Changes in Musculoskeletal Pain Among Computer Workers When Working From Home

During the SARS-CoV-2 pandemic, factors in the changing work environment affected musculoskeletal health of computer workers. Understanding these factors is important for preventing the onset or exacerbation of musculoskeletal pain in employees who work from home and provides the basis for prevention work.

Abstract

Objectives

This study evaluated the impact of increased working from home on musculoskeletal pain before and after the SARS-CoV-2 pandemic.

Methods

In an online survey (September 2023–April 2024), pain was rated on a numeric rating scale. The impact of working from home and covariates on new pain onset and exacerbation was modeled using logistic regression analyses.

Results

Of 1064 participating computer workers, 968 were working from home. Working from home compared to office-only work showed a trend of increasing pain. Longer remote working hours and poorer workstation setups also increased the risks for pain (eg, new neck or upper back pain; odds ratio 2.02, 95% confidence interval 1.08–3.76).

Conclusions

Preventing musculoskeletal pain should involve improving remote workstation ergonomics, promoting healthy lifestyles, participation in regular occupational health screenings, and supporting employees with anxiety or depression symptoms.

LEARNING OUTCOMES

• Assessment of teleworking characteristics in a survey of computer workers in Germany

• Impact of working from home during the SARS-CoV-2 pandemic on new onset as well as exacerbation of musculoskeletal pain

• Recommendation of preventive measures to improve health when working from home

Remote working, especially working from home (WFH), has increased worldwide, particularly due to lockdown restrictions during the SARS-CoV-2 pandemic. In Germany, WFH almost doubled between 2019 und 2020, from 12.8% to 21.0% of the work force.¹ Alongside the known positive aspects of home-based telework from prepandemic experiences (eg, better work-life balance, increased flexibility, and autonomy), WFH may also be associated with blurring of boundaries between work and leisure time, overwork, social isolation, lack of support, or inadequate equipment.² A key difference of WFH during the pandemic compared to the period before the pandemic was that the switch from office work to WFH was hasty and unplanned. As a result, workstations at home had to be set up in due time, which oftentimes resulted in workplaces that did not meet ergonomic requirements.³

Because of the special working conditions associated with remote working at changing locations inside and outside the home, employees in Germany are generally responsible for designing an ergonomic working environment at home. In contrast to regular home-based telework, which is usually permanently set up by the employer in the employees’ home, companies have less control over remote workplaces’ designs. In general, to enable users to adopt a safe and healthy posture, a workstation should be ergonomically designed including, among others, a height-adjustable chair and desk, foot support (if necessary under certain circumstances), an external keyboard and mouse, high-quality screens and their ergonomic position on the desk, as well as sufficient illumination.⁴ In the case of impaired visual eyesight, adjusted computer glasses should also be worn.⁵

In the wake of the pandemic, several studies and reviews investigated the effects of increased WFH with respect to various outcomes, such as physical and mental health, health behaviors, quality of life, or productivity.^6,7 A recent review found inconsistent results with respect to the prevalence of musculoskeletal pain (MSP) and the impact of WFH on musculoskeletal health.⁸ Although most studies indicate increased MSP among employees WFH, few studies, in contrast, reported improvements during government-imposed lockdowns in 2020.^3,9 However, certain trends among WFH employees, such as associations of MSP with poor workplace designs or longer working hours seem to dominate the discussion.⁸

In general, lockdown restrictions during the pandemic and the subsequent increase of WFH led to an increased proportion of sedentary activities, while physical activity decreased at the same time.¹⁰ For example, a large Brazilian online survey of 43,000 adults found that lower physical activity and high computer use were associated with new onset of back pain during the pandemic.¹¹

Because prior literature on the effect of WFH on MSP is inconsistent, we conducted this study to investigate the prevalence of MSP in office workers in Germany before and after the SARS-CoV-2 pandemic. The situation before the pandemic was assessed retrospectively. We address the following research questions: (1) How have pain symptoms in the lower and upper back, neck, shoulders, arms, and hands of office workers changed in relation to the increase in remote work during the pandemic? (2) How do intensity and ergonomic properties of WFH influence musculoskeletal pain? (3) What other factors (eg, changes in physical activity, weight gain, mental distress) influence a possible change in pain symptoms?

METHODS

Study Design and Study Population

This online survey was conducted between September 1, 2023, and April 15, 2024. Office workers were recruited by the German Social Accident Insurance and the workers’ compensation boards for the woodworking and metalworking industries (“Berufsgenossenschaft (BG) Holz und Metall”), for the administrative sector (“Verwaltungs-BG”), for the public sector in Hesse (“Unfallkasse Hessen”), for the trade and logistics industry (“BG Handel und Warenlogistik”), and the German Social Accident Insurance institution of the Federal Government and for the railway services (“Unfallversicherung Bund und Bahn”), and received a participation link to the online survey via their employers. The previously mentioned institutions also published the study on their online channels and distributed information about the study via mailing lists. In addition, the study was promoted on the social media channels of the German Social Accident Insurance. Eligible subjects for participation were employees of legal age with a computer workstation, and a minimum of 4 hours of daily computer work, based on previous studies’ inclusion criteria.^12,13 Hence, exclusion criteria were a level of computer work of less than 4 hours per day and a birth year after 2005. People born before 1955 were also excluded to focus on employees before retirement. Subjects with pain caused by cancer or with preexisting unstable conditions like multiple sclerosis, gout, intermittent rheumatism, fibromyalgia, or mental illnesses (eg, depression, burnout, bipolar disorder) were also excluded.

Individual participation was on a voluntary base. All participants agreed to the study’s privacy policy and provided informed consent. The study was approved by the ethics committee of the Ruhr University Bochum (Reg. No. 23-7884). We followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines¹⁴ (Strobe Supplementary Digital Content, http://links.lww.com/JOM/B841).

Questionnaire Survey

At the beginning of the questionnaire, participants were informed about the research objectives, the privacy statement, and the option of withdrawing from participation. They were asked about possible exclusion criteria (ie, age, computer time per day, and preexisting conditions associated with pain), followed by six sections of questions: In section 1 sociodemographic data (eg, sex, education, marital status), height, weight, and weight change were solicited. The latter was assessed by the question “Has your weight changed during the pandemic?” (no or little [maximum 1–2 kg]; yes, weight gain; yes, weight loss). Section 2 dealt with occupational activities, characteristics of the workspace, remote working hours and in the office, social interactions among employees, and work-privacy conflicts. The ergonomic characteristics of the remote workstation in terms of furniture and equipment were examined in Section 3. Based on the information on seating (sofa or bed [0]; regular chair [1]; height-adjustable chair without use of adjustment functions [2]; height-adjustable chair with optimal adjustment by another person or own knowledge [3]), table (no table [0], desk without height adjustment, dining or kitchen table [1]; height-adjustable desk with optimal adjustment [2]; use of height adjustment for occasional standing work [3]), computer screen (laptop screen only [0]; at least one separate, adjustable screen [1]), keyboard and computer mouse (laptop keyboard and touchpad [0]; external keyboard and/or mouse [1]), a total ergonomic score (0–8 points) was calculated. The total score for ergonomic remote workstation equipment (WFH score) was categorized as follows: poor (0–1 points), less good (2–4 points), good (5–7 points), and optimally equipped (8 points).

Sections 4 and 5 dealt with the severity of MSP, pain medication use, general health, headaches, eye problems, and physical activity. The German translation of the Brief Pain Inventory (BPI) was used for a comprehensive assessment of pain symptoms.¹⁵ The BPI records both the intensity of pain and the impairment caused by the pain in seven domains of life (general activity, mood, ability to walk, normal work, relationships with other people, sleep, enjoyment of life). This study surveyed the most common pain localizations among WFH employees: low back, upper back, neck, shoulder, forearm/elbow, wrist/hand, as well as a free text variant for other localizations.^12,16,17 Average pain during the last week (t1) and at the time before the pandemic at the beginning of 2020 (t0, retrospectively) was classified by the respondents on a numerical rating scale from 0 (no pain) to 10 (most severe pain imaginable). Because only threshold values for moderate and severe pain levels in patients after surgery exist,¹⁸ and this study involved subjects able to work with presumably less pain, we used the lowest suggested pain thresholds for categorization: none (BPI = 0), mild (BPI 1–2), moderate (BPI 3–5), and severe (BPI 6–10). The BPI interference score describes the average impairment caused by pain in seven domains of life mentioned above, on a scale from 0 (no interference) to 10 (complete interference) for each domain. The total BPI interference score was calculated as the mean value from the seven domains.¹⁹ The severity of headaches and eye problems at time points t0 and t1 was rated by the respondents on a numerical rating scale from 0 (no complaints) to 10 (most severe complaints imaginable). General health was assessed on a 5-point Likert scale with the response options excellent, very good, good, fair or poor. Because of low numbers in some categories, categories were combined as good to excellent and fair or poor for the statistical analysis.

In Section 6, the brief 4-item Patient Health Questionnaire (PHQ-4) was used to rate mental distress in terms of depression and anxiety symptoms. PHQ-4 scores range from 0 to 12, with higher scores indicating more severe symptoms.²⁰ Section 7 assessed physical activity with the questions “Do you exercise or play sports regularly?” (yes; no; don’t know)” and “Do you do more sports or exercise than before the pandemic?” (yes, more; unchanged; no, less; don’t know).

Statistical Analysis

The sample size was estimated using the Fleiss method with continuity correction available on OpenEpi (https://www.openepi.com/SampleSize/SSCohort.htm), applying a two-sided significance level of 95% and 80% power. Estimates were based on prior literature, with a 35% MSP prevalence,¹¹ 50% exposure to inadequate ergonomics in the home office,¹⁷ and a relative risk of 1.3 for pain worsening when working from home,¹¹ resulting in a required sample size of 722 subjects.

Continuous variables were characterized by median and interquartile range (IQR) and categorical variables by number and percent. Scores obtained at different time points were compared using Wilcoxon signed-rank tests (WSR). McNemar tests or Bowker tests for symmetry were applied to paired categorical data, that is, when comparing data from t0 and t1. Multiple logistic regression was used to explore the relationship between a new onset of MSP at t1 and an exacerbation of MSP from t0 to t1 with WFH conditions and other risk factors. Because of the low prevalence proportions of MSP in the forearm or elbow, hand or wrist, or other regions, statistical models were only calculated for low back, upper back, neck, and shoulder MSP. Results of the regression models are presented as odds ratios (OR) with 95% confidence intervals (95% CIs). A directed acyclic graph (DAG, http://www.dagitty.net/) was used to determine minimally sufficient adjustment sets of covariates to estimate the effect of ergonomic characteristics of WFH on MSP (Fig. S1, http://links.lww.com/JOM/B842). To estimate a direct effect, according to the DAG, adjustment for age, sex, body mass index (BMI)/weight change, physical activity, mental health, eye problems, headaches, analgesic use, and occupational health screening is sufficient. Analyses were stratified by mental distress, BMI, sex, and improvement of remote workstation equipment.

We presented the results of all exposure variables in the tables but focused on general trends in the main text. Because of the exploratory nature of the study, no adjustments were made for multiple comparisons.²¹ Post hoc power analyses were conducted using G*Power, version 3.1.9.6 (Heinrich-Heine-Universität Düsseldorf, Germany).²² All other statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC). Graphs were prepared with GraphPad Prism, version 10 (GraphPad Software, La Jolla, CA).

RESULTS

By April 2024, 1274 subjects had registered for participation, of whom 164 subjects (12.9%) were excluded based on one of the exclusion criteria, including computer work for less than 4 hours per day (n = 66), being born after 2005 (n = 2), being of retirement age (n = 9), and suffering from preexisting pain-related conditions (n = 87). In addition, 40 subjects had to be excluded because of missing information on key variables, that is, MSP (n = 4), remote working status (n = 7), use of WHO level 3 pain medication (n = 1), suffering from relevant occupational diseases (n = 24), or making critical comments, suggesting that the questions were not answered truthfully (n = 5). Subjects who changed jobs during the pandemic were also excluded (n = 6). Thus, for this analysis, data from 1064 subjects were available.

Table 1 depicts the characteristics of the study population. Slightly more men than women participated (49.9% vs 49.1%), and more than 50% of study subjects were university graduates. The median age was 48 years. The proportion of participants with poor or fair general health and with more severe symptoms of depression and anxiety was statistically significantly higher at t1 compared with t0. During the pandemic, 14% of office workers experienced a reduction in physical activity and 29% experienced an increase and 7% a reduction in body weight.

TABLE 1.

Sociodemographic Characteristics of the Study Population

		t0		t1
Characteristics		n	%*	n	%*	P Value
Age, yr	Median (IQR)				48 (39–56)
Sex	Male			531	49.9
	Female			522	49.1
	Other			4	0.38
Education	≤10 yr of schooling			224	21.1
	>10 yr of schooling			262	24.6
	University degree			576	54.1
Body mass index (kg/m2)	Median (IQR)				25.7 (23.1–28.9)
Weight change since t0	None/barely			671	63.1
	Weight gain			312	29.3
	Weight loss			77	7.2
Physical activity	None			237	22.3
	Unchanged since t0			451	42.4
	Less since t0			145	13.6
	More since t0			204	19.2
General health	Excellent	144	13.5	103	9.7	<0.0001
	Very good	454	42.7	332	31.2
	Good	408	38.3	408	38.3
	Fair	47	4.4	182	17.1
	Poor	3	0.28	32	3.0
Depression and anxiety	Normal (0–2)	692	65.0	665	62.5	<0.0001
symptoms (PHQ-4 score)	Mild (3–5)	313	29.4	271	25.5
	Moderate (6–8)	26	2.4	83	7.8
	Severe (9–12)	16	1.5	25	2.3
Work-privacy conflicts	High			65	6.1
	Moderate			252	23.9
	Low			735	69.1
Suffered from reduced	Yes			229	21.5
contact with colleagues	No			594	55.8
	No reduced social interaction			229	21.5
Remote working (ever n = 968)		418	39.3	922	86.7	<0.0001
Weekly working hours	Median (IQR)	1,053	40 (38–41)	1,059	40 (37–41)	0.0186
Weekly remote working hours	Median (IQR)	418	10 (7–20)	922	18 (10–24)	<0.0001
Ergonomic remote	Poor			14	1.52
Workstation equipment	Less good			220	23.9
	Good			581	63.0
	Optimal			102	11.1

P values were assessed with McNemar tests (prevalence) or Wilcoxon signed-rank test.

t0, before the pandemic at the beginning of 2020; t1, time of survey; IQR, interquartile range.

*Percentages do not always add up to 100%, because some respondents did not answer all questions.

At the time of the survey (t1), 922 subjects (87%) worked remotely at least some of the time, with a median of 18 hours per week (IQR 10–24). Before the pandemic (t0), considerably fewer subjects (39%) had engaged in remote work and spent fewer working hours at home (median 10 hours per week, P_WSR < 0.0001). A total of 96 participants (9%) had not worked remotely at all. The ergonomic remote workstation equipment according to the WFH score could be considered as very good at t1. About three quarters of the 968 subjects who had ever worked remotely were well or even optimally equipped. Specifically, 69% of WFH respondents reported optimal seating, that is, using an optimally height-adjusted chair (seating score = 3), 76% used an additional monitor, and 78% used an external keyboard and/or computer mouse. A total of 271 participants (28%) stated that they were able to work standing up or at a desk with optimal height adjustment. During the pandemic, about half of the WFH employees improved their remote workstation equipment.

Overall, 15% of participants reported new MSP and further 19% reported worsening MSP during the pandemic in at least one of the body regions. Usually, new or worsening pain occurred in one or a few, but not all, body regions. Table 2 shows the prevalence and incidence of MSP as measured by the BPI. Pain severity is also depicted in Figure 1. Over time, the prevalence of low back pain decreased from 49% to 42% (P_McNemar < 0.0001). However, BPI severity increased for all body regions except for hand or wrist pain. Likewise, the BPI interference score increased in all body regions over time. At t1, the prevalence of headache was 41% (t0 = 54%) and the prevalence of eye complaints was 29% (t0 = 29%), but with higher intensities compared to t0.

TABLE 2.

Prevalence and Severity of Musculoskeletal Pain Measured by the Brief Pain Inventory

	Prevalencea (%)			Incidenceb (%)	Severity Score			BPI Interference Scorec
Body Region	t0	t1	P d	t1	t0	t1	P d	t0	t1	P e
Low back	49.1	42.4	<0.0001	27 (2.5)	4 (2–5)	4 (2–6)	0.0009	1.00 (0.43–2.00)	1.71 (0.71–3.43)	<0.0001
Upper back	24.6	24.2	0.5002	26 (2.4)	3 (2–5)	4 (3–6)	0.0036	1.14 (0.43–2.29)	2.00 (0.86–3.79)	<0.0001
Neck	40.4	40.4	1.0000	40 (3.8)	3 (2–5)	4 (3–6)	<0.0001	1.14 (0.43–2.00)	1.71 (0.71–3.43)	<0.0001
Shoulder	27.1	28.6	0.0881	52 (4.9)	3 (2–5)	4 (2–6)	0.0003	1.29 (0.57–2.14)	2.00 (0.86–3.86)	<0.0001
Forearm or elbow	6.9	8.4	0.0136	29 (2.7)	3 (2–5)	4 (2–5)	0.0267	1.29 (0.71–2.00)	2.14 (1.14–3.57)	<0.0001
Hand or wrist	12.1	12.5	0.7009	33 (3.1)	3 (2–5)	3 (2–6)	0.3614	1.29 (0.71–2.21)	2.14 (1.29–3.71)	<0.0001
Other	8.6	10.6	0.0008	30 (2.8)	3 (2–5)	4 (3–6)	<0.0001	1.71 (0.86–2.57)	2.71 (1.29–4.00)	<0.0001
Headaches	53.6	41.1	<0.0001	10 (0.9)	4 (2–5)	4 (2–6)	<0.0001
Eye problems	28.9	29.4	0.3621	33 (3.1)	3 (2–5)	4 (3–6)	<0.0001

Values of severity score and interference score are expressed as medians (interquartile range) among participants with pain.

^aPrevalence is defined as BPI > 0.

^bIncident pain is defined as new pain symptoms at t1 not present at t0.

^cBPI interference describes the average impairment caused by pain in the seven domains of life on a scale from 0 to 10.

^dMcNemar test.

^eWilcoxon signed-rank test.

t0, before the pandemic at the beginning of 2020; t1, at the time of survey.

FIGURE 1.

Musculoskeletal pain measured with the BPI in 1064 computer workers. Higher BPI scores indicate higher symptom burden. The numbers in the bar plots indicate the percentage of workers in each BPI category.

The study subjects’ pain severity was influenced by many different factors as shown by the estimates for the adjustment variables according to the DAG for the association between remote work (independent variable) and MSP (dependent variable). Factors associated with an increased risk of new or worsening pain were female sex, weight change, irregular attendance at occupational health screenings, more severe symptoms of anxiety and depression, and eye complaints (Table 3 and Table S1, http://links.lww.com/JOM/B842). MSP exacerbation was additionally associated with a more pronounced headache severity and a decrease in physical activity. Participants who practiced relaxation techniques or certain exercises to strengthen stressed areas of the body (eg, yoga, specific neck and shoulder exercises, or back exercises) showed an increase of new or worsening pain. The effect of the adjustment variables on MSP varied considerably between different parts of the body. For example, weight gain had a greater effect on worsening low and upper back pain than neck or shoulder pain.

TABLE 3.

Estimates for the Adjustment Set for New Onset of Musculoskeletal Pain at t1

		Low Back			Upper Back			Neck			Shoulder
		OR	95% CI		OR	95% CI		OR	95% CI		OR	95% CI
Regular participation in occupational health screening	No	4.61	1.03	20.58	2.16	0.71	6.59	1.58	0.70	3.58	1.28	0.64	2.58
	Yes	1			1			1			1
Sex	Female	3.13	1.08	9.05	0.92	0.37	2.28	0.65	0.32	1.34	1.61	0.82	3.15
	Male	1			1			1			1
Body mass index	≥ Median (25.66 kg/cm2)	0.32	0.11	0.95	0.52	0.20	1.32	0.74	0.36	1.54	0.68	0.35	1.34
	< Median	1			1			1			1
Weight change	Loss	1.55	0.31	7.81	2.32	0.46	11.78	1.19	0.33	4.34	1.22	0.34	4.33
	Gain	2.10	0.77	5.73	2.44	0.94	6.35	1.02	0.45	2.33	1.22	0.59	2.51
	None/barely	1			1			1			1
Change of physical activity	Less	1.11	0.23	5.21	1.61	0.24	10.94	1.31	0.38	4.54	1.76	0.59	5.24
	Unchanged/unknown	1.45	0.41	5.13	5.68	1.21	26.76	1.92	0.73	5.04	1.44	0.58	3.56
	No physical activity	1.39	0.35	5.48	1.25	0.19	8.16	0.74	0.21	2.61	1.08	0.38	3.10
	More	1			1			1			1
Analgesics use (t1)	Yes	0.68	0.21	2.20	1.44	0.53	3.94	1.00	0.39	2.55	1.95	0.92	4.13
	No	1			1			1			1
Age (per 10 yr)		0.87	0.59	1.29	0.75	0.51	1.11	0.89	0.65	1.21	1.03	0.78	1.38
PHQ-4 score (t1) (0–12)		1.20	1.03	1.41	1.08	0.91	1.28	1.15	1.01	1.31	1.05	0.93	1.18
Headache severity (t1) (0–10)		0.94	0.77	1.14	1.02	0.84	1.23	0.92	0.78	1.09	0.98	0.85	1.13
Severity of eye complaints (t1) (0–10)		1.10	0.93	1.31	1.19	1.01	1.39	1.13	0.99	1.30	1.05	0.92	1.19

Multiple logistic regression estimates for each body region.

CI, confidence interval; OR, odds ratio; t1, time of survey.

Only a few occupational risk factors showed a statistically significant association on the development of new MSP (Table 4 and Table S2, http://links.lww.com/JOM/B842) or MSP exacerbation (Table S3, http://links.lww.com/JOM/B842) in the adjusted models. Overall, WFH in comparison with office-only workers showed a trend to impact the outcomes in a negative manner in all body regions studied, albeit failing to reach the formal level of statistical significance (eg, exacerbation of neck pain (OR 1.62, 95% CI 0.83–3.19, Table S3, http://links.lww.com/JOM/B842). Stratified analyses by age showed a statistically significant effect of WFH on neck pain exacerbation in younger subjects, that is, those <50 years of age (OR 5.15, 95% CI 1.45–18.38, data not shown). In addition, long periods of daily computer work of more than 6 hours had a negative effect, particularly on incident neck pain. The type of workplace in the office had also a negative effect on MSP. In all body regions studied, working in a desk sharing arrangement was associated with the highest risk of new or worsening MSP compared to working in other office types with fixed workplaces (eg, single offices, open-plan offices).

TABLE 4.

Occupational Risk Factors for the New Onset of Musculoskeletal Pain at t1

			Low Back			Upper Back			Neck			Shoulders
		n	OR	95% CI		OR	95% CI		OR	95% CI		OR	95% CI
Daily computer work	>6 h	655	1.26	0.44	3.59	1.98	0.65	6.03	2.36	0.96	5.81	1.10	0.55	2.19
	4–6 h	284	1.00			1.00			1.00			1.00
Flex-office or desk	Yes	63	4.38	1.43	13.42	3.67	1.22	11.00	1.96	0.71	5.38	2.96	1.22	7.18
sharing in office	No	876	1.00			1.00			1.00			1.00
Working from	Yes	854	2.32	0.30	18.17	0.87	0.24	3.21	1.19	0.35	4.04	1.15	0.39	3.35
home	Office only	85	1.00			1.00			1.00			1.00
Weekly remote working hours (8 hr)		848	1.16	0.80	1.69	1.12	0.77	1.62	1.04	0.79	1.38	1.27	0.98	1.63
Ergonomic remote	Poor, less good	235	1.00	0.36	2.80	2.09	0.83	5.26	1.93	0.94	3.94	0.67	0.31	1.45
workstation equipment	Good, optimal	615	1.00			1.00			1.00			1.00
Additional screen	No	203	1.19	0.43	3.29	2.39	0.94	6.11	2.02	0.97	4.18	0.74	0.33	1.65
	Yes	650	1.00			1.00			1.00			1.00
Keyboard and	Laptop only	182	1.02	0.34	3.04	2.26	0.87	5.88	1.55	0.72	3.37	0.58	0.24	1.42
computer mouse	Additional equipment	669	1.00			1.00			1.00			1.00
Smartphone use	Yes	280	3.04	1.21	7.62	1.96	0.78	4.98	1.50	0.73	3.07	0.77	0.37	1.61
	No	571	1.00			1.00			1.00			1.00
Tablet computer use	Yes	58	0.80	0.09	7.02	2.71	0.70	10.49	1.11	0.32	3.91	2.22	0.80	6.17
	No	790	1.00			1.00			1.00			1.00
Touchpad use	Yes	21	-			-			1.43	0.17	12.00	4.33	1.09	17.26
	No	648	-			-			1.00			1.00
Separate room for	No	274	1.56	0.62	3.91	1.56	0.62	3.95	1.15	0.55	2.41	0.84	0.42	1.68
computer work	Yes	576	1.00			1.00			1.00			1.00
Employer provision of a home-based	Yes	118	0.28	0.04	2.21	0.29	0.04	2.34	0.57	0.17	1.93	0.61	0.21	1.76
workstation	No	735	1.00			1.00			1.00			1.00
Improvement of	Yes	429	0.93	0.38	2.30	0.85	0.34	2.11	0.94	0.47	1.91	1.04	0.55	1.97
remote workstation	No	422	1.00			1.00			1.00			1.00
Optimal screen	No	205	0.71	0.23	2.16	1.13	0.41	3.07	1.18	0.52	2.65	0.97	0.46	2.05
position	Do not know	39	0.50	0.05	4.57	0.50	0.05	4.63	1.89	0.50	7.07	0.37	0.05	2.89
	Yes	604	1.00			1.00			1.00			1.00

Each risk factor was modeled in a single logistic regression model and adjusted by the adjustment set as described in the methods section and shown in Table 3.

CI, confidence interval; n, number of subjects included per model by exposure category varying slightly between body regions depending on missing MSP data; OR, Odds ratio; t1, time of survey; Ergonomic equipment of remote workstation, Based on the information on the seat, desk, screen, keyboard and mouse, an overall ergonomic score (0–8 points) was calculated with the categories poor (0–1 points), less good (2–4 points), good (5–7 points), and optimal equipped (8 points).

Among WFH employees, longer remote working hours per week were associated with an increased risk of pain exacerbation in the upper back (OR 1.23, 95% CI 1.01–1.49, Table S3, http://links.lww.com/JOM/B842) and shoulder (OR 1.22, 95% CI 1.03–1.45, Table S3, http://links.lww.com/JOM/B842). The ergonomic remote workstation equipment, as measured by the WFH score, had an impact on newly onset of neck and upper back pain (OR 2.02, 95% CI 1.08–3.76, Table S2, http://links.lww.com/JOM/B842). Among participants with a normal PHQ-4 score, higher ORs for exacerbation of low back, upper back, and neck pain were observed when subjects reported poorer equipment (Table S4, http://links.lww.com/JOM/B842). The absence of additional equipment such as an additional screen, keyboard or computer mouse increased the risk of new pain and pain exacerbation, especially neck and upper back pain (Table 4 and Table S2, http://links.lww.com/JOM/B842). The absence of such equipment also increased the risk of low back pain exacerbation if the BMI was high (OR 2.09, 95% CI 1.03–4.23, Table S5, http://links.lww.com/JOM/B842).

Touchpad use was uncommon when working from home, with only 22 employees surveyed using one regularly. However, we observed increased ORs for shoulder pain (new onset OR 4.33, 95% CI 1.09–17.26, Table 4; exacerbation OR 2.76, 95% CI 0.86–8.83, Table S3, http://links.lww.com/JOM/B842). Using a smartphone or a computer tablet for work-related activities other than making phone calls was more common (33% and 6%, respectively) and was generally associated with a higher risk of new or worsening MSP (Table 4 and Table S3, http://links.lww.com/JOM/B842). The risk of an exacerbation of MSP if working with a smartphone was generally higher in women (Table S6, http://links.lww.com/JOM/B842).

Most participants (68%) worked in a separate room at home. Over 80% of these had good or optimal remote workstation equipment, compared to 51% of subjects without a separate room. The regression models showed an increased risk of new or more severe pain if not having a separate room. Employer provision of a home-based workstation indicated a protective effect on MSP. Stratified analyses showed a more positive effect for women, younger workers, and those with a higher BMI. A nonoptimal screen position (ie, eyes and top of screen not at the same height) was associated with an increased risk of exacerbated pain in all body regions studied (eg, exacerbation of neck or upper back pain OR 1.42, 95% CI 0.94–2.15, Table S3, http://links.lww.com/JOM/B842). Nonoptimal screen position appears to be particularly relevant in relation to MSP exacerbation among participants with higher BMI, showing increased ORs in all body regions (Table S5, http://links.lww.com/JOM/B842). In contrast, the risk of pain exacerbation in the absence of a separate room for computer work when working from home was higher in workers with a lower BMI (eg, upper back pain OR 2.03, 95% CI 1.04–3.97, Table S5, http://links.lww.com/JOM/B842).

Improvement of the remote workstation during pandemic had little impact on MSP onset or exacerbation. However, stratification showed that for workers without improvement, nonoptimal ergonomic remote workstation equipment and longer weekly remote working hours had a greater impact on pain exacerbation (Table S7, http://links.lww.com/JOM/B842). The impact of longer weekly remote working hours on worsening low back, neck, and shoulder pain was also stronger for men than for women (Table S6, http://links.lww.com/JOM/B842).

DISCUSSION

Consistent with existing data,¹ our results confirm that the pandemic led to an increase in WFH among office workers in Germany. Our results also indicate that in Germany WFH remains a very frequent model among the new forms of working. We observed a rise in MSP prevalence and a trend toward worsening MSP before and after the pandemic, except for low back pain. At the time of the survey, low back and neck pain were most frequently reported. Despite improvements in remote workstation equipment, which was rated good according to WFH scores at the time of the survey, WFH conditions contributed significantly to new or exacerbating MSP. This was especially evident in the upper back and neck among the workers studied.

This analysis had the advantage of using data from a large survey of computer workers either working in the office or remotely. Validated scales were used to assess pain severity as well as depression and anxiety symptoms at two time points.

However, this study also has limitations. First, data at t0 were collected retrospectively. Therefore, this study may be subject to well-known biases associated with retrospective designs, such as recall bias. Second, participation was voluntary, so only individuals with a particular interest were included, raising the possibility of selection bias. Thus, in our random study sample predominantly well-educated office workers were included. Hence, the results may not be generalizable to the entire population of office workers in Germany. Third, the recruitment of employees from participating companies was conducted through participating institutions and promotion via social media channels. Hence, it is unclear how many individuals were reached by the study invitation, making the response rate indeterminable. Fourth, despite the large study population surpassing the calculated required sample size, some analyses were statistically underpowered. For example, modeling the effect of poorly equipped remote workstations on new-onset MSP resulted in a post hoc power of 46% for upper back pain and 53% for neck pain. Fifth, the number of exposure variables, the body regions studied, and the different stratifications resulted in many regression models being run.

Overall, the prevalence of MSP was rather high in this study. Similar to a study among 2000 Polish office workers,²³ the two most frequently reported complaints were low back and neck pain. However, we observed an MSP prevalence in the low back of 43% and in the neck of 41%, compared with 17% each in a Polish study.²³ Other studies observed similar¹⁷ or even higher prevalence proportions.²⁴ The observed decline in the prevalence of low back pain during the pandemic contracts with the overall trend of increasing pain that had already started prior to the pandemic.²⁵ Possible reasons for these differences may toile in the definition of prevalent pain or in the characteristics of the study population examined, such as age, sex, exposure, or other factors.

In line with other surveys, we observed an association between MSP, working conditions and ergonomics or remote workstations.^17,23,24 In general, our data showed a trend toward a higher risk of both new onset and exacerbation of MSP in the group of employees who had ever worked from home compared to those who never worked from home. Consistent with a recent review,⁸ we identified a negative effect of long periods of daily computer work as well as longer remote working hours on MSP. Like a cross-sectional study of computer workers in Japan,²⁶ we demonstrated associations between MSP exacerbation and duration of work, particularly in poor mobile working environments. However, in our study, this applied to neck and upper back pain rather than low back pain. For those using inadequate ergonomic remote workstation equipment (WFH score <5), the OR for neck and upper back pain was increased for daily computer work exceeding 6 hours compared to 4–6 hours (OR 2.02, 95% CI 0.96–4.23, data not shown). This association was not observed for employees with an ergonomically good or optimal remote workstation (OR 0.96, 95% CI 0.58–1.60).

In general, the ergonomic equipment for remote work was very good for 72% of the participants. For example, optimal seating was reported by 69% of the participants, which was higher compared to 45% in an Italian study conducted at the beginning of the pandemic.¹⁷ The use of sit-stand desks was also more common than in a 2020 Australian study (14% vs 5%).²⁷ However, technical equipment, like additional screens or keyboards, were similarly frequent.^17,27 In our investigation, about half of the respondents improved their remote workstation equipment during the pandemic. It can be assumed that improvements were made either to address serious deficiencies or to provide a more comfortable working environment, as well-equipped home offices are associated with positive outcomes like reduced technostress and fewer work-family conflicts in remote work settings.²⁸ The WFH score was higher for those who had improved their workstations than for those who had not. Stratification showed that nonoptimal remote workstation equipment and longer weekly remote working hours had a greater impact on MSP exacerbation in workers without improvement than for those with improvement. This is consistent with the findings of a review demonstrating the positive impact of ergonomic interventions on employees’ well-being.⁸

We found that the lack of a separate room when WFH had a negative impact on MSP, confirming observations from previous investigations.^29,30 This may be caused by a noisier work environment with more distractions when there is no dedicated room for WFH, and concentration is more likely to be reduced.^29,30 On the other hand, we observed that workplaces in a separate room were more often well-equipped according to the WFH score. This aligns with findings in our study and others, showing that working in office environments without a fixed workstation increases the risk of new or worsening MSP. The difficulty in setting up the workstation correctly each day likely contributes to this risk.³¹

Low back pain is associated with higher BMI and it was found to be a major cause of loss of disability-adjusted life years attributable to high BMI.^32,33 In this study we observed that the absence of additional equipment (eg, additional screen, keyboard, or mouse) and poor screen positioning increased the risk of MSP exacerbation in subjects with higher BMI. On the other hand, changes in weight affected MSP (especially on MSP exacerbation of the back). This was true for weight gain and weight loss, compared to no weight change. The latter is in contrast to observations that weight loss is beneficial for people with chronic pain.³⁴ The increase in MSP among participants who practiced relaxation techniques or exercises like yoga likely occurred because they started these activities after experiencing MSP. Thus, our observation does not contradict the beneficial effects of yoga noted in various intervention studies during the pandemic.^35,36

Next to the decreased physical well-being of the study subjects, we also observed a decreased mental well-being which is in line with an online study among 988 office workers with transition to WFH during the pandemic.³⁰ The association between psychological factors such as depression, anxiety or stress and MSP³⁷ was also confirmed in this investigation. Impairments caused by WFH including social isolation, lack of contact with colleagues, and blurred boundaries between work and personal life^16,38,39 may lead to an increase in psychological complaints.⁴⁰ However, as we found in a previous study, countermeasures such as strengthening social interactions between employees and reducing work-privacy conflicts tended to protect mental health even after the pandemic ended.⁴¹ Findings on the association between occupational psychosocial exposures, such as job support or job stress, and MSP may also differ. For instance, a recent meta-analysis identified no link with low back pain,⁴² whereas another review showed an association between neck pain and factors like perceived job demands or low support from supervisors and coworkers.⁴³ Furthermore, a recent analysis of the General Social Survey of the US population found associations between nonspecific back pain and several psychosocial factors in a pre-pandemic analysis.⁴⁴ In our study, we observed a trend toward lower MSP among employees who were provided with a remote workstation by their employer. This could be due to better remote workstation equipment or a psychological effect. Free-text comments from our questionnaire revealed that employees who are well-equipped by their employer feel more valued. This aligns with results from a UK study showing a positive effect of organizational and coworker support on job satisfaction and happiness.⁴⁵

In conclusion, we showed that even among a relatively well-equipped study population, longer computer working hours as well as poorer equipment of the remote workstation in terms of seating, desk, and technical equipment (external monitor, keyboard, and computer mouse) are relevant occupational risk factors for MSP. Therefore, to prevent the onset or exacerbation of MSP, employers should ensure that employees have the knowledge to correctly set up their remote workstation, for example, self-sufficiently applying an optimal screen, chair, and desk position. Working with a smart phone, a touch pad, or a computer tablet should be avoided in favor of working at a fully equipped workstation. Beyond raising awareness about ergonomics and work organization, workplace prevention of MSP should also emphasize promoting healthy lifestyles. For instance, according to these results, the risk of low back pain exacerbation in the absence of appropriate ergonomic remote working equipment was particularly high for employees with a higher BMI. Promoting regular attendance at occupational health screenings could help to reduce the risk of MSP onset. In addition, employers should provide appropriate support for employees with known symptoms of anxiety or depression.