Prolonged and uninterrupted sitting is increasingly recognised as a risk factor for ill health. As a result, it is imperative to limit the volume of sitting in populations that engage in high levels of sitting to improve the health of these populations [7, 39, 40]. White collar and professional occupation groups currently engage in high levels of occupational sitting [33, 41, 42] and the high level of sitting is likely a function of the social and environmental characteristics of the workplaces these groups occupy . Thus, it is important to understand how these factors influence sitting within the office environment and to do so within behaviour and setting-specific Ecological Models [19, 43, 44].
This study described the development and measurement properties of a self-report instrument, the OFFESS, to examine spatial characteristics of office environments as potential correlates of occupational sitting. Outcomes of this study suggest that the individual scales of the OFFESS hold acceptable levels of internal consistency and substantial to excellent levels of test-retest reliability. Additionally, it appears that the OFFESS has good construct validity as the individual scales vary by office type in theoretically expected directions and the correlations between individual scales are comparable in magnitude to those observed when using objective measures of Space Syntax [9, 12]. Furthermore, significant associations between all OFFESS scales and either one or both of the occupational sitting behaviours were observed in all office types. Significant associations between local connectivity and the frequency of breaks (private enclosed and shared office types), overall connectivity and the duration of sitting (open plan office types), co-worker visibility and the frequency of breaks (open plan offices) and the proximity of co-workers and the frequency of breaks (private enclosed office types) were observed in hypothesised directions. The magnitude of these associations were modest which is consistent with way in which environments are thought to influence activity behaviours - small changes to the behaviours of all people in the environment over long periods of time [44–46]. The associations observed between office spatial configuration and sitting behaviours highlights the need to continue to examine these factors to better understand how to modify occupational sitting behaviours to improve the health of workers.
The internal consistency, test-retest reliability and associations between the OFFESS and occupational sitting behaviour suggest that the OFFESS is a useful instrument to examine spatial configurations of office environments as potential correlates of occupational sitting. However, it must be noted that the direction of several of the significant associations observed were counter to that expected under the hypothesis that greater connectivity (local and proximity) and greater visibility and proximity of colleagues would drive reduced duration of occupational sitting and more frequent breaks in sitting. Examples of the counter expected relationships include associations between frequency of breaks in sitting and local connectivity in open plan offices, sitting duration, break frequency and overall connectivity in private enclosed offices, and sitting duration and co-worker proximity in open plan offices. Potential reasons for these counter intuitive associations may be related to the cultural norms of workplaces or the inherent job characteristics. For example, inverse associations between break frequency and local connectivity in open plan offices could be impacted by employees in these offices minimising breaks to reduce disruptions caused to other employees, a commonly reported annoyance in open plan offices . Alternatively, cultural norms such as the perceived need to be at the workstation in order to be viewed as productive or a mutual surveillance effect may have impacted the associations [8, 48]. Workplace norms and policy, towards electronic communication may also impact these associations in workplaces where electronic forms of communication are encouraged for record-keeping purposes, which may be a stronger influence on behaviour than spatial configuration. Cultural norms and communication mode preference and policies were not examined in this study and highlight the need to examine occupational sitting in a way that is cognisant of the potential multiple influences of sitting behaviour. Private offices are typically assigned to higher status employees or employees required to conduct more individualised job tasks. While our analyses adjusted for occupational level to minimise the impact of these issues, the associations between the duration and frequency of breaks in sitting and overall connectivity in counter expected directions suggest that these issues may have still confounded these associations and that more refined measures of job tasks should be used in future research.
Proximity of co-workers was positively associated with duration of sitting in open plan office types. This is both in agreement and in contrast to previous studies examining proximity and movement or face-to-face interaction [8, 16, 18]. Rashid and colleagues reported that closeness, a variable related to proximity, was inversely associated with frequency of face-to-face interactions and positively associated with sedentary workstations activities . A potential explanation for this was mutual surveillance between co-workers , and this may partially explain the results observed in the current study. Alternatively, it may be that co-workers in very close proximity communicate without leaving their workstations; this is supported by evidence that conversation related distractions are frequent in open plan offices . This issue should be explored further, as a threshold effect may occur as very proximally located employees can interact face-to-face without moving from their workstation, while after some (as yet unknown) distance face-to-face interaction requires one or all parties to move from their workstation. The way in which the breaks in sitting variable was operationalized may have also impacted the associations observed, as the interruptions to sitting driven by office spatial configuration may include breaks shorter than one minute in duration. The one minute threshold was used in this study to enable comparison to breaks in sedentary time measured by accelerometer  and to assist in recall. Since this item was developed, a measure of breaks in sitting without a minimum time threshold has been developed and tested , which may be more suited to understanding how office environments influence sitting behaviour. Research examining how spatial configuration impacts sitting behaviour using alternate measures of sitting behaviour and in more diverse populations will assist in clarifying the way in which office configuration influences sitting behaviours. Objective measures of sitting could be useful in further understanding these relationships as they will allow more detailed and accurate measures of sitting behaviour to be examined.
A key requirement in the development of the OFFESS was to take four spatial configuration characteristics of office environments (connectivity, integration, visibility of co-workers, and proximity of co-workers) that are related to occupant movement patterns as measured by Space Syntax methodologies and create four self-report scales that reflected each spatial configuration characteristic. These four Space Syntax measured constructs are frequently observed to be lower in Private Enclosed offices compared to Open Plan Offices [9, 22], and the respective OFFESS scales, all displayed lower average scores in Private Enclosed offices compared to Open Plan Offices. The differences between OFFESS scales and office types were statistically significant with the exception of Overall Connectivity; however, in all instances differences were in the theoretically expected directions. Additionally, the internal consistency of individual OFFESS scales observed in the phase two population approached a level considered to be acceptable, providing some indication of the robustness and generalisability of the scales. So, while the criterion validity of the OFFESS was not examined, the pattern of differences between office types, associations between the OFFESS and sitting behaviours, and the level of internal consistency demonstrated in a separate sample provides some evidence of the construct validity. Comparison to a criterion measure of OFFESS constructs was not possible in the current study as detailed floor plans required for this analysis could not be obtained due to restricted access to some offices.
The OFFESS scales demonstrated substantial to excellent levels test-retest reliability (ICC = 0.70 - 0.87) indicating that its scales are stable over time; this is important for an instrument assessing environmental characteristics that are relatively static over time. The ICCs exhibited by the OFFESS are lower than those observed by the Perceived Workplace Environment Policy Scale (ICC = 0.97) that assessed individual, social, environmental, and policy-level characteristics of the workplace . Although both the Perceived Workplace Environment Policy Scale and the OFFESS examine workplace environments, given the differences in the characteristics assessed by each of these instruments, the level of test-retest reliability between these instruments is not directly comparable. However, when comparing OFFESS to surveys that assess neighbourhood environments, examining similar spatial characteristics such as street connectivity, and service proximity, (e.g. access to shops), the OFFESS demonstrates higher levels of test-retest reliability than the Neighbourhood Environment Walkability Scale (ICC = 0.58 – 0.80) and the Physical Activity Neighbourhood Environment Survey (ICC = 0.64 – 0.84) [21, 51]. This suggests that the OFFESS has levels of test-retest reliability comparable to other self-report instruments assessing related environmental characteristics.
Although the measurement properties of the OFFESS examined in the current study are encouraging, further examination of its measurement properties in larger samples and those working in varying office types would provide greater confidence in the measurement properties of the OFFESS. The classification of office types; used in the current study did not encompass offices that have a mixture of office types, for example, offices with open plan configurations in the centre and private or shared offices on the perimeter of the office. It is unknown how respondents of the current study who worked in this type of office responded to the item used to classify office type. Additionally, we acknowledge the time between repeat surveys in phase two was relatively small and future studies may seek to examine the reliability of the instrument over longer time periods and also its sensitivity to changes in the workplace environment. The OFFESS does not capture information on the unique characteristics of an individual workstation, such as seated, standing, or height adjustable; this is important to capture in future studies examining this topic as use of height adjustable workstations increases and is associated with reduced sitting . Nor does the OFFESS capture information on the presence of destinations in the office environment (i.e. amenities, kitchens, toilets, cafeterias or meeting rooms). The availability, distance, and frequency of travel are factors that may impact upon how office destinations influence sitting behaviour. Greater insight is needed regarding the interrelationships of these factors to inform the development and implementation of such a measure.
The current study used an online administration of the instrument, yet we are confident that the instrument could be successfully administered in either paper-based or telephone-based surveys. Associations between sitting behaviours and the potential behavioural correlates should be examined within ecological models that are domain-specific [19, 44]. Thus, a strength of the OFFESS is that it can be used to provide domain-specific associations between environments and behaviours for individual’s employed in office settings. Sitting time items were developed specifically for this study and demonstrated acceptable psychometric properties, and the level of test-retest reliability for the duration of sitting time item is superior to that compared to other measures of occupational sitting time . The measure of breaks in sitting time used a different recall period (daily vs. hourly) compared to a similar measure , yet the frequency of breaks in sitting during the work day appears similar. Comparison of the sitting time measures used in this study to objective measures of sitting time (i.e. ActivPAL) or direct observation should be conducted to establish the validity of these measures.