Flow chart of this study
A total of 1791 individuals participated in the online questionnaire survey, with 498 (27.81%) office workers included, and 1293 individuals excluded. Those participants who were excluded were 563 (31.34%) individuals who did not complete the questionnaire, 218 (12.17%) students, 78 (4.36%) building workers, and 434 (24.23%) other non-office workers. The flow chart can be seen in Fig 1.
Sociodemographic and clinical characteristics of the sample.
The baseline characteristics of all patients are shown in Table 1. The results of the t-test showed no significant differences among the gender groups for age, working age, and NDI scores (P>0.05), and the DCF of males was higher than that of females (P<0.05). Pearson's chi-squared test showed that differences in the number of educational degrees in the male and female groups were significant (P<0.05). The activities ratio of smart phone use, using computers, reading books, and using other electronic devices was 63.71%/78.07%, 30.65%/18.72%, 1.61%/2.14%, and 4.03%/1.07% in the male and female groups, respectively. Pearson’s chi-square tests showed differences were statistically significant (P<0.05). The ratio of participants with neck pain (74.19% male, 66.31% female) and the incidence of low back pain (73.39% male, 78.07% female) were high; the difference between males and females was not significant (P>0.05).
Crude correlation associations of DCF, covariates, and NDI of the sample
As seen in Table 2, single factor correlation analysis showed that age and working age did not have a correlation with NDI scores for participants (P>0.05). DCF had a positive correlation with NDI scores (P<0.05). Compared with other activities, smart phone use had no positive correlation with NDI scores (β=0.83, 95%CI= -0.07 to 1.73, P>0.05), while low back pain had a strong correlation with NDI scores (P<0.05).
Multivariate logistic regression model for DCF and NDI of the sample
Low back pain factor was the covariates of NDI as seen in supplementary table; we also selected age, gender and working age as covariates, on the basis of their associations of outcome of interest. as. After adjusting for age, working age, and sex covariates, DCF had a positive correlation with NDI scores (β=0.28, 95%CI=0.13 to 0.43, P<0.05), and after adjusting for low back pain DCF had a positive correlation with NDI scores (β=0.26, 95%CI=0.12 to 0.40, P<0.05). (See Table 3)
Curve line correlation between the DCF and NDI of the sample
Generalized additive models were used to visually assess the DCF and NDI relationships. We adjusted for age, sex, working age, and low back pain factors. The DCF had a curve line correlation with NDI - a monotone increasing relationship. (See Fig 2)
Analysis of threshold saturation effect between the DCF and NDI of the sample
As seen in Table 4, we performed threshold saturation effect analysis between DCF and NDI scores in participants. The logarithmic likelihood ratio test showed that there was a fold point (K=6) between DCF and NDI scores, and the differences were statistically significant (P<0.05). When DCF was less than 6 hours (K<6), the estimated change in NDI was 0.53, 95%CI was 0.26 to 0.81, and the differences were statistically significant (P<0.05). When DCF was greater than 6 hours (K>6), the estimated change in NDI was -0.03, 95%CI was -0.33 to 0.26, and the differences were not statistically significant (P>0.05). The logarithmic likelihood ratio test showed that this fold point was statistically significant (P<0.05).