Dear Dr. Karolina Szerencsi:
We really appreciate your comments to improve our manuscript. We have taken
these comments into account and herewith submit a revised version of our paper.
In the following pages are our responses to your comments. We hope that our
explanations and revisions are satisfactory and the revised version of our paper is now
suitable for publication.
REVIEWER COMMENTS
The comments listed below can be categorized as ‘Major Compulsory
Revisions’.
Comment: Some issues need clarification, such as the rationale, the definition
of health status, whether the author checked the assumptions before using
regression analyses. Furthermore, certain choices need clarification, e.g. why
exclude those> 50 years of age?, more detailed description of the assessment of
all variables used in the study. In addition, some explanations given in the
discussion are not supported (by the data): e.g. smoking could clarify the
association between worse blood pressure, lipid and glucose and fixed term
contract, while in the analyses adjustment has been performed for smoking. The
strobe checklist for observational studies (cross sectional) might be helpful for
the reporting of the data and the description of the study.
Response: We excluded workers older than 50 years of age because most
workers were employed on permanent contracts at these age groups; 609 (95%)
were permanent workers and 33 (5%) were fixed-term workers over the age of
50. We analyzed the data using all samples, including workers older than 50,
and the results did not change much. We therefore decided to present the results
as a comparison between permanent and fixed-term workers not categorized by
average age.
More detailed description of all variable were added to our revised manuscript
as you also suggested us in 1-e).
Also, as we will describe to your comment of 5-a), our results were implied
with caring the adjustment of smoking in multiple regression analysis.
The STROBE checklist for cross-sectional studies has been included in the
1
revised manuscript.
1. Is the question posed by the authors well defined?
a) The purpose of the study is defined as ‘to compare obesity, lipid and glucose
metabolism and health related factors between permanent and fixed term
workers employed in the manufacturing industry’ in the abstract, whereas in
the paper (in the last sentence of the last paragraph of the background) it is
formulated as ‘to compare health status in terms of obesity and lipid and
glucose metabolism by contract types’. In the latter, the population is missing.
Furthermore, how do you define your outcome? Please include a definition.
Response: The purpose of the study has been revised. The purpose now reads:
“The current study compared BMI, blood pressure, and lipid and glucose
metabolism, and health-related lifestyle factors between permanent workers
and fixed-term workers employed in the manufacturing industry.” (page 6, line
10)
The definition of “health” used in the study included factors related to health
(obesity, blood pressure levels, and metabolic index). The outcome has been
described directly in the revised manuscript.
b) Furthermore, to define health status as obesity and lipid and glucose
metabolism is strange to my opinion. Obesity and lipid and glucose
metabolism are risk factors of health status. If you wish to study health status,
why do you exclude hypertension, hyperlipidemia and diabetes (last
sentence paragraph methods-study subjects)? These are more closer to
health status.
Response: Use of the vague term “health” has been corrected. The revised
manuscript has been changed to explain the outcome as risk factors of health
rather than using “health status”.
We excluded workers who had already been diagnosed with hypertension,
hyperlipidemia, and diabetes, because these people might be on medication to
improve their blood pressure and lipid and glucose metabolism, and medication
might change the results.
c) Rationale of the study -> do want to gain insight into whether fixed term and
permanent workers differ in health status? Or specifically in indicators of
2
health status, and if so why?
Response: The rationale of the study was a comparison of health risk factors
between fixed-term and permanent workers.
The term “health status” is confusing and has been changed to “risk factors of
health” in the revised manuscript.
2. Are the methods appropriate and well described?
a) I do not agree with the statement ‘this factory treats permanent and fixed term
workers identically in terms of physical safety because both are provided
protective equipment. Thus, fixed term workers may experience similar
working conditions as permanent workers, except for contract length, salary
and working hours’. What about job related tasks and psychosocial work
characteristics, such as job demands, job control, role and task clarity etc (see
Kompier et a 2009 ‘Employment contracts: cross sectional and longitudinal
relations with quality if work life, health and well being’). How do the two
groups differ with regard to the mean number of working hours per week? In
the article only overwork (yes/ no) is given. Does the statement imply that fixed
term and permanent workers have identical jobs (and tasks) within this
factory? Do you have data to support this statement? Do you know anything
about work history of the two groups and whether they differ? This information
could offer insight into why there are differences between fixed and permanent
workers in BMI, blood pressure, blood glucose and triglyceride.
Response: These sentences are to provide information on safety equipment
only, as the conditions leading up to injury are often investigated in
occupational health studies.
Differences in working hours are described in the Results section, and job
description is discussed in the Discussion section, because these factors might
influence the risk factors of health.
b) Why did you exclude the workers aged 50-60 years? Do you expect
differences in ‘health risks’ of fixed term contracts for those > 50 and < 50? If
yes, why?
Response: Analysis including 50–60 year olds did not change the results. Since
the number of fixed-term workers aged 50–60 was small (n=33) compared with
3
permanent workers (n=609), it was possible that this difference might have
distorted the results. Therefore, we chose workers younger than 50.
c) Do you know anything about the 119 employees for whom anthropometric and
blood pressure data was missing? Did they differ in the distribution of
contract type? Did they differ in working conditions or age, gender, smoking
status etc?
Response: Among the 119 employees, 118 were male and one was female.
Almost all were 20–30 years old, non-smokers, and permanent workers. It is
possible that this population might have decided not to attend or miss
check-ups because they thought they were healthy, and having only an annual
health check-up was sufficient.
d) Section ‘health data’, second paragraph: how was blood pressure assessed?
Which instrument, more than one measure to reduce variability?
Response: Blood pressure was measured using an UM-15P (Parma Tech) by
skilled public health nurses. Measurements were only taken once. The data
were from an annual health check-up.
e) Section ‘health related lifestyles’: Please describe how all variables were
measured, e.g. formulate the questions how smoking status, exercise and
alcohol consumption were inventoried. Please describe all categories and how
the variables were adjusted for in the analyses? E.g. It is not clear from the
text how many categories you have for alcohol: no drinker, drinker, heavy
drinker. Do you use the same groups when you adjust for alcohol in the
analyses or do you use the continuous answer scale (g ethanol per week)?
Response: How variables were measured has been included in the revised
manuscript in the “Health-related lifestyle variables” section.
When we adjusted for alcohol in the analysis, we used a dummy variable
referred to as “non-drinker”.
f) Please provide some information about the validity of the questions used to
assess health data and health related lifestyle variables (in the discussion
section).
Response: Our study questionnaire used questions commonly used in
4
occupational health settings in Japan, and by reference to questions from the
National Health and Nutrition Survey by the Ministry of Health, Labour and
Welfare of Japan. This information has been added to the Method section.
(page 10, line 1)
g) Did you measure socio-economic status? If yes, how?
Response: We did not measure socioeconomic status. However, the study
population was all high school graduates, and the level of lowest educational
attainment was the same between permanent and fixed-term workers.
h) Section statistical analysis, formulation of the sentence ‘multiple regression
analysis was used to estimate and compare the BMI, blood pressure and
glucose metabolism data.’ Strange sentence, what did you compare?
Response: This sentence has been changed in the revised manuscript. The
sentence now reads: “Multiple regression analysis was used to estimate BMI,
blood pressure, and lipid and glucose metabolism data.” (page 11, line 3)
i) Were the assumptions for using regression analyses justified? Are the
variables normally distributed? If not, what have you done with the data? Have
you examined whether there is a high correlation between the various
independent variables for which you adjust? This is important because you
want to prevent multi-colinearity.
Response: Since some of our data on risk factors were not normally distributed,
we took the logarithm for our multi-regression analysis. Therefore, the
statistical analysis (p-value) was the result of multi-regression analysis using
logarithmic transformation. This information was added to methods section.
(page 11, line 6)
With regard to multicollinearity, we calculated correlation coefficients among
explanatory variables. Age and blood pressure, age and blood glucose, age and
HbA1c were correlated (r=0.30). However, age should be taken into account,
and we did not remove it from the final model.
j) Why do you adjust for lifestyle related variables in the results as presented in
Table 4? What is the theory or hypothesis in this field on the role of smoking,
alcohol, exercise, eating habits, sleep quality on the mechanism underlying the
association between precarious work and health? Within research on work
5
stress and health, these variables are considered as potential intermediates.
Could this also apply to the association between fixed term work and health
status (=obesity, lipid and glucose metabolism). If so, this should be discussed
in the discussion section as well.
Response: Previous studies have also adjusted for lifestyle factors as
confounding variables. We considered that lifestyle factors were associated
with the health risks of the study (BMI, blood pressure, etc.), and contract types
might be associated with lifestyle factors, as shown in Table 2. The
observations of the authors including an occupational medical doctor’s view,
which helped to decide our analysis.
3. Are the data sound?
a) I question the validity of the health related lifestyle variables and the definition
of the main outcome of this study. Please provide some data on the validity of
the items used to assess the lifestyle variables.
Response: The questions asked are questions commonly used in occupational
health check-ups in Japan, and also in the National Nutrition and Health Survey
conducted by the Ministry of Health, Labour and Welfare.
b) The title of table 4 seems incorrect ‘regression model showing association of
precarious employment and change in BMI, BP, lipid and glucose metabolism’.
In the article nothing is mentioned about examining change in these variables.
You use a cross sectional design.
Response: The table title has been changed in the revised manuscript. The
title now reads: “Association of precarious employment and levels of BMI,
blood pressure and lipid and glucose metabolism” (Table 4).
c) I think it is strange that you use BMI as a continuous variable in table 4 and
mention obesity in the title of the paper. I think you should examine BMI as
continuous variable and mention BMI in the title or you should examine
Obesity (BMI > 25 versus < 25) and then mention obesity in the title.
Response: The title has been changed in the revised manuscript. The title of the
paper now reads: “Body mass index, blood pressure, and glucose and lipid
metabolism among permanent and fixed-term workers in the manufacturing
industry: A cross-sectional study”.
6
d) The presentation of the variables in table 2 should be clarified. It should be
clear, without having to go back to the methods section or the description given
in table 1, what alcohol means, for instance replace ‘alcohol’ by ‘alcohol
consumption, yes’ and ‘sleep satisfaction’ by ‘sleep satisfaction is good’ etc.
Response: The names of the variables in Table 1 and Table 2 have been
changed for clarity in the revised manuscript.
e) Table 4, in the footnote the meaning of *, ** is missing.
Response: The description of * and ** have been added to Table 4 in the
revised manuscript.
4. Does the manuscript adhere to the relevant standards for reporting and
data deposition?
See comments above that relate to reporting of the data
Response: Thank you for your comments. Please see above responses that
relate to the reporting of data.
5. Are the discussion and conclusions well balanced and adequately
supported by the data?
a) I miss the discussion of whether or not to adjust for lifestyle related variables
(are they confounders or intermediates) and whether adjustment for these
variables strongly attenuated the association and what this could mean.
Response: We think that lifestyle-related variables are confounders, as previous
studies have reported. Adjustment for these variables attenuated associations
because these lifestyle factors might influence any association with health risk
factors also. Comment has been added to the Discussion section on this point.
b) Please make clear the contribution is of this cross sectional study to this
research field, what do we know more due to this study?
Response: Although this is a cross-sectional study, and we cannot show causal
relationships between contract types and health risks, we do report differences
in health risk factors between fixed-term and permanent workers by contract
7
types. Differences were found (especially higher levels of blood pressure and
blood glucose in fixed-term workers) between fixed-term workers and
permanent workers. This is the first study, to the best of the authors’ knowledge,
to report such findings using objective measurements.
c) Section ‘Discussion- limitations’. I do not understand the reasoning of the
author. First possible differences in work content and psychosocial
characteristics between permanent and fixed term workers are mentioned to
explain why permanent workers show worse blood pressure, lipid and glucose,
then it is followed by the sentence ‘although these disadvantageous conditions
exist…. they remain healthy to perform their jobs’ -> I do not understand this
latter sentence in relation to the explanation of confounding?
Response: The sentence has been removed in the revised manuscript.
d) 6th paragraph of discussion ‘in terms of workers background, we did not
examine SES or family in detail’ - Are all workers in the factory of high school
graduate level? Or are there differences in educational level (as a measure for
socio-economic status) among the workers and especially between
permanent and fixed term workers?
Otherwise the explanation given in the 7th paragraph of the discussion for the
difference in BMI between permanent and fixed workers (which may be
attributed to e.g. SES) is less likely. Or if the author means that it can be due to
SES because fixed term workers have a lower salary (as indicator of SES), the
author should discuss this interesting issue in more detail, since both
(educational level and income) are often used as indicators of SES in the
literature.
Response: All workers in the factory were high school graduates. Therefore,
their educational attainment was not different by contract type; however,
salaries were different by contract type. We mean that the difference can be
due to a lower salary. This is discussed in more detail in the Discussion
section.
e) The issue of longitudinal assessments of outcome and life style variables and
contract type should be discussed in more detail, since the lack of information
on duration of contract type, duration of exposure to other work related factors
(psychosocial work characteristics etc) limit the understanding of the findings.
Could differences in exposure in contract type and lifestyle factors between
the two groups contribute to the differences between the two groups? And if so,
8
how?
Response: As mentioned, lack of information on duration of contract type,
duration of exposure to other work-related factors, such as psychosocial work
characteristics, limit the understanding of the findings in our analysis. Different
contract types might influences different stress types. For example, permanent
workers might experience job pressure due to longer work hours, while
fixed-term workers might experience pressure in the form of job security
worries. Exposure to these different pressures might exist in our study
participants. This information has been added to the Discussion section in the
revised manuscript.
Lifestyle factors were different between fixed-term workers and permanent
workers (Table 2). Permanent workers worked longer hours, did less exercise,
and drank more heavily than fixed-term workers, while fixed-term workers
smoked more than permanent workers. These different patterns of exposure
might cause different health risks after a long period of time.
f) In the 6th paragraph of the discussion it is stated ‘therefore, young permanent
workers in our population…only recently’. The mean age was around 31 for
permanent and 30 for fixed term workers? In the Netherlands, workers with
high school education are 18 years old when they start working. If the same
applies to Japan, the permanent workers cannot be considered as started
working recently, since they already have at least 10 years of work
experience.
Response: The mean age was around 31 years for this study population. The
range was 20–49 years. High school graduates start work when they are 18
years old in Japan. Young permanent workers in the manuscript mean workers
aged 20. These workers should have less work experience. This information
was added to revised manuscript. (page 17, fifth line from the bottom)
g) One of the explanations that is given for the finding that fixed term workers
showed unhealthy tendencies in terms of blood pressure, lipid and glucose
metabolism although BMI was lower as compared to permanent workers, is
smoking. The prevalence of smoking was higher among fixed workers,
however in Table 4, the results are adjusted for smoking and the associations
remain significant.
Response: Thank you very much for your comments. This has been corrected.
9
The finding that fixed-term workers showed higher unhealthy tendencies in
terms of blood pressure, lipid and glucose metabolism was because of their
unstable employment or socioeconomic status rather than just smoking.
h) In the introduction two studies were mentioned that used objective data on
blood pressure and BMI and examined the association with precarious work.
What did these studies show? Was this in line with your findings? What did
other studies on lifestyle factors (self reported) and precarious work find?
Response: The previous studies are not directly related to our study because
the two studies were on natural killer cells and mental health and circadian
cortisol levels.
With regard to lifestyle factors, higher alcohol consumption was associated
with workers who changed jobs into precarious employment; and female
workers who changed jobs into precarious employment showed reduced
frequency in physical activity. These results have been added to the
Background section. (page 5 line 7)
i) in the first paragraph of the discussion you refer to a study that showed an
association between low SES and diabetes. Why? Diabetes is not part of your
outcome, this was excluded, as well as hypertension and hyperlipidemia.
Response: Since our outcome includes blood glucose and HbA1c levels, which
are related to diabetes, we referred to the previous study to provide information
on associations between low SES and diabetes. We thought diabetes to be one
of the lifestyle-related diseases.
6. Are limitations of the work clearly stated?
Confounding by other factors which are not measured are stated in fifth
paragraph (differences in other working conditions). Selection of manufacturing
workers from one factory -> why do you think that the findings might not be
generalizable to all fixed term workers working in that industry? Were the
working circumstances in the factory different from the working circumstances in
other factories in that industry?
Response: We thought this factory to be representative of the standard level in
Japan. However, as the information in the original manuscript may have been
misleading, we have removed this sentence from the Discussion section in the
10
revised manuscript.
Overall, for various limitations mentioned, I miss the implication of the limitations
on the results of the study. E.g. how would data on marital status influence the
association between contact type and obesity and lipid glucose metabolism? E.g.
it is reasoned that higher mental pressure among permanent workers would
negatively impact mental health among permanent workers. What is the
expected influence on lifestyle factors and BMI etc?
Response: We have added information about the implications of the limitations
of the study from lack of the data on marital status and mental pressure to the
Discussion section in the revised manuscript. (page 16, fourth line from the
bottom)
I miss the discussion of using a cross sectional design in examining the
association between precarious work and health status.
Response: We only observed associations, and causal relationships could not
be verified. A comment on this point has been added to the Discussion section
in the revised manuscript. (page 18, line 3)
7. Do the authors clearly acknowledge any work upon which they are
building, both published and unpublished?
Background, third paragraph: there are two studies that have objectively
assessed BMI and blood pressure and examined the association with fixed term
contract type, what was the outcome of these studies? (MINOR revision)
Response: The outcomes of these studies were change of BMI and
cross-sectional data of BMI and blood pressure. We have added some
information on these studies to the Background section in the revised
manuscript. (page 5, fourth line from the bottom)
Background, fourth paragraph, Is the hypothesis based on results of previous
studies or theories/ models? Please include them.
Response: The hypothesis was based on the results of previous studies and our
observation of factory workers. This information has been added to the
Background section of the revised manuscript.
11
As mentioned under 5h) I miss the results of other studies for comparison of the
results of the current study with the results of other studies regarding BMI, diet,
exercise, blood pressure etc among fixed term workers as compared to
permanent workers.
Response: Please refer to our response to 5-h) above.
8. Do the title and abstract accurately convey what has been found?
Title : In the aim of the study, as formulated at the end of the background, health
status is defined as obesity and lipid and glucose metabolism, whereas the title
also contains blood pressure.
Response: We have included blood pressure data as part of the aim of the study
in the Background section of the revised manuscript.
Abstract: Background: the rationale of the study should be more clearly given,
why is objective clinical data important? Why is this study important, what does
this study add to the current knowledge on this topic?
Response: Without objective data (as was the case in previous studies), we
cannot capture the real situation with regard to health risks, especially
asymptomatic diseases. Subjective data might underestimate the real health
risks to participants because some fixed-term workers may be afraid of
providing true information for fear of losing their job. It is therefore important
to measure factors using objective clinical data. In this regard, our study is
important as it measures health risks using objective data.
Method: Design of the study. How did you perform the analyses and for which
factors did you adjust?
Response: We performed multi-regression analysis adjusted by age and
lifestyle characteristics. We have included information in the Abstract of the
revised manuscript. (page 2, second line from the bottom)
Conclusion: You did adjust your analyses for smoking, so why do you think it
might cause the association?
Response: This has been removed from our conclusion in the revised
manuscript. We now only focus on contract types and socioeconomic status
12
based on employment status.
9. Is the writing acceptable?
The writing should be improved as there are several strangely formulated
sentences and grammatical errors.
E.g. in abstract, background ‘A precarious form of employment is recognized as
a factor of social determinants of health’ I did not understand the meaning of this
sentence, only after reading the first sentence of the introduction.
In the abstract ‘Evidences examined by objective clinical measurements were
quite limited’? Do you mean ‘Evidence that supports precarious employments as
a risk factor for health, is mainly based on subjective data for health, whereas
evidence using objective clinical measurements for the assessment of health
status is limited ’
E.g. in the main paper – Methods section: ‘study subjects were manufacturing
industry workers in a company in the Chobu region of Japan, which employs
both permanent and fixed workers’. ‘whereas fixed term workers employment
term with the targeted company was 6 months’. ‘thus, fixed term workers may
experience working conditions similar to those of permanent workers’
Above Results section: ‘Finally a multiple regression analysis was used to
estimate and compare the BMI , BP and metabolism data’
Results section, 4th paragraph ‘table 4 shows the …to examine the relationship
of BMI, blood pressure and metabolism to contract type’.
Title table 3 ‘comparison of … between permanent and fixed term workers’
However, overall, the writing does not strongly hinder understanding of the
content of the paper.
Response: The grammar has been corrected in the revised manuscript and the
manuscript has been edited by Edanz (an English-editing company).
13