Theory-driven, web-based, computer-tailored advice to reduce and interrupt sitting at work: development, feasibility and acceptability testing among employees

In the present study, the development of a new theory-driven web-based, computer-tailored intervention aimed at influencing sitting at work was described. Furthermore, the feasibility of reaching employees and the acceptability of the intervention among its users was examined.

Concerning the feasibility of reaching employees with this intervention, positive findings are the relatively high response to the email invitation to receive personal advice (63 %) and the high proportion of interested individuals actually requesting the advice (80 %). In previous studies on computer-tailored physical activity advice, the initial response rates among the general population were 79 % when recruiting via telephone [59], 46 % when recruiting face-to-face [60], 6 % when recruiting via flyers [60] and 6 % when recruiting via general practitioners [38]. Actual advice request was also lower for these previous web-based computer-tailored interventions on physical activity (46–53 %) [59, 61] compared to the present study. Further research should confirm whether recruiting in workplaces via email, without any telephone or face-to–face contact, is indeed most effective in reaching employees for web-based, computer-tailored interventions. The present study group could possibly be a sample of already engaged employees as they previously participated in another study. Because of the select sample here, the actual number of employees open to participate and receive advice may be overestimated.

It was also encouraging that men and women, young and old individuals, those with low and high sitting levels, as well as those who scored low and high for the psychosocial correlates (self-efficacy, social support and ‘intention to change’) towards interrupting sitting at work visited the website. In a computer-tailored physical activity program, more women than men and more adults of medium socio-economic status (SES) than adults of low SES participated in the program [62]. Here, advice request was however predicted by two socio-demographic variables (education and employment status), one behavioural variable (interruption in sitting) and one psychosocial correlate (attitudes). The fact that employees with a college or university degree were more likely to request the advice is not surprising as highly educated individuals may have more knowledge about and interest in health-related behaviours [63]. It is also possible that higher educated individuals are more digital media savvy or have access to the internet through more devices (e.g. laptop, tablet, smartphone), however most employees have computer access at work. Therefore everyone was able to request the advice. Still, particular attention is needed to reach more low educated individuals in the future. Especially as some may argue that access, affordability, usability and appropriateness of e-health interventions and digital health literacy may be lower among vulnerable or underserved groups, which may lead to increased health inequalities [64]. Maybe when more attention in the media [58] is directed towards the damaging health consequences of too much sitting, this group will also be interested in receiving advice. For physical activity for example, media attention has led to increases in population awareness and knowledge about this public health problem [65].

In the present study, full-time employees were less likely to request the advice compared to part-time workers, so it may be that ‘time’ was an issue. Focus group interviews among employees and managers also revealed that productivity concerns and loss of time were barriers for the implementation of intervention strategies to reduce or interrupt sitting at work [58]. Nevertheless, as full-time employees should be reached in the future, special attention to this group is also needed. Maybe it should be explicitly stressed that requesting and applying the advice is not time-consuming. The time investment to request the general advice (section 1) was no more than 8 min. For the other sections, no more than 4 min was needed to complete the assessment questionnaires. However, it is not clear whether only this amount of time was needed to complete the questionnaires properly or whether participants rushed through the questions to get the advice as quickly as possible.

Further, there was a trend that those who had less than 2 standing interruptions per hours were more likely to request the advice, which is positive. Finally, those with positive attitudes towards interrupting sitting were more likely to request the advice. A cross-sectional study among Australian employees showed that those who perceived more advantages of sitting less at work, reported more sitting at work [55]. This would mean that the group for whom the intervention is intended (employees with high and prolonged occupational sitting time) was indeed reached, which is promising. However, Australian part-time employees reported less total and work-related sitting time than full-time employees [66], making the latter group an at-risk group. So, in the future, also those with less positive attitudes and full-time employed individuals should be reached, which may also be achieved with more (media) attention for the concept of sitting too much [58].

Regarding the acceptability of the advice, the assessment questionnaire prior to receiving the advice was evaluated positively in terms of understandability, logic, length, credibility and lay-out, however attention is needed concerning the ease of answering the assessment questionnaires. The questions were based on validated instruments, however sitting may be an automated and habitual behaviour [67], which may be hard to recall.

Further, while the authors initially assumed it would be a strength to split up the advice into separate sections to be requested when suitable for the users, only 8.5 % of the employees thought it was useful to provide the advice in separate sections. No additional questions were asked about the reason for this response. It was assumed that the freedom of choice in whether and when to request additional sections would support an individual’s feeling of autonomy, one of the three SDT basic needs allowing an individual to act out of autonomous motivation [42, 43]. In addition, providing all information at once would make the assessment questionnaire prior to receiving the advice very long (a total of maximum 116 questions would have to be completed) and the advice itself would also be much longer (while the length of the advice in its current format was evaluated positively). As such, it’s likely that even fewer users would read the (whole) advice. Still, it should be acknowledged that only 39 % requested other sections of the advice. More efforts are needed to motivate people to request the other sections, maybe by stressing the difference from the general section and the additional value of the sections. Perhaps the number and frequencies of the reminder emails should be increased as well. Alternatively, the different parts of the advice could be released on a weekly basis, rather than all at once. However, this undermines the principles of the SDT [42, 43], as the users have no autonomy in deciding when to request other sections. In addition, a structured weekly release runs the risk of few people returning to the website on multiple occasions. In the current study, about half of the participants (49 %) visited the website on more than one occasion.

It is also not promising that only 21 % of the users completed an action plan, as the SRT assumes that transforming implementation-intentions into specific actions (through an action plan) results in behaviour change [56, 57]. Nonetheless, the action plan scored the highest on relevance and motivation among those completing the acceptability questionnaire. In the future, it should be good to stress the usefulness of an action plan to employees willing to change their behaviour. In other studies, the proportion of users completing an action plan was not reported, but the acceptability of it was also found to be high [36, 37]. Future studies can test whether the effects in behaviour change are indeed larger in employees completing an action plan compared to those who did not.

Concerning the use of the general advice (section 1), 83 % of the employees reported they read it and 20 % discussed the advice with others. This is lower compared to other computer-tailored physical activity (96–100 % read it, 59–64 % discussed it with others) and step (96 % read it, 42 % discussed it with other) advice [37, 38]. Today, physical activity promotion programs, such as the 10,000 Steps, are well-known and popular in the general population, which does not yet apply for the ‘sitting-too-much-is-bad-for-you’-concept in Flanders [58]. It is not encouraging that only one out of five interacted with others about the advice, as group coherence and social support have been expressed as essential in order to transform challenging strategies into action and goal realization [68]. One would expect that the possibilities for peer support at work would increase through this type of intervention, however, this was not the case, maybe because of the fact that employees individually requested the advice. Discussing the advice with others could also overcome barriers, such as the awkwardness of standing or the fear of disturbing others while standing, which were reported by employees and executives in focus group interviews [58].

Overall, the majority of the sample evaluated the advice sections as credible, interesting and relevant, which is comparable to previously developed computer-tailored physical activity and step advice [37, 38]. The low scores on practicability (except for section 4, sitting during commuting) may be due to the fact that interventions to influence sitting are fairly new. Employees may perceive the suggested strategies as hard to apply because they are not familiar with this concept, nor are their managers. The lowest applicability score was found for section 3 (replacing sitting at work by standing). This was also reflected in the ‘intention to change’ scores reported after receiving the advice, as only 14 % was already replacing sitting by standing and only 16 % was certain they could succeed in starting/continuing this. In contrast, more than half reported (58 %) they were already frequently interrupting their sitting at work and another 33 % intended to do so. About 80 % also believed they were able to continue/accomplish this. These important findings suggest that interrupting sitting is more achievable to implement than replacing longer periods of sitting by standing at work. The health promotion message “good STUFF” (Stand Up For Fitness) suggested by Rutten et al. (2013) [52] is also mainly encouraging interruptions in long sitting periods by short breaks. However, the majority of interventions in this area have focussed on reductions in total sitting time [16‐23]. For example, the intervention “Stand Up, Sit Less, Move More”, encourages to use workstations in order to have equal amounts of sitting and standing time, to stand up at least every 30 min, and to increase incidental physical activity [19, 23]. The latter intervention showed promising short-term effects for sitting and standing outcomes. Further research should investigate which strategies to influence sitting at work are most feasible, acceptable and effective among employees.

The overall strength of this research is the web-based approach of influencing sitting at work. This computer-tailored approach was theory-driven and the advice was based on current evidence available in the literature. This type of intervention has not been used or tested for feasibility and acceptability before. Second, the assessment questionnaires were based on or similar to existing validated questionnaires [40, 46‐48, 53‐55] and the questionnaire evaluating the acceptability of the advice was also based on existing measures [36, 38]. There are, however, some limitations to take into account. First, the mainly female and highly educated sample was recruited within only one company that had already participated in another study and may consequently be more interested in health-related behaviours and research. Furthermore, the ‘intention to change’ towards short standing breaks was also high. For these reasons, this sample may not be representative for the broader working population, which compromises generalizability of the findings. In the future, feasibility and acceptability studies should be done in several companies, covering a large variety of employees. Second, all data were self-reported which may cause recall and social desirability biases. The use of objective measures, such as accelerometers or inclinometers, could limit these biases. The self-reported sitting at work of the present sample was on average circa 6.8 h/day, which is in line with a previous study using self-reports in an Australian convenience sample (6.5 h/day) [46]. However, this is higher than occupational sitting reported in a Danish population-based sample (4.4 h/day) [69] and in a sample of Australian workers (3.3 h/day) [70], and higher than objectively measured occupational sitting time in Australian office employees (4.3 h/day; 5.6 h/day) [19, 71]. Third, qualitative methods were not applied during intervention development, while including input from the target group, could have enhanced the intervention. For example, the perceived uselessness of separating the advice in several sections could have been discovered earlier. Also for the feasibility and acceptability testing, a mixed methods approach, including focus groups or interviews, could have resulted in more in-depth information about certain aspects, such as the high initial response rate and the relatively low proportion completing an action plan. In addition, through qualitative methods, the potential benefits of this advice in a multi-component intervention could have been discussed. Finally, for the feasibility study, the data were collected in November 2012. Some outcomes may have changed at the time the present feasibility study was conducted (February 2014). Future studies should use more recent data when predicting advice request.