.72; 2-26) | 11.87 (7.99; 1-30) | 11.78 (6.86; 1-25) | ||||
|
|||||||
|
Work | 55.40 (26.64; 0-100) | 19.09 (10.43; 0-100) | 54.42 (28.85; 0-100) | |||
|
Home | 24.92 (22.42; 0-100) | 51.07 (28.23; 0-100) | 14.41 (15.55; 0-67) | |||
|
Leisure | 18.79 (20.11; 0-92) | 22.73 (22.62; 0-100) | 25.99 (22.27; 0-86) | |||
|
Transport | 0.89 (3.03; 0-15) | 7.38 (13.78; 0-75) | 5.19 (9.82; 0-50) | |||
|
|||||||
|
Alone | 49.92 (24.90; 0-100) | 37.36 (30.61; 0-100) | 45.86 (27.73; 0-100) | |||
|
With colleagues | N/Aa | 12.18 (19.44; 0-83) | 17.94 (24.66; 0-100) | |||
|
With friends | N/A | 29.59 (29.06; 0-100) | 22.21 (22.26; 0-100) | |||
|
With family | N/A | 23.51 (29.30; 0-100) | 11.31 (19.02; 0-100) | |||
|
With strangers | N/A | 1.96 (7.82; 0-60) | 6.85 (13.33; 0-56) | |||
|
With others | N/A | 1.13 (4.80; 0-33) | 2.48 (6.95; 0-33) |
aFrequency percentage on a participant level for each study.
bN/A: not applicable.
This paper introduced sedentary behavior–triggered EMA as an innovative methodological advancement in the field of sedentary behavior research and assessed the accuracy of sedentary behavior–triggered EMA in 3 different studies of healthy adults. The results indicated that sedentary behavior–triggered EMA captured 82.77% (3339/4034) of all possible sedentary bouts from the different studies. Compared to simulations of random triggered prompts, our data revealed that the sedentary behavior–triggered EMA system triggered more prompts during moderate-to-long sedentary bouts, and thus it increases the chance for getting social and environmental context information more often, especially during these kinds of sedentary bouts. Overall, the results indicate that sedentary behavior–triggered EMA is an accurate method and allows the capture of “just-in-time” social and environmental context information of sedentary behavior bouts.
Sedentary behavior has received much attention in the scientific community over the past decade. However, in-depth knowledge about this invisible behavior is still missing [
In general, EMA is an established procedure for the assessment of intrapersonal and social and environmental contextual information, and it has been widely used in previous studies, for example, in the field of physical activity research [
These are the first studies that used a sedentary behavior–triggered EMA and that assessed social and environmental contextual factors during prolonged sedentary bouts. Sedentary behavior–triggered EMA enables researchers to gather relevant information related to the behavior in real-time. Moreover, sedentary behavior–triggered EMA can also be used to unravel dynamic associations. In particular, future researchers may be interested in discovering dynamic associations between sedentary behavior and possible antecedents and consequences, such as the association between sedentary behavior and time-varying constructs like mood, stress, or working memory. In such a study, it may be reasonable to combine triggered and random prompts to maximize the outcome variance. Furthermore, sedentary behavior–triggered EMA can be modified as a methodological system in a JITAI [
There are also some challenges when using sedentary behavior–triggered EMA. The accuracy depends on both technical stability and user compliance when participating. In particular, technical issues (such as the accelerometer stopping data recording, or the accelerometer and the smartphone losing their BLE connection or not reconnecting with each other) may hinder a functional system. Furthermore, the compliance and reliability of the participant with regards to carrying the smartphone throughout the study period is a critical aspect. For example, if the participant leaves the smartphone at home while he is going to work, the BLE connection would not be available, and the trigger system would not work. This may explain why the accuracy for some participants was very low in our studies. However, short-term disconnections might be a minor issue for future studies since the next generation of accelerometers can store temporary, online, calculated data and transfers that data to the smartphone after a reconnection. Another issue is that if the participant does not wear the accelerometer and puts the sensor on its side (for example, when in a sitting or lying position), this may lead to the incorrect detection of a prolonged sitting bout. A similar problem may occur if the participant did not wear the accelerometer according to the manufacturer's instructions. However, this could be corrected with valid nonwear time algorithms during offline calculations [
The results of 3 independent studies revealed that sedentary behavior–triggered EMA is an accurate method for collecting contextual information in daily life. The accuracy of this approach can vary as a function of the study design (eg, time-out triggers), technical stability (eg, connection between the smartphone and accelerometer), and compliance of the participants (eg, following study instructions). Given the growing interest in sedentary behavior research and the lack of knowledge about social and environmental circumstances surrounding sedentary behavior, this sophisticated approach can offer real advancement. Sedentary behavior–triggered EMA can be used to collect social and environmental contextual information or to unravel dynamic associations. Furthermore, it can be modified to develop sedentary behavior–triggered mHealth interventions.
Bluetooth Low Energy
Ecological Momentary Assessment
g-force
just-in-time adaptive intervention
movement acceleration intensity
metabolic equivalent
milli g-force
Sedentary Behavior Research Network
UEP receives consultancy fees from Boehringer-Ingelheim. MK, CN, and MG have no conflicts of interest to declare.