Cultural adaptation, translation and validation of the Spanish version of Past-day Adults’ Sedentary Time

The PAST questionnaire was developed by Clark et al. to provide contextual information about SB by domains [26]. This instrument has seven items and can be completed in approximately 20 min. The seven domains of SB are work, transportation, television, computers/Internet/video games (excluding work), reading (excluding work), hobbies, other purposes (not reported previously in the other items). Respondents are requested to provide the time in hours and minutes spent sitting or lying in those domains. The continuous time reported in the seven items is added to estimate total sedentary time.

These are some examples of the questions included in the original PAST questionnaire [26]: 1) “How long were you sitting at your workplace or working from home yesterday, including during meal and snack breaks?”; 2) “Thinking again of yesterday, please estimate the total time that you spent sitting to travel from one place to another. Please include sitting and waiting for transport. Do not include any time you were standing up while travelling or waiting”; 3) “Please estimate the total time you spent sitting or lying down to watch TV or DVDs or play games on the TV, such as play station yesterday? This includes if you watch TV in bed”; The other questions may be accesed from the original questionnaire in English in the following link http://​links.​lww.​com/​MSS/​A252.

The PAST has demonstrated moderate to good validity in Australian adult population (range = 0.57–0.78) compared with ActivPAL [26, 27]. It has also reported an average difference in total SB of 5–25 min per day vs the ActivPAL (0.8–5.2%) [26, 27], which is less than other methods (4–5 h) [23, 28].

The PAST questionnaire was adapted culturally and translated from English to Spanish according to standardised criteria [29]. Stage 1 of translation from English to Spanish was done by an informed translator (t1) and one not informed (t2). Then in a meeting (stage 2) versions t1 and t2 were synthesised, and any discrepancies resolved by the research team. This new version was back-translated to English (stage 3) from Spanish by native English translators (bt1 and bt2). The research group reviewed the translation to English (bt1 and bt2) and compared these results to the synthesis in Spanish (t1 and t2) to develop a preliminary version (Stage 4). This version was tested (stage 5) in a small sample of adults (n = 40, university staff and students), but not included in the sample for the validation process. General observations of the participants from the test were made for the review, analysis, and final design of the version in Spanish (stage 6). Regarding the cultural adaptation, examples in some of the questions were added to clarify to what they referred, without altering the understandability of the items. As a result, we included the following examples or clarifications at the end of the questions for some domains: 1) Item 4, Transport: “if you cycle for transportation, do not include that time either. If your job includes sitting in a vehicle, do not include that time (e.g., taxi driver).”; 2) Item 5, Watching television: “Only include the time when this was your main activity (e.g., if you watched TV and used your computer or ate, report only one).”; and 3) Item 7, Reading: “Remember, do not include reading time that has been part of any previously reported.” The final version of the PAST in Spanish is publicly available in this link https://​doi.​org/​10.​6084/​m9.​figshare.​13539056.​v1.

For this stage, workers over 18 years were invited to participate through printed news, telephone calls, and mailings to workplaces in Temuco, Chile. The only exclusion criterion being the use of a wheelchair. The validation sample included participants from different working sectors: public transport, schools, childcare, healthcare, higher education, sales and administration, and others.

The participants were visited twice in their place of work. The first visit was arranged to collect sociodemographic characteristics (age, weight (Seca 803, Germany), height (Seca 213, Germany), educational level, occupation, and self-perceived health), answer the PAST questionnaire and deliver an ActivPAL 3™μ accelerometer (ActivPAL, Pal Technologies Ltd., UK) (T1).

To assess the reliability of the PAST, the participants responded to the questionnaire during the first visit (T1), and a second visit (T2) that was completed a week later. The concurrent validity of the PAST was tested versus the measurement from the ActivPAL. Each participant received oral and written instructions on the wear of the ActivPAL for the following 7 days. Still, they were asked to perform their usual activities during the week of monitoring. The participants were also asked to complete a diary of accelerometer use and hours of sleep for the following 7 days, where they detailed their work schedules. The reference instrument for the validation was the ActivPAL. The ActivPAL is small (23.5 × 43 × 5 mm) and light (10 g). This device presents minimal differences with direct observation (0.1–2.8%) [13, 30]. Also, it is more sensitive to changes than other objective measuring instruments, like the ActiGraph GT3X [31, 32]. The device was sealed with a latex protector and adhered to the participant’s front right thigh with a transparent and hypoallergenic tape (Tegaderm™ Roll, 3M™) to provide a waterproof barrier. This allows its continuous 24-h wear for 7 days without needing to remove it for activities like showering or sleeping, for example.

After a week, the second visit was made, where the participant’s ActivPAL was removed, completed the PAST, and they were asked if they had difficulties or discomfort using the accelerometer (T2). Also, the participants were asked to report their most preferred instrument between the PAST and ActivPAL to measure SB in a future study.

The ActivPALs were started up and downloaded with ActivPal™ Professional Software, v7.2.32 Research Edition (Pal Technologies Ltd., UK). The data were extracted from the ActivPAL events file, which includes continuous intervals in seconds of activity, for subsequent integration with the participant’s diary. A semiautomatic filter was used to extract the time of use and the hours in which the participant was awake [33]. To be considered valid, the participant’s data had to include at least 10 h of daily use for 5 days of the week [34‐36]. Average times were calculated for seated/lying (i.e., SB), standing, stepping, and total transitions from sitting to standing for an average day, the previous day, and at work. Also, relative estimates for time spent in SB at work were calculated for both the ActivPAL and PAST. To be able to compare the data reported with the PAST, the SB time of the previous day was used.

To describe the study sample, measures of central tendency were used for continuous variables and percentages for the categorical variables. Comparisons of continuous variables between sexes were performed with T-test or Mann–Whitney U test, while categorical variables were compared with chi-square and proportion tests. For the analysis, stratifications by sex were performed due to women and men may accumulate SB differently throughout the day, thus affecting the self-reported estimations in the different contexts [37], including work [38]. Test-retest reliability of the PAST instrument was assessed by comparing the results obtained in T1 and T2 with the intraclass correlation coefficient (ICC). The criteria for ICC values classification were: < 0.50 as poor, 0.50- < 0.75 as moderate, 0.75- < 0.90 as good, and ≥ 0.90 was considered as excellent [39]. The concurrent validity between the PAST questionnaire and the ActivPAL was evaluated with Pearson’s (total time in SB and at work) and Spearman’s (SB percentage from work) correlations and the Bland-Altman method to estimate mean bias and limits of agreement (LoA) [40]. The criteria to evaluate the correlation between sitting time estimated by the PAST and the reference were: 0.00–0.19 as very weak, 0.20–0.39 as weak, 0.40–0.59 as moderate, 0.60–0.79 as strong, and 0.80–1.0 as very strong [41]. Paired T-test and Wilcoxon signed-rank test were used to compare total SB estimates between the PAST and ActivPAL.

The Kappa method was used to determine the agreement between tertiles, quartiles, and quintiles of SB as classified by the PAST and the ActivPAL. The following criteria were used to assess the agreement: 0.01–0.20 as none to poor, 0.21–0.40 as low, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.00 as almost perfect agreement [42]. All the calculations were done with a 95% confidence level. All data cleaning and analysis were done with Stata, version 15 (StataCorp. College Station, TX, USA).

The PAST questionnaire showed moderate reliability for estimating total SB in the total sample (Table 4), while the reliability was good in men and excellent in women. The reliability was moderate when assessing time spent sedentary at work.

The SB estimates derived from both the PAST and the ActivPAL showed no differences for the total time per day, and at work, including its per cent contribution from SB to the whole day (Table 4). The PAST showed a non-significant correlation with ActivPAL for the total SB (Table 4), with an average bias of about 1 h (54.9 min/day, 95% LoA -484.3, 594.2) with the ActivPAL (Fig. 1). The women presented a mean bias of almost 1 h more than that observed in the men (85.3 [95% LoA: − 413.6, 584.2] vs 28.8 [95% LoA: − 544.8, 601.4] min/day). The agreement between the PAST and ActivPAL to classify the mean sedentary time into tertiles, quartiles, and quintiles was not significant (Table 4).

Overall, the PAST showed a weak correlation with the ActivPAL for the total sedentary time at work (Table 4). The mean bias between the PAST and ActivPAL was about half an hour per day for SB during working hours, with similar differences for men and women (Fig. 1). The agreement between the PAST and ActivPAL to classify the time spent in SB during working hours was poor to low when tertiles, quartiles and quintiles were used (Table 4).

When comparing percentages of work time spent sedentary between the PAST and ActivPAL, a weak correlation was observed (Table 4). The average bias was 2.0% per day between the PAST and the ActivPAL (Fig. 2). No agreement between the PAST and ActivPAL was found to classify the proportion of time spent in SB at work into percentile categories (Table 4).

This study was the first to validate the PAST questionnaire in Spanish. Although the recruitment was not probabilistic, it managed to include participants from various work environments and with a wide age range. This improves the external validity of the findings reported in this study about a working adult population. A major limitation of the study was that despite 101 participants were recruited, only 77 were included in the analyses, mainly for not wearing the accelerometer (i.e., enough data on at least 5 days). Nevertheless, there were no differences in the demographic characteristics between those included and not included in the study. This instrument only evaluates the SB from the last day, which could limit its use to estimate common SB that could be more variable, particularly in free time [47]. Although the PAST assesses SB in different domains, apart from the whole day, we were able to explore only the time spent sedentary at work. This remains an issue in the field due to the difficulty of assessing SB in different contexts [14], so the use of cameras or logs as reference standards may be useful tools to fill this gap in future studies.