871730 HFSXXX10.

1177/0018720819871730Human Factors<sc>Driving</sc> P<sc>osture and Low Back Pain</sc>research-article2019

Sitting Posture During Occupational Driving

Causes Low Back Pain; Evidence-Based Position or
Dogma? A Systematic Review
Sarah Tinitali , Kelly-Ann Bowles, Jennifer L. Keating, and Terry Haines,
Monash University, Melbourne, Victoria, Australia

Objective: In this review, we determine if there is Introduction

evidence to demonstrate a relationship between occupa-
tional driving posture and low back pain. Low back pain has been described as the
Background: The burden of low back pain is increas- greatest cause of disability worldwide when
ing. An understanding of this relationship is required to compared with any other condition (Hoy et al.,
enable the development of recommendations for clinicians 2014) and is a burden that is continuing to
and policy-makers for the driving industry. increase (GBD 2017 Disease and Injury Inci-
Method: Five databases were searched up to March dence and Prevalence Collaborators, 2018). In
12, 2018. Study quality was assessed using the National
Heart, Lung, and Blood Institute’s Quality Assessment
2010, the global point prevalence of low back
Tool for Observational Cohort and Cross-Sectional Stud- pain was estimated to be 9.4% (Hoy et al.,
ies, followed by a GRADE analysis to consider the evidence 2014). In the most recent Australian economic
as a whole. A narrative, critical synthesis was completed burden study, the estimated cost of low back
that considered the methods by which driving posture and pain in a 1-year period was Aus$9.17 billion
low back pain were measured and analyzed. (Walker, Muller, & Grant, 2003). Aus$8.15
Results: There were 653 articles identified, with
billion of this burden of disease was driven by
seven eligible for review. Four articles identified an asso-
ciation between occupational driving posture and low back indirect costs such as loss of earnings and pro-
pain, yet this was based on the use of measurement tools ductivity, where people were not able to engage
lacking validity. Although a relationship may exist, the spe- in their usual occupations (Walker et al., 2003).
cific driving postures associated with low back pain and “Awkward posture,” defined as body position-
the strength of this relationship have not been confirmed. ing that deviates from the neutral or natural posi-
Conclusion: Future research needs to employ vali- tion (Moore, Torma-Krajewski, & Steiner, 2011),
dated and reliable, real-time qualitative methods for mea-
suring occupational driving posture to advance our under-
has been discussed as a risk factor for low back
standing of this relationship. pain (Burdorf & Sorock, 1997; Johanning, 2000).
Application: Clinical and policy recommendations The proposition that awkward sitting posture may
regarding driving posture and low back pain should be contribute to low back pain is also widely dis-
used with caution, as they are guided by evidence incor- cussed (Lis, Black, Korn, & Nordin, 2007; Manek
porating bias. Future studies are required to confirm the & MacGregor, 2005; Neumann et al., 2001;
specific postures assumed while occupational driving and
Yilmaz & Dedeli, 2012). The ability to engage in
their relationship with low back pain, before recommen-
dations can be made. occupations that require sitting may therefore be
impaired by low back pain attributable to awk-
Keywords: occupation, occupational health, workplace ward sitting postures. There is a substantial evi-
injury, musculoskeletal, spine dence base that has examined postures adopted in
static (i.e., stationary) sitting and relationships
with low back pain (Claus, Hides, Moseley, &
Address correspondence to Sarah Tinitali, Department Hodges, 2009; Dankaerts, O’Sullivan, Burnett,
of Physiotherapy, Monash University, McMahons Road, & Straker, 2006; Park, Tsao, Claus, Cresswell, &
Melbourne, Victoria 3199, Australia; e-mail: sarah.
Hodges, 2013; Williams, Hawley, McKenzie, &
[email protected].
HUMAN FACTORS
van Wijmen, 1991). Comparatively less work,
Vol. XX, No. X, Month XXXX, pp. 1­–13 however, has been undertaken on the sitting
DOI: 10.1177/0018720819871730 behavior that may be required as part of occupa-
Article reuse guidelines: sagepub.com/journals-permissions tional activities that require movement, such as
Copyright © 2019, Human Factors and Ergonomics Society. driving an automobile.
2 Month XXXX - Human Factors

Between 2004 and 2015, 3.5 million workers’ experienced by the drivers in the review by Lis
compensation claims were accepted in Austra- et al. are likely to differ considerably across par-
lia, 120,000 of which were made by truck driv- ticipant occupations, particularly in regard to the
ers, and a total of 151,000 claims were made by movements of the spine and upper limbs when
delivery drivers, bus drivers, and rail drivers operating various vehicle functions. The trunk
(Xia, Iles, Newnam, Lubman, & Collie, 2018). and upper limb movements required of a heli-
Musculoskeletal conditions, such as non-spe- copter pilot or garbage truck driver are likely to
cific low back pain, neck pain, and injury to the differ to those required of a bus or truck driver,
limbs, were found to be the most common rea- due to the need to operate additional or alterna-
son for claims made by transport workers and tive controls within the vehicle, or to perform
were responsible for 60.2% of all claims by peo- additional tasks such as repeated rotation of the
ple in the transport sector (Xia et al., 2018). Bus trunk to ensure correct position of a lever pick-
drivers and delivery drivers were responsible for ing up a garbage bin. Trunk movements in auto-
the largest proportion of these claims (Xia et al., mobile drivers are also likely to differ when an
2018). Data on the number of claims attributed individual is driving on an uneven surface, or at
to low back pain are not currently available; a high speed, in comparison with a standard
however, these data demonstrate that people public road. It is therefore difficult to make spe-
employed in the transport sector have a higher cific conclusions related to drivers of standard
risk of developing a work-related injury or dis- road vehicles based on a review of such a hetero-
ease when compared with workers in other fields geneous group of drivers. Furthermore, Lis et al.
(Xia et al., 2018; Xia, Iles, Newnam, Lubman, & provide limited discussion on the specific driv-
Collie, 2019). ing postures associated with low back pain and
The link between sitting posture and low report the need for future research once valid
back pain during driving may have first been and reliable measurement tools are developed.
postulated by Troup (1978), although they pro- Papers exist that demonstrate an association
vided no direct evidence to support this claim. A between low back pain and the interplay of driv-
later review by Lyons (2002) discussed a poten- ing posture and other occupational factors, such
tial mechanism; that low back pain may be as whole-body vibration and manual material
caused by the adoption of awkward driving pos- handling (Bovenzi et al., 2006; Lis et al., 2007;
tures due to the poor ergonomic design of vehi- Okunribido, Magnusson, & Pope, 2006, 2008).
cles. Lyons, however, referenced data from only Although it is important to acknowledge this
one empirical study by Krause et al. (1997) who interplay, some of these risk factors are out of
reported an increased risk of back or neck pain scope for modification by clinicians or employ-
for drivers who experienced “ergonomic prob- ers; a physiotherapist is not able to modify the
lems,” such as difficulty seeing out of the vehi- vibration present in a patient’s car. It is therefore
cle, or problems with operating vehicle controls important to have an understanding of risk fac-
(p. 2121). Krause et al. did not directly measure tors as discrete, independent outcomes, and their
driving posture beyond self-report. relationship with low back pain in order to
More recently, a review by Lis et al. (2007) develop recommendations for clinicians working
considered the association between occupational with occupational drivers and to enable policy
sitting and low back pain while assuming awk- development for the driving industry. This there-
ward postures and/or while exposed to whole- fore poses the question, Is there evidence to dem-
body vibration. The review considered a range onstrate a relationship between occupational
of driving populations, including drivers of driving posture and low back pain?
buses, tractors, cranes, helicopters, and rubbish
collection trucks. Four included studies reported
a moderate positive association between low Materials And Methods
back pain and sitting in a vehicle when com- The methods and reporting of this review
pared with control (non-driving) populations were guided by the Preferred Reporting Items
(average odds ratio [OR] = 2.1). The conditions for Systematic Reviews and Meta-Analyses
Driving Posture and Low Back Pain 3

Figure 1. Complete list of search terms and resulting search strategy.

(PRISMA) guidelines (Moher, Liberati, Tet- confounding factors, papers were excluded if
zlaff, & Altman, 2009). they included people with spinal anomalies,
genetic conditions associated with the spine,
Search Strategy insidious or serious pathologies of the lumbar
Searching was performed by the lead author. spine (e.g., cancer, inflammatory diseases, frac-
The complete holdings of five databases ture), conditions or injuries involving the spinal
(AMED, CINAHL Plus, Cochrane Central Reg- cord, or pregnant women. Studies that considered
ister of Controlled Trials, Cochrane Database sitting posture in non-standard forms of occupa-
of Systematic Reviews, and Ovid MEDLINE) tional automobiles were also excluded, due to
were searched up to March 12, 2018. Keywords factors associated with each vehicle type: differ-
used included “sit,” “automobile,” “low back,” ing driving surfaces in comparison with a stan-
“posture,” and their synonyms. Truncations dard public road (e.g., tractors), differing forces
were used where appropriate, and the Boolean acting through the body of the driver (e.g., rally
operators of “OR” and “AND” were applied cars, due to their extreme speed), and the differing
in line with the Population/Problem, Interven- requirements of the tasks being performed by the
tion/Exposure, Comparison, Outcomes (PICO) driver, such as operating additional vehicle con-
model (Richardson, Wilson, Nishikawa, & Hay- trols (e.g., forklifts, helicopters, street cleaners,
ward, 1995). Figure 1 displays the complete list garbage trucks). Studies that measured driving
of search terms and the search strategy imple- posture in simulated environments or measured
mented. Search results were pooled in biblio- proxies of posture (e.g., seat inclination) were also
graphic management software, duplicates were excluded, to ensure an understanding of the pos-
removed, and ineligible papers were screened tures assumed while reacting to traffic and road
and deleted based on title and abstract. All conditions, and to avoid the assumption that people
remaining full-text articles were then reviewed, sit in a static position with their back against the
with the final yield determined by the eligibility seat at all times.
criteria listed later. Reference lists of eligible
publications and relevant systematic reviews Quality Assessment
were then screened for additional papers not A critical analysis was performed by two
found in the primary search. Author and cita- independent researchers using the National
tion checks were performed for all first authors Heart, Lung, and Blood Institute’s (2014) Qual-
of eligible papers via the Scopus Database and ity Assessment Tool for Observational Cohort
Google Scholar. and Cross-Sectional Studies to consider sources
Papers were included in the review if they of bias in the individual studies. Conflicts were
reported data on the relationship between occupa- resolved by a third researcher. This tool was
tional sitting posture while driving a car, bus, chosen as it was the most appropriate quality
truck, or van in association with pain in the lum- assessment tool available for the study designs
bar spine region, and if they measured both driv- of the included papers. The tool assists users to
ing posture and low back pain. Only articles determine the degree of bias within a paper by
written in English were included. To minimize considering 14 items and then establishing a rat-
4 Month XXXX - Human Factors

ing of “poor,” “fair,” or “good” quality. A Grad- tions in the eligible articles were occupational
ing of Recommendations Assessment, Develop- drivers, including bus drivers, taxi drivers, police
ment, and Evaluation (GRADE) analysis (Grade officers, and drivers of trucks and vans. All seven
Working Group, 2004) was then performed by articles used an observational study design. The
two independent researchers to enable an assess- mean participant age across the included articles
ment of the body of evidence as a whole. The was 43.4 years. Sample sizes varied, ranging
GRADE is a tool used to assess the effectiveness from 61 to 1,242 participants. Participants were
of interventions in randomized or observational predominantly male due to the eligibility criteria
studies, including papers of narrative or statisti- for the individual studies or a lower number of
cal synthesis (Grade Working Group, 2004). females in the source population. Two of the
articles by Okunribido et al. (2006, 2008) pre-
Data Extraction and Synthesis sented data from the same source population. It
Data extracted from eligible papers included was assumed that the third article published by
study design, participant characteristics, meth- the same lead author included bus drivers from
ods used to assess driving posture and low back the same sample as the two previous papers, as
pain, methods of data analysis, and the reported population characteristics were identical (Okun-
statistical or textual description of the associa- ribido et al., 2007). These papers were handled
tion between occupational driving posture and as three separate studies, as they reported differ-
low back pain. Any evidence provided as to the ent relationships between driving posture and
validity or reliability of assessment and analysis low back pain.
methods was also extracted. Data extraction
was performed by the lead author. Checking Quality Assessment
and analysis of data were performed by all four The critical analysis of the included articles
authors. A meta-analysis was planned pending identified several domains that may give rise to
extraction of sufficient data that could be pooled. bias in the published results and related conclu-
sions drawn. An overall strength of the included
Results articles was the participation rates greater than
A total of 653 articles were identified by the 50%, indicating that the sample populations
search strategy. Thirty-three duplicates were were likely to be an adequate representation of
removed. Of the remaining 620 papers, 555 the target driving populations. However, only
were excluded based on title and abstract, and two of the seven studies provided sufficient
the remaining 65 underwent full-text review. information on the occupational driving popu-
Sixty-one articles were excluded based on lations from which participants were selected
full-text, with the remaining four publications (Chen et al., 2005; Tamrin et al., 2007), with
included (Chen, Chang, Chang, & Christiani, detail on location and the recruitment time not
2005; Okunribido, Shimbles, Magnusson, & provided in the remaining papers. This reduces
Pope, 2007; Sakakibara, Kasai, & Uchida, 2006; the reproducibility of the research. None of the
Tamrin et al., 2007). Three additional articles papers satisfied the “exposure measure” domain
were included following hand searching of eli- (National Heart, Lung, and Blood Institute,
gible reference lists, and a Scopus and Google 2014; see item 11b, Table 2), whereby none of
Scholar search of first authors from eligible the papers demonstrated posture measurement
articles (Bovenzi et al., 2006; Okunribido et al., methods that had undergone validity or reli-
2006, 2008). Figure 2 presents a PRISMA flow ability testing. A summary of the results can be
chart describing the search and screening results viewed in Table 2.
(Moher et al., 2009). When assessing the quality of the overall
body of evidence, the GRADE analysis (Grade
Working Group, 2004) indicated that this
Demographic Characteristics research was of low quality. Although many of
Table 1 presents the sample characteristics the articles reported an association between
and design of the included articles. The popula- occupational driving posture and low back pain
Driving Posture and Low Back Pain 5

Figure 2. Flow of studies into the review. LBP = low back pain.

(Bovenzi et al., 2006; Chen et al., 2005; Okunri- through direct observation by filming participants
bido et al., 2006, 2008), the methodological with a video camera as they completed their usual
flaws within each study design (e.g., a lack of work shift. This video footage was then scored
data on the reliability and validity of the posture by an unspecified number of raters to count the
measurement tools) result in poor generalizabil- number of times that various movements (bend-
ity for the occupational driving population. ing forward, leaning, sitting straight, twisting)
were performed. The participants in the study
Posture Measurement by Okunribido et al. (2007) were observed by an
All of the papers reported measurements of assessor while driving. This assessor noted the
posture while driving during a workday, using posture configuration (“torso against backrest,
a variety of measurement techniques. Tamrin torso straight, torso bent, torso twisted, and/or
and colleagues (2007) measured driving posture torso bent and twisted simultaneously”) assumed
6
Table 1: Study Design and Population Characteristics
Mean (SD)
Length of
Mean (SD) Mean (SD) Time in
Sample Size Sample Size Driving Mean (SD) Days of Current
Head Author Driver at Start of at End of Mean (SD) Gender Ratio Mean (SD) Distance per Driving Time Driving Each Profession
(Year) Country Type Study Study Age (Years) (M:F) BMI (kg/m2) Day (km) per Day (Hr) Month (Years)

Bovenzi et al. Italy B — 171 43.6 (6.6) 1:0 26.6 (3.2) — 6.0 (0.8) — 16.1 (8.5)
(2006)
Chen, Chang, Taiwan Ta 1,355 1,242 44.5 (8.7) 1,193:49 — — 9.8 (2.8) 26.2 (2.6) 11.4 (7.8)
Chang, and
Christiani
(2005)
Okunribido, United P P: 75 P: 58 P: 34.5 (5.9) Few women, P: 26.0 (2.7) — — — P: 12.9 (7.24)
Magnusson, Kingdom T/V T/V: 110 T/V: 64 T/V: 46.9 data T/V: 27.7 (4.5) T/V: 10.5 (8.6)
and Pope B B: 80 B: 61 (11.0) pooled B: 28.3 (4.4) B: 16.1 (11.7)
(2006) Ta Ta: 90 Ta: 30 B: 47.6 (10.4) Ta: 28.3 (4.8) Ta: 11.1 (8.7)
Ta: 49.3 (8.3)
Okunribido, United B 80 61 LBP: 48.1 (9.7) 19:1 LBP: 28.5 (4.0) — LBP: 7.5 (1.4) — LBP: 14.3 (7.8)
Shimbles, Kingdom No LBP: 46.8 No LBP: 28.1 No LBP: 7.6 No LBP: 17.9
Magnusson, (11.5) (4.9) (1.8) (12.2)
and Pope
(2007)
Okunribido, United P P: 75 P: 58 P: 34.5 (5.9) Few women, P: 26.0 (2.7) — — — P: 12.9 (7.24)
Magnusson, Kingdom T/V T/V: 110 T/V: 64 T/V: 46.9 data pooled T/V: 27.7 (4.5) T/V: 10.5 (8.6)
and Pope B B: 80 B: 61 (11.0) B: 28.3 (4.4) B: 16.1 (11.7)
(2008) Ta Ta: 90 Ta: 30 B: 47.6 (10.4) Ta: 28.3 (4.8) Ta: 11.1 (8.7)
Ta: 49.3 (8.3)
Sakakibara, Japan C — 551 LBP: 37 (8.5) 530:21 — LBP: 111 (54) LBP: 3.8 (1.8) — LBP: 14.2 (8.8)
Kasai, and No LBP: 35.8 No LBP: 115 No LBP: 3.5 No LBP: 12.8
Uchida (2006) (8.7) (104) (1.8) (8.7)
Tamrin et al. Malaysia B — 760 43 (8.64) — — — 10.5 (0.1) — —
(2007)

Note. M:F = male:female; BMI = body mass index; B = bus; Ta = taxi; P = police car; T/V = truck/van; C = car; LBP = low back pain.
 Table 2: Results of National Heart, Lung, and Blood Institute’s Quality Assessment

Okunribido,
Chen, Chang, Okunribido, Shimbles, Okunribido,
Chang, and Magnusson, Magnusson, Magnusson, Sakakibara,
Bovenzi et al. Christiani and Pope and Pope and Pope Kasai, and Tamrin et al.
(2006) (2005) (2006) (2007) (2008) Uchida (2006) (2007)

   1. Research question       
   2. Population identified       
   3. Participation rate       
   4. Participant recruitment and eligibility       
   5. Sample size and/or power calculation       
   6. Exposure of interest       
   7. Timeframe of exposure       
   8. Levels of exposure       
   9. Exposure measure defined, valid,       
reliable, and consistent
10. Repeated measure of exposure NA      
11a. Outcome measure (low back pain)       
defined, valid, reliable, consistent
11b. Outcome measure (driving posture)       
defined, valid, reliable, consistent
12. Assessor blinding       
13. Follow-up NA NA NA NA NA NA NA
14. Potential confounding variables       

Note. ✓ = Fulfills item criteria; ✗ = does not fulfill item criteria; NA = not applicable.

7
8 Month XXXX - Human Factors

by the driver at 1-min intervals (p. 30). The reli- were used to elicit posture assessment data
ability of these observation approaches were (Sakakibara et al., 2006).
not specified nor was whether the same asses-
sor made the observation on each occasion or Low Back Pain Measurement
if the assessor was trained to enhance reliability All seven papers measured low back pain
or validity. There was no measurement of the using a self-assessed questionnaire. Sakakibara
magnitude of movement in each direction or the et al. (2006) developed a questionnaire, while
length of time spent in each movement position the other authors used or modified existing
or posture. Bovenzi et al. (2006) performed direct questionnaires (Bovenzi et al., 2006; Chen et al.,
observation by taking video and still photogra- 2005; Okunribido et al., 2006, 2008; Okunribido
phy footage of participants during a work shift et al., 2007; Tamrin et al., 2007). The surveil-
but did not provide detail on the methods with lance questionnaire by Pope et al. (2002) was
which these data were interpreted. Bovenzi et al. also used to measure low back pain in the papers
and Okunribido et al. (2007) also assessed driv- by Okunribido et al. (2006, 2008; Okunribido
ing posture with a self-assessed questionnaire, in et al., 2007). Bovenzi et al. (2006) reported
addition to direct observation. the use of a modified Nordic Questionnaire yet
Six of the papers measured occupational referenced the standardized version and did not
driving posture using a self-assessed question- provide information regarding the modifica-
naire (Bovenzi et al., 2006; Chen et al., 2005; tions. Chen et al. (2005) based a questionnaire
Okunribido et al., 2006, 2008; Okunribido et al., on items from the Nordic Questionnaire and the
2007; Sakakibara et al., 2006). One paper asked Job Contents Questionnaire in combination with
participants to select a posture that they usually other items, but did not report the final item set.
adopted while driving from a polychotomous Tamrin et al. (2007) used the Standardised Nor-
list (“straight, slightly slouched, or slouched”; dic Questionnaire translated into Malay; how-
Sakakibara et al., 2006, p. 495). Participants ever, it is unclear if the best practice crosscheck,
were not permitted to specify a combination of as specified by Kuorinka et al. (1987), was
these postures or relative time spent in each pos- performed following translation. Sakakibara
ture, indicating an underlying assumption in this et al. developed a questionnaire with items that
measurement approach, that drivers adopt only measured the presence, severity, duration, and
one posture. Five of the questionnaires asked previous history of low back pain and the effect
participants to rate the frequency with which of driving on low back pain. They reported all
specific driving postures were adopted (Bovenzi questionnaire items.
et al., 2006; Chen et al., 2005; Okunribido et al.,
2006, 2008; Okunribido et al., 2007). None of
these papers reported data demonstrating the The Association of Driving Posture and
reliability or validity of this measurement Low Back Pain
approach. Okunribido et al. (2006, 2008; Okun- Four of the seven papers reported a sig-
ribido et al., 2007) reported the use of a previ- nificant association between low back pain
ously validated questionnaire developed by and driving with the back in a bent or twisted
Pope et al. (2002); however, the reference pro- posture (Bovenzi et al., 2006; Chen et al., 2005;
vided was for a draft questionnaire that did not Okunribido et al., 2006, 2008). This posture was
undergo investigation of its psychometric prop- associated with current low back pain (Okunri-
erties, nor has such investigation occurred since bido et al., 2006) and history of low back pain in
publication. Modifications to the questionnaire the past 12 months (Bovenzi et al., 2006; Chen
by Pope et al. were also apparent; Okunribido et al., 2005; Okunribido et al., 2008). Driving
et al. (2006, 2008; Okunribido et al., 2007) with the back bent and twisted simultaneously
described the consideration of five driving pos- was also significantly associated with history of
tures, yet the referenced questionnaire lists only low back pain in the past 12 months (Okunribido
three (Pope et al., 2002). Only one included et al., 2008). Okunribido et al. also reported a
study reported the wording of the items that significant association between driving with
Driving Posture and Low Back Pain 9

the torso against the backrest and both current score” for each participant by assigning severity
low back pain (2006) and history of low back “points” to specific postures (“torso against
pain in the past 12 months (2008). In contrast backrest—1, torso straight—2, torso bent—3,
to the findings of these four papers, Tamrin torso twisted—4, torso bent and twisted
et al. (2007), in their study that involved direct simultaneously—5”) and multiplying each by
observation of bus drivers (subgroup of posture frequencies of occurrence (“never—0, occasion-
measured, n = 143), reported no significant ally—1, often—2”), and then deriving their sum
association between driving postures adopted (Okunribido et al., 2007, p. 30). Drivers of police
during the observed period and self-reported cars and taxi drivers had low posture scores,
low back pain in the last 12 months. Sakakibara while truck, van, and bus drivers had medium
et al. (2006; n = 551) found no significant dif- posture scores (Okunribido et al., 2006). The
ference between the self-reported driving pos- earlier paper by Okunribido et al. (2006) found a
ture (classified as straight, slightly slouched, or significant association between posture score
slouched) of participants who have or have not and low back pain (previous and current) when
“had any low back pain lately” (p = .67; p. 495). posture score was analyzed as a continuous vari-
Okunribido et al. (2007; n = 61) did not report a able, while the other two papers by Okunribido
relationship between low back pain and driving et al. (2008) and Okunribido et al. (2007) found
posture as a discrete outcome. no significant association between these out-
Studies by Bovenzi et al. (2006) and Okunri- comes. Okunribido et al. (2006) reported that the
bido et al. (2006, 2008; Okunribido et al., 2007) posture score was yet to be validated.
also examined the relationship between a summa- Meta-analysis was not possible due to the
tive score incorporating occupational driving pos- heterogeneity in participants, measurement
ture (± other factors) and low back pain. Bovenzi methods, and study designs.
et al. calculated a self-developed “physical load
index” for each participant based on exposure to Discussion
awkward postures and heavy physical work. This review has found that there is evidence
Observational and questionnaire data were used to demonstrate a relationship between occupa-
to determine exposures. Awkward posture expo- tional driving posture and low back pain; how-
sure was calculated through consideration of each ever, this evidence is based on studies incorpo-
posture and how long participants assumed each rating bias, due to the use of measurement tools
posture during a shift (never, less than 1 hr, 1–2 lacking validity and reliability. Four of the seven
hr, or more than 2 hr). Exposure to heavy physical articles reported an association between driving
work was determined by rating the frequency posture and low back pain, and thus, it could
with which physical activities were performed be implied that a relationship exists between
(“not at all, 1–10 times, more than 10 times”; these outcomes; however, the specific postures
Bovenzi et al., 2006, p. 518). The average physi- associated with low back pain and the strength
cal workload index for the sample was then of the relationship between these outcomes are
divided into quartiles to represent grades of phys- not yet known.
ical load (mild, moderate, hard, and very hard). A Many of the included studies reported the use
very hard physical load index was associated with of validated tools to measure low back pain and
an increased risk of low back pain disability (OR driving posture; however, on closer inspection,
2.57, 95% confidence interval [CI] [1.25, 5.26]) it was apparent that this was not the case. Sev-
and history of low back pain within the previous eral papers made modifications to previously
12 months (OR 2.25, 95% CI [1.39, 3.64]; formed questionnaires without reporting the
Bovenzi et al., 2006). The physical workload validity of the modified tool or formed new
index incorporated both driving and non-driving questionnaires that did not undergo validity test-
work postures. Information was not provided as ing. The study by Tamrin et al. (2007) may have
to how the index was developed. used a validated questionnaire, yet the lack of
Okunribido et al. (2006, 2008; Okunribido information provided regarding the process by
et al., 2007) calculated a “personal posture which the Standardised Nordic Questionnaire
10 Month XXXX - Human Factors

was translated into Malay language prevents the measurements of lumbar angle, such as degrees
validity of this tool from being known. The risk of lumbar flexion, would prevent the use of pos-
of using unvalidated tools is the possibility of ture categories that may be interpreted differently
measuring a different outcome than is initially by readers, and provide an understanding of the
intended (Franklin & Lee, 2014). In the case of specific postures assumed while driving, and their
these studies, it is not known whether the mea- relationship with low back pain to be determined.
surements of driving posture are, in fact, mea- Furthermore, six of the eligible papers exam-
suring this outcome or another outcome such as ined multiple measurements of driving posture
fatigue. The paper by Okunribido et al. (2007) and/or analysis approaches with no a priori
reported a discrepancy between the findings of statement as to which measurement of driving
the self-reported questionnaire and the direct posture was primary and what cut off thresholds
observation of participants working. The partici- would be employed in the analyses. These
pants reported that “torso against backrest” was design features increase the risk of type 1 errors
assumed more commonly than “torso straight”; being made and inflate the strength of associa-
however the direct observation found the reverse tion between driving posture and low back pain
to be true. This indicates that the self-reported reported by the eligible papers (Haines, Hill,
questionnaire is not a reliable tool for the mea- Walsh, & Osborne, 2007).
surement of driving posture, as the results were The observational study design has an inher-
converse in their findings. To enable an under- ent risk of bias in comparison with randomized
standing of the relationship between occupa- controlled studies (Petrisor & Bhandari, 2007).
tional driving posture and low back pain, these This is not to say that observational designs are
outcomes must first be measured using validated not completely without merit. Observational
and reliable measurement tools. Real-time mea- designs can identify temporal sequencing that
surement of driving posture while a participant may help satisfy the Bradford Hill criteria of
completes their shift in their occupational vehi- causality (Hill, 1965). For this, prospective stud-
cle may be beneficial, to limit bias associated ies that demonstrate first, the adoption of par-
with retrospective self-report. ticular postures, followed by the development of
Four of the seven eligible articles reported a low back pain, are required. These studies would
relationship between some dimension of driving provide strong justification for intervention
posture and low back pain. The postures reported studies that could be used to demonstrate
to be associated with low back pain were, at whether changing postures assumed while driv-
times, contradictory, with both “driving with the ing for occupation can be useful for reducing the
back in a bent or twisted posture” and “driving burden of low back pain. No studies of this
with the torso against the backrest” associated nature were identified through this review.
with current low back pain and history of low Despite these methodological flaws, given
back pain in the last 12 months. This therefore that four of the seven eligible papers reported a
brings to question, are all of these driving pos- relationship between occupational driving pos-
tures associated with low back pain, or is further ture and low back pain, a relationship may exist
clarification or classification required about driv- between these outcomes. However, further
ing posture (e.g., Does “torso against backrest” research using more robust and reproducible
mean the back is not bent or twisted? Are there measurement approaches needs to be under-
differing degrees of torso bending that are associ- taken to confirm the specific postures assumed
ated with an increased risk of low back pain? while driving and the strength of the association
And, does “back bent” refer to lumbar flexion, between these postures and low back pain.
lumber extension, or both?)? The diverse methods Our findings are similar to the review by Lis
by which posture has been measured and ana- et al. (2007), who concluded that awkward pos-
lyzed by the eligible articles and the varied pos- tures, “sitting forward,” and “trunk flexed, bent,
tures reported to be associated with low back pain and twisted” were associated with an increased
demonstrate the need for tools that measure driv- risk of low back pain, and acknowledged that
ing posture in a quantitative way. Quantitative further studies using validated and reliable tools
Driving Posture and Low Back Pain 11

are required in this area (p. 289). This conclu- needed to guide the appropriate development of
sion by Lis et al. was determined by four studies future strategies to minimize the burden of low
(Bovenzi & Betta, 1994; Bovenzi & Zadini, back pain associated with occupational driving.
1992; Bridger, Groom, Jones, Pethybridge, & Measurement tools are now available that are
Pullinger, 2002; Massaccesi et al., 2003) that capable of quantifying an outcome such as sit-
were excluded from this review, as they were not ting posture that could be used for this purpose
among cars, buses, trucks, or vans, or included (Charry, Umer, & Taylor, 2011; Ronchi, Lech,
driving and non-driving postures in the same Taylor, & Cosic, 2008; Wartski, Williams,
data set. Despite similar results, due to the use of Broughton, & Bowles, 2019). These tools are
measurement tools lacking validity and reliabil- capable of measuring outcomes such as degrees
ity, it is unclear whether the inclusion of vehicles of lumbar flexion in real time, have been inves-
other than cars, buses, trucks, and vans could tigated for validity and reliability (Ronchi et al.,
limit the transferability of future research find- 2008), and are being used for clinical applica-
ings for standard road vehicles. tions and for monitoring low back movement
A limitation of this review is that the data patterns of workers during a typical workday
available did not permit a nuanced analysis of (Kent et al., 2015; Wartski et al., 2019). Future
the relationship between posture and low back studies using tools such as this could establish
pain across the potential moderating effect of the specific postures assumed while driving for
different vehicle types. The decision to exclude occupation, and their relationship with low back
articles that were not written in English was a pain. This would enable evidence-based recom-
further limitation of this study. Three of the mendations for clinicians and policy-makers
seven included articles were from Asian coun- working with people in the transport industry.
tries. It is not known how much research has The aim of this review was to determine if
been done and published in non-English lan- there is evidence demonstrating a relationship
guages. Another limitation of this study was the between occupational driving posture and low
choice to consider the relationship between driv- back pain. Four of the seven eligible papers
ing posture and low back pain, while ignoring reported an association, yet the inherent risk
the potential confounding effect of other factors, within the design of these studies limits the abil-
such as whole-body vibration or manual materi- ity to confidently understand the effect driving
als handling. The findings of this review, how- posture has on low back pain. A true understand-
ever, demonstrate the need for studies that uti- ing of this relationship cannot be known until
lize valid and reliable measurement tools to further studies using validated, reliable, quanti-
enable an understanding of the relationship tative, and real-time methods for posture mea-
between occupational driving posture and low surement are performed.
back pain, before the effect of the interplay of
posture and other factors can be considered.
Policies and recommendations that offer Acknowledgments
advice for the prevention and management of This systematic review was completed as part of
low back pain via modification of occupational a master’s degree, without a funding source. It is an
driving posture should be viewed with some advisory paper for a study funded by the Institute
caution, given the evidence available is based on for Safety, Compensation and Recovery Research
methods lacking validity and reliability, and a (ISCRR). Although this paper was written inde-
lack of understanding as to the specific postures pendently from the study, the authors would like to
associated with low back pain. Previous research acknowledge the support provided by ISCRR. T.H.
has shown that posture modification in patients and K.A.B. were paid from funding provided by the
with low back pain can be an economically effi- Victorian State Government and dorsaVi for a differ-
cient (Haines & Bowles, 2017) and a clinically ent publication related to low back movement moni-
effective method of improving pain and activity toring technology. The authors analyzed the results of
limitation (Kent, Laird, & Haines, 2015). An this research independently of both funders and nei-
understanding of the relationship between occu- ther funder had any influence on how these data were
pational driving posture and low back pain is presented or conclusions reached. T.H. has previously
12 Month XXXX - Human Factors

been paid to undertake analyses and report prepara- Franklin, A. E., & Lee, C. S. (2014). Effectiveness of simulation
for improvement in self-efficacy among novice nurses: A meta-
tion for the pilot randomized controlled trial that this
analysis. Journal of Nursing Education, 53, 607–614.
economic evaluation was based upon. For the remain- GBD 2017 Disease and Injury Incidence and Prevalence Col-
ing authors, no conflicts of interest were declared. The laborators. (2018). Global, regional, and national incidence,
authors also wish to acknowledge Megan Jepson for prevalence, and years lived with disability for 354 diseases and
injuries for 195 countries and territories, 1990–2017: A sys-
her contribution toward the quality assessment.
tematic analysis for the Global Burden of Disease Study 2017.
Global Health Metrics, 392, 1789–1858.
Grade Working Group. (2004). Grading quality of evidence
Key Points
and strength of recommendations. British Medical Journal,
•• This paper reviews the current literature pertain- 328(7454), 1490–1494.
ing to the relationship between occupational driv- Haines, T., & Bowles, K. A. (2017). Cost-effectiveness of using
a motion-sensor biofeedback treatment approach for the man-
ing posture and low back pain. agement of sub-acute or chronic low back pain: Economic
•• The review found that four out of the seven eligible evaluation alongside a randomised trial. BMC Musculoskeletal
articles reported a relationship between these out- Disorders, 18, Article 18.
comes; however, this was based on measurement Haines, T., Hill, K., Walsh, W., & Osborne, R. (2007). Design-
related bias in hospital fall risk screening tool predictive
approaches with unknown validity and reliability. accuracy evaluations: Systematic review and meta-analysis.
•• Further studies using valid and reliable, quantitative The Journals of Gerontology, Series A: Biological Sciences &
and real-time measurement techniques are required Medical Sciences, 62, 664–672.
before policy recommendations can be made. Hill, A. B. (1965). The environment and disease: Association or
causation? Proceedings of the Royal Society of Medicine, 58,
295–300.
ORCID iD Hoy, D., March, L., Brooks, P., Blyth, F., Woolf, A., Bain, C., . . .
Buchbinder, R. (2014). The global burden of low back pain:
Sarah Tinitali https://orcid.org/0000-0001-9054 Estimates from the Global Burden of Disease 2010 study.
-6215 Annals of the Rheumatic Diseases, 73, 968–974.
Johanning, E. (2000). Evaluation and management of occupational
low back disorders. American Journal of Industrial Medicine,
References 37, 94–111.
Bovenzi, M., & Betta, A. (1994). Low-back disorders in agricul- Kent, P., Laird, R., & Haines, T. P. (2015). The effect of chang-
tural tractor drivers exposed to whole-body vibration and pos- ing movement and posture using motion-sensor biofeedback,
tural stress. Applied Ergonomics, 25, 231–241. versus guidelines-based care, on the clinical outcomes of peo-
Bovenzi, M., Rui, F., Negro, C., D’Agostin, F., Angotzi, G., Bian- ple with sub-acute or chronic low back pain—A multicentre,
chi, S., . . . Stacchini, N. (2006). An epidemiological study of cluster-randomised, placebo-controlled, pilot trial. BMC Mus-
low back pain in professional drivers. Journal of Sound and culoskeletal Disorders, 16, Article 131.
Vibration, 298, 514–539. Krause, N., Ragland, D. R., Greiner, B. A., Fisher, J. M., Holman,
Bovenzi, M., & Zadini, A. (1992). Self-reported low back symp- B. L., & Selvin, S. (1997). Physical workload and ergonomic
toms in urban bus drivers exposed to whole-body vibration. factors associated with prevalence of back and neck pain in
Spine, 17, 1048–1059. urban transit operators. Spine, 22, 2117–2126.
Bridger, R. S., Groom, M. R., Jones, H., Pethybridge, R. J., & Kuorinka, I., Jonsson, B., Kilbom, A., Vinterberg, H.,
Pullinger, N. (2002). Task and postural factors are related to Biering-Sorensen, F., Andersson, G., & Jorgensen, K. (1987).
back pain in helicopter pilots. Aviation Space & Environmental Standardised Nordic questionnaires for the analysis of muscu-
Medicine, 73, 805–811. loskeletal symptoms. Applied Ergonomics, 18, 233–237.
Burdorf, A., & Sorock, G. (1997). Positive and negative evidence Lis, A. M., Black, K. M., Korn, H., & Nordin, M. (2007). Asso-
of risk factors for back disorders. Scandinavian Journal of ciation between sitting and occupational LBP. European Spine
Work, Environment & Health, 23, 243–256. Journal, 16, 283–298.
Charry, E., Umer, M., & Taylor, S. (2011, December). Design and Lyons, J. (2002). Factors contributing to low back pain among pro-
validation of an ambulatory inertial system for 3-D measure- fessional drivers: A review of current literature and possible
ments of low back movements. Paper presented at the Meeting ergonomic controls. Work, 19, 95–102.
of Intelligent Sensors, Sensor Networks and Information Pro- Manek, N. J., & MacGregor, A. J. (2005). Epidemiology of back
cessing (ISSNIP), Adelaide, South Australia. disorders: Prevalence, risk factors, and prognosis. Current
Chen, J. C., Chang, W. R., Chang, W., & Christiani, D. (2005). Opinion in Rheumatology, 17, 134–140.
Occupational factors associated with low back pain in urban Massaccesi, M., Pagnottaa, A., Soccettia, A., Masalib, M.,
taxi drivers. Occupational Medicine, 55, 535–540. Masieroc, C., & Grecoa, F. (2003). Investigation of work-
Claus, A. P., Hides, J. A., Moseley, G. L., & Hodges, P. W. (2009). related disorders in truck drivers using RULA method. Applied
Is “ideal” sitting posture real? Measurement of spinal curves in Ergonomics, 34, 303–307.
four sitting postures. Manual Therapy, 14, 404–408. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred
Dankaerts, W., O’Sullivan, P., Burnett, A., & Straker, L. (2006). reporting items for systematic reviews and meta-analyses: The
Differences in sitting postures are associated with nonspecific PRISMA statement. Annals of Internal Medicine, 151, 264–269.
chronic low back pain disorders when patients are subclassi- Moore, S. M., Torma-Krajewski, J., & Steiner, L. J. (2011). Report
fied. Spine, 31, 698–704. of investigations 9684: Practical ergonomic principles.
Driving Posture and Low Back Pain 13

Pittsburgh, PA: National Institute for Occupational Safety and Walker, B. F., Muller, R., & Grant, W. D. (2003). Low back pain in
Health. Australian adults: The economic burden. Asia-Pacific Journal
National Heart, Lung, and Blood Institute. (2014). Quality assess- of Public Health, 15, 79–87.
ment tool for observational cohort and cross-sectional studies. Wartski, M., Williams, C. M., Broughton, N. S., & Bowles, K. A.
Retrieved from https://www.nhlbi.nih.gov/health-topics/study- (2019). Quantifying the lumbar spine movements of surgeons
quality-assessment-tools during surgical lists in a teaching hospital. ANZ Journal of Sur-
Neumann, W. P., Wells, R. P., Norman, R. W., Kerr, M. S., Frank, gery, 89, 153–158.
J., & Shannon, H. S. (2001). Trunk posture: Reliability, accu- Williams, M. M., Hawley, J. A., McKenzie, R. A., & van Wijmen,
racy, and risk estimates for low back pain from a video based P. M. (1991). A comparison of the effects of two sitting pos-
assessment method. International Journal of Industrial Ergo- tures on back and referred pain. Spine, 16, 1185–1191.
nomics, 28, 355–365. Xia, T., Iles, R., Newnam, S., Lubman, D. I., & Collie, A. (2018).
Okunribido, O. O., Magnusson, M., & Pope, M. H. (2006). Low Driving Health Report No 1: Work-related injury and disease
back pain in drivers: The relative role of whole-body vibration, in Australian transport sector workers. Retrieved from https://
posture and manual materials handling. Journal of Sound and www.monash.edu/__data/assets/pdf_file/0006/1457448/Driv
Vibration, 298, 540–555. ing-Health-Report-1.pdf
Okunribido, O. O., Magnusson, M., & Pope, M. H. (2008). The Xia, T., Iles, R., Newnam, S., Lubman, D. I., & Collie, A. (2019).
role of whole body vibration, posture and manual materials Work-related injury and disease in Australian road transport
handling as risk factors for low back pain in occupational driv- workers: A retrospective population based cohort study. Jour-
ers. Ergonomics, 51, 308–329. nal of Transport & Health, 12, 34–41.
Okunribido, O. O., Shimbles, S. J., Magnusson, M., & Pope, M. Yilmaz, E., & Dedeli, O. (2012). Effect of physical and psycho-
H. (2007). City bus driving and low back pain: A study of the social factors on occupational low back pain. Health Science
exposures to posture demands, manual materials handling and Journal, 6, 598–609.
whole-body vibration. Applied Ergonomics, 38, 29–38.
Park, R. J., Tsao, H., Claus, A. P., Cresswell, A. G., & Hodges, P.
W. (2013). Recruitment of discrete regions of the psoas major
Sarah Tinitali completed a bachelor of physiotherapy
and quadratus lumborum muscles is changed in specific sitting (honors) at Monash University in Melbourne, Victo-
postures in individuals with recurrent low back pain. Journal ria, Australia, in 2015. She is currently completing
of Orthopaedic & Sports Physical Therapy, 43(11), 833–840. her master’s of philosophy in the Department of
Petrisor, B. A., & Bhandari, M. (2007). The hierarchy of evidence:
Levels and grades of recommendation. Indian Journal of
Physiotherapy, Monash University.
Orthopaedics, 41(1), 11–15.
Pope, M. H., Magnusson, M. L., Lundstrom, R., Hulshof, C., Ver- Kelly-Ann Bowles is the director of the Paramedic
beek, J., & Bovenzi, M. (2002). Guidelines for whole-body Health and Wellbeing Research Unit for the Depart-
vibration health surveillance. Journal of Sound and Vibration,
253, 131–167. ment of Community Emergency Health and Para-
Richardson, W. S., Wilson, M. C., Nishikawa, J., & Hayward, R. medic Practice at Monash University, Melbourne,
S. (1995). The well-built clinical question: A key to evidence- Victoria, Australia. She received her PhD in biome-
based decisions. American College of Physicians Journal
chanics at the University of Wollongong in 2013.
Club, 123(3), A12–A13.
Ronchi, A. J., Lech, M., Taylor, N. F., & Cosic, I. (2008, August). A
reliability study of the new back strain monitor based on clini- Jennifer L. Keating is an emeritus professor in the
cal trials. Paper presented at the Conference Proceedings of the Department of Physiotherapy at Monash University,
Institute of Electrical and Electronics Engineers Engineering in
Melbourne, Victoria, Australia. She received her PhD
Medicine and Biology Society, Vancouver, British Columbia,
Canada. in physiotherapy at La Trobe University in 1998.
Sakakibara, T., Kasai, Y., & Uchida, A. (2006). Effects of driving
on low back pain. Occupational Medicine, 56, 494–496. Terry Haines is the head of the School of Primary
Tamrin, S. B. M., Yokoyama, K., Jalaludin, J., Aziz, N. A., Jemoin,
N., Nordin, R., . . . Abdullah, M. (2007). The association
and Allied Health Care at Monash University, Mel-
between risk factors and low back pain among commercial bourne, Victoria, Australia. He received his PhD in
vehicle drivers in Peninsular Malaysia: A preliminary result. physiotherapy at the University of Melbourne in
Industrial Health, 45, 268–278. 2005.
Troup, J. D. G. (1978). Driver’s back pain and its prevention: A
review of the postural, vibratory and muscular factors, together
with the problem of transmitted road-shock. Applied Ergonom- Date received: February 21, 2019
ics, 9, 207–214. Date accepted: July 25, 2019