Applied Ergonomics 68 (2018) 230–239

Contents lists available at ScienceDirect

Applied Ergonomics
journal homepage: www.elsevier.com/locate/apergo

Review article

The effects of breaks on low back pain, discomfort, and work productivity in T
office workers: A systematic review of randomized and non-randomized
controlled trials
Pooriput Waongenngarm, Kantheera Areerak, Prawit Janwantanakul∗
Department of Physical Therapy, Faculty of Allied Health Sciences, Chulalongkorn University, Bangkok, Thailand

A R T I C L E I N F O A B S T R A C T

Keywords: The purpose of this study was to evaluate the effectiveness of breaks on low back pain, discomfort, and work
Break productivity in office workers.
Spinal pain Publications were systematically searched in several databases from 1980 to December 2016. Relevant
Musculoskeletal disorders randomized and non-randomized controlled trials were retrieved and assessed for methodological quality by two
Computers
independent reviewers. Quality of evidence was assessed and rated according to GRADE guidelines.
Eight randomized controlled trials and three non-randomized controlled trials were included in this review, of
which 10 were rated as high-quality studies. The break programs were highly heterogeneous with work duration
ranging from 5 min to 2 h and break duration ranging from 20 s to 30 min. The results showed low-quality
evidence for the conflicting effect of breaks on pain and low-quality evidence for the positive effect of breaks on
discomfort. When stratified by type of breaks, moderate-quality evidence was found for the positive effect of
active breaks with postural change for pain and discomfort. Moderate-quality evidence indicated that the use of
breaks had no detrimental effect on work productivity.
More high-quality studies are needed before recommendations can be given. Within a number of methodo-
logical limitations that are present in the published studies, active breaks with postural change may be effective
in reducing pain in workers with acute low back pain and to prevent discomfort in healthy subjects.

1. Introduction subtle and regular spinal movements, while sitting (Dankaerts et al.,
2006; O'Sullivan et al., 2012). The prolonged postural loading of the
One common health problem experienced by office workers is low spine while sitting can reduce joint lubrication, fluid content of inter-
back pain (LBP). Approximately between 34% and 51% of office vertebral discs, and increase stiffness, which can be detrimental to back
workers experienced LBP in the preceding 12 months (Ayanniyi et al., health (Beach et al., 2005; Chan et al., 2011). Prolonged muscle acti-
2010; Janwantanakul et al., 2008) with the 1-year incident rate for LBP vation in static sitting may lead to localized muscle tension, muscle
at about 14–23% (Juul-Kristensen et al., 2004; Sitthipornvorakul et al., strains, muscle fatigue, and other soft-tissue damage, causing impair-
2015). Furthermore, the 1-year prevalence of chronic LBP has been ment of motor coordination and control as well as increased mechanical
reported to range from 15% to 45%, with a point prevalence of 30% stress on ligaments and intervertebral discs (Granata et al., 2004).
(Manchikanti et al., 2009). Low back pain causes personal suffering, Prolonged sitting also induces low back discomfort (Waongenngarm
disability, and impaired quality of life and work in general, which can et al., 2015), which is a strong predictor of LBP (Hamberg-van Reenen
pose a great socioeconomic burden for both patients and society et al., 2008).
(Manchikanti et al., 2014). Breaks are recommended for alleviating the adverse effects of pro-
Office workers are usually required to sit for long hours working on longed sitting with poor postures. Scheduled breaks can prevent the
a computer while spending most of their time in a sitting position. onset or progression of cumulative trauma disorders in the computer-
Occupational groups exposed to poor postures while sitting for longer ized workstation environment (Balci and Aghazadeh, 2004; Barredo
than half a day have a considerably increased risk of experiencing LBP and Mahon, 2007; Sheahan et al., 2016). A break is generally defined as
(Lis et al., 2007). Subjects with LBP are likely to be in sustained pos- the cessation of computer work tasks and can be either passive or ac-
tures and have large and infrequent spinal movements, rather than tive. For a passive break, operators leave their computer tasks to sit and

∗
Corresponding author. Department of Physical Therapy, Faculty of Allied Health Sciences, Chulalongkorn University, Bangkok 10330, Thailand.
E-mail addresses: golff[email protected] (P. Waongenngarm), [email protected] (K. Areerak), [email protected] (P. Janwantanakul).

https://doi.org/10.1016/j.apergo.2017.12.003
Received 23 April 2017; Received in revised form 1 December 2017; Accepted 4 December 2017
Available online 08 December 2017
0003-6870/ © 2017 Elsevier Ltd. All rights reserved.
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

relax during this period, while during active breaks, operators are re- comprising the characteristics of participants, intervention parameters,
quired to perform specific movements, exercises, or change their pos- outcomes, and results. The consensus method was used to resolve dis-
ture (Nakphet et al., 2014). Previous studies compared the beneficial agreements between the two reviewers. A third reviewer (PJ) was
effects of passive and active breaks, by assessing oxygenation in mus- consulted to achieve a final judgment if disagreement persisted.
cles, muscle activity, and discomfort in the neck and upper extremity.
The results showed that breaks – regardless of the type of activities
during the breaks – had a positive effect on the recovery of muscle 2.5. Data analysis
discomfort (Crenshaw et al., 2006; Nakphet et al., 2014). However, due
to the impracticalities and potential impact on work productivity of Conclusions were reached on the effectiveness of breaks based on
breaks, it is difficult to implement the breaks in an office setting the reported outcome of pain, discomfort, or work productivity using
without working being continued. Thus, standing breaks while per- the GRADE (Grades of Recommendation, Assessment, Development and
forming computer work have been recently introduced as an option to Evaluation) system, which was used to evaluate the overall quality of
reduce discomfort and pain in the low back while still maintaining the evidence and the strength of the recommendations (Furlan et al.,
worker productivity (Thorp et al., 2014). 2015). For each outcome, an a priori ranking of ‘high’, or ‘low’ was
To date, there have been no studies on the effects of the type of assigned depending on whether the majority of studies were categor-
breaks on pain and discomfort in the low back as well as work pro- ized as randomized controlled trials or non-randomized controlled trials
ductivity. Thus, the primary aim of this study was to systematically (Swinton et al., 2017). Five domains of quality were rated for each
review randomized and non-randomized controlled trials (RCTs) to comparison: (1) limitations of study design; (2) inconsistency; (3) in-
gain insights into the effectiveness of breaks on low back pain, dis- directness; (4) imprecision; (5) publication bias across all trials (Furlan
comfort, and work productivity in office workers. The secondary aim et al., 2015; Guyatt et al., 2011). A four-point rating scale ranging from
was to identify the type of breaks effective in reducing pain and pre- ‘high quality’ on one end to ‘very low quality’ on the other was em-
venting discomfort in the low back. ployed. The quality of the summary of findings was rated as moderate if
one, low if two, and very low if three of the criteria were not met. The
2. Methods following definitions of quality of evidence were applied (Balshem
et al., 2011):
2.1. Search strategy
• High quality: We are very confident that the true effect lies close to
Online searches were conducted on the Web of Science, PubMed, that of the estimate of the effect.
ScienceDirect, the Cochrane Library, PEDro, and Scopus databases from • Moderate quality: We are moderately confident in the effect esti-
1980–December 2016. The following keywords were used: back pain, mate: The true effect is likely to be close to the estimate of the effect,
low back pain, chronic low back pain, LBP, break, pause, rest, rest but there is a possibility that it is substantially different.
break, micro-break, active break, passive break, and postural change. • Low quality: Our confidence in the effect estimate is limited: The
The search and full inclusion process was performed by two reviewers true effect may be substantially different from the estimate of the
(PW and KA). After the inclusion of articles based on the selection effect.
criteria, references were searched for additional articles. • Very low quality: We have very little confidence in the effect esti-
mate: The true effect is likely to be substantially different from the
2.2. Selection of studies estimate of the effect.

The selection criteria of relevant articles were:

2.6. Subgroup analysis
(1) The study design was a RCT or a non-RCT that employed the break
as a primary intervention. All relevant studies were stratified by type of breaks. Breaks were
(2) The study population was office workers, or those working with classified as 1) an active break with postural change, 2) an active break
computers, visual display units, or visual display terminals. without postural change, 3) a passive break, and 4) a standing break
(3) Low back pain, discomfort, or work productivity was assessed in the while performing computer work. An active break with postural change
study. Studies on LBP due to specific underlying pathology, such as was defined as operators being required to change their postures (i.e.
tumors, fractures, infection, dislocation, or osteoporosis were ex- from siting to standing) and perform specific movements or exercises in
cluded. the low back. An active break without postural change was defined as
(4) The article was a full report published in English. Letters, abstracts, operators being required to perform specific movements or exercises in
books, conference proceedings, and posters were excluded. the low back in the sitting position. A passive break was defined as
operators leaving their computer tasks to sit and relax during this
2.3. Quality assessment of studies period. A standing break while performing computer work was defined
as operators being required to change their posture (from siting to
The articles were evaluated for methodological quality by two re- standing) while still performing computer work.
viewers (PW and KA). Risk of bias was assessed using the Cochrane
Back and Neck Review Group expanded 13-item criteria (Furlan et al.,
2015). A high-quality study was defined as scoring positive in at least 2.7. Sensitivity analysis
50% (7/13) of the items. Disagreements between the reviewers were
discussed in an attempt to achieve consensus. If agreement could not be Sensitivity analysis was conducted to assess how sensitive the re-
reached, a third reviewer (PJ) was consulted to achieve a final judg- sults of the review were in relation to the way it was performed. For the
ment. results of qualitative analysis (using the GRADE approach), the effect of
the cut-off point used in the methodological quality assessment for
2.4. Data extraction qualification as a high-quality study on the synthesized results was
assessed by shifting the cut-off point from ≥50 to ≥60%, or shifting
Data extraction was performed by two reviewers (PW and KA). The the cut-off point from ≥50 to ≥70%.
reviewers independently extracted the data using a standardized form,

231
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

were non-RCTs (Table 2). Follow-up periods ranged from 48 min to 4

Computerized search of
databases (n=2,895 citations) months. Four RCTs were conducted in field settings (Galinsky et al.,
22,876 abstracts excluded because 2000, 2007; Henning et al., 1997; Lanhers et al., 2016). The remaining
tthey did not meet the selection four RCTs and three non-RCTs were conducted in laboratory settings.
ccriteria based on the screening of Only one of eleven studies was conducted in acute LBP subjects
aabstracts and titles
(Sheahan et al., 2016). The remaining studies were conducted in
healthy subjects who reported no LBP at baseline. Six studies compared
19 full-text articles retrieved break programs to no break programs. Four studies compared different
for closer inspection break programs among each other. The remaining study compared
by two reviewers 9 articles excluded because they did
nnot meet the selection criteria based on break programs with and without feedback. Of the eleven included
ffull-text articles studies, eight studies showed a positive effect of breaks and three stu-
• 1 article was a protocol study dies reported no effect of breaks.
• 1 article was duplicated
• 1 article studied in the neck and The break interventions of included studies were classified into 4
sshoulder types: active breaks with postural change, active breaks without pos-
• 3 articles did not evaluate the effects tural change, passive breaks, and standing breaks while performing
oof breaks
• 3 articles were not conducted in
computer work (Table 3). Of the included studies, four trials examined
ooffice workers the effectiveness of active breaks with postural change (Davis and
Kotowski, 2014; Galinsky et al., 2000; McLean et al., 2001; Sheahan
et al., 2016). Three trials investigated active breaks without postural
change (Balci and Aghazadeh, 2004; Galinsky et al., 2007; Lanhers
Reference checking and manual
R et al., 2016). One trial investigated passive breaks (Henning et al.,
ssearch found 1 additional article 1997). Two trials examined the effectiveness of standing breaks while
11 studies were included for performing computer work (Davis and Kotowski, 2014; Thorp et al.,
methodological quality 2014). The remaining two trials did not clearly specify which types of
assessment break protocol were examined (Henning et al., 1994, 1996). The work
duration ranged from 5 min to 2 h or their own discretion, and the
break duration from 20 s to 30 min or their own discretion.
0 articles were excluded because
tthey did not compare break
pprograms to no break programs 3.4. Summary of effectiveness of breaks

11 studies were included for The summary of evidence for the effectiveness of breaks and type of
GRADE analysis breaks on pain, discomfort, and work productivity are presented in
Tables 4 and 5.
Fig. 1. Flow diagram of the searching and screening process.

3.4.1. Evidence of the effectiveness of breaks for pain reduction

3. Result Two high-quality RCTs investigated the effectiveness of breaks in
reducing low back pain (Lanhers et al., 2016; Sheahan et al., 2016). The
3.1. Search strategy results indicated low-quality evidence (2 RCTs, n = 208; inconsistency,
imprecision) for the conflicting effect of breaks on LBP. Lanhers et al.
A total of eleven articles were judged to meet the selection criteria (2016) found no significant effect for an active break without postural
(Fig. 1). In one included study, the authors described their study design change program on LBP reduction compared to a control group of
as a quasi-experimental design (Davis and Kotowski, 2014). However, healthy subjects (Lanhers et al., 2016). On the other hand, Sheahan
both reviewers of this systematic review (PW and KA) identified such et al. (2016) reported a significant effect of an active break with pos-
study as an RCT design study because the participants in the study were tural change program on LBP reduction compared to a no break pro-
randomly assigned to groups and the control group was the conven- gram in acute LBP subjects (Sheahan et al., 2016).
tional workstation condition. As a result, the study was included in this When stratified by break type, moderate-quality evidence (1 RCT,
systematic review. All eleven articles were assessed for methodological n = 8; imprecision) was found for the positive effect of an active break
quality and data extraction. with postural change for pain reduction (Sheahan et al., 2016). Mod-
erate-quality evidence (1 RCT, n = 200; imprecision) was found for no
3.2. Methodological quality assessment effect of an active break without postural change on LBP (Lanhers et al.,
2016). No evidence existed concerning the effectiveness of passive
Eleven articles were evaluated by two reviewers (PW and KA) and breaks and standing breaks while performing computer work on LBP
the scoring of both reviewers before discussion had an agreement rate reduction.
of 96.5% (138/143). The overall inter-rater agreement resulted in a
kappa value of 0.93 with a standard error of measurement of 0.05. 3.4.2. Evidence of the effectiveness of breaks for discomfort prevention
Following discussion, the two reviewers reached full consensus (100%; Five high-quality RCTs (Davis and Kotowski, 2014; Galinsky et al.,
143/143). Thus, no article was evaluated by the third reviewer (PJ). 2000, 2007; McLean et al., 2001; Thorp et al., 2014), three high-quality
The scores for the methodological quality of the studies ranged from 6 non-RCTs (Balci and Aghazadeh, 2004; Henning et al., 1994, 1996),
to 9 points (Table 1). The median score was 8 points. Ten of the eleven and one low-quality RCT (Henning et al., 1997) investigated the ef-
included studies were rated as high-quality studies. All studies were fectiveness of breaks for preventing low back discomfort. The results
rated negative for items 3 (blinding of all participants), 4 (blinding of indicated low-quality evidence (6 RCTs and 3 non-RCTs, n = 273; in-
all therapists), and 5 (blinding of all assessors). consistency, imprecision) for the positive effect of break programs on
low back discomfort. Five high-quality RCTs (Davis and Kotowski,
3.3. Study characteristics 2014; Galinsky et al., 2000, 2007; McLean et al., 2001; Thorp et al.,
2014) and two high-quality non-RCTs (Balci and Aghazadeh, 2004;
Eight of the eleven trials were RCTs and the remaining three trials Henning et al., 1994) indicated that break programs significantly

232
 P. Waongenngarm et al.

Table 1
Methodological quality score of the 11 included studies.

Author Scores on the Cochrane Back and Neck Review Group expanded 13-item criteria Total Quality of
score study
Randomization Allocation Worker Care Outcome Drop- Intention to Selective Similar at Co-interventions Compliance Timing of the Other bias
concealed blinded provider assessor out treat outcome baseline avoided or similar acceptable outcome
blinded blinded rate reporting assessment

Davis and + ? – – – + + + + + + + + 9/13 High

Kotowski
(2014)
Galinsky et al. + ? – – – + + + + + + + + 9/13 High
(2000)
Lanhers et al. + + – – – + + + ? + + + + 9/13 High
(2016)
McLean et al. + ? – – – + + + + + + + + 9/13 High
(2001)

233
Sheahan et al. + ? – – – + + + + + + + + 9/13 High
(2016)
Balci and - - - - - + + + + + + + + 8/13 High
Aghazadeh
(2004)
Galinsky et al. + ? – – – + – + + + + + + 8/13 High
(2007)
Henning et al. – ? – – – + + + + + + + + 8/13 High
(1994)
Henning et al. – ? – – – + + + + + + + + 8/13 High
(1996)
Thorp et al. + ? – – – + – + + + + + + 8/13 High
(2014)
Henning et al. + ? – – – – – + ? + + + + 6/13 Low
(1997)
Positive (%) 73 9 0 0 0 91 73 100 82 100 100 100 100

+, positive; -, negative; ?, unclear.

Applied Ergonomics 68 (2018) 230–239
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

Table 2
Characteristics and results of the 11 included studies.

Author Study design (follow-up Study population Intervention Outcome Result

period;
study setting)

Davis and Kotowski, RCTa 37 call center employees I1: conventional workstation Discomfort I1 vs I2 = +
2014 (4 weeks; laboratory I2: conventional workstation with reminder I1 vs I3, I4 = +
study) software I2 vs I3, I4 = +
I3: sit-stand workstation Work I1 vs I2 vs I3 vs
I4: sit-stand workstation with reminders productivity I4 = 0
Galinsky et al., 2000 RCT: Pre-post 42 data-entry operators I: supplementary break Discomfort I vs C = +
experiment C: conventional break Work I vs C = 0
(4 weeks; productivity
field study)
Galinsky et al., 2007 RCT 51 employees VDU I1: stretching + conventional break Discomfort I1 vs I2 = +
(8 weeks; I2: stretching + supplementary break I3 vs I4 = +
field study) I1 vs I3 = 0
I2 vs I4 = 0
I3: non-stretching + conventional break Work I1 vs I2 = +
I4: non-stretching + supplementary break productivity I3 vs I4 = +
I1 vs I3 = 0
I2 vs I4 = 0
Henning et al., 1997 RCT 26 VDU operators for I1: breaks only Discomfort I1 vs I2 vs
(4 weeks; discomfort and C=0
field study) 34 VDU operators for work I2: breaks and exercises Work I1 vs I2 vs
productivity (larger site) C: no breaks or exercises productivity C=0
Lanhers et al., 2016 Cluster randomized trial 200 employees VDU I: received I-Preventive program Pain I vs C = 0
(4 months; field study) C: received no I-Preventive program
McLean et al., 2001 RCT: Pre-post 15 office workers I1: micro-breaks at 20 min intervals Discomfort I1 vs C = +
experiment I2: micro-breaks at 40 min intervals I2 vs C = +
(4 weeks; laboratory I3 vs C = +
study) I1 vs I2 vs
I3 = 0
I3: micro-breaks at their own discretion Work I1 vs C = 0
C: no breaks productivity I2 vs C = 0
I3 vs C = 0
I1 vs I2 vs
I3 = 0
Sheahan et al., 2016 RCT: Pre-post 8 university students with acute I1: 5 min of standing break every 30 min Pain I1, I2, I3 vs
experiment LBP C=+
(4 days; laboratory (Sig for MCID)
study) I2: 2.5 min of standing break every 15 min Work I1 vs I2 vs I3 vs
I3: 50 s of standing break every 5 min productivity C=0
C: no break
Thorp et al., 2014 RCT 23 overweight/obese office I1: SIT condition Discomfort I1 vs I2 = +
(5 days; laboratory workers I2: STAND-SIT condition Work I1 vs I2 = 0
study) productivity
Balci and Aghazadeh, Non-RCT (150 min; 10 male students I1: 60-min work/10-min break Discomfort I1 vs I2 vs
2004 laboratory study) I3 =+
I2: 30-min work/5-min break Work I1 vs I2 vs
I3: 15- min work/30-sec break (3 times) and 3 min productivity I3 =0
for the fourth time
Henning et al., 1994 Non-RCT 38 undergraduate psychology I1: regimented break Discomfort I1 vs I2 = +
(48 min; laboratory students I2: compensatory break Work I1 vs I2 = 0
study) productivity
Henning et al., 1996 Non-RCT 31 undergraduate psychology I1: break with feedback Discomfort I1 vs I2 = 0
(65 min; laboratory students I2: break without feedback Work I1 vs I2 = 0
study) productivity

Positive (+) if a break intervention was demonstrated to be statistically more effective than a control group.
Negative (−) if a break intervention was demonstrated to be statistically less effective than a control group.
Neutral (0) if a break intervention did not statistically differ from a control group.
I, intervention group; I1, intervention group 1; I2 intervention group 2; I3, intervention group 3; I4, intervention group 4; C, control group.
LBP, low back pain; RCT, randomized controlled trial; Non-RCT, non-randomized controlled trial; VDU, visual display unit; MCID, minimal clinically important difference.
a
The study stated that their study was a quasi-experimental design study. However, both reviewers of the present systematic review (PW and KA) identified such study as an RCT
design study because participants were randomly assigned to conditions and the conventional workstation condition in this study was considered the control group.

reduced discomfort of the low back compared to control groups. 2014; Galinsky et al., 2000; McLean et al., 2001). Moderate-quality
However, one low-quality RCT (Henning et al., 1997) and one high- evidence (1 RCT and 1 non-RCT, n = 61; imprecision) was found for
quality non-RCT (Henning et al., 1996, 1997) reported no significant the positive effect of active breaks without postural change for dis-
difference in low back discomfort between healthy workers who re- comfort reduction (Balci and Aghazadeh, 2004; Galinsky et al., 2007).
ceived and did not receive breaks. Low-quality evidence (1 RCT, n = 26; limitation in study design, im-
When stratified by break type, moderate-quality evidence (3 RCTs, precision) was found for no effect of passive breaks on low back dis-
n = 94; imprecision) was found for the positive effect of active breaks comfort (Henning et al., 1997). Moderate-quality evidence (2 RCTs,
with postural change for discomfort reduction (Davis and Kotowski, n = 60; imprecision) was found for the positive effect of standing

234
 P. Waongenngarm et al.

Table 3
Details of break interventions (n = 11).

Author Work duration Break duration Break protocol Type of break

Davis and Kotowski, 2014 30 min Their own discretion For the conventional workstation with software reminders, software was loaded onto the computers Active breaks with postural change
to track computer activity and participants requested to make a postural change at 30-min intervals
(stood up and moved around). Participants used discretion when taking a break but were encouraged
to follow the prompts as much as possible.
30 min Their own discretion For the sit-stand workstation condition with software reminders, participants were instructed how to Standing breaks while performing
use the adjustment controller. All participants were encouraged to change the workstation height computer work
during the workday. The height of the workstation could be adjusted between 61.0 cm and 130.8 cm.
However, the actual height of the workstation during sit-stand conditions was self-selected and was
then specifically measured.
Galinsky et al., 2000 1h 5 min Participants were encouraged to get up and took at least a short walk away from their workstations Active breaks with postural change
during each break.
Galinsky et al., 2007 1h 5 min Participants were asked to perform stretching exercise. They included brief stretches targeting the Active breaks without postural
neck, shoulders, back, and upper extremities, and required no more than a total of 2 min to perform. change
Henning et al., 1997 1h 30 s (3 times) every 15 min and VDU operators were asked to cease all computer operations by removing their hands from the Passive breaks
3 min for the last time keyboard, avoiding looking at the VDU display, and also sitting back in their chairs and relaxing for
the duration of the break.

235
Lanhers et al., 2016 2h 30 s - 1.30 min each At regular intervals, a visual signal on the screen prompted VDU workers to take active breaks. The Active breaks without postural
exercises were easy to perform and most could be carried out while seated without needing specific change
equipment.
McLean et al., 2001 20 or 40 min or their own 30 s Participants were prompted to get out of their chairs during each micro-break. They were not asked Active breaks with postural change
discretion to perform any particular activity during their break, but were asked to walk away from their
workstation.
Sheahan et al., 2016 30 min, 5 min, Participants were required to stand in the middle of the room without leaning or stretching. They Active breaks with postural change
15 min, 2.5 min, were permitted to access their mobile phones or converse with the researcher.
5 min 50 s
Thorp et al., 2014 30 min sitting work 30 min standing work Participants were asked to perform their usual computer and/or telephone-based work tasks in a Standing breaks while performing
simulated office setting for 8 h/day in either a seated work posture (SIT condition) or while computer work
systematically interchanging every 30 min between a standing and seated work posture (STAND-SIT
condition).
Balci and Aghazadeh, 2004 60 min, 10 min, Participants performed simple exercises in the break periods. The exercises included hand, head, Active breaks without postural
30 min, 5 min, neck, back, and shoulder regions and were easily applicable in the offices. change
15 min 30 s (3 times) and 3 min for the last
time of each hour
Henning et al., 1994 5 min 20 s No information is available. Cannot specify
Henning et al., 1996 10 min 30 s No information is available. Cannot specify

VDU, visual display unit.

Applied Ergonomics 68 (2018) 230–239
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

Table 4
Summary of evidence for the effectiveness of breaks on pain, discomfort, and work productivity.

Outcome n Risk of Inconsistency Indirectness Imprecision Publication bias Quality of

(# of studies) bias evidence
(GRADE)

Pain 208 (2) No serious Serious No serious Serious None ++OO low a,b
Discomfort 273 (9) No serious Serious No serious Serious None ++OO low a,b
b
Work productivity 289 (10) No serious No serious No serious Serious None +++O moderate

a
Serious inconsistency (e.g., opposite direction of effects).
b
Serious imprecision (e.g., fewer than 400 participants were included or only one study included).

breaks while performing computer work for discomfort prevention indicated that three conclusions would alter. First, the level of evidence
(Davis and Kotowski, 2014; Thorp et al., 2014). for the effectiveness of breaks on discomfort would change from “low”
to “very low” and on work productivity would change from “moderate”
3.4.3. Evidence of the effectiveness of breaks on work productivity to “low”. Second, the level of evidence for the effectiveness of active
Six high-quality RCTs (Davis and Kotowski, 2014; Galinsky et al., breaks without postural change on discomfort would change from
2000, 2007; McLean et al., 2001; Sheahan et al., 2016; Thorp et al., “moderate” to “low” and on work productivity would change from
2014), three high-quality non-RCTs (Balci and Aghazadeh, 2004; “low” to “very low”. Third, the level of evidence for the effectiveness of
Henning et al., 1994, 1996), and one low-quality RCT (Henning et al., standing breaks while performing computer work on discomfort would
1997) investigated the effectiveness of break programs on work pro- change from “moderate” to “low” and on work productivity would
ductivity. Moderate-quality evidence (7 RCTs and 3 non-RCTs, change from “moderate” to “low”.
n = 289; imprecision) indicated no effect of break programs on work
productivity. 4. Discussion
When stratified by break type, moderate-quality evidence (4 RCTs,
n = 102; imprecision) indicated no effect of active breaks with postural The break interventions reported in this review included active
change on work productivity (Davis and Kotowski, 2014; Galinsky breaks with postural change, active breaks without postural change,
et al., 2000; McLean et al., 2001; Sheahan et al., 2016). Low-quality passive breaks, and standing breaks while performing computer tasks.
evidence (1 RCT and 1 non-RCT, n = 61; inconsistency, imprecision) This review summarized the results of seven high-quality RCTs, three
indicated no effect of active breaks without postural change on work high-quality non-RCTs, and one low-quality RCT investigating break
productivity (Balci and Aghazadeh, 2004; Galinsky et al., 2007). Low- interventions on low back pain, discomfort, and work productivity in
quality evidence (1 RCT, n = 34; limitation in study design, impreci- office workers. We found heterogeneity among studies as to specific
sion) indicated no effect of passive breaks on work productivity aspects such as study population, type of break, break protocol, method
(Henning et al., 1997). Moderate-quality evidence (2 RCTs, n = 60; of outcome assessment, and data presentation. Thus, the analysis of the
imprecision) indicated no effect of standing breaks while performing results was limited to a qualitative summary.
computer work on work productivity (Davis and Kotowski, 2014; Thorp
et al., 2014). 4.1. Methodological considerations

3.5. Sensitivity analysis Of the eleven included studies, none blinded the participants,
therapists who administered the therapy, and assessors, whereas nine of
Changing the cut-off point from ≥50 to ≥60% would not have the eleven studies were unclear about the concealment of treatment
altered our conclusions at all. With a cut-off point of ≥70%, the results allocation. Participant blinding ensures that the apparent effect (or lack

Table 5
Summary of evidence for the effectiveness of active breaks with/without postural change, passive breaks, and standing breaks while performing computer work on pain, discomfort, and
work productivity.

Outcome n Risk of Inconsistency Indirectness Imprecision Publication bias Quality of

(# of studies) bias evidence
(GRADE)

For active break with postural change

c,d
Pain 8 (1) No serious No serious No serious Serious None +++O moderate
c
Discomfort 94 (3) No serious No serious No serious Serious None +++O moderate
c
Work productivity 102 (4) No serious No serious No serious Serious None +++O moderate
For active break without postural change
c,d
Pain 200 (1) No serious No serious No serious Serious None +++O moderate
c
Discomfort 61 (2) No serious No serious No serious Serious None +++O moderate
Work productivity 61 (2) No serious Serious No serious Serious None ++OO low b,c
For passive break
a,c,d
Discomfort 26 (1) Serious No serious No serious Serious None ++OO low
a,c,d
Work productivity 34 (1) Serious No serious No serious Serious None ++OO low
For standing break while performing computer work
c
Discomfort 60 (2) No serious No serious No serious Serious None +++O moderate
c
Work productivity 60(2) No serious No serious No serious Serious None +++O moderate

a
Serious limitations of study design (e.g., > 25% of participants from studies with low quality methods, risk of bias recommended by Cochrane Back and Neck Group < 7 points).
b
Serious inconsistency(e.g., opposite direction of effects).
c
Serious imprecision (e.g., fewer than 400 participants were included or only one study included).
d
Consistency for only one study cannot be evaluated.

236
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

thereof) of treatment is not due to the placebo or Hawthorne effects. 4.3. Evidence of the effectiveness of breaks for pain, discomfort and work
Expectations are an important factor in placebo effects (Price et al., productivity
1999). Participants in the control group would have had no expecta-
tions, but the intervention group was prone to expectations. Blinding of All included studies investigated the effectiveness of breaks on pain,
all therapists and assessors is also important to guarantee that the ap- discomfort, or work productivity. Concerning the effect of breaks on
parent effect of treatment is not due to the therapist's/assessor's en- pain and discomfort in the low back, the current review showed that
thusiasm or lack of enthusiasm for the intervention or control condition breaks seem to be effective in discomfort prevention. Conflicting evi-
(Portney and Watkins, 2009). Participant, therapist, and assessor dence was found for the effect of breaks on low back pain reduction.
blinding are important for the internal validity of a study. However, it is However, when stratified by type of break, active breaks with postural
very difficult – perhaps impossible – to blind participants or therapists change was found to be effective in reducing low back pain and dis-
in studies regarding break interventions and to blind assessors in self- comfort. Active breaks without postural change or standing breaks
reported outcomes (e.g. pain and discomfort). while performing computer work were found to be effective in pre-
Concealed treatment allocation is important in preventing sys- vention of low back discomfort. Passive breaks were found to be in-
tematic and selection bias. Concealed treatment allocation ensures that effective in reducing both low back pain and discomfort. The findings
the sequence in which subjects would be allocated to treatment is not are consistent with previous research showing that active breaks are
disclosed before random allocation. If treatment allocation is not con- better than passive breaks (Asmussen and Mazin, 1978).
cealed, the decision of whether or not to include a person in the trial Active breaks with postural change require participants to change
could be influenced by knowledge of whether or not the subject is to their posture during breaks, leading to improvement in blood circula-
receive treatment (Portney and Watkins, 2009). However, concealment tion in the lumbar region, change in spinal curvature, delay in the onset
of treatment allocation was mentioned in only two of the eleven in- of any specific musculoskeletal discomfort, and increase in the flow of
cluded studies. In fact, concealment of treatment allocation is relatively synovial fluid to lubricate and nourish the intervertebral disc (Marras
easy to implement and describe in a published report (Elkins, 2013). et al., 1995; Thorp et al., 2014). Deconditioning from prolonged and
Future research should consider the concealment of treatment alloca- awkward positions, sustained postures, and repetitive movements may
tion to reduce bias and ensure that it is stated in the reports. lead to a reduction in the length of soft tissues, which consequently
limits the ranges of available motion in joints. Limited joint motion will
distort normal body biomechanics. Such distortions can contribute to
4.2. Study characteristics the risk of injury (Main et al., 2008). Thus, active breaks with postural
change may hypothetically be an effective intervention in the preven-
There was heterogeneity among the studies in terms of the popu- tion and treatment of LBP.
lation studied and break protocols. The majority of the included studies Breaks – either active/passive breaks or standing breaks while
(91%) investigated the effect of breaks in healthy subjects. Although performing computer tasks – appear to have no adverse effect on work
previous studies showed that both healthy and LBP subjects received productivity. A previous study showed that breaks did not affect per-
benefits from breaks by reducing low back pain and discomfort formance on skill-based tasks (i.e. typing and arithmetic) (Lee and
(McLean et al., 2001; Sheahan et al., 2016), back pain among office Duffy, 2015). Breaks have been found to promote concentration,
workers is unlikely to originate from identical causes. Implementing the alertness, motivation, and activity at work (Thorp et al., 2014). Feeling
same intervention for everyone would not be appropriate. Thus, ex- relaxed and refreshed after a break has been identified as having a
trapolation of results from one group of subjects to another should be positive effect on work productivity (Epstein et al., 2016).
undertaken with caution. Further research should attempt to in-
vestigate the effectiveness of breaks by selecting a more specific group 4.4. Sensitivity analysis
of subjects who would theoretically benefit from breaks for the study.
Different break protocols in terms of break type, work duration, and The methodological quality of the included studies ranged between
break duration were employed among the included studies. Thus, the 6 and 9. In this review, a priori cut-off point of ≥50% was used, which
current state of the literature limits comparability between trials. The might have influenced the level of evidence and potentially the results
type of break found to be effective in reducing both low back pain and of the review. Since all high-quality studies had total scores of greater
discomfort was that of active breaks with postural change by reducing than 60%, changing the cut-off point from ≥50 to ≥60% would not
LBP symptoms, musculoskeletal discomfort, back muscle fatigue, and have altered our conclusions. However, shifting the cut-off point from
mental fatigue in prolonged sitting tasks (Davis and Kotowski, 2014; ≥50 to ≥70% would mean only five studies qualify as high-quality
McLean et al., 2001; Sheahan et al., 2016). Active breaks have been studies. Several conclusions about the effectiveness of breaks, active
found to lead to a more variable muscle activity pattern and increase breaks without postural change, and standing breaks while performing
muscle oxygenation during computer work than passive breaks computer work on discomfort and work productivity would subse-
(Crenshaw et al., 2006; Samani et al., 2009). quently alter.
The work and break durations varied considerably, ranging from This variation in the level of evidence reflects the fact that there has
5 min to 2 h or their own discretion for work duration and from 20 s to been a small number of very good quality studies investigating the ef-
30 min or their own discretion for break duration. Optimal break fectiveness of breaks on discomfort and work productivity in office
scheduling is the appropriate combination of task demands (e.g. work workers. Thus, further study is required before firm conclusions can be
duration) and break duration (Kopardekar and Mital, 1994). Previous drawn.
studies showed that frequent (i.e. at least once every hour) and short
(i.e. less than 10 min) breaks lead to significant improvements in the 4.5. Strengths and limitations of the study
musculoskeletal disorders in office workers (Balci and Aghazadeh,
2003; Kopardekar and Mital, 1994). Taking a break every 2 h seems to The major strength of this review is that the studies were system-
be insufficient for adequate musculoskeletal recovery (Lanhers et al., atically searched, evaluated for their methodological quality by two
2016). Therefore, future studies should take into account work and independent reviewers, extracted and synthesized based on the number
break durations when setting the break protocols in the study of the of studies and the quality score of the studies. However, there are three
effectiveness of breaks. main methodological limitations of note. First, the search strategy was
limited only to full published reports in English. There is the possibility
that language bias may have affected the results of the review. Second,

237
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

almost two thirds of the included studies were conducted in laboratory sense during computer mouse work. Impact of active versus passive pauses. Eur. J.
settings and follow-up periods were relatively short, ranging from Appl. Physiol. 97, 59–67.
Dankaerts, W., O'Sullivan, P., Burnett, A., Straker, L., 2006. Differences in sitting postures
48 min to 4 months. Therefore, generalization of the results from this are associated with nonspecific chronic low back pain disorders when patients are
review to real working situations or to the long-term effects of breaks subclassified. Spine 31, 698–704.
should be made with caution. Third, the researchers summarized the Davis, K.G., Kotowski, S.E., 2014. Postural variability: an effective way to reduce mus-
culoskeletal discomfort in office work. Hum. Factors 56, 1249–1261.
results from studies with substantial heterogeneity in study character- Elkins, M., 2013. Concealed allocation in randomised trials. J. Physiother. 59, 134–136.
istics. This may explain the observed variation in the results among the Epstein, D.A., Avrahami, D., Biehl, J.T., 2016. Taking 5: work-breaks, productivity, and
studies. Future research is required to indicate whether differences in opportunities for personal informatics for knowledge workers. In: Proceedings of the
2016 CHI Conference on Human Factors in Computing Systems. ACM, Santa Clara,
these aspects affect the effectiveness of breaks on pain, discomfort, and California, USA, pp. 673–684.
work productivity. Furlan, A.D., Malmivaara, A., Chou, R., Maher, C.G., Deyo, R.A., Schoene, M., Bronfort,
G., van Tulder, M.W., Editorial Board of the Cochrane Back, N.G., 2015. 2015
Updated Method Guideline for Systematic Reviews in the Cochrane Back and Neck
5. Conclusions
Group. Spine (Phila Pa 1976). vol. 40. pp. 1660–1673.
Galinsky, T.L., Swanson, N.G., Sauter, S.L., Hurrell, J.J., Schleifer, L.M., 2000. A field
Eight RCTs and three non-RCTs investigating the effectiveness of study of supplementary rest breaks for data-entry operators. Ergonomics 43,
break programs for pain, discomfort, and work productivity in office 622–638.
Galinsky, T., Swanson, N., Sauter, S., Dunkin, R., Hurrell, J., Schleifer, L., 2007.
workers were reviewed and analyzed. The findings revealed that low- Supplementary breaks and stretching exercises for data entry operators: a follow-up
quality evidence supported the effectiveness of breaks on discomfort field study. Am. J. Ind. Med. 50, 519–527.
prevention. Moderate-quality evidence was found to support there Granata, K.P., Slota, G.P., Wilson, S.E., 2004. Influence of fatigue in neuromuscular
control of spinal stability. Hum. Factors 46, 81–91.
being no adverse effect of breaks on work productivity. The type of Guyatt, G., Oxman, A.D., Akl, E.A., Kunz, R., Vist, G., Brozek, J., Norris, S., Falck-Ytter, Y.,
breaks that may be effective in reducing low back pain and discomfort Glasziou, P., DeBeer, H., Jaeschke, R., Rind, D., Meerpohl, J., Dahm, P., Schunemann,
while maintaining work productivity was identified, namely active H.J., 2011. GRADE guidelines: 1. Introduction-GRADE evidence profiles and sum-
mary of findings tables. J. Clin. Epidemiol. 64, 383–394.
breaks with postural change. More high-quality studies are needed Hamberg-van Reenen, H.H., van der Beek, A.J., Blatter, B.M., van der Grinten, M.P., van
before recommendations can be given. Literature with respect to the Mechelen, W., Bongers, P.M., 2008. Does musculoskeletal discomfort at work predict
effect of break programs on pain, discomfort, and work productivity in future musculoskeletal pain? Ergonomics 51, 637–648.
Henning, R.A., Kissel, G.V., Maynard, D.C., 1994. Compensatory rest breaks for VDT
office workers was heterogeneous. The design of future studies may be operators. Int. J. Ind. Ergon. 14, 243–249.
improved by conducting studies in real working situations with long- Henning, R.A., Callaghan, E.A., Ortega, A.M., Kissel, G.V., Guttman, J.I., Braun, H.A.,
term follow-up periods and being more specific regarding study popu- 1996. Continuous feedback to promote self-management of rest breaks during com-
puter use. Int. J. Ind. Ergon. 71–82.
lation, break type, and break protocol.
Henning, R.A., Jacques, P., Kissel, G.V., Sullivan, A.B., Alteras-Webb, S.M., 1997.
Frequent short rest breaks from computer work: effects on productivity and well-
Authors' contributions being at two field sites. Ergonomics 40, 78–91.
Janwantanakul, P., Pensri, P., Jiamjarasrangsri, V., Sinsongsook, T., 2008. Prevalence of
self-reported musculoskeletal symptoms among office workers. Occup. Med. 58,
The authors have contributed in the following ways: PW provided 436–438.
the concept/research design, data collection, data analysis and manu- Juul-Kristensen, B., Søgaard, K., Strøyer, J., Jensen, C., 2004. Computer users' risk factors
script writing. KA provided data collection, analysis and manuscript for developing shoulder, elbow and back symptoms. Scandinavian Journal Work.
Environ. Health 30, 390–398.
writing. PJ provided concept/research design, data analysis and Kopardekar, P., Mital, A., 1994. The effect of different work-rest schedules on fatigue and
manuscript writing. All authors read and approved the final manu- performance of a simulated directory assistance operator's task. Ergonomics 37,
script. 1697–1707.
Lanhers, C., Pereira, B., Maublant, C., Garde, G., Coudeyre, E., 2016. Evaluation of i-
Preventive: active prevention digital tool for musculoskeletal disorders among com-
Conflicts of interest puter workers. Ann. Phys. Rehabil. Med. e38.
Lee, B.C., Duffy, V.G., 2015. The effects of task interruption on human performance: a
study of the systematic classification of human behavior and interruption frequency.
None.
Hum. Factors Ergon. Manuf. Serv. Ind. 25, 137–152.
Lis, A., Black, K., Korn, H., Nordin, M., 2007. Association between sitting and occupa-
Acknowledgements tional LBP. Eur. Spine J. 16, 283–298.
Main, C., Sullivan, M., Watson, P., 2008. Pain Management: Practical Applications of the
Biopsychosocial Perspective in Clinical and Occupational Setting, second ed.
This work was supported by The Thailand Research Fund through Churchill livingstone.
the Royal Golden Jubilee Ph.D. Program (PHD/0180/2558). Manchikanti, L., Singh, V., Datta, S., Cohen, S.P., Hirsch, J.A., American Society of
Interventional Pain, P, 2009. Comprehensive review of epidemiology, scope, and
impact of spinal pain. Pain Physician 12, E35–E70.
References Manchikanti, L., Singh, V., Falco, F.J., Benyamin, R.M., Hirsch, J.A., 2014. Epidemiology
of low back pain in adults. Neuromodulation 17 (Suppl. 2), 3–10.
Asmussen, E., Mazin, B., 1978. Recuperation after muscular fatigue by "diverting activ- Marras, W.S., Lavender, S.A., Leurgans, S.E., Fathallah, F.A., Ferguson, S.A., Allread,
ities". Eur. J. Appl. Physiol. Occup. Physiol. 38, 1–7. W.G., Rajulu, S.L., 1995. Biomechanical risk factors for occupationally related low
Ayanniyi, O., Ukpai, B., Adeniyi, A., 2010. Differences in prevalence of self-reported back disorders. Ergonomics 38, 377–410.
musculoskeletal symptoms among computer and non-computer users in a Nigerian McLean, L., Tingley, M., Scott, R.N., Rickards, J., 2001. Computer terminal work and the
population: a cross-sectional study. BMC Musculoskelet. Disord. 11, 177. benefit of microbreaks. Appl. Ergon. 225–237.
Balci, R., Aghazadeh, F., 2003. The effect of work-rest schedules and type of task on the Nakphet, N., Chaikumarn, M., Janwantanakul, P., 2014. Effect of different types of rest-
discomfort and performance of VDT users. Ergonomics 46, 455–465. break interventions on neck and shoulder muscle activity, perceived discomfort and
Balci, R., Aghazadeh, F., 2004. Effects of exercise breaks on performance, muscular load, productivity in symptomatic VDU operators: a randomized controlled trial. Int. J.
and perceived discomfort in data entry and cognitive tasks. Comput. Ind. Eng. 46, Occup. Saf. Ergon. JOSE 20, 339–353.
399–411. O'Sullivan, K., O'Keeffe, M., O'Sullivan, L., O'Sullivan, P., Dankaerts, W., 2012. The effect
Balshem, H., Helfand, M., Schunemann, H.J., Oxman, A.D., Kunz, R., Brozek, J., Vist, of dynamic sitting on the prevention and management of low back pain and low back
G.E., Falck-Ytter, Y., Meerpohl, J., Norris, S., Guyatt, G.H., 2011. GRADE guidelines: discomfort: a systematic review. Ergonomics 55, 898–908.
3. Rating the quality of evidence. J. Clin. Epidemiol. 64, 401–406. Portney, L., Watkins, M., 2009. Foundations of Clinical Research : Application to Practice.
Barredo, R.D.V., Mahon, K., 2007. The effects of exercise and rest breaks on muscu- Pearson education, New Jersey.
loskeletal discomfort during computer tasks: an evidence-based perspective. J. Phys. Price, D.D., Milling, L.S., Kirsch, I., Duff, A., Montgomery, G.H., Nicholls, S.S., 1999. An
Ther. Sci. 19, 151–163. analysis of factors that contribute to the magnitude of placebo analgesia in an ex-
Beach, T.A., Parkinson, R.J., Stothart, J.P., Callaghan, J.P., 2005. Effects of prolonged perimental paradigm. Pain 83, 147–156.
sitting on the passive flexion stiffness of the in vivo lumbar spine. Spine J. 5, Samani, A., Holtermann, A., Sogaard, K., Madeleine, P., 2009. Active pauses induce more
145–154. variable electromyographic pattern of the trapezius muscle activity during computer
Chan, S.C., Ferguson, S.J., Gantenbein-Ritter, B., 2011. The effects of dynamic loading on work. J. Electromyogr. Kinesiol 19, e430–437.
the intervertebral disc. Eur. Spine J. 20, 1796–1812. Sheahan, P.J., Diesbourg, T.L., Fischer, S.L., 2016. The effect of rest break schedule on
Crenshaw, A.G., Djupsjobacka, M., Svedmark, A., 2006. Oxygenation, EMG and position acute low back pain development in pain and non-pain developers during seated

238
P. Waongenngarm et al. Applied Ergonomics 68 (2018) 230–239

work. Appl. Ergon. 53 (Pt A), 64–70. Thorp, A.A., Kingwell, B.A., Owen, N., Dunstan, D.W., 2014. Breaking up workplace
Sitthipornvorakul, E., Janwantanakul, P., Lohsoonthorn, V., 2015. The effect of daily sitting time with intermittent standing bouts improves fatigue and musculoskeletal
walking steps on preventing neck and low back pain in sedentary workers: a 1-year discomfort in overweight/obese office workers. Occup. Environ. Med. 765–771.
prospective cohort study. Eur. Spine J. 24, 417–424. Waongenngarm, P., Rajaratnam, B.S., Janwantanakul, P., 2015. Perceived body dis-
Swinton, P.A., Cooper, K., Hancock, E., 2017. Workplace interventions to improve sitting comfort and trunk muscle activity in three prolonged sitting postures. J. Phys. Ther.
posture: a systematic review. Prev. Med. 101, 204–212. Sci. 27, 2183–2187.

239