Article (PDF Available)

A replicated field intervention study evaluating the impact of a highly adjustable chair and office ergonomics training on visual symptoms

National Institute for Occupational Safety and Health, Analytic and Field Evaluations Branch, 1095 Willowdale Dr Road, MS-1811, Morgantown, WV 26505, USA.
Applied ergonomics (Impact Factor: 2.02). 10/2011; 43(4):639-44. DOI: 10.1016/j.apergo.2011.09.010
Source: PubMed

ABSTRACT

Examine the effects of two office ergonomics interventions in reducing visual symptoms at a private sector worksite.
A quasi-experimental study design evaluated the effects of a highly adjustable chair with office ergonomics training intervention (CWT group) and the training only (TO group) compared with no intervention (CO group). Data collection occurred 2 and 1 month(s) pre-intervention and 2, 6 and 12 months post-intervention. During each data collection period, a work environment and health questionnaire (covariates) and daily health diary (outcomes) were completed. Multilevel statistical models tested hypotheses.
Both the training only intervention (p<0.001) and the chair with training intervention (p=0.01) reduced visual symptoms after 12 months.
The office ergonomics training alone and coupled with a highly adjustable chair reduced visual symptoms. In replicating results from a public sector worksite at a private sector worksite the external validity of the interventions is strengthened, thus broadening its generalizability.

A replicated eld intervention study evaluating the impact of a highly adjustable
chair and ofce ergonomics training on visual symptoms
q
Cammie Chaumont Menéndez
a
,
b
,
*
, Benjamin C. Amick III
b
,
c
, Michelle Robertson
d
, Lianna Bazzani
e
,
Kelly DeRango
f
, Ted Rooney
e
, Anne Moore
g
a
National Institute for Occupational Safety and Health, Analytic and Field Evaluations Branch, 1095 Willowdale Dr Road, MS-1811, Morgantown, WV 26505, USA
b
The University of Texas School of Public Health, Southwest Center for Occupational and Environmental Health, 1200 Herman Pressler, Houston, TX 77030, USA
c
The Institute for Work and Health, 481 University Avenue, Toronto, Ontario M5G2E9, Canada
d
Liberty Mutual Research Institute for Safety, 71 Frankland Road, Hopkinton, MA 01748, USA
e
Health and Work Outcomes, Brunswick, ME, USA
f
The Upjohn Research Institute, 300 S. Westnedge Avenue, Kalamazoo, MI 49007, USA
g
York University, 349 Bethune, Toronto, Ontario M3J1P3, Canada
article info
Article history:
Received 29 April 2011
Accepted 26 September 2011
Keywords:
Ofce ergonomics intervention
Chair
Training
Visual symptoms
Replication
Multi-level modeling
abstract
Objective: Examine the effects of two ofce ergonomics interventions in reducing visual symptoms at
a private sector worksite.
Methods: A quasi-experimental study design evaluated the effects of a highly adjustable chair with ofce
ergonomics training intervention (CWT group) and the training only (TO group) compared with no
intervention (CO group). Data collection occurred 2 and 1 month(s) pre-intervention and 2, 6 and 12
months post-intervention. During each data collection period, a work environment and health ques-
tionnaire (covariates) and daily health diary (outcomes) were completed. Multilevel statistical models
tested hypotheses.
Results: Both the training only intervention (p < 0.001) and the chair with training intervention (p ¼ 0.01)
reduced visual symptoms after 12 months.
Conclusion: The ofce ergonomics training alone and coupled with a highly adjustable chair reduced
visual symptoms. In replicating results from a public sec tor worksite at a private sector worksite the
external validity of the interventions is strengthened, thus broadening its generalizability.
Published by Elsevier Ltd.
1. Introduction
According to a 1992 national survey of optometrists, computer-
related visual symptoms occurred in at least 14% of patients
(Sheedy, 1996). Furthermore, computer user surveys conducted at
work consistently reported that vision and eye problems were the
most frequently self-identied health problem (Collins et al., 1991;
Smith et al., 1981; Dain et al., 1988). It has been demonstrated that
the health effects associated with computer use impact produc-
tivity (Daum et al., 2004; DeRango et al., 2003). The use of
computers has grown exponentially in the last 25 years. Ofce
ergonomics intervention research suggests interventions designed
to improve upper extremity musculoskeletal symptoms likely share
their effectiveness with visual symptoms (Aaras et al., 1998, 2001,
2005; Dainoff et al., 20 05a,b; Horgen et al., 2004, 2005; Konarska
et al., 2005). A comprehensive ofce ergonomics intervention
designed to improve both musculoskeletal and visual health, work
performance, and productivity would be an invaluable tool as half
of the American workforce is estimated to regularly use computers
at work (US BLS, 2002). To date the ofce ergonomics literature
remains sparse (Brewer et al., 2006) with no replication of
successful interventions suggesting it works in more than one
unique employer willing to participate. The research reported in
this paper replicates an intervention designed to improve both
visual and musculoskeletal health in a public sector organization
(Amick et al., 2003). The intervention targeted improving upper
extremity musculoskeletal health. The theory of change (Fig. 1)
conceptualized the interventions as affecting visual health and,
q
The ndings and conclusions in this report are those of the authors and do not
necessarily represent the views of the Centers for Disease Control and Prevention,
National Institute for Occupational Safety and Health.
*
Corresponding author. National Institute for Occupational Safety and Health,
Analytic and Field Evaluations Branch, 1095 Willowdale Dr, MS-1811, Morgantown,
WV 26505, USA. Tel.: þ1 304 285 6233; fax: þ1 304 285 6235.
E-mail address: cammiemenendez@yahoo.com (C.C. Menéndez).
Contents lists available at SciVerse ScienceDirect
Applied Ergonomic s
journal homepage: www.elsevier.com/locate/apergo
0003-6870/$ e see front matter Published by Elsevier Ltd.
doi:10.1016/j.apergo.2011.09.010
Applied Ergonomics xxx (2011) 1e6
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 1
ultimately, productivity through ofce ergonomics knowledge and
behavioral changes. Specically, the highly adjustable chair allows
the ofce worker to make the postural and behavioral changes
already desired to reduce visual symptoms related to ofce
computer use. Additionally, the ofce ergonomics training by itself
is hypothesized to increase ergonomics knowledge, facilitating
adoption of postures and behaviors in the unchanged workstation
amenable to improved visual health. A combination of the highly
adjustable chair and ofce ergonomics training is thought to
maximize the ability of the ofce worker through learning about
correct postures and behaviors and how to accomplish them using
the chair as they relate to improved visual health. Furthermore,
improved visual health is expected to indirectly improve produc-
tivity at work through increased functional health. The end product
of improved productivity is also a direct result of correct computer
use postures and behaviors as well as satisfaction from having
a highly adjustable chair. The highly adjustable chair combined
with the ofce ergonomics training has already demonstrated
improved worker musculoskeletal and visual health with a 17%
increase in productivity in a public sector workplace (Amick et al.,
2003; Amick et al., submitted for publication; DeRango et al., 2003).
In this replication in a private sector workplace, we hypothesize
receiving a new highly adjustable chair and ofce ergonomics
training reduces worker visual symptoms at the end of the workday
compared to workers receiving only training and workers in
a control group. Furthermore, we hypothesize workers only
receiving training will have reduced visual symptoms at the end of
the workday compared to workers in a control group. Similarly, we
hypothesize the interventions reduce visual symptoms growth over
the workweek.
2. Methods
2.1. Study participants
Full-time ofce workers at an insurance company spending at
least 4 h a day computing and at least 6 h a day sitting in an ofce
and had not led a workers compensation claim within the past six
months were invited to participate. Participants had internet access
at work and worked in sedentary, computer-intensive ofce jobs.
All participants worked in one U-shaped building characterized by
two short arms and a long base.
2.2. Study design
A non-randomized quasi-experimental eld study was used to
evaluate the effect of two ofce ergonomics interventions on visual
health. Participants were divided into three study groups: those
receiving a highly-adjustable chair with ofce ergonomics training
(CWT), those receiving the ofce ergonomics training only (TO),
and a control group receiving training upon completion (CO).
Participants located at both ends (short arms) of the building were
assigned to the control group to best minimize ofce ergonomics
knowledge contamination through workplace interactions from
the study groups receiving the interventions. Supervisory groups in
the long arm were randomized to either the training only group or
the chair with training group. Information on daily visual health
symptoms, demographics, overall health status, and computing-
related habits and ofce conditions were collected at two months
and one month before interventions and took place again at two
months, six months, and twelve months post-intervention. The
study protocol was approved by the Liberty Mutual Research
Institute for Safety Human Subjects Internal Review Board.
2.3. Study interventions
The highly adjustable chair (see Fig. 2) was designed to give the
ofce worker control in improving how he or she ts in the ofce
workspace (Bush and Hubbard, 1999; Faiks and Allie, 1999).
Specically, fully-adjustable arm rests, seat pan, and back support
afford the ofce worker some leeway in movement while still
maintaining optimal gaze angle and distance crucial for visual
health related to computer use. The ofce ergonomics training was
developed for this workplace while applying instructional systems
Fig. 1. Theory of change.
Fig. 2. Highly adjustable chair.
C.C. Menéndez et al. / Applied Ergonomics xxx (2011) 1e62
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 2
design principles and adult learning theories to the training design
(Robertson et al., 2002, 2009). The training goals involved under-
standing ofce ergonomic principles, performing ergonomic self-
evaluation of workspace, and adjusting and rearranging ones
own workspace. From these goals nine training instructional
objectives were specied: recognizing work-related musculoskel-
etal disorders and risk factors, understanding the importance of
varying work postures, knowing how to rearrange the workstation
to maximize the comfort zone, recognizing and understanding
visual issues in the ofce environment and reducing visual
discomfort, understanding computing habits (rest breaks) and
knowing how to change work-rest patterns, being aware of the
companys existing health and ergonomic programs, and knowing
how to obtain ergonomic accessories through the companys
programs. Two co-facilitators delivered the hour and 45-min long
training.
2.4. Data collection
The Work Environment and Health questionnaire was admin-
istered a week following the web-based symptoms survey at each
administration period and consisted of questions on over thirty
covariates and potential confounders (a complete list is available
from the second author). Visual symptoms data administered by
daily symptoms diaries were collected at the end of the workday
over a one-week period. Participants completing less than four
days worth of questions were asked to complete it again the
following week. Respondents answered yes or no to having any of
the following 11 symptoms: stinging, itching, feeling gritty,
becoming red, tearing, feeling dry, burning, aching, feeling sensitive
to light, blurry vision, and difculty focusing. A visual symptoms
scale was constructed out of a count of number of symptoms
present (ranging from 0 meaning no symptoms present to 11
meaning all symptoms present). If information was missing on any
one question the scale was set to missing. The scale had an average
Cronbachs alpha of 0.98 across the ve survey rounds (range:
0.96e0.99).
2.5. Covariate selection
Covariate selection occurred through a pre-dened process.
Covariates meeting any of the following criteria were considered
for inclusion: (1) demonstrating an association with the outcome
variable, (2) not being highly correlated with other covariates, (3)
not being evenly distributed among study groups either pre- or
post-intervention, and (4) having substantively meaningful differ-
ences within the study groups across intervention periods.
A p-value less than or equal to 0.05 was used as a cutoff to assess
the signicance of the Pearsons correlation in meeting criterion 1,
and a correlation coefcient of <0.65 for criterion 2. Potential
confounders more signicantly associated with the outcome vari-
able were chosen among the variables highly correlated with each
other. A two-level variance components model with a potential
time-varying continuous covariate outcome was used to determine
if criterion 3 was met. In this model, the independent variables
were the study groups, intervention period, and their two-way
interactions. The joint chi-square statistic was used to determine
if there was a signicant difference between the study groups pre-
or post-intervention (p ¼ 0.10). A one-way ANOVA with a time-
invariant continuous variable was used to assess study group
differences (p ¼ 0.10). Cross-sectional time series logistic regression
models were used to assess study group differences of dichotomous
potential confounders (p ¼ 0.10). Before proceeding to criterion 4,
potential covariates meeting any of the rst three criteria were
placed into a multilevel linear model and a stepwise backwards
selection procedure was conducted where a non-signicant model
log likelihood reduction of p > 0.20 signaled removal from the
model. Evaluating whether or not criterion 4 was met required the
research team to examine mean differences for each potential
confounder by study group and intervention periods. The goal was
to obtain the most reasonably parsimonious and meaningful model
still correctly specied. Thus, each potential covariate was exam-
ined individually to establish meaningful differences appropriate to
that variable. Each candidate variable not demonstrating mean-
ingful differences was not included in the nal model.
2.6. Statistical analyses
A multilevel linear model was used to test the xed effect of each
study group, compared to the control group, in the context of inter-
vention period on visual health. Additional terms included in the
model were three potential confounders and a random error term.
The multilevel statistical model was formed by including the selected
cov ariat es, an indicator v aria ble for each of the two intervention
groups (with the control group as referent), a variable representing
study phase (0 ¼ pre-intervention, 1 ¼ post-intervention), and two
two-wa y interaction terms for study phase and group. The following
index ed variance components model w as used to explain the effect of
study phase, intervention group, andselectedcovariatesonvisual
symptoms.
y
ij
¼
b
0
þ
b
1
ð
chair
Þ
þ
b
2
ð
training
Þ
þ
b
3
ð
intervention
Þ
þ
b
4
ðchair*interventionÞþ
b
5
ðtraining*interventionÞ
þðcovariatesÞþu
j
þ ε
ij
where y
ij
¼ visual symptom measurement on the ith occasion for the
jth subject;
b
0
¼ overall mean for the control group;
b
1
¼ differential
effect of membership in CWT group;
b
2
¼ differential effect of
membership in TO group;
b
3
¼ differential effect of intervention;
b
4
¼ effect of receiving the CWT intervention;
b
5
¼ effect of receiving
the TO intervention (covariates) denotes terms correcting for selected
cov ariat es; u
j
¼ variation due to subjects; ε
ij
¼ variation due to occa-
sions of measurement within subjects.
It was assumed (subject to verica tion) that the random variables
m
j
and e
ij
are independentl y and normally distributed, each with mean
0 and constant variances
s
m
2
and
s
e
2
, respectiv el y. More specicall y, the
level 2 random variable,
m
j
, has a multivariat e normal distribution
with a constant covariance matrix. It was also assumed that the
dependent variable, y
ij
, is normally distributed with a mean equal to
the x ed part of the model with variance at level 1 and at level 2 of the
model (y w N(XB,
U
)).Finally,adistributionassumptionforunbal-
anced data in longitudinal analysis is the probability of being missing
was independent of an y of the random variables in the model.
The Wald Z statistic of two -way interaction term
chair * intervention was used to test the hypothesis that the
CW T group experienced a reduction in visual s ymptoms
compared to t he control group; similarly, the Wald Z statistic for
the two-way interaction term training * intervention tested the
hypothesis that the TO group experienced a reduction in visual
symptoms compared to the control group. A jo int chi-squared
with two degrees of freedom tested the hypothesis that the
CW T group experienced a reduction in visual symptoms signif-
icantly different from that of the TO group.
A residuals analysis was performed at level 1 and level 2 to
determine if model assumptions were violated or if systematic
variation was present among the residuals. Upon examination of
the residuals, it was determined there were no unwanted patterns
(no systematic variation) and no appreciable deviation from
normality (no violation of distributional assumptions).
C.C. Menéndez et al. / Applied Ergonomics xxx (2011) 1e6 3
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 3
Also tested was the hypothesis that visual symptoms increased
over the workweek: mean pre-intervention visual symptoms scores
were plotted by day of week to determine if there was a trend. No
clear trend was discernible, but introducing a day of week variable
into the model with just the intervention groups (study phase ¼ 0for
pre-interv ention) suggested, on average, a reduction in symptoms
over the workweek (
b
dayofw e ek
¼0.06, z ¼ 2.03, p ¼ 0.04). However,
there was no pattern of symptoms across the workweek to suggest
aspecic trend that the interv entions could affect and, therefor e, no
h ypothesis testing was conducted.
The statistical modeling and residual analysis were conducted
using MLwiN 1.1, summary statistics were carried out with Stata 8,
and the covariate selection process was conducted using both
software packages (MLwiN, 2001; StataCorp, 2003). All analyses are
available from the rst author.
3. Results
3.1. Study population
A workforce of 309 eligible persons was invited to participate
and 250 completed electronic informed consent (81% participation
rate). For the visual symptoms analysis, 181 workers provided
sufcient baseline data to conduct the analyses (66 workers in the
CWT group, 51 in the TO group, and 64 in the C group). At
12-months post-intervention, 154 participants completed the
questionnaire (85% retention, Table 1). A participant responding
every day of the workweek for all ve intervention periods
completed 25 symptoms surveys; the minimum considered as
completing the study was 15 (3 symptoms surveys a week).
Participants were nearly all white or Caucasian, the average age was
38 years, 90% were female, and the average time spent in an ofce
chair and computing was over 5e6 h/day.
Out of 59 non-responders,16 (representing 10e15% sample of non-
responders) wer e chosen to contact: seven did not answer the phone
after three attempts and one declined to participate. The eight who
agreed were on averag e seven years younger , aver age a year less of
education, rat ed their health and chair comfort better, had less bodily
pain relative to those who participated in the main study and spent
more time in the ofce chair and at the ofce computer but less time
using home computers on the weekday and weekend.
3.2. Covariates selected
After undergoing the same pre-dened and structured covariat e
selection process designed in the original intervention study at
a previous wor ksite, six covariates met criteria at all three levels of the
process: social support at work, decision latitude, distracted at work,
chair comfort, medication strength, and exer cise during the workday
to relieve pain and discomfort. In the interest of model parsimony,
only those that demonstrated meaningful changes pre- versus post-
interven tion and across groups were considered for inclusion as
cov ariates. The nal cov ariat e set was selected based on model
robustness, ease of interpretation, and comparability to the model
used in the public sector study (Amick et al., 2003). The nal cova-
riates were: chair comfort (1 ¼ strongly disagree chair is comfortable
to 4 ¼ strongly agree), medication strength (0 ¼ no pain medication
taken at all to 3 ¼ very strong prescription pain medications taken),
and proportion ex ercisin g during the workday to relieve pain or
discomfort. Table 2 show s the means or pro portions for each covariate
by study group and intervention period, in addition to the outcome
variable. Appendix 1 lists all of the potential study covariates.
3.3. Analytic results
Table 3
displays the multilevel model results. Two models were
constructed to demonstrate the contribution of the intervention
groups in reducing visual symptoms. Model 1 and Model 2 differ only
in that Model 2 includes the two-wa y interaction terms that test the
main study h ypotheses: the chair * intervention term tests the effect
of the chair and training intervention on visual symptoms and the
training * intervention term tests the effect of the training only
intervention on visual symp toms. The log likelihood difference
between the two models is statistically signicant (X
2
(2)
¼ 12.22,
p < 0.005), suggesting the two interventions (chair with training and
training only) improve model t. Specically, Model 2 illustrat es both
the CWT and the TO group experienced an improvement in visual
health through a statistically signicant reduction in visual sympt oms
(
b
chair*intervention
¼0.40, z ¼ 2.50, p ¼ 0.0 1;
b
training*intervention
¼0.53,
z ¼ 3.35, p < 0.001). There was a statistically signicant difference in
effect between the CWT group and the TO group (X
2
(2)
¼ 12.25,
p < 0.0 1). Fig. 3 is a graphical representation of the signicant effects of
both intervention groups and further illustrates the lack of meaningful
difference in effect between the two groups.
Each time period (2, 6, and 12 months) was considered separatel y
to determine if the interventions effect w as time-dependent (see
Fig. 4). At 2 months post-int erv en tion the TO group experienced
Table 1
Number of participants/number of completed visual symptom surveys by inter-
vention group and measurement period.
a
Intervention group Measurement period
12345
Chair and training (n ¼ 66) 62/234 58/238 61/261 62/259 56/246
Training-only (n ¼ 51) 48/178 45/198 45/190 49/219 47/213
Control (n ¼ 64) 61/236 61/244 53/212 61/251 51/238
Periods 1 and 2 are pre-intervention and 3e5 are post-intervention.
a
One participant responding every day for every period has 25 completed
surveys.
Table 2
Distribution of covariates and outcome by intervention period and intervention
group.
a
Intervention period
Pre Post
12345
Chair comfort
b
CWT 2.45 2.53 3.26 3.21 3.19
TO 2.48 2.60 2.73 2.72 2.80
Control 2.47 2.44 2.50 2.52 2.52
Medication strength
c
CWT 0.83 0.76 0.76 0.65 0.88
TO 1.11 0.77 0.72 0.79 0.83
Control 0.74 0.63 0.70 0.65 0.76
Exercised to ease pain or discomfort
d
(%)
CWT3634475247
TO 44 51 54 51 55
Control 42 47 37 44 39
Visual symptoms score
e
CWT 2.58 2.45 1.98 1.29 1.52
TO 2.90 2.62 2.06 1.78 2.00
Control 2.56 2.32 2.52 1.84 2.64
a
CWT refers to the chair-with-training group and TO refers to the training only
group.
b
Chair is comfortable ranges from 1 indicating strongly disagree to 4 indicating
strongly agree.
c
Medication strength ranges from 0 indicating taken no medicine at all for pain to
3 indicating taken very strong prescription pain medications such as Darvocet,
Lortab, Percocet, or Darvon.
d
Proportion who exercise during the workday to relieve pain or discomfort.
e
Visual symptoms ranges from 0 representing no visual symptoms to 11 repre-
senting having all visual symptoms listed.
C.C. Menéndez et al. / Applied Ergonomics xxx (2011) 1e64
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 4
a statisticall y signicant reduction in visual symptoms compared to
the contro l grou p (
b
training*int erve ntion
¼0.51; z ¼ 2.43; p ¼ 0.02),
while the CWT group was no t signicantly different fro m the contr ol
group (
b
chair*intervention
¼0.24; z ¼ 1.12; p ¼ 0.26). Additionall y, the
CWT and TO groups do differ signicantly from each other
(X
2
(2)
¼ 5.84, p ¼ 0.05). At 6 months post-interven tion neither the
CWT nor the TO group differed signicantly from the control group
(
b
chair*intervention
¼0.18, z ¼ 0.89, p ¼ 0.36;
b
training*intervention
¼0.23,
z ¼ 1.21, p ¼ 0.22). In addition, the CWT and TO groups did not differ
signicantl y from each other (X
2
(2)
¼ 1.62; p ¼ 0.45). However, at 12
months post-intervention both the CWT and TO groups demonstrated
a statisticall y signicant reduction in visual symptoms compared to
the control gr oup (
b
chair*intervention
¼0.52, z ¼ 2.46, p ¼ 0.01;
b
training*intervention
¼0.71, z ¼ 3.45, p < 0.001). Consistent with the
overall study ndings, there was a statistically signicant difference
between the two intervention groups (X
2
(2)
¼ 12.67; p < 0.0 1).
Due to sample size restrictions independent analyses for each
visual symptom could not be conducted.
4. Discussion
The ndings presented here are based on two ofce ergonomics
interventions originally conducted in a public sector (PU) worksite
(Amick et al., 2003) now replicated at a private sector (PR) worksite.
At both sites, the CWT group reported signicantly reduced visual
symptoms compared to the control group. However, at the PR site,
but not the PU site, the TO group experienced a signicant reduc-
tion in visual symptoms as well. At both sites the secondary
hypothesis for a workweek effect was tested and no workweek
trend in symptoms was found to exist.
The CWT group at the PU site experienced a greater reduction in
visual symptoms overall: the PU site effect w as larger than the
replication (PR) site effect (b
PU
¼0.49, SE ¼ 0.15; b
PR
¼0.40,
SE ¼ 0.16). Similarly in both groups the effect of the chair with training
was insignicant at 6 months post-intervention (b
PU
¼0.20,
p ¼ 0.32; b
PR
¼0.18, p ¼ 0.36) and then had its largest effect at 12
months (b
PU
¼0.67, p ¼ 0.001; b
PR
¼0.52, p ¼ 0.01). Overall, both
sites experienced a statistically signicant reduction in visual symp-
toms in the CWT group [b
PU
¼0.49,p < 0.00 1; b
PR
¼0.40, p ¼ 0.0 1).
There were differences in ndings between the priv ate sector and
public sector worksites for the TO group. Although the effect was in the
same direction for both sites, the effect at the PR site was larger than
that for the PU site (b
PR
¼0.53, SE ¼ 0.1 6; b
PU
¼0.13, SE ¼ 0.17). For
the PU site, the training-onl y effect never resulted in a signicant
visual symptoms reduction although the effect grew larger
throughout follow-up. At 2 months, there was no change in visual
symptoms (b
PU
¼ 0.08; p ¼ 0.72), at six months the effect started to
grow but was in the wrong direction (b
PU
¼ 0. 13; p ¼ 0.5 7), while at 1 2
months there was a larger reduction in visual symp toms, though
insignicant (b
PU
¼0.30; p ¼ 0.20). How ever, in the PR site the
training-onl y intervention led to a signicant reduction in visual
sympt oms both at 2 months post-intervention (b
PR
¼0.51; p ¼ 0.01)
and at 12 months (b
PR
¼0.71; p ¼ 0.001). A t 6 months there w as no
signicant reduction in visual symptoms reported by the training-only
groupattheprivatesectorworksite(b
PR
¼0.23; p ¼ 0.22).
The hypotheses stated the effect of the CWT group would be
larger than the TO and control groups, and that of the TO group
would be larger than the control group. The PU sector data sup-
ported the hypothesis that the effect of the CWT group was larger
than the TO and control groups but there were no differences
between the TO and control groups. However, at the PR site, both
the CWT and TO groups experienced signicant visual symptoms
reduction compared to the control groups, but there was no
difference in visual symptom reduction between the CWT and the
TO groups. Data on workstations during the training are being
analyzed to better describe how the proposed framework impacted
visual health. Potential differences in workstation adjustability
between the two sites may account for the lack of a difference in
visual symptoms reduction between the two study groups.
There was a time-dependent effect for both intervention groups
at both worksites in that the largest effect size was seen at 12
months. Interestingly, at 6 months both worksites and study groups
experienced a drop in effect size accompanied by an insignicant
visual symptoms reduction. It is not clear why the intervention
effect experienced this dip at 6 months, then regained its signicant
effect at 12 months. Changes in daylight and seasonal allergies are
not likely as both worksites had very limited natural lighting and
seasons did not really coincide with the timing of the study to
produce these changes. One reason for this time-dependent effect
could be a honeymoon effect that ended after the early post-
intervention time period where the novelty of the training and
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
erocSsmotpmySlausiVniegnahC
Study Group
Chair With Training Training Only Control
Fig. 3. Absolute change in visual symptoms score by study group.
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
sr
etemaraPtceffE
Study Group
Control Training only Chair with Training
Fig. 4. Effect parameters by study group and time period.
Table 3
Multi-level model including and excluding main effect parameters.
Variable Model 1
(std. error)
Model 2
(std. error)
Chair comfort 0.41 (0.09)* 0.34 (0.10)*
Exercise due to pain 0.31 (0.09)* 0.29 (0.09)*
Strength of medications 0.20 (0.06)* 0.19 (0.06)*
b
1
: Chair-and-training group 0.26 (0.36)
ns
0.05 (0.37)
ns
b
2
: Training-only group 0.03 (0.38)
ns
0.28 (0.40)
ns
b
3
: Intervention phase 0.44 (0.07)* 0.17 (0.11)
ns
b
4
: Chair-intervention e 0.40 (0.16)*
b
5
: Training-intervention e 0.53 (0.16)*
b
0
: Intercept term 3.68 (0.33) 3.36 (0.36)
ε
ij
: Level 1 variance 2.95 (0.08) 2.94 (0.08)
m
j
: Level 2 variance 3.95 (0.43) 3.97 (0.43)
Model 2 ln (likelihood) 13526.02 13513.80
Difference in 2 ln (likelihoods) ¼ 12.22*
* ¼ p < 0.05; ns ¼ p > 0.05.
C.C. Menéndez et al. / Applied Ergonomics xxx (2011) 1e6 5
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 5
chair diminished and a personal commitment to workstation
behaviors had yet to be made. Continuing in this logic, somewhere
between 6 and 12 months the behavior changes were adopted and
a signicant reduction in visual symptoms was seen.
Replication studies make crucial contributions in evaluating the
generalizability of an intervention. Each worksite offered its own
unique challenges with respect to pre-existing workstations, ofce
layout, and expectations from management as to how the study
would be conducted. External validity represents the extent to
which a causal relationship holds over variations in persons,
setting, interventions, and outcomes (Shadish et al., 2002).
Observing the reduction in visual symptoms occur for the chair
with training group and training only group at ofce worksites that
are completely different from each other serves only to raise the
external validity of the causal relationship between the interven-
tions and the visual symptoms. Successful interventions such as
these are easier and more practical to implement in other ofce
environments because of the demonstrated generalizability.
Acknowledgments
This research was funded by grants from Steelcase, Inc. to the
University of Texas, The Upjohn Research Institute, York University
and Health and Work Outcomes and through the support of the
Liberty Mutual Research Institute for Safety. At the time of the study
Cammie Chaumont Menéndez was a doctoral student at the
University of Texas supported by NIOSH Occupational Injury
Prevention Training Grant T42 0H008421.
Appendix 1
Comprehensive list of potential study covariates.
References
Aaras, A., Horgen, G., Bjorset, H., et al., 1998. Musculoskeletal, visual and psycho-
social stress in VDU operators before and after multidisciplinary ergonomic
interventions. Appl. Ergon. 29, 335e354.
Aaras, A., Horgen, G., Bjorset, H., et al., 2001. Musculoskeletal, visual and
psychosocial stress in VDU operators before and after multidisciplinary
ergonomic interventions. A 6 years prospective studydPart II. Appl. Ergon.
32, 549e557.
Aaras, A., Horgen, G., Ro, O., Loken, E., Mathiasen, G., Bjorset, H.H., Larsen, S.,
Thoresen, M., 2005. The effect of an ergonomic intervention on musculoskel-
etal, psychosocial and visual strain of VDT data entry work: the Norwegian part
of the international study. Int. J. Occup. Saf. Ergon. 11, 25e47 .
Amick III, B.C., Roberston, M.M., DeRango, K., et al., 2003. Effect of ofce ergo-
nomics intervention on reducing musculoskeletal symptoms. Spine 28,
2706e2711 .
Amick, B.C., III, Chaumont Menéndez, C., Bazanni, L., et al. A eld intervention
examining the impact of an ofce ergonomics training and a highly adjustable
chair on visual symptoms. Appl. Ergon., submitted for publication.
Brewer, S., Van Eerd, D., Amick III, B.C., et al., 2006. Workplace interventions to
prevent musculoskeletal and visual symptoms and disorders among computer
users, a systematic review. J. Occup. Rehabil. 16, 325e358.
Bush,T.,Hubbard,R.,1999.Anevaluationofposturalmotions,chair
motions and contact in four ofce seats. In: Proceedings of the Human
Factors and Ergonomics S ociety 4 3rd Annual Meeting, Houston, Texas,
pp. 589e593.
Collins, M.J., Brown, B., Bowman, K.J., et al., 1991. Task variables and visual
discomfort associated with the use of VDTs. Optom. Vis. Sci. 68, 27e33.
Dain, S.J., McCarthy, A.K., Chan-Ling, T., 1988. Symptoms in VDU operators. Am. J.
Optom. Physiol. Opt. 65, 162e167.
Dainoff, M.J., Aaras, A., Horgen, G., Konarska, M., Larsen, S., Thoresen, M.,
Cohen, B.G., MEPS Study Group, 2005a. The effect of an ergonomic intervention
on musculoskeletal, psychosocial and visual strain of VDT entry work: organi-
zation and methodology of the international study. Int. J. Occup. Saf. Ergon. 11,
9e23.
Dainoff, M.J., Cohen, B.G., Dainoff, M.H., 2005b. The effect of an ergonomic inter-
vention on musculoskeletal, psychosocial, and visual strain of VDT data entry
work: the United States part of the international study. Int. J. Occup. Saf. Ergon.
11, 4 9e63.
Daum, K.M., Clore, K.A., Simms, S.S., et al., 2004. Productivity associated with visual
staring among computer users. Optometry 75, 33e47 .
DeRango, K., Amick, B.C., III, Robertson, M.M., et al., 2003. The productivity conse-
quences of two ergonomic interventions. Institute for Work and Health (Tor-
onto, Ontario) Working Paper Series, #222.
Faiks, F., Allie, P., 1999. Supporting the torso through seated articulation. In:
Proceedings of the Human Factors and Ergonomics Society 43rd Annual
Meeting, Houston, Texas, pp. 574e578.
Horgen, G., Aaras, A., Thoresen, M., 20 04. Will visual discomfort among visual
display unit (VDU) users change in development when moving from single
vision lenses to specially designed VDU progressive lenses? Optom. Vis. Sci. 81,
341e349.
Horgen, G., Aaras, A., Dainoff, M.J., Konarska, M., Thoresen, M., Cohen, B., MEPS
Study Group, 2005. A cross-country comparison of short- and long-term effects
of an ergonomic intervention on musculoskeletal discomfort, eyestrain and
psychosocial stress in VDT operators: selected aspects of the international
project. Int. J. Occup. Saf. Ergon. 11, 77e92.
Konarska, M., Wolska, A., Widerszal-Bazyl, M., Bugajska, J., Roman-Liu, D., Aaras, A.,
2005. The effect of an ergonomic intervention on musculoskeletal, psychoso-
cial, and visual strain of VDT data entry work: the Polish part of the interna-
tional study. Int. J. Occup. Saf. Ergon. 11, 65e76.
MLwiN [software]. Version 1.10, 2001. Multilevel Models Project Institute of
Education, London.
Robertson, M.M., Amick III, B.C., DeRango, K., et al., 2002. The i mpact of an
of
ce ergonomics training on worker knowledge, behavior and musculo-
skeletal risk. In: Proce edings of the 6th International Sci enticConference
on Work with Display Units. ERGONOMIC Institut für Arbeit-und
Sozialforschung Forschungsgesellschaf t mbH, Berchtesgaden, Germany,
pp. 376e378.
Robertson, M.M., Amick III, B.C., DeRango, K., Rooney, T., Bazzani, L., Harrist, R.H.,
Moore, A., 2009. The effects of an ofce ergonomics training and chair inter-
vention on worker knowledge, behavior and musculoskeletal risk. Appl. Ergon.
40, 124e135.
Shadish, W.R., Cook, T.D., Campbell, D.T., 2002. Experimental and Quasi-
Experimental Designs for Generalized Causal Inference. Houghton Mifin
Company, Boston.
Sheedy, J.E., 1996. The bottom line on xing computer-related vision and eye
problems. J. Am. Optom. Assoc. 67, 512e516.
Smith, M.J., Cohen, B.G.F., Stammerjohn Jr., L.W., 1981. An investigation of health
complaints and job stress in video display operations. Hum. Factors 23,
387e400.
Stata [software]. Release 8.0, 2003. Stata Corporation, College Station, Texas.
US Bureau of Labor Statistics, 2002. Computer and internet use at work in 2001,
USDL 02-601. News Release retrieved December 12, 2004. Available from:
http://www.bls.gov/news.release/ciuaw.toc.html.
Hours spent sitting in ofce chair in typical week
Hours spent working at ofce computer in typical week
Total computer use
Number of breaks taken in the past week
Repetitive hand and wrist activity
How much force used with hand and wrists in ofce
Using skills at work
Authority at work
Decision latitude at work
Psychological demands at work
Social support
Workstation layout
Level of glare produced by lighting
Level of distraction by noise at work
Satisfaction level of workplace privacy
Chair comfort
Chair satisfaction
Body mass index
General (Poor) Health
Taken any medicine at all for pain
Frequency of prescription pain medications taken in past month
Medication strength
Exercise to relieve pain
Type of eyeglasses lenses worn
Education level
Level
Age
Job tenure
Gender
Disability status
Marital status
Number of persons living in household
Racial/ethnic background
Did job demands prevent rest breaks from computer
C.C. Menéndez et al. / Applied Ergonomics xxx (2011) 1e66
Please cite this article in press as: Menéndez, C.C., et al., A replicated eld intervention study evaluating the impact of a highly adjustable chair
and ofce ergonomics training on visual symptoms, Applied Ergonomics (2011), doi:10.1016/j.apergo.2011.09.010
Page 6
  • Source
    • "Deney, gönüllü olarak seçilen 6 denekle (3 bay, 3 bayan) gerçekleştirilmiş ve sonuç olarak sandalye arkalığına yaslanmanın ve hareketliliğin lumbar bölgedeki ağrıyı azaltmaya yardımcı olduğunu savunmuşlardır [19] kullananlar ve ofis ergonomisi eğitimi alanlar, sadece ofis ergonomisi eğitimi alanlar ve bunların ikisini de gerçekleştirmeyenler olarak 3 farklı gruba ayırmışlardır. Çalışmanın sonucunda hem ofis ergonomisi eğitimi alan hem de yüksek ayarlanabilir ofis sandalyesi kullanan grup üzerinde 12 ay sonunda görünen semptomların azaldığını tespit etmişlerdir [25]. "
    [Show abstract] [Hide abstract] ABSTRACT: Çalışmada, insanların birebir ilişkide olduğu oturma mobilyalarının tasarımında, ergonomik kriterlerin belirlenmesi amacıyla bir " insan odaklı ürün geliştirme " yazılımının (Anybody Technology) kullanılabilirliği araştırılmıştır. Yapılan örnek uygulamada, oturma derinliği, oturma yüksekliği ve arkalık eğim açısı farklı olan 2 adet sandalyede oturan insanın kas– iskelet sisteminde oluşan yükler niceliksel olarak analiz edilmiştir. Ayrıca, 10 erkek deneğin katılımıyla uygulanan bir anket yardımı ile de bu iki farklı sandalyede oturan bireylerin çeşitli vücut bölgelerinde hissettikleri duygular niteliksel olarak belirlenmiştir. Çalışma sonucunda, ürün geliştirme yazılımından alınan veriler, anketlerden alınan veriler ile tutarlı olup, insan odaklı ürün geliştirme yazılımının tasarlanacak mobilyaların ergonomik kriterlerinin belirlenmesinde kullanılabilir olduğunu doğrulamaktadır. Anahtar Kelimeler: Ergonomi, mobilyada ergonomi, oturma mobilyası, Anybody Technology
    Full-text · Article · Apr 2015
  • Source
    • "Şirket çalışanlarını, yüksek ayarlanabilir ofis sandalyesi kullananlar ve ofis ergonomisi eğitimi alanlar, sadece ofis ergonomisi eğitimi alanlar ve bunların ikisini de gerçekleştirmeyenler olarak 3 farklı gruba ayırmışlardır. Çalışmanın sonucunda hem ofis ergonomisi eğitimi alan hem de yüksek ayarlanabilir ofis sandalyesi kullanan grup üzerinde 12 ay sonunda görünen semptomların azaldığını tespit etmişlerdir [25]. "
    [Show abstract] [Hide abstract] ABSTRACT: Ergonomi birçok alanda olduğu gibi mobilyanın da önemli birleşenlerinden bir tanesini teşkil etmektedir. Özgün endüstriyel mobilya tasarımında ergonomi konusunda çekilen sıkıntılar; ilgili alanda derinlemesine analiz ve araştırmaların yeterince yapılmamış olmasına ve mevcut analizlerin niceliksel verilere dayandırılamamasına bağlanabilir. Çalışmada, insanların birebir ilişkide olduğu oturma mobilyalarının tasarımında, ergonomik kriterlerin belirlenmesi amacıyla bir “insan odaklı ürün geliştirme” yazılımının (Anybody Technology) kullanılabilirliği araştırılmıştır. Yapılan örnek uygulamada, bir adet ergonomik kriterlere uygun, bir de uygun olmayan oturma elemanlarında oturan insanın kas–iskelet sisteminde oluşan yükler analiz edilmiştir. Ürün geliştirme yazılımından alınan veriler kabul edilebilir düzeyde olup, yazılımın tasarlanacak mobilyaların ergonomik kriterlerinin belirlenmesinde kullanılabilir olduğunu doğrulamaktadır. This paper discusses the chair design in regards to ergonomics. Different chair prototypes are manufactured parametrically defining the design space. Design alternatives are assessed for comfort using questionnaire methods. Computer Aided Ergonomics software Anybody Modelling System is used to obtain quantitative values for defining comfort. Our results show that CAE Ergonomics matches questionnaire based methods and provides detailed information beyond EMG measurements.
    Full-text · Article · Apr 2015
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: Dry eye syndrome in the workplace is associated with new ways of working, with increasing use of screens and electronic devices and environmental conditions encountered in modern office designs and other environments. Also affect occupational exposure to ionizing radiation, chemicals or atmospheric dust with increased ocular dryness. The study of pathophysiological aspects and laboral causality of the dry eye, must be to develop joint task in Occupational Health, Public Health in coordination with and responsible for the national health system, which would involve primary and secondary preventive measures more effective and proper diagnosis, control and monitoring of the disease, A better knowledge of occupational hazards and actions agreed and coordinated between occupational physicians, preventers, primary care physicians and specialist physicians, such as ophthalmology, will get results much more effective when earlier and optimize available resources.
    Full-text · Article · Aug 2013 · SEMERGEN - Medicina de Familia
Show more