Introduction

Globally, low back pain (LBP) is the leading cause of years lived with disability and accounts for more lost work days than any other musculoskeletal disorder; yet only 58% of people with LBP seek care [1,2,3]. Further, over 85% of the United States (US) population will experience low back pain (LBP), accounting for more than $100 billion per year in indirect and direct costs [1,2,3,4]. The majority of LBP cases are classified as “non-specific” LBP, in which there is no direct known cause, and thus, no direct long-term treatment [4, 5]. There are several recognized lifestyle factors contributing to LBP. Primarily, prolonged periods of uninterrupted sitting result in neuromuscular and biomechanical postural changes that lead to compensation in the lumbo-pelvic hip complex (LPHC), eventually leading to LBP [4]. This phenomenon has been classified as Lower Crossed Syndrome which results in muscular imbalances and increased stress on the spine and associated soft tissues [6, 7]. The average adult spends over 50% of their waking hours sitting and the average working adult spends about 75% of their hours sitting during an office job (i.e., 6 h of an 8 h work-day) [8]. Furthermore, recent studies indicate that the COVID-19 pandemic, transition to remote work, and “shelter-in-place” policies have significantly increased sitting time [9, 10]. A recent review of 39 studies in working adults found that 76% of the studies reported an overall decrease in physical activity and an increase in sedentary behavior, as compared to pre-COVID-19 observations [11]. Not only has sitting time increased, employees who now work remotely likely have a less favorable ergonomic work set-up which may compound the likelihood of poor posture and subsequent LBP [12, 13].

Current studies, in several different countries, examining workplace sitting time have primarily focused on metabolic outcomes (e.g., cholesterol, glucose, and blood pressure) [14,15,16,17,18,19,20,21,22]. Limited cross-sectional studies have shown an association between LBP and prolonged sitting, but have used self-report measures, for either one or both outcomes (i.e., LBP and/or sitting time). Additionally, none of these studies have examined the effects of sitting or LBP on the potential biomechanical mechanism of posture [23, 24]. Self-report, particularly for sitting time, has shown to have large bias, poor precision, and weak correlation with objective measures of sitting [25]. Since the COVID-19 pandemic, few studies reported objective sitting time in sedentary employees and have focused only on prevalence of sitting behaviors rather than health outcomes [26, 27]. Further, examination of postural compensation in the LPHC is needed for prevention of non-specific LBP. Hence, there is a need to establish an understanding of the precursors for LBP (sitting time and postural compensations) using accurate and objective measures. There is also a need to understand sitting and physical activity behaviors of individuals now working in different work environments (e.g., remote work from home) to optimize preventative care, health programs, and therapeutic interventions.

Thus, the purpose of this cross-sectional study was to examine the relationship between objectively measured sitting time, posture, and low back pain in adult (24–64 years) employees with full-time (≥ 40 h/wk) sedentary (i.e., “desk”) jobs. Secondarily, the present study reports the activity behaviors (physical activity and sedentary behavior) of employees and compares these behaviors between work environments (i.e., those who work from home with those who work from an office during COVID-19). The methodology is highly innovative, as objective data were used for posture assessment as well as measurement of sitting time with the gold standard measurement device, the ActivPAL™ [28, 29]. As remote employment increases following COVID-19, this study is timely and foundational for developing behavioral health interventions and programs for employees with sedentary jobs in various environments (home remote vs. office). Future health behavior programs will be able to effectively develop targeted programs to prevent the many adverse musculoskeletal and metabolic health outcomes associated with high amounts of sitting.

Methods

Study design and sample

This study was a cross-sectional design in adults working full-time in sedentary occupations. Inclusion criteria were (a) 24–64 years, (b) self-reported full-time sedentary occupation, (c) no history of acute LBP trauma. These criteria were set for the feasibility to carry out the study in the target population. LBP is most prevalent in those 40–69 years and contributes the most to the number of years lived with disability in adults 24–64 years [4]. Thus, this age range was selected to target working adults who were likely to benefit most from a future workplace behavioral intervention. This study was approved by the university Institutional Review Board for the protection of human subjects.

Procedures

Due to the COVID-19 pandemic and social distancing policies, all data were collected remotely. Participants were recruited via e-mail and social media flyer announcements at various university functions and business contacts. If interested, participants were scheduled for a Zoom video call to complete eligibility screening (~ 5 min). If eligible, participants read and signed an informed consent. Participants then continued to complete a survey (~ 10 min) and a posture assessment (~ 10 min), and directions regarding receiving the ActivPAL™ via mail and wearing the device. Other than mailing the device, these measures were all performed on Zoom, allowing for research assistants to answer any questions posed by participants regarding informed consent or survey items.

Measures

Demographics, work environment, and background information were gathered via a self-report online survey.

Posture compensation at the LPHC was assessed with the Thomas Test, which is used to measure the length and functionality of the hip flexors. With the participant laying on the ground flat on their back, they hug both knees into their chest and then release one knee and allow the leg to hang towards the ground in a relaxed manner. This test identified a pass (no compensation) or a fail (compensation). A recording was saved to analyze the results offline. One researcher, an expert in biomechanics, scored all recordings to ensure reliability between participant assessments. Prior to data collection, the research team also conducted validity testing by comparing in-person assessment to the video assessment.

Low back pain was assessed via a self-report survey item asking if the participant had experienced low back pain in the past 12 months, lasting for longer than 3 months, to classify them as having “chronic low back pain.”

Physical Activity (PA) levels for moderate-to-vigorous aerobic, muscular strengthening, and flexibility PA were assessed using the self-report Godin Physical Activity Questionnaire [30]. Aerobic activity was summarized to minutes per week, while resistance and flexibility training were summarized as days per week. The Godin Physical Activity Questionnaire has been shown to have excellent validity and reliability in this population and has been used in several studies assessing physical level during COVID [31]. For each type of physical activity (aerobic, muscular, flexibility), participants were asked if this amount was the same, less, or more than the physical activity they engaged in prior to the COVID-19 “shelter in place” orders.

Sitting Behavior was measured objectively via the gold standard ActivPAL™ device (ActivPAL™ micro 3, Glasgow, Scotland)[25]. The participant identified a “typical work week” time frame (within two weeks of the posture assessment) and the ActivPAL™ device was mailed with instructions as well as a link to video demonstration on how to attach and wear the device for seven full days. Prior to mailing, the device was waterproofed (sealed polytubing). Participants attached the ActivPAL™ with adhesive (Hypafix and Tegaderm™) to the mid-thigh. Follow-up communication (email or phone) was given to confirm participants were wearing the device properly. The device is the gold standard for SB measurement and has evidence of acceptability, reliability, and validity in adults [28]. The device was returned with a pre-stamped return envelope to the research team for data processing.

ActivPAL™ processing

As per previous studies with wearable devices, a minimum of 4 days (≥ 10 h/day) as a valid wear time criterion to initialize the devices [24, 25]. Once returned, the data were downloaded and processed with the CREA algorithm in the ActivPAL™ software. Next the “Events File” (comma separated values file (csv)) was exported using the Pal Analysis software, then imported and processed in RStudio (R version 4.1.0). Once in R, the data were processed to summarize the number of sit-to-stand (STS) transitions, number of steps, and time spent stepping, sitting, and standing for each day worn for each participant. Invalid days (< 10 waking hours) were removed and participants with less than four valid days were excluded [28]. Then, before analysis, each observation was verified by a separate research assistant and compared with the PAL file to ensure validity. After data cleaning, each ActivPAL™ variable (i.e., STS transitions, sitting/standing time, stepping time, steps) was averaged across the number of valid days and used for analysis.

Data analysis

Data were entered and analyzed using SPSS (IBM Statistics Version 28.0) and RStudio (R version 4.1.0). Descriptive statistics, frequency counts, tests of normality, and outlier identification (> 3.3 SDs from mean) were checked for background and outcome variables. Binomial logistic regressions were used to assess predictor variables on outcomes including sitting time on posture, sitting time on LBP, and posture on LBP. One way Analysis of Variance (ANOVA) was used to assess differences between work environment (office vs. home vs. hybrid) and ActivPAL™ behavior variables and physical activity variables. An alpha level of 0.05 was used for tests of statistical significance.

Sample size consideration

There have been few studies since COVID with objectively measured sitting time in remote workers. These studies had small samples of N < 24 [27, 28]. Further, to compare environments (home vs. office), a power analysis in R was completed. With a power of 0.8 (p < 0.05) and moderate effect size of d = 0.3, there would need to be a sample of 11 participants per group. Because the study goal was to achieve a representable sample of adult full-time employees, the aim was to complete the present study with over 45 participants for descriptive statistics.

Results

Sixty-one participants expressed interest in the study and were contacted to schedule an eligibility screening. Eight of those potential participants did not start the study due to schedule conflicts or no longer had interest. Thus, fifty-three completed the eligibility screening and were eligible to start the study. Three had unusable questionnaire data, three others had unusable device data, and six had unusable posture data due poor video quality (n = 2) or inability to perform the test during the Zoom call (n = 4). For each analysis, we used all available data for each test. Flow is shown in Fig. 1.

Fig. 1
figure 1

Study flow of participants

Full size image

Demographic and outcome variable overview

Descriptive statistics and frequency counts for demographic and outcome variables are shown in Tables 1, 2, 3. Tests of normality identified three outliers (> 3.3 SDs from mean total) for sitting and stepping with the ActivPAL™; these were removed for follow-up tests [32].

Table 1 Baseline descriptives of participants background and demographics (N = 53)
Full size table
Table 2 Means, SDs, F-Statistic, and p-value of self-report and objective activity variables by work environment
Full size table
Table 3 Comparison of physical activity during and prior to COVID-19
Full size table

The study sample (N = 50; Mage(SD) = 20 (± 12)years) was nearly half female (56%, n = 28), mostly Caucasian (80%, n = 37) and has worked full-time in their “desk job” for an average of 10.6 years (± 10.3). More participants worked from home (40%, n = 19) than from the office (30%, n = 14) or a mix (i.e., hybrid) of home and office (22%, n = 11). Further, the sample had an average BMI or 28(± 7) kg/m2 and most participants rated their health as excellent or good (66%, n = 31). Only 32% (n = 15) indicated adequate hip posture and nearly 60% (n = 29) reported having low back pain in the last 12 months.

Notably, for objective activity level, the study sample accumulated an average almost 9.0 ± 1.5 h/day of sitting and 7016 ± (2811) steps/day (Table 2). However, participants self-reported their sitting at 6.6 ± 1.4 h/day. Further, participants self-reported that they completed an average of 109 ± 88 min/week of moderate-to-vigorous aerobic physical activity and about 1.7 ± 1.9 days/week or resistance training physical activity (Table 2). Also, 34% (n = 16) and 20.5% (n = 9) reported they were doing less aerobic and muscular strengthening physical activity, respectively (Table 3). While 28% (n = 13) and 25% (n = 11) reported they were doing more aerobic and muscular strengthening physical activity, respectively. Other objective and self-report means ± SDs are outlined in Tables 1, 2, 3.

Sitting, hip posture compensation, and low back pain

A binomial logistic regression showed sitting time (min/day) was a statistically significant predictor of poor hip posture (β = − 0.01, p < 0.046). However, a binomial logistic regression showed sitting time (min/day) was not a statistically significant predictor of low back pain (β = 0.002, p < 0.43). Additionally, a binomial logistic regression showed posture was not a statistically significant predictor of low back pain, (β = − 1.1, p < 0.11).

Environment and activity variables

One-way Analysis of Variance tests examined differences between work environments (home, office, hybrid) and activity variable outcomes (see Table 2). There were significant main effects for environment on sitting time (F = 4.2(2,37), p < 0.02) and self-report muscular strengthening physical activity (F = 3.6(2,40), p < 0.04). Follow-up comparison tests showed that those who worked from home had significantly higher sitting times, higher step count, and more days per week of muscular strengthening activity than those who worked from the office. However, there were no statistically significant differences in any variables between those that worked from a hybrid (home/office) and those that worked exclusively from home or office (all p > 0.05). There were no main effects for the work environment on other variables listed in Table 2. F statistics and p-values are shown in Table 2.

Discussion

Overview

The current study is the largest study, thus far, with objective sitting data (n = 50) to report sitting in employees working desk jobs through the COVID-19 pandemic. Further, it is the first study to use objective sitting and postural assessments to examine the relationship to low back pain in “desk” job employees. The study also compares the objective sedentary behavior and self-reported physical activity between working from home, an office, or a hybrid of home/office. The present study showed that higher amounts of sitting (min/day) was predictive of postural compensation at the hip. While not statistically significant, compensated hip posture showed a trend towards predicting low back pain in sedentary employees.

Previous studies and LBP

Previous studies have shown that sedentary time is associated with low back pain [23, 24]. Gupta et al. (2015) reported a positive association between high amounts of sedentary hours (> 8.3 h/d) and self-reported LBP level of > 4/9 (N = 33;16% of their sample) in blue collar workers, as compared to those that had less average sedentary hours [23]. However, their study included working adults from occupations involving regular lifting and carrying (e.g., manufacturing, assemblers, cleaners) with no consideration of a posture measure. Further, they included all valid days from their devices, with an average of 2.5 days/person. Whereas, the present study only included participants with greater than four valid days to reflect habitual behavior [23]. A strength to the present study was to specifically examine habitual sitting behavior with a thigh-worn accelerometer and use objective posture data as a predictor of LBP.

Given the biomechanical compensations (i.e., Jandas Lower Crossed Syndrome) in the hip with long periods of sitting, we expected to find a correlation between sitting time, hip posture, and low back pain. Although the current study did not show statistical significance between sitting or posture and low back pain, several possible reasons for this finding were considered. First, it is likely that those who currently experience pain change their behavior to alleviate their pain and maintain work. Since the average severity of pain (88% indicated a pain level of < 4/10) was relatively low, it is likely that participants are at a low threshold of dealing with low intensity pain. Thus, participants may simply find ways to manage low nagging pain through moving around more throughout the day or starting to engage in therapy exercises/stretches to alleviate pain that were not accounted for. While these remedies are beneficial in the short term, strategies to treat or maintain the appropriate biomechanical and postural health are needed for long-term prevention and maintenance. For example, previous studies have shown correlations between hamstring mobility and low back pain [33]. Further, some researchers have suggested implementation of a feasible stretching intervention combined with a proper ergonomic set up in the workplace to alleviate back pain [34]. Shariate et al. (2018) found significant improvement in low back pain after four months of a workplace stretching intervention compared to a control group that did not experience improvements in pain [34]. Another study reported the efficacy of a simple exercise (i.e., Bruegger’s exercises) in office workers to improve pain and increase hip range of motion over four weeks [35]. While physical therapy has shown to be effective to restore health, only 58% seek care and, of those who seek care, nearly 50% are non-adherent to prescribed protocols [5, 13]. Further, with low intensity chronic pain, people often routinely self-manage the pain until further severity persists and the problem is exacerbated before seeking help. Thus, without proper prevention and maintenance of hip posture with workplace behavioral interventions, relapse is inevitable and the vicious cycle will continue.

Behavior and environment

Another strength of the current study is that there are differences reported in behavior between participants in different work environments, reporting significantly higher sitting time for those who work from home than those who work from the office. Advances in technology have drastically changed the work environment and the COVID-19 pandemic has further led to more remote workers than before. Due to this increase in remote worker prevalence, there is now a specific need to understand behaviors of remote workers. This evidence-based groundwork allows for the development of proper home-based work interventions in which healthy behaviors are promoted. Working from home has decreased commute time (or entirely eliminated), flexibility in household or other activities done throughout the day, as well as decreased office “mingle” time that can cause homework to be efficient, adaptable, and productive, yet highly sedentary. This study also showed that those who worked from home had higher sitting time but also self-reported significantly higher days per week that they completed muscle-strengthening exercises; there were also higher step counts (not statistically significant; 7289 vs. 5984 steps/day, p = 0.29) than those that worked at the office. Thus, while sitting for long periods of time has many adverse outcomes, perhaps the flexibility, ability to complete household activities, and lesser commute time has also contributed to participants adding beneficial activity behaviors. For example, home-based employees may add subtle stepping throughout the day for household tasks and more days per week that they complete muscle-strengthening exercises from a home gym.

Recent studies report sedentary and physical activity behavior in remote employees since the start of the COVID-19 pandemic (most used surveys to self-report activity measures). The majority of these studies reported an overall decrease in physical activity (76.4% of studies) and an increase in sedentary behavior (75% of studies) in workers comparing before and after the start of the pandemic [11]. Two previous studies used accelerometer measures and both had smaller sample sizes (N = 11; N = 22) than the current study (N = 50) [27, 28]. Interestingly, both studies reported higher total sitting times (> 10 h/d) in their sample than our study (9 h/d). These differences may have been due to COVID-19 policies in other countries (i.e., Brazil, Sweden vs. US). Hallman et al. (2020) also used a thigh-worn device in office workers (N = 27) in Sweden [27]. They reported no significant difference in sitting, standing, or moving time between days participants worked from home compared to days worked from the office. While these are different results than our study, the analyses were intra-individual, and thus most comparable to our group that worked both a “hybrid” of days between the office and home. This study did not collect data on which days our participants worked from home or office and could not compare the variable between days in our participants.

While it is important to bring awareness to the adverse effects of sitting, the objective posture measure is also a unique strength to our study. Behavioral interventions have previously demonstrated efficacy to reduce sitting time in the workplace. However, if proper biomechanics to the hip complex are not first realigned, this may exacerbate LPB when standing with compensated alignment. For example, [37] found a reduction in sitting and a non-significant decrease in low back pain in their intervention for office workers [36]. This non-significant result may be caused by a miscommunication regarding proper alignment of their spine for standing. As mentioned earlier, this realignment can be achieved with a feasible yet intentional stretch routine [34]. Over time, consistent realignment will likely achieve long-term maintenance of proper posture and reduce risk of LBP in the workplace. Thus, interventions need to incorporate both physical therapy and behavior change professionals to achieve long-term improvements and maintenance in healthy behaviors.

Study strengths and future research

This study is the largest known study to use objective sitting data with the gold standard (i.e., ActivPAL™) measurement of employees with desk jobs during COVID-19. Further, this study builds on previous studies which have used self-reported low back pain to also include objectively measured hip posture. Posture assessments help further explain the relationship between sitting and musculoskeletal pain and will help future programs and health behavior interventions target precursors to chronic pain. This study is also timely, as more employees have moved to long-term remote work since the beginning of the COVID-19 pandemic. There is an urgent need to develop behavioral interventions and programs to target specific work environments, which will help reduce sitting and improve posture to combat the effects of low back pain. These programs may include education on proper in-home ergonomics, short stretches to improve posture, and a reduction in sitting with sit-stand home workstations.

Conclusions

In conclusion, this study demonstrates a relationship between high amounts of sitting and postural compensations that may lead to low back pain. Higher amounts of sitting in employees that work from home are reported. Future studies and Public Health programs need to target specific work environments with behavioral interventions and recommendations to combat the adverse effects of high sedentary behavior and posture.