Social Science & Medicine 227 (2019) 56–62

Contents lists available at ScienceDirect

Social Science & Medicine

journal homepage: www.elsevier.com/locate/socscimed

Occupied with classification: Which occupational classification scheme T

better predicts health outcomes?
Emily Eyles∗, David Manley, Kelvyn Jones
School of Geographical Sciences, University of Bristol, University Road, BS8 1SS, Bristol, United Kingdom

ARTICLE INFO ABSTRACT

Keywords: Health inequalities continue to grow despite continuous policy intervention. Work, one domain of health in-
Occupational health equalities, is often included as a component of social class rather than as a determinant in its own right. Many
Classifications social class classifications are derived from occupation types, but there are other components within them that
Class mean they may not be useful as proxies for occupation. This paper develops the exposome, a life-course exposure
Work
model developed by Wild (2005), into the worksome, allowing for the explicit consideration of both physical and
Worksome
Exposome
psychosocial exposures and effects derived from work and working conditions. The interactions between and
Social exposure within temporal and geographical scales are strongly emphasised, and the interwoven nature of both psycho-
social and physical exposures is highlighted. Individuals within an occupational type can be both affected by and
effect upon occupation level characteristics and health measures. By using the worksome, occupation types are
separated from value-laden social classifications. This paper will empirically examine whether occupation better
predicts health measures from the European Working Conditions Survey (EWCS). Logistic regression models
using Bayesian MCMC estimation were run for each classification system, for each health measure. Health
measures included, for example, whether the respondent felt their work affected their health, their self-rated
health, pain in upper or lower limbs, and headaches. Using the Deviance Information Criterion (DIC), a measure
of predictive accuracy penalised for model complexity, the models were assessed against one another. The DIC
shows empirically which classification system is most suitable for use in modelling. The 2-digit International
Standard Classification of Occupations showed the best predictive accuracy for all measures. Therefore, ex-
amining the relationship between health and work should be done with classifications specific to occupation or
industry rather than socio-economic class classifications. This justifies the worksome, allowing for a conceptual
framework to link many forms of work-health research.

1. Introduction when they may be related to one another, similar to the interconnected
‘whole person’ view of physical and mental health (Carter et al., 2015).
Health research has looked at a variety of domains of inequalities. However, results from research on employment conditions and health
One of these, work, has been generally neglected, though the re- can be inconsistent. For example, nonstandard labour contracts were
lationship between it and health has increasingly been highlighted, not associated with adverse health effects by Scott-Marshall and Tompa
particularly in terms of its psychosocial conditions (see Siegrist et al., (2011), but Benach and Muntaner (2007) report that those in insecure
2010; Benach et al., 2012; Lewchuk et al., 2003; Braveman et al., 2005). jobs have higher self-reported morbidity. Differences in results can be
The workplace psychosocial environment is generally thought to be a attributed to the diversity of the outcomes researched in terms of the
consequence of employment relations rather than solely of external forms of employment examined, the composition of the sample, which
social determinants of health (Benach et al., 2014). It is not only the health measures are included, and the location or context of the work
work itself or the technologies around it, but the structure and order of (see Kauskamp et al., 2013; Virtanen et al., 2005). Indeed, Hoven and
the workplace that may create both physical and mental health effects Siegrist (2013) review mediation and moderation studies on adverse
(Canaan, 1999). Psychosocial hazards in the workplace have often been working conditions and health outcomes, noting that studies feature “a
considered separate to physical ones (Karasek and Theorell, 1990), high degree of heterogeneity of core measurements.”

DOI of original article: https://doi.org/10.1016/j.socscimed.2019.01.025

∗
Corresponding author.
E-mail address: [email protected] (E. Eyles).

https://doi.org/10.1016/j.socscimed.2018.09.020
Received 13 January 2018; Received in revised form 10 August 2018; Accepted 14 September 2018
Available online 15 September 2018
0277-9536/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/BY/4.0/).
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

This can be exacerbated as the difference between some of these (1999) argues that changing patterns of employment, generally thought
measurements and contexts used in studies may be unclear (e.g. part- to be a consequence of the late 20th century neoliberal shift, are re-
time work hour thresholds), and new forms of work, such as flexible flected in new patterns of the production and distribution of risk and
employment, can be difficult to classify, particularly with respect to hazard, namely the transfer of the costs and risks of employment from
terminologies which may be unclear; temporary, non-permanent, pre- the employer to the employee (see also Standing, 2011; Kim et al.,
carious, non-standard, insecure, contingent, fixed-term, atypical, ca- 2012). Flexibility is not just found in the technical systems of work, but
sual, and unregulated represent similar concepts in various studies also more abstract elements thereof, such as tasks, status, and sche-
(Peck, 1996; Kim et al., 2012; Kauskamp et al., 2013; Benach and duling (Ross, 2009; Benach and Muntaner, 2007). It also has filtered
Muntaner, 2007; EMCONET, 2007; Scott-Marshall and Tompa, 2011). A through to work where these conditions may have once been thought
variety of perspectives have sought to address inconsistency in this unthinkable. In general, work has also been intensified with a pressure
field: two of the most commonly used schemata of the work-health to do the same or more work in less time, or to expand tasks and ex-
relationship are the job strain or job-demand-control model (Karasek pectations beyond what they were before (McNamara et al., 2011).
and Theorell, 1990) and the effort-reward imbalance model (Siegrist, These conditions are not equally distributed amongst occupations, and
1996). Benach and Muntaner (2007) suggest that these frameworks, perhaps even individuals, so it follows, then, that inequalities in health
though, may not be able to incorporate “more distal social and orga- should be also examined occupationally.
nizational determinants of health.” This paper will bring new per- Clougherty et al. (2010) assert that “occupational classifications
spectives in to bear on issues with these conceptual models, in order to used in many epidemiological studies (i.e., manual/nonmanual dis-
take into account structural and social inequalities, geographical con- tinctions, professional grade, and census job classifications) have
text, and time. Looking towards the life-course approach and through proven too coarse to capture fine-scale status differences […]”. Some-
the lens of exposure, a framework linking concepts in epidemiology, times, for example, there is a lack of clarity: Hallerod and Gustafsson
occupational health, and inequalities research has been developed – the (2011) argue that occupations can be used to create ‘economical
worksome. classes,’ but occupation only forms part of these classes. Moreover,
Hallerod and Gustafsson (2011) use ‘occupational position’, ‘employ-
2. Background ment position’, ‘economical classes,’ and ‘social classes’ almost inter-
changeably, possibly based on that argument, which can lead to some
Health research can sometimes confound occupation and class, confusion when it comes to interpreting results. The UK NS-SEC (Na-
often by using either as a proxy for the other (Shaw et al., 2000). That is tional Statistics Socio-Economic Classification), for example is gener-
to say, occupational classifications are sometimes used as class and vice ated with the UK version of the ISCO 2008, the SOC2010, but it con-
versa (see Clougherty et al., 2010), but class is more complex than just tains other inputs relating to status (ONS, 2010). For example, Corna
occupation, and this is reflected in how class classifications are created and Sacker (2013) convert from the SOC1990 to the NS-SEC to assign
(see ONS, 2010). Occupation is a component of class, and occupation is ‘occupational class,’ and refer to it as such, when the SOC codes are
simply the job or work someone does; class is, in essence, a hierarchical themselves an occupational classification. The European social class
measure of socioeconomic positioning, and on a higher scale, a measure measure for the European Social Survey is composed in a very similar
of social structure. Savage (2015, p35) asserts that occupationally- fashion from multiple items, including occupation (Erikson and
based measures of class are “actually a way of making for cultural Jonsson, 2001; Almeida et al., 2006). Almeida et al., 2006 claim that
judgements about the ranking and social importance of jobs.” “class structures significantly mark the value patterns found in the
MacDonald et al. (2009), in a review of epidemiological studies, found populations analysed.” We assert, then, that using class can also limit
that while many collected occupational measures, most work used these the transferability of results due to variation in contexts. Occupations,
data to inappropriately represent socioeconomic class. Class is almost while socially mediated, are not, like class, socially defined, and are
always unsuitable to examine the work-health relationship as it has more readily conceptually transferable between contexts.
historically been articulated in a variety of ways. Class contains an Different occupations are associated with varying conditions, risks,
implied hierarchy, which already imposes a relationship that may be prospects, and outcomes, these are not given across or even within
unsuitable and inappropriate. Further, with class, it is difficult to un- occupations. There is thus considerable heterogeneity within and be-
derstand axial differences (e.g. skilled versus unskilled white and blue- tween occupations. Therefore, the relationship between working con-
collar workers). Using occupation as a proxy for class or vice versa can ditions and health should be analysed with respect to this hetero-
mask the nuances between or within occupations with respect to geneity, looking both between and within occupations. As such, the
working conditions and exposures. Occupation can indeed be articu- hierarchy implicit in class classifications may confound the examina-
lated as part of a class definition, but it is not simply a component of it; tion of these already complex relationships.
it can be a social determinant of health in its own right. To that end, this paper will examine several classification systems
While work such as the Whitehall studies (see Clougherty et al., empirically, namely ISCO as an occupational one, NACE as an eco-
2010) have created the basis for examining the relationship between nomic/occupational system, and NSSEC as a socioeconomic class one,
work and health, it is important to remove implied hierarchies or as it is commonly used in the literature, to provide a base for using
grades of occupation to discover further information about these re- occupational classifications over socio-economic class ones by de-
lationships. It is important to remove implied hierarchies or grades termining which classification has the best predictive accuracy with
from occupation to discover further information about these relation- respect to a range of health measures from a specifically occupational
ships, in addition to the evidence on social class gradients of health. dataset. Further, it will argue that finer-scale versions of these classi-
There has been a vast array of research into employment status, or fication systems perform better in general even when parsimony is
grade and health (e.g., Marmot et al., 1991; Benach et al., 2014; considered. This will also forward the worksome framework by em-
MacDonald et al., 2009; Mackenbach et al., 2008), but with a changing pirically demonstrating that occupational classifications are the most
world of work and employment relations, precarious or ‘flexible’ con- appropriate for work-health research, when often class is used to proxy
ditions are filtering through to jobs where it would have been incon- occupation or vice versa, which is not always the right approach. There
ceivable before. Socioeconomic characteristics like class may interact is therefore a need to bridge what people are exposed to and what
with flexible working conditions vis a vis health outcomes, but re- people say or believe they are exposed to, including social exposures.
flecting on the percolation of these conditions to other jobs, refined This paper will introduce the worksome in order to provide a frame-
occupational categories change less over time, and may be more ap- work for justifying the use of occupational classifications over class, and
propriate (see Benach et al., 2014; Hoven and Siegrist, 2013). Daykin the importance of occupation as a social determinant of health. The

57
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

worksome will be underpinned empirically by an examination of oc- exposure and the interactions both within and between scales. Scale,
cupational, social, and economic classification systems. used here in the sense of ‘level’, can include individuals, work groups,
firms, industries, and so on, with other geographic and contextual
(geocontextual) factors existing at the same or different levels, such as
3. Theoretical framework: the worksome the workplace, the city, or the regulatory regime at varying levels of
government. This does not mean that scales are rigid. Delaney and
The worksome is an expansion of the exposome. The exposome was Leitner (1997), argue that scale is often constructed, and so the work-
developed by Wild (2005) in response to the sequencing of the human some takes scale as a fluid, interactive concept of levels, while keeping in
genome, and to incorporate the life-course approach (see Ben-Shlomo mind that scale is often socially and politically mediated. The physical-
and Kuh, 2002) to exposure into epidemiology. The exposome includes social aspects of exposure are represented by the social gradient linking
three separate-but-overlapping domains, the internal, specific external, the physical to the geocontextual and the workplace, in order to en-
and general external (Wild, 2005, 2012) whilst also capturing both compass largely physical exposures (such as chemical handling, see for
nature and nurture (Miller and Jones, 2014). This sort of life-course instance Arif and Delclos, 2012), predominantly social exposures (in-
approach is appropriate for work (which we can define as a ‘general cluding social support, see Niedhammer et al., 2012), and exposures
external’ element) as it accounts for a large proportion of time in a life- which are inherently both physical and social and fall between the
course (Bambra, 2011; Payne, 1999; Peck, 1996), and it can impact extremes, such as working time (see Dembe et al., 2005; Kivimaki et al.,
how lives are lived outside the workplace (Kleiner and Pavalko, 2013). 2015) (see Fig. 1). Working time is both; as a basic concept, it is phy-
Working consumes a large part of any life course, regardless of whether sical: the time spent exerting oneself at work, but it too is social, in the
it is formal or informal. The general external elements, like work, of the sense that it is also the time spent being exposed to a variety of (phy-
framework are, in the general version of the exposome assumed rather sical and social) working conditions. Social exposures have a certain
than measured, as work with the exposome is predominantly top-down, intangibility to them, something which is emphasised in the social-
focusing on physically measurable exposures (Rappaport, 2011). The physical gradient of the worksome, though it is an exposure type not
exposome has been adapted for health inequality research, notably by emphasised by the exposome. A social-physical gradient of exposure
Juarez et al. (2014) who created ‘the public health exposome,’ which allows for flexibility in analysis as it provides a framework within the
focuses primarily on environmental health. Research creating various worksome for disparate and similar-but-different measures of exposure
types of exposome, for instance see the exposomics project (see Vineis to be compared. Moreover, individual-level exposures and workplace
et al., 2017), the public health exposome (Juarez et al., 2014), and the level exposures interact: individuals within a workplace are affected
occupational exposome (Faisandier et al., 2011), focuses on the use or and effect upon workplace-level characteristics. Individuals, therefore,
adaption of the exposome more with respect to biological analyses and cannot be considered solely as discrete entities with respect to the
issues which may arise thereof, without realising that other approaches work-health relationship. There are also factors above the individual
using survey data may also be suitable under the paradigm (Brunekreef, and the workplace.
2013). Workplaces are also located within geographical contexts, be it in
The worksome expands on the idea of exposure to include a social- relation to other firms, related industries, as well as in social and reg-
physical gradient (see Fig. 1). It is necessary to consider work explicitly ulatory contexts. Geocontextual influences are an undercurrent and
to draw out lower-level scale (micro/meso) exposures, vectors, and require consideration in work-health research. Time is also considered
effects. The worksome emphasises the importance of the scale of in the worksome – exposures continue across the life course. Again,
interactions within and between all of these domains must be empha-
sised – people exist at multiple scales simultaneously: ‘echoes’ of past
actions or consequences are reflected in these interactions as well. A
given individual's contribution can prevail and the residual impacts
remain with people for a long time after the initial exposure, as well as
influencing their and others' behaviours. By including the interactions
between scales, individuals, times, and geographies in the worksome,
we further our understanding of the complexities of this landscape. As
work too consumes a large part of any given individual's life, the life
course approach is key to understanding work as a social determinant of
health. With respect especially to time, the life trajectory approach
allows the worksome to also cover those who are unemployed or en-
gaged in informal work. The former is incorporated as they move in and
out of the workforce. The latter is encompassed as the worksome does
not distinguish between formal and informal work, in the sense that
they are both considered equally under the framework. Indeed, there
are a number of papers examining life trajectories and career typologies
with respect to occupational mobility, for example, and these ap-
proaches, often using sequence analysis or latent class analysis, can and
should be emulated in work that examines the relationships between
working conditions and health (Haapakorva et al., 2017; Anders and
Dornett, 2017; Scott and Zeidenberg, 2016; Corna and Sacker, 2013).
Movement between occupation types, such as from manufacturing to
the low-paid service sector, has been connected with poorer health
using these approaches (Kampanellou and Houston, 2016). Employing
latent class models, Corna and Sacker (2013) modelled the lifetimes of
older British adults, particularly around the labour market and family
experiences, finding significant differences in the mental health do-
main. The worksome is useful over the exposome as it adds specificity
Fig. 1. The worksome. and interaction between the domains, has a social-physical exposure

58
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

gradient, and emphasises scale more strongly. Table 1

Both qualitative and quantitative forms of research are key to Health measures to be modelled and classification systems.
forming a better picture of the work-health relationship. Within the Health Measures
quantitative approaches multilevel models can be used to approximate
the proposed structures (see Hox, 2010). For qualitative research, the Yes
Work-effect on health 39%
effects people have on systems and scales and how they are affected by
Self-rated health Good: 77%
them could be elucidated through interviews, or participatory work In the last 12 months …
where the participants guide the research journey. Skin problems 7%
Using the language of biomedical epidemiology is key to this ap- Hearing problems 6%
proach; the goal is to not only forward a more clear and comparable set Backache 45%
Muscular pains in lower limbs 33%
of social research projects but also to develop clearer research findings
Muscular pains in shoulders, neck and/or upper limbs 44%
for policymakers and other scientists. The worksome makes explicit the Anxiety 13%
elements that the exposome treats as givens, allowing for the use of Fatigue 41%
language familiar to policymakers while including effects that may not Headache and eye strain 39%
Injury(ies) 7%
be considered explicitly in the biomedical approach. This framework
Classification Systems
can help fit disparate pieces of research together and contextualise NS-SEC Analytic Groups, 8 categories
them to form a wider collective of research. Flexibility is important, as NACE Revision 2 Letter groups (1 character), 8 categories
for research involving people, a complete body of research is impossible ISCO 2008 Major groups (1 digit), 10 categories
as society is constantly changing, so gaps in research are to be expected, ISCO 2008 Sub-major groups (2 digit), 42 categories
NACE Revision 2 2 digit groups, 88 categories
and can be filled.
For the empirical portion of this paper, the objective is to distin-
guish work, or occupation, from class, and to set out which system of and ‘No.’. Manor et al., (2000) found in their analysis of self-rated
classification is most appropriate for use in quantitative analysis. This health that both dichotomised and ordered categorical models showed
will advance the argument that occupation and class should be ex- similar results with only small differences in power and efficiency. The
amined separately, as well as supporting the usefulness of the work- health measure variables are also self-reported, which may not be ideal.
some in underpinning work-health research. This will use the European Miilunpalo et al. (1997) assessed subjective measures of health, and
Working Conditions Survey to see which classification system has the found that, in relation to objective health measures, they are valid for
best predictive accuracy for a variety of health measures including use in population health research. They also argued that perceived
backache, self-rated health, and fatigue. health measures were stable due to a small rate of major changes in that
status (Miilunpalo et al., 1997). Burstrom and Fredlund (2001) found a
4. Methods strong relationship between poor self-rated health and mortality, im-
plying that self-rated health is a suitable predictor of mortality, and
4.1. Data therefore ‘a useful outcome measure.’ DeSalvo et al. (2005) found that,
compared to multi-item measures of self-reported health and co-
The European Working Conditions Survey (EWCS) is a repeated morbidity, a single-item measure is as good at prediction. It is therefore
cross-sectional quinquennial survey started in 1991. It is administered acceptable, then, to use the health measures collected in the EWCS, to
by the European Foundation for the Improvement of Living and examine them in relation to working conditions, or, in the case of this
Working Conditions (Eurofound) for the European Union (EU). Waves paper, to see which classification system better predicts them.
were conducted in 1991, 1995/6, 2000/1, 2005, 2010, and 2015. This Given the argued importance of time for understanding the work-
paper uses data from the 2010 and 2015 waves, due to the presence of some, specific health problems were defined to look at issues that had
occupational class variables. All EU countries and European Economic occurred within the last year while the general measures were defined
Area (EEA) countries were included, with a number of EU candidate contemporaneously at the time the questionnaire was answered.
members in some waves, therefore not all countries are in all waves
(Eurofound, 2015). The target sample in each country was between 500
and 1500 individuals. The EWCS data were obtained from the UK Data 4.2. Models
Service (Eurofound, 2017). The EWCS data has individuals classified
both by the Statistical Classification of Economic Activities in the 60 Logistic regression models were run using MLWiN 3.01. Logistic
European Community (NACE) and the International Standard Classifi- regression is used here as the measures are binary (yes/no, or good/
cation of Occupations (ISCO). The NACE is an industry classification, bad). Separate models for each health measure as the independent
and the ISCO an occupational one. The National Statistics Socio-Eco- variable (n = 12) with each classification system (n = 5) as dependent
nomic Classification (NS-SEC) is a British system of socio-economic variables were run using a Markov Chain Monte Carlo (MCMC)
classification based on the UK occupational classification system Bayesian framework (see Browne, 2015). This provides a Deviance In-
(SOC2010), employment status, and firm size (ONS, 2010). The formation Criterion (DIC), a Bayesian version of the Akaike information
SOC2010 was derived from the ISCO 2008 2-digit version, and the criterion (AIC). The DIC is a measure of predictive accuracy that is the
employment status and firm size variables were derived from survey badness of fit between the observed and modelled measures penalised
questions in the EWCS. for model complexity (Spiegelhalter et al., 2002). The number of ca-
A new dataset was created with the occupational classifications and tegories in any given system should not be a factor in determining
relevant health measure variables (see Table 1) for the years 2010 and which model has the best predictive accuracy as the DIC operates by
2015. Only 2010, 2015 were included as the expanded 2-digit ISCO and estimating model complexity (the so-called degrees of freedom con-
NACE classifications were only in those waves (n = 81115). In the sumed in the fit) and automatically penalizes models that do not show
analysis the data for 2010 and 2015 are combined as this study is not an improvement in the badness of fit over and above model complexity;
concerned with change and it means that there is a larger sample size to that is, it the DIC privileges parsimony. As such it is an ideal procedure
detect meaningful effects; the classifications and outcome variables for comparing models with different specifications involving different
were consistent over this relatively short period. The health measure classifications. The DIC can be compared within the same health mea-
variables were dichotomised in order to fit the logistic regression sures, but not between health measures, i.e., the DIC for the NS-SEC for
model. The original responses were ‘Yes, positively’, ‘Yes, negatively’, skin problems cannot be compared to the DIC for backache for the ISCO

59
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

Fig. 2. DIC, a measure of model fit, by health measure. A smaller DIC indicates better predictive accuracy. For a fuller description of the outcomes, see Table 1.

1-digit system. In terms of the specifics of MCMC estimation we fol- different due to the varying measures, as discussed earlier, but the
lowed the good-practice recommendation of Draper (2008). Thus, we comparison of classification systems should be considered within a
use likelihood approach to estimate an initial model, specify default measure rather than between a measure. It is not the specific value of
priors to impose as little information as possible on the estimates, a DIC which is important, but which has the lowest DIC within a measure.
burn-in of 500 simulations to get away from these initial (potentially The ISCO 2-digit schema best predicts whether an individual's work
poor) estimates, and a monitoring chain of some further 5000 simula- may affect their health. Indeed, the ISCO 2008 2-digit classification has
tions to characterise the parameter estimates and calculate the DIC. the highest predictive accuracy for all health measures across the data,
not only for those questions which referred to the work-health re-
lationship specifically. The 2-digit NACE classification outperformed
5. Results
the 1-digit ISCO 2008 for some outcomes, though for self-rated health,
backache, lower muscular pain, upper muscular pain, and injury it was
Fig. 2 presents the results of all 60 models, by question or health
surpassed by the 1-digit ISCO. The 1-digit ISCO, therefore, did not al-
measure and the classification scheme. Only the DIC for each model
ways perform as consistently as the 2-digit version of the classification.
measure/classification pair is reported. Each outcome has the in-
The NS-SEC in this study borrows some predictive power from the ISCO
dividual classification models sorted by DIC, so that the classification
2-digit classification in this dataset as it is partially derived from it, and
system with the best parsimony (lowest DIC) is on the left. The colour
this may be why NS-SEC showed higher predictive accuracy than both
on the graph is consistent for each system. The y-axes of the graphs are

60
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

the 1- and 2-digit NACE classifications for backache and lower muscular study.
pain, as well as over the 1-digit version of the NACE for upper muscular For clarity in research, especially when interested in either class or
pain and injury. The NS-SEC also had somewhat higher predictive ac- occupation, it is necessary to separate out class and occupation as de-
curacy over the 1-digit ISCO and 1-digit NACE in terms of fatigue. It terminants of health. This rationale supports the use of the worksome, a
seems then, that the NS-SEC may be slightly better at predicting mea- conceptual framework developed in this paper, for the examination of
sures relating to general or muscular health than the NACE. the work-health relationship. The 2-digit ISCO 2008 occupational
Nonetheless, the ISCO 2-digit classification remains the most empiri- classification is the most appropriate when examining the relationship
cally appropriate for predicting health measures in the EWCS dataset, between work and health, compared with the NACE and NS-SEC.
as it had the lowest DIC for all health measures. Theoretically, this Therefore, there is also empirical justification for the use of the work-
indicates that work should be considered separately from class when some as a framework, and examining occupation as a separate domain
examining health measures, and that the worksome is an appropriate of health inequalities and as a separate determinant of health.
model for enquiry into this relationship. With both empirical and theoretical justification, the worksome
therefore can provide a transferable framework for research into work
6. Discussion and conclusions and health. Through its flexibility, it can accommodate research from a
variety of scales and contexts, allowing for the conceptual linking of
There is a clear need to focus both theoretically and empirically on disparate yet related studies. It is an expansion of a familiar concept,
work and occupation in and of itself rather than as a component of class the exposome (Wild, 2005, 2012), and encompasses a life-course ap-
or a feature that can be proxied by class. While many socio-economic proach, as work is something which generally consumes a large part of
classification systems, like the NS-SEC do use occupation as their base, any given individual's time. The exposome was explicitly chosen as a
they are not a ready substitute for occupational classifications them- base, as its biomedical language and approach is well understood by
selves. Furthermore, class contains an implied hierarchy, something policymakers. The worksome reorients the way in which the relation-
which may confound results, as it is a hierarchical system of social or ship between occupation and health is understood – as an interactive,
cultural value partially based in occupation. Social classification sys- multi-scalar framework of exposures set along a social-physical gradient
tems are informed by their social contexts, as the cultural value of oc- (see Fig. 1). By integrating scales, times, individuals, and geographies
cupations change through time. For example, around a quarter of oc- and their interactions, the complexities of these relationships become
cupations in the UK Registrar General's Class Classification changed clearer. Separating occupation from class, and justifying it empirically,
between classes from 1951 to 1961 (Liberatos et al., 1988). A system is necessary to forward the worksome, as occupation is at its core.
with an implied hierarchy may not be appropriate for occupational Again, as class is defined through sociocultural values, this makes it less
research, particularly with a changing world of work where flexible or suitable for examining the relationship between work and health,
precarious conditions have filtered even to ‘standard’ occupations. especially compared with more refined, less time-variant occupational
Further, some occupational classification schemes (such as manual/ classifications. This is not to say that there should be no research on
non-manual) are too simple or coarse to examine fine-scale detail in class and health, but merely to allow for a more thorough and em-
terms of health measures (Clougherty et al., 2010), and it has been pirically appropriate interrogation of the complex relationship between
shown that the 2-digit level of the ISCO 2008 performs better. This work and health.
means, then, that occupation, and therefore, the worksome, is con-
ceptually valid as a separate and distinct social determinant of health. Acknowledgements
Theoretically, the expansion of the exposome into the worksome pro-
vides a framework for both qualitative and quantitative work. The authors would like to acknowledge the assistance of the
Empirically, the analysis in this paper has shown that for examining Economic and Social Research Council Advanced Quantitative Methods
the health of workers (through the European Working Conditions studentship (S126111-120).
Survey), occupational classifications such as the ISCO are generally the
most appropriate. The more detailed 2-digit level provides better pre- References
dictive accuracy, whereas the 1-digit levels may be more practical for
certain analyses and data collection practises. However, some issues Almeida, J.F. de, Machado, F.L., da Costa, A.F., 2006. “Social classes and values in
remain with the 1-digit ISCO when it comes to predictive accuracy for Europe” Portuguese. J. Soc. Sci. 5 (2), 95–117.
Anders, J., Dorsett, R., 2017. What young English people do once they reach school-
certain health measures, where it is outperformed by the NACE 2-digit leaving age: a cross-cohort comparison for the last 30 years. Longitudinal Life Course
classification. In some cases, the NS-SEC did not have the least pre- Stud. 8 (1).
dictive accuracy compared to the other systems, primarily the NACE. Arif, A.A., Delclos, G.L., 2012. Association between cleaning-related chemicals and work-
related asthma and asthma symptoms among healthcare professionals. Occup.
One reason for this could be that the SOC2010, used to derive the NS- Environ. Med. 69 (1), 35–40.
SEC, in the case of this data, was derived itself from the ISCO 2008 2- Bambra, C., 2011. Work, Worklessness and the Political Economy of Health. Oxford
digit version, and therefore could have borrowed some statistical power University Press, Oxford.
Ben-Shlomo, Y., Kuh, D., 2002. A life course approach to chronic disease epidemiology:
from the ISCO 2008 2-digit. Another could be that the NACE is a conceptual models, empirical challenges and interdisciplinary perspectives. Int. J.
classification of industries or economic activities rather than occupa- Epidemiol. 31, 285–293.
tions and may not be completely suited to this sort of analysis. The Benach, J., Muntaner, C., 2007. Precarious employment and health: developing a research
agenda. J. Epidemiol. Community Health 61 (4), 276–277.
NACE, though, is formed so as not to distinguish by the ownership,
Benach, J., Puig-Barrachina, V., Vives, A., Tarafa, G., Muntaner, C., 2012. The challenge
legality, modes of operation, or formality of economic activities of monitoring employment-related health inequalities. J. Epidemiol. Community
(Eurostat ND). This may be nonetheless helpful, as the EMCONET Health 66 (12), 1085–1087.
(2007) research agenda includes non-standard forms of work beyond Benach, J., Vives, A., Amable, M., Vanroelen, C., Tarafa, G., Muntaner, C., 2014.
Precarious employment: understanding an emerging social determinant of health.
precarious or flexible work, including informal work and slavery. The Annu. Rev. Publ. Health 35, 229–253.
worksome too allows for non-standard forms of work. The ISCO, for Braveman, P.A., Cubbin, C., Egerter, S., Chideya, S., Marchi, K., Metzler, M., Posner, S.,
example, does not necessarily have provisions for these, so in those 2005. Socioeconomic status in health research: one size does not fit all. J. Am. Med.
Assoc. 294 (22), 2879–2888.
cases, the NACE may be more appropriate depending on the nature of Browne, W.J., 2015. MCMC estimation in MLwiN Version 2.32. http://www.bris.ac.uk/
the work. The ISCO 2008 2-digit version nonetheless does allow for the cmm/media/software/mlwin/downloads/manuals/2-32/mcmc-web.pdf, Accessed
vast majority of occupations to be classified as it does not discriminate date: 10 December 2017.
Brunekreef, B., 2013. Exposure science, the exposome, and public health. Environ. Mol.
by conditions, so therefore flexible and modern working conditions can Mutagen. 54 (7), 596–598.
be accounted for as long as they are acknowledged explicitly in the

61
E. Eyles, et al. Social Science & Medicine 227 (2019) 56–62

Burstrom, B., Fredlund, P., 2001. Self rated health: is it as good a predictor of subsequent Nordin, M., Oksanen, T., Pejtersen, J.H., Pentti, J., Rugulies, R., Salo, P., Siegrist, J.,
mortality among adults in lower as well as in higher social classes? J. Epidemiol. Steptoe, A., Suominen, S., Theorell, T., Vahtera, J., Westerholm, P.J.M., Westerlund,
Community Health 55, 836–840. H., Singh-Manoux, A., Jokela, M., 2015. “Long working hours, socioeconomic status,
Canaan, J.E., 1999. In: Daykin, N., Doyal, L. (Eds.), The Hand or in the Head? and the risk of incident type 2 diabetes: a meta-analysis of published and unpublished
Contextualising the Debate about Repetitive Strain Injury (RSI). Health and Work: data from 222 120 individuals. Lancet Diabetes Endocrinol. 3 (1), 27–34.
Critical Perspectives. Macmillan, Basingstoke, pp. 143–160. Kleiner, S., Pavalko, E.K., 2013. Double time: is health affected by a spouse's time at
Carter, J., Zawalski, S., Sminkey, P.V., Christopherson, B., 2015. Assessing the whole work? Soc. Forces 92 (3), 983–1007.
person: case managers take a holistic approach to physical and mental health. Prof. Lewchuk, W., De Wolff, A., King, A., Polyani, M., 2003. From job strain to employment
Case Manag. 20 (3), 140–146. strain: health effects of precarious employment. Just Labour 3, 23–35.
Clougherty, J.E., Souza, K., Cullen, M.R., 2010. Work and its role in shaping the social Liberatos, P., Link, B.G., Kelsey, J.L., 1988. The measurement of social class in epide-
gradient in health. Ann. N. Y. Acad. Sci. 1186, 102–124. miology. Epidemiol. Rev. 10, 87–121.
Corna, L., Sacker, A., 2013. “A lifetime of experience: modeling the labour market and MacDonald, L.A., Cohen, A., Baron, S., Burchfiel, C.M., 2009. Occupation as socio-
family histories of older adults in Britain” Longitudinal and Life Course. Studies 4 (1). economic status or environmental exposure? A survey of practice among population-
European Foundation for the Improvement of Living and Working Conditions based cardiovascular studies in the United State. Am. J. Epidemiol. 12, 1411–1421.
(Eurofound), 2017. European Working Conditions Survey Integrated Data File, 1991- Mackenbach, J.P., Stirbu, I., Roskam, A.-J.R., Schaap, M.M., Menvielle, G., Leinsalu, M.,
2015. [data Collection], fourth ed. UK Data Service SN: 7363. http://doi.org/10. Kunst, A.E., 2008. Socioeconomic inequalities in health in 22 European countries. N.
5255/UKDA-SN-7363-4. Engl. J. Med. 358 (23), 2468–2481.
Daykin, N., 1999. Introduction: critical perspectives on health and work. In: Daykin, N., Manor, O., Matthews, S., Power, C., 2000. Dichotomous or categorical response?
Doyal Basingstoke, L. (Eds.), Health and Work: Critical Perspectives. Macmillan Press, Analysing self-rated health and lifetime social class. Int. J. Epidemiol. 29 (1),
pp. 1–20. 149–157.
Delaney, D., Leitner, H., 1997. The political construction of scale. Polit. Geogr. 16 (2), Marmot, M.G., Stansfeld, S., Patel, C., North, F., Head, J., White, I., Brunner, E., Feeney,
93–97. A., Marmot, M.G., Davey Smith, G., 1991. Health inequalities among British civil
Dembe, A.E., Erickson, J.B., Delbos, R.G., Banks, S.M., 2005. The impact of overtime and servants: the Whitehall II study. Lancet 337 (8754), 1387–1393.
long work hours on occupational injuries and illnesses: new evidence from the United McNamara, M., Bohle, P., Quinlan, M., 2011. Precarious employment, working hours,
States. Occup. Environ. Med. 62 (9), 588–597. work-life conflict and health in hotel work. Appl. Ergon. 42 (2), 225–232.
DeSalvo, K.B., Fan, V.S., McDonell, M.B., Fihn, S.D., 2005. Predicting mortality and Miilunpalo, S., Vuori, I., Oja, P., Pasanen, M., Urponen, H., 1997. Self-rated health status
healthcare utilization with a single question. Health Serv. Res. 40 (4), 1234–1246. as a health measure: the predictive value of self-reported health status on the use of
Draper, D., 2008. Bayesian multilevel analysis and MCMC. In: de Leeuw, J., Meijer, E. physician services and on mortality in the working-age population. J. Clin.
(Eds.), Handbook of Multilevel Analysis, pp. 77–139 New York, NY, US. Epidemiol. 50 (5), 517–528.
EMCONET, 2007. Employment Conditions and Health Inequalities: Final Report to the Miller, G.W., Jones, D.P., 2014. The nature of nurture: refining the definition of the ex-
WHO Commission on Social Determinants of Health (CSDH). J. Benach, C. Muntaner posome. Toxicol. Sci. 137 (1), 1–2.
and V. Santana. Barcelona. Niedhammer, I., Chastang, J.F., Sultan-Taieb, H., Vermeylen, G., Parent-Thirion, A.,
Erikson, R., Jonsson, J.O., 2001. “European Social Survey Core Questionnaire 2012. Psychosocial work factors and sickness absence in 31 countries in Europe. Eur.
Development – Chapter 2: How to Ascertain the Socio-structural Position of the J. Publ. Health 23 (4), 622–629.
Individual in Society. European Social Survey, City University London, London. Official for National Statistic (ONS), 2010. SOC2010 volume 3: the National Statistics
Eurofound, 2015. 6th EWCS- methodology. http://www.eurofound.europa.eu/surveys/ Socio-economic classification (NS-SEC rebased on SOC2010). http://webarchive.
european-working-conditions-surveys/sixth-european-working-conditions-survey- nationalarchives.gov.uk/20160106042025/http://www.ons.gov.uk/ons/guide-
2015/6th-ewcs-methodology, Accessed date: 10 December 2017. method/classifications/current-standard-classifications/soc2010/soc2010-volume-3-
Eurostat (ND). “RAMON: Reference and Management of Nomenclatures; Your Companion ns-sec–rebased-on-soc2010–user-manual/index.html, Accessed date: 11 December
Guide to International Statistical Classifications: Section IV – Description of the Main 2017.
Economic Classifications” From: http://ec.europa.eu/eurostat/ramon/ Payne, S., 1999. Paid and unpaid work in mental health: towards a new perspective. In:
miscellaneous/index.cfm?TargetUrl=DSP_GENINFO_CLASS_4 [Accessed 11/12/ Daykin, N., Doyal Basingstoke, L. (Eds.), Health and Work: Critical Perspectives.
2017]. Macmillan Press, pp. 35–53.
Faisandier, L., Bonneterre, V., De Gaudemaris, R., Bicout, D.J., 2011. Occupational ex- Peck, J., 1996. Work-place: the Social Regulation of Labor Markets. The Guilford Press,
posome: a network-based approach for characterizing Occupational Health Problems. New York.
J. Biomed. Inf. 44 (4), 545–552. Rappaport, S.M., 2011. Implications of the exposome for exposure science. J. Expo. Sci.
Haapakorva, P., Ristikari, T., Gissler, M., 2017. “The impact of parental employment Environ. Epidemiol. 21 (1), 5–9.
trajectories on the children's early adulthood education and employment trajectories Ross, A., 2009. Nice Work if You Can Get it: Life and Labor in Precarious Times. New York
in the Finnish birth cohort 1987. Longitudinal Life Course Stud. 8 (4). University Press, New York.
Hallerod, B., Gustafsson, J.E., 2011. A longitudinal analysis of the relationship between Savage, M., 2015. Social Class in the 21st Century. London, Pelican.
changes in socio-economic status and changes in health. Soc. Sci. Med. 72 (1), Scott, M., Zeidenberg, M., 2016. Order or chaos? Understanding career mobility using
116–123. categorical clustering and information theory. Longitudinal Life Course Stud. 7 (4).
Hoven, H., Siegrist, J., 2013. Work characteristics, socioeconomic position and health: a Scott-Marshall, H., Tompa, E., 2011. The health consequences of precarious employment
systematic review of mediation and moderation effects in prospective studies. Occup. experiences. Work 38 (4), 369–382.
Environ. Med. 70 (9), 663–669. Shaw, M., Dorling, D., Gordon, D., Davey Smith, G., 2000. The Widening Gap: Health
Hox, J., 2010. Multilevel Analysis: Techniques and Applications. Routledge, Abingdon. Inequalities and Policy in Britain. The Policy Press, Bristol.
Juarez, P.D., Matthews-Juarez, P., Hood, D.B., Im, W., Levine, R.S., Kilbourne, B.J., Siegrist, J., 1996. Adverse health effects of high-effort/low-reward conditions. J. Occup.
Langston, M.A., Al-Hamdan, M.Z., Crosson, W.L., Estes, M.G., Estes, S.M., Agboto, Health Psychol. 1 (1), 27–41.
V.K., Robinson, P., Wilson, S., Lichtveld, M.Y., 2014. The public health exposome: a Siegrist, J., Benach, J., McKnight, A., Goldblatt, P., Muntaner, C., 2010. Employment
population-based, exposure science approach to health disparities research. Int. J. Arrangements, Work Conditions and Health Inequalities. Strategic Review of Health
Environ. Res. Publ. Health 11 (12), 12866–12895. Inequalities Post 2010. UCL Institute of Health Equity, London.
Kampanellou, E., Houston, D., 2016. The health impacts of the contemporary manu- Spiegelhalter, D.J., Best, N.G., Carlin, B.P., Van Der Linde, A., 2002. Bayesian measures of
facturing and service sectors on men and womern. Longitud. Life Course Stud. 7 (4), model complexity and fit. J. Roy. Stat. Soc. B 64, 583–639.
368–385. Standing, S., 2011. The Precariat: the New Dangerous Class. Bloomsbury, London.
Karasek, R., Theorell, T., 1990. Healthy Work: Stress, Productivity, and the Vineis, P., Chadeau-Hyam, M., Gmuender, H., Gulliver, J., Herceg, Z., Kleinjans, J.,
Reconstruction of Working Life. Basic Books, New York City. Kogevinas, M., Kyrtopoulos, S., Nieuwenhuijsen, M., Phillips, D.H., Probst-Hensch,
Kauskamp, D., Ziersch, A.M., Baum, F.E., LaMontage, A.D., 2013. Precarious employ- N., Scalbert, A., Vermeulen, R., Wild, C.P., 2017. The exposome in practice: design of
ment, psychosocial working conditions, and health: cross-sectional associations in a the EXPOsOMICS project. Int. J. Hyg Environ. Health 220 (2A), 142–151.
population-based sample of working australians. Am. J. Ind. Med. 56, 838–844. Virtanen, M., Kivimaki, M., Joensuu, M., Virtanen, P., Elovainio, M., Vahtera, J., 2005.
Kim, I.-H., Muntaner, C., Vahid Shahidi, F., Vives, A., Vanroelen, C., Benach, J., 2012. Temporary employment and health: a review. Int. J. Epidemiol. 34 (3), 610–622.
Welfare states, flexible employment, and health: a critical review. Health Pol. 104 Wild, C.P., 2005. Complementing the genome with an "exposome": the outstanding
(2), 99–127. challenge of environmental exposure measurement in molecular epidemiology.
Kivimaki, M., Virtanen, M., Kawachi, I., Nyberg, S.T., Alfredsson, L., Batty, G.D., Bjorner, Cancer Epidemiol. Biomark. Prev. 14 (8), 1847–1850.
J.B., Borritz, M., Brunner, E.J., Burr, H., Dragano, N., Ferrie, J.E., Fransson, E.I., Wild, C.P., 2012. The exposome: from concept to utility. Int. J. Epidemiol. 41 (1), 24–32.
Hamer, M., Heikkila, K., Knutsson, A., Koskenvuo, M., Madsen, I.E.H., Nielsen, M.L.,

62