Back pain study July 2012 measuring biopsychosocial risk factors

advertisement
WorkCoverSA
Work-related back pain study:
measuring biopsychosocial risk
factors
Discussion paper
July 2012
Authors: Radek Stratil, Clinical Psychology Consultant &
Margaret Swincer, Manager, Research
Disclaimer
Free information support services are available for:
TTY (deaf or have hearing/speech impairments) call (08) 8233 2574.
Languages other than English call the Interpreting and Translating Centre (08) 8226 1990 and ask for an
interpreter to call WorkCoverSA on 13 18 55.
Braille, audio, or e-text call 13 18 55.
The information in this publication is compiled by WorkCover Corporation of South Australia. The data and facts
referred to are correct at the time of publishing and provided as general information only. It is not intended that
any opinion as to the meaning of legislation referred to is to be relied upon by readers who should seek
independent advice as to any specific issues relevant to you, your workplace or organisation.
This publication is not intended as a substitute for the requirements of the Workers Rehabilitation and
Compensation Act 1986 or the Occupational Health Safety and Welfare Act 1986.
© WorkCover Corporation of South Australia, 2010, published [insert month] 2010.
© WorkCoverSA, 20
Page 2
Contents
Executive summary ...............................................................................................................................................4
Background ............................................................................................................................................................8
Aim of study .............................................................................................................................................................8
The biopsychosocial model .....................................................................................................................................8
Risk assessment ..............................................................................................................................................9
The flags model ................................................................................................................................................9
Psychosocial study outline ............................................................................................................................... 11
Methodology ......................................................................................................................................................... 11
Participants .................................................................................................................................................... 11
Assessment ................................................................................................................................................... 11
Worker supplied information .......................................................................................................................... 12
Clinician supplied information ........................................................................................................................ 13
WorkCover database information .................................................................................................................. 13
Results ................................................................................................................................................................. 13
Demographic data on participants and non-participants ...................................................................................... 13
Patterns of recovery ...................................................................................................................................... 13
Accuracy of comprehensive assessment tools..................................................................................................... 15
Discussion ........................................................................................................................................................... 18
Who should do the assessment and when? ......................................................................................................... 19
Which psychometric instruments to use? ............................................................................................................. 20
Brief screening instrument: applicability of Orëbro ........................................................................................ 22
Guidance on potential risks and potential interventions ................................................................................ 22
Conclusion and recommendations ....................................................................................................................... 24
Appendix A Psychosocial studies conducted in WorkCoverSA ........................................................................... 25
Appendix B Examples of key conceptual models representing the domains of psychosocial risk. ..................... 27
Appendix C Screening principles .......................................................................................................................... 28
Appendix D Psychometric instruments used and key domains measured .......................................................... 29
Appendix E Who is in the study? .......................................................................................................................... 30
Bibliography ........................................................................................................................................................ 31
© WorkCoverSA, 20
Page 3
Executive summary
This report presents the rationale for and findings from a multistage, prospective and longitudinal study of backinjured workers which investigated the risk factors present at acute stages of injury affecting disability, recovery,
return to work outcomes and claim costs. The study also obtained information on the frequency and type of
medical treatment, claims management and workplace rehabilitation. Provided are recommendations to improve
screening for and management of key risk issues in back-related work injuries.
Aim of study
The objective of this study was to improve the early identification of specific psychosocial risk factors that could
be targeted by evidence-based medical and vocational management. The current study extends and broadens
the findings of a 1999-2002 WorkCover study that demonstrated the validity of early risk screening using the
Orëbro low-back questionnaire (Linton and Hallden 1998) at acute stages of a work-related back injury to detect
general risks of long term disability.
Study design
Unique features of the current study include peer reviewed prospective design, use of ‘gold standard’ screening
and comprehensive psychosocial risk assessment tools with injured workers beginning at an acute stage of
injury with periodic follow-up assessment (3, 6, 12 and 24 months post injury), and enhanced by the integration
of multiple data sources on risk issues impacting on health, work and compensation outcomes.
The participants were over five hundred injured workers of employers registered with WorkCover SA, recruited
at the acute stages of back related work injury. The participation rate was high; fifty percent completed risk
assessment tools and follow-up questionnaires. The instruments used were those recommended by
international and Australian expert panels for use to assess major domains of risk and outcomes of pain related
injuries, selected for practical applicability, and being in use by health clinicians in South Australia.
Key to the study was the comparison of the effectiveness between multiple measures of risk, disability and
treatment progress against outcomes relevant to the local compensation environment. The psychometric risk
information was cross-linked with the WorkCover database information and also with the medical information
provided by the workers’ treating general practitioners and physiotherapists regarding the worker’s capacity,
health status and treatment. The study also used qualitative information provided by workers to provide
additional data on disability, treatment and rehabilitation outcomes.
As the follow-up was conducted with all participants for two years post-injury, irrespective of subsequent
compensation status, this study has for the first time in Australia provided accurate information on injured
workers’ long term vocational, health, and functional outcomes and information on re-injury rates. Furthermore
the integrated individual and treatment data collected in this study provides a useful and comprehensive
baseline of current treatment, management strategies and long term health and vocational outcomes useful to
measure the impact of new strategies targeting outcomes.
Key messages
General study findings
The tools used in the study identified a small number of discrete risk factors present at acute stages of injury
which accurately identified two thirds of the ten percent of injured workers who subsequently developed a
chronic pain related work disability. Those identified as ‘at risk’ had risk scores on the psychosocial assessment
tools at levels equivalent to scores reported in chronic pain patients seeking treatment and these scores
decreased only marginally with time. Those at ‘low risk’ had significantly lower risk scores which rapidly
decreased within two months post-injury. This finding supports the view that psychological factors are present
© WorkCoverSA, 20
Page 4
early (Dunn 2010) and reinforce the need to identify those at risk early and act to address these issues as
recommended in relevant guidelines.
By contrast, the study found that the reported treatment by doctors and physiotherapists participating in the
study changed marginally since the 2002 WorkCover study with lower rates of ‘passive’ forms of treatment in
line with evidence based guides, however the participating clinicians when asked to assess the psychosocial
risk in their patients were unable to correctly identify more than 5% of those with ‘non-organic’ presentation at
subacute stages of injury and only 13% accurately identified “moderate risk of chronic pain”. Thus these study
findings suggested that clinical judgment if used alone at early stages to inform medical treatment will need to
be better informed by the significantly more accurate psychosocial assessment provided by the tools validated
in this study. This is important as a more specific management is recommended in the relevant guidelines
(National Health and Medical Research Council (Australia) 1999) for managing those “at risk”.
Specific study findings
The psychosocial tools used identified at acute stages of injury six key and overlapping dimensions of
psychosocial risk that accurately predicted compensation and return to work outcomes. These included:
1.
high pain intensity and quality
2.
high perceived disability which was impacting on most home, work, recreational and social activities
3.
excessive pain focus and fear of aggravating pain through normal daily activity
4.
beliefs about having a negative future prognosis and belief that work activities were unsafe
5.
poor pain related self-efficacy, poor belief in own self-management of pain
6.
prominent psychological distress including anxiety and depressive symptoms which further adversely
affected their ability to cope.
The perception of the workplace supervisor as being non-supportive added to the risk in a sub-group of injured
workers. Most of the identified risk issues are potentially modifiable and this paper proposes interventions to
address these risks at the acute and subacute stages of compensable work injury.
Follow-up assessment with relevant tools found changing patterns of psychosocial risk during transition to
recovery or chronic disability. As shown Figure 1, the proportion of workers not fully cleared for work decreased
rapidly to eight weeks then tapered off from 26 to 52 weeks reinforcing the opportunity for early intervention to
maximise positive outcomes.
However the medical, physiotherapy treatment pattern did not change over time to address risk issues as
recommended in evidence guides. For example, as shown in Figure 1, physiotherapists who provide up to 80%
of treatments provided this at a relatively constant rate for all workers irrespective of risk status and outcome.
Vocational rehabilitation involvement increased gradually from 30% at 8 weeks post injury to over 80% by the
end of first year. By contrast, psychology practitioners were not significantly involved within twelve months of
injury. Given the evidence obtained in this study that psychosocial assessment can identify risk issues which do
not appear to be noted by the clinicians involved highlights the need to develop strategies that help inform these
practitioners and provide a better model of care as outlined in the recommendations.
Recommendations
This paper proposes an intervention strategy using early risk assessment tools to assist and inform clinical
judgement to significantly improve compensation outcomes. Figure 2 summarises the proposed process, tools
and procedures to be used during the three phases of injury from acute to chronic to improve screening and
management of key risk factors by clinicians. Two intervention strategies proposed are: 1) Encourage risk
screening by doctors and physiotherapists at early stages of injury to improve the current injury management
model. 2) Develop risk screening involving a two-stage process to help identify subgroups needing specific
interventions. Such interventions may involve specifically trained psychologists and other clinicians to provide
targeted brief interventions focussing on key factors amenable to clinical intervention. See Figure 2 for a
summary with details discussed within this report.
© WorkCoverSA, 20
Page 5
Figure 1. Key activities at various stages after
injury
Recovery rates of back injured workers
70
and servicing patterns at each phase
Percentage not fully fit on medical certificate
60
50
40
30
20
10
Claim
lodged
for
recent
soft
tissue
injury
© WorkCoverSA, 20
4 to 8 weeks
üMore medical
provided
üMore physical
therapy
üLittle voc rehab
0 to 4 weeks üLittle
psychology
ü1st medical
visits
üPhysical
therapy
starts
12 to 26 weeks
üMore medical provided
8 to 12 weeks
üMore physical therapy
üMore medical provided
on more then 80% of
üMore physical therapy
claims
ü 1/3 of workers have
üVoc rehab on up to
voc rehab at 8 wks,
82% of claims
2/3 by 12 wks
üLittle psychology
üLittle psychology
52 weeks plus
üMore medical provided
26 to 52 weeks
üPhysical therapy on
üMore medical provided
90% of claims
üMore physical therapy üVoc rehab on more
provided
than 80% of claims
üVoc rehab on more
üPsychology on 42% of
than 80% of claims
claims
üPsychology on 14% of
claims
0
at 4 wks or
more
at 8 wks
or more
at 12 wks
or more
Acute phase
at 52 wks
or more
at 26 wks or
more
Sub-acute phase
Chronic phase
Page 6
Figure 2. Screening for psychosocial issues
Who screens whom, when and for what? How to and who should intervene?
Risk assessments
Acute
phase
Use brief screeners like
Orebro and pain intensity
used by GP at 2nd visit for
these high risk factors
(i) function
(ii) global prediction of RTW
(iii) fear/avoidance
behaviour
(iv) serious re-injury fears
Intervention type
Early intervention
Reassurance by GP and
physio on:
üpain
üseverity of injury
© WorkCoverSA, 20
Sub-acute
phase
Chronic
phase
Two tiered screening
(i) Brief screener with GP for high level
risks then
(ii) comprehensive assessment to
identify key psychosocial issues by
psychologist include Pain Disability
Index, Pain Self-efficacy Q’re, K-10
Comprehensive screening plus clinical input
Using Pain Disability Index, Pain Self-efficacy Q’re
Pain Catastrophising Scale, Depression Anxiety
Stress Scale
Cognitive behaviour techniques to
avoid chronic pain syndrome
addressing
ü fear/avoidance behaviour at work
ü high pain focus
ü distress
Multi disciplinary team (or virtual team of
experienced clinicians)
General practitioner/occupational physician
Psychologist
Occupational therapist
Workplace rehabilitation provider
or rehabilitation RTW coordinator
Cognitive behaviour techniques to
counter
ü distress, depression, anxiety
ü self perceived disability
ü pain management techniques
Pain management techniques via
ü physical exercise regime
ü coaching/motivational techniques
Page 7
Background
This section describes the rationale for and the outcomes of the longitudinal psychosocial risk assessment
study (also known as Back Pain Study) of over five hundred injured workers. The study aimed to provide
high quality practical information to improve early management of back-related injuries in the South
Australian workers’ compensation scheme. The study involved the early assessment of key psychosocial,
health, disability and work related issues with periodic follow up at 3, 6, 12, 24 months post injury.
Aim of study
The range of goals of this multistage longitudinal study included:

Identify key psychosocial risk drivers of chronicity that are potentially modifiable to early
management.

Extend and improve the current risk assessment beyond the actuarial data analysis by integrating
this information with relevant psychosocial information.

Develop a practical and testable model of risk assessment and intervention applicable to individuals;
expand on the findings from an earlier WorkCover study conducted in 1999-2002 which
demonstrated the accuracy of using a brief early screening tool recommended as a part of the NZ
ACCC ‘Yellow flags’ project (Kendall 1999).

Develop more comprehensive measures of baseline risk indicators, to help identify subgroups
requiring specific targeted treatment and management.

Evaluate the nature and relevance of current treatment and management strategies to improving
injured workers’ long term health and vocational outcomes, integrate this with information in the
WorkCover database to help develop more effective and targeted management specific to relevant
subgroups of injured workers.

Provide a more comprehensive and accurate integrated baseline to better measure the impact of
new strategies on long term health and vocational outcomes of injured workers and their employers.
The biopsychosocial model
It has been recognised for some time that long-term musculoskeletal, sprain and strain injury claims involve
a complex interaction of biomedical, psychosocial and legal factors and increasingly the clinical and applied
research evidence indicated that psychosocial and compensation issues are as important as medical factors
in determining recovery from work-related musculoskeletal injuries (Burton, Tillotson et al. 1995; Turk and
Okifuji 2002; Linton, Gross et al. 2005). The biopsychosocial nature of disability was recognised by the
Australian government together with 200 other nations in the internationally agreed WHO International
Classification of Functioning, Disability and Health (ICF) (Kuijer, Brouwer et al. 2006) now being adapted for
use in conceptualizing pain related disability (Lakke, Soer et al. 2009; Roe, Sveen et al. 2009).
Despite considerable advances in this area many gaps exist, particularly in applying this information to the
specifics of the compensation environment and in changing health provider behaviour. Early screening has
now been shown to be effective in identifying a ‘general risk of future disability’ and a large number of
screening tools have been validated (Linton and K 1996; Shaw, Pransky et al. 2005; Hill, Dunn et al. 2010;
Linton, Nicholas et al. 2011) including those validated within the WorkCover Psychosocial Indicators of
Chronicity Study (PICS) project 1999-2002.
Page 8
However, these brief screeners are not intended to identify specific risk issues or to select people into
subgroups based on shared risk. Having a general risk score therefore does not enable targeted treatment
by primary health clinicians or rehabilitation or help workplace management (Dunn and Croft 2005; Foster,
Hill et al. 2011). Furthermore studies have identified a large number of potential ‘psychosocial’ factors
comprising ’risk’, however, these differ with the nature of the compensation scheme, work environment,
country or stages of the disability; few studies are directly relevant to workplace-based rehabilitation and
recovery. Some of the identified factors appear to contribute to disability, others appear as consequences
(sequelae), others appear to be associated in more complex ways with the development of long-term chronic
compensation claims (Gatchel, Peng et al. 2007).
The nature and interaction of some of these risk factors can change during the transition from acute to
chronic stages but only a few factors have been consistently shown to be amenable to intervention as shown
in the flags project, making it difficult to develop effective treatment and management strategies. Only a few
studies such as (Linton 2001; Boersma and Linton 2005), have used a prospective design or were
conducted in the compensation arena but even these were not directly applicable to South Australian
scheme (Schultz, Crook et al. 2005; Nicholas 2010).
The strategies to-date mainly addressed the education of workers, treatment providers and employers. A key
focus was placed on provider education and guidelines to help improve medical and physiotherapy clinicians’
assessment and management of psychosocial issues (National Health and Medical Research Council
(Australia) 1999). Other strategies attempted to prevent chronic outcomes by positively influencing beliefs
held by the general public (including clinicians) about pain and early return-to-function through mass
education campaigns such as by the Victorian WorkCover Authority in 2005 (Buchbinder 2008). Despite
these strategies, the long term impact on provider behaviour and on disability rates has been seen as
minimal at the level of the individual (Foster, Dziedzic et al. 2009). Many compensation schemes including
WorkCover have also attempted to provide specific education and guidelines for health providers (TREAT
and other projects).
Risk assessment
Actuarial and clinical models
Two key risk-screening approaches used in the past were the actuarial/administrative approach, and clinical
judgment approach (Waddell, Burton et al. 2003). While many compensation schemes use actuarial risk and
outcome assessment tools, most have limited application to the risk assessment and management outcomes
of specific individuals or groups of individuals, leaving the clinicians to identify and manage. However, risk
assessment done by medical and rehabilitation professionals has not been shown to follow available
guidelines as it is an unreliable process subject to many potential biases (Main, Foster et al. 2010).
The flags model
One of the earliest and the most influential models of risk assessment and management was the ‘Flags’
model of biopsychosocial risk assessment developed in 1997 as a part of the NZ ACCC Guides project
(ACCC) then extended by an international working group of researchers in 2007 and 2010 (‘Decade of the
Flags’) to form a consensus classification using different ‘flags’ to denote various risk domains.(Shaw, van
der Windt et al. 2009). Different representations are shown in Appendix B and illustrated in Figure 3.
Page 9
Figure 3 Representation of key psychosocial factors incorporating the ‘flags’ concept
Individual issues - ‘Yellow’ and ‘Orange’ flags
Perception of workplace - ‘Blue’ flags’
Coping skills
Severity of disorder, recurrence, symptoms
Mood and emotion
Clinical history, comorbidity
Thoughts and beliefs
Behaviour and activity
Quality of workplace contact
Availability of suitable duties
Job satisfaction, co-worker support
Management support, culture
Supervisor involvement, belief that work tasks are
harmful. Belief that worker will not RTW
Health care system - treatment issues
Systemic issues - ‘black’ flags
Delays in treatment
Non-evidence based treatment
Contradictory treatments
Uncoordinated treatment
Focus exclusively on physical injury
Lack of specialised training in pain-related arenas
Compensation and variable legislative requirements
Financial and legal
Social issues
Financial support, dispute - hardship
Legal issues, proof, claim validity
Reluctance to accept validity of psychosocial issues
*Red flags were used to indicate the rare physical conditions requiring immediate attention are not shown
here.
The ‘flags’ concepts illustrated in Figure 3 emphasise that risk assessment may need to take into account
the interaction of many issues, some are beyond an individual. For example, the systemic issues where
workers may not be used to dealing with large bureaucracies, paperwork and professional groups such as
doctors and lawyers (Schultz, Crook et al. 2000; Loisel, Durand et al. 2003). Other similar conceptual models
include those of Loisel (Loisel, Buchbinder et al. 2005; Loisel, Hong et al. 2009) (also shown on Appendix B).
While the ‘flags’ models have been used as aids to management (Kendall, Burton et al. 2009), a need exists
to develop structured and practical risk-screening and management procedures to change provider
behaviour to target relevant issues.
Therefore, despite the acceptance of the biopsychosocial models of disability and considerable and
expensive education strategies undertaken worldwide, very little of the increasing knowledge about workrelated disability prevention has been effectively translated to modify actual health provider behaviour, to
improve the effectiveness of workplace rehabilitation management or to an improvement in disability rates
(Shaw, van der Windt et al. 2009; Main, Foster et al. 2010).
Page 10
Psychosocial study outline
The current psychosocial study proposed to develop a specific assessment and intervention strategy that
combines accurate and cost effective risk-screening assessment that could be linked to targeted treatment
and rehabilitation management by appropriately informed and trained providers. Similar strategies are now
under-way overseas (Hill, Dunn et al. 2008; Schultz, Crook et al. 2008), however, none are directly relevant
to the SA Scheme. The current study was therefore undertaken to help address this gap and to be able to
integrate with actuarial and individual/clinical data. Furthermore, this study was also intended to provide an
accurate baseline for measuring outcomes and the impact of other strategies at both individual, group and
scheme level (for the first time in Australia). The principles of screening are discussed in Appendix C.
Methodology
The study was designed in 2005 as a prospective cohort study following an extensive review of 1200 studies
on early risk-assessment and intervention. The design was approved by the University of South Australia
(UniSA) Ethics Committee with oversight and assistance from key Australian and international clinicians and
researchers in this field including A/Professor Michael Nicholas. The prospective study has been conducted
under the WorkCover research agenda in cooperation with UniSA from 2006 to 2010.
Participants
Injured workers fitting the study criteria were identified via the WorkCoverSA database at two weeks postinjury and were invited to participate in the study. Each participant was asked to complete psychosocial riskscreening questionnaires at each stage of injury. Over 600 injured workers agreed to join the study from
among the 1300 injured workers approached.
The main selection criteria were: no work-related compensation claim in the previous year; a primary injury
to the back; and not fully cleared to return-to-work. In addition, the study tracked the database outcomes of
the comparison group of 700 workers who were eligible and fit the study criteria but did not participate.
Information sources
The information from the mailed questionnaires was matched to additional data sources including from the
workers’ treating clinicians and the WorkCover claims management database on services by providers of
healthcare, rehabilitation and diagnostic services and chemists prescriptions. The worker’s treating medical
and physiotherapy clinicians (with permission from their patient) were asked to independently outline their
clinical management and provide a risk-assessment.
Assessment
Outcomes assessed included database information on return-to-work, the extent and duration of income
maintenance and disability ratings and written qualitative information provided by the worker at 6, 12 and 24
months after injury.
The study used key proven risk, health and return-to-work measures in domains relevant to the
biopsychosocial model of injury. Assessment data was collected in the acute injury phase (2-8 weeks) with
individual assessment data provided by workers at 3, 6, 12, 24 months post-injury. Weekly data on individual
outcomes and treatment was obtained from WorkCover database.
Page 11
Worker supplied information
Workers supplied quantitative and qualitative data collected directly via mailed self-report questionnaires and
included short screening instruments and comprehensive measures covering disability, function, distress,
fear-avoidance, self-efficacy, and perceptions of the work environment.
The comprehensive psychometric measures included ‘gold standard’ instruments recommended by key
practitioners in the field (Deyo, Battie et al. 1998; Dworkin, Turk et al. 2005; Turk, Dworkin et al. 2008).
Instruments included measures of work-related issues such as the perception of the workers that workplace
accommodation was available and that supervisors and co-workers were supportive of the rehabilitation
process. These measures are now seen as major risk issues during rehabilitation as highlighted in recent
evidence-based guidelines (Shaw, Pransky et al. 2005; Shaw, van der Windt et al. 2009). System-based
issues described as ‘black’ flags are also seen as potential risks by the ‘Flags’ group of researchers (Kendall,
Burton et al. 2009). Table 1 lists the key screening and comprehensive measures used.
The workers also provided information on long-term outcomes including durable RTW, absenteeism and
presenteeism1 rates and qualitative assessment information on satisfaction with care, long-term impact of
injury and quality of life.
Table 1 Key psychometric instruments used to measure psychosocial risk factors
Screening type
Risk domain of
individual
Instrument
Brief screeners
Pain intensity
Numerical rating score out of 10
Pain quality
Single scale question
Durable RTW
Single scale question
Brief screeners
Orëbro (Linton and Hallden 1998)
Occupational screening questionnaire (Shaw, Pransky et al. 2005); Beliefs
about return to work and function (Dunn and Croft 2005; Hill, Dunn et al.
2008) and possible system issues (black flags)
Perceptions of workplace support (blue flags) (Shaw, Pransky et al. 2005;
Shaw, van der Windt et al. 2009)
Comprehensive
instruments
Function
Pain Disability Index (PDI
Oswestry Disability Index (Oswestry) ) (Gronblad, Jarvinen et al. 1994;
Fairbank and Pynsent 2000)
Fear-avoidance beliefs
and pain catastrophising
Pain Catastrophising Scale (PCS) (Sullivan, SR et al. 1995)
Confidence about coping
with pain
Pain Self Efficacy Questionnaire (PSEQ)(Asghari and Nicholas 2001)
Emotional distress
Depression Anxiety and Stress Scale (DASS) (Lovibond and Lovibond
Tampa Scale of Kinesiophobia (Tampa) (Kori, Miller et al. 1990)
Fear of work activities from the Fear Avoidance Beliefs Questionnaire
(FABQ) (Waddell, Newton et al. 1993)
1995)
Kesslers or K10 (measures psychological distress)(Furukawa, Kessler et
al. 2003)
Positive and Negative Affective Scales (PANAS provides a measure of
depression and anxiety traits)(Watson, Clark et al. 1988)
1
at work with diminished capacity
Page 12
Appendix D details the particular risk domains covered by the brief screening instruments and more
comprehensive measures.
Clinician supplied information
Treating general practitioners and physiotherapists completed mailed questionnaires (based on the 2005
Shaw, Pransky et al format) providing relevant biomedical information on the injury, incapacity, treatment
regime and the clinician’s predictions of outcome and an assessment of worker’s level of depression.
WorkCover database information
Quantitative and qualitative data from the WorkCover database were used including income maintenance (or
discontinuance), treatment patterns, medications used, scans and rehabilitation outcome data. Information
included other qualitative data regarding the claim. That information is not reported on in detail here.
Results
Demographic data on participants and non-participants
The 550 study participants broadly represent the age and gender distribution of the South Australian working
population. Sixty-eight percent of the participants were male which aligns to the demographics of other
musculoskeletal injury types. Additional de-identified data was also obtained from the WorkCover database
regarding the characteristics of the 600 injured workers with back injuries who did not complete the
questionnaires. This analysis was done to assess the potential differences between the two groups to
indicate potential bias. The general recommendations for studies are to have a participation rate of 60-80%.
However the participation rate in this study is very high for this population. Previous brief mailed surveys as a
part of WorkCover reviews had much lower participation rates. Furthermore the information required was
sought at acute stages of injury where a majority of people were recovering and as indicated in the
telephone follow-up many of the non-participants felt that as they had “recovered” and did not need to
participate. Furthermore given that the questionnaire had over 100 questions the rate of 50% was deemed
as extremely good. None the less the analysis comparing demographics, return to work, medical intervention
and injury recurrence rates did not indicate significant difference. Key differences were female gender and
older ages marginally were more likely to participate and workers who at the time of being approached had
fully returned to work were less likely to participate. Importantly the educational data obtained from the
participants closely fitted the distribution of educational qualifications in the general employed population,
reducing the possibility of bias based on this criterion. See Appendix E for more information on
demographics of the study sample.
Patterns of recovery
The recovery trend (Figure 1) shows that at four weeks after claim lodgement, two-thirds of the workers were
still not fully recovered. The medical certificate data shows the workers were not fit for their pre-injury job by
being either unfit or fit for alternative duties. The trend then shows a significant increase in return-to-work
whereby at three months the proportion not fully recovered has almost halved to be 35 per cent. The rate of
return-to-work then decreases, by six months 25 per cent are still not recovered, reducing to 16 percent by
12 months.
Page 13
Treatment patterns
The outcome data shows that most workers return to work within three months of injury (Figure 1). Claims
management data (Figure 4) on these back-injured workers shows that physiotherapists provide up to 80%
of treatments from early stages; similar treatment frequency is maintained for all workers irrespective of risk
status and outcome. Vocational rehabilitation which is seen as critical to early management increased
gradually from 30% at 8 weeks post-injury to over 80% by the end of first year. Psychology practitioners
were not significantly involved within six to twelve months of injury.
Information from the clinicians and WorkCover database shows that there has been a small reduction since
2002 in the use of the more passive forms of treatment as recommended by evidence based guides,
however the treatment pattern continues to have a biomedical focus as illustrated by the frequency of
physiotherapist services and by the number of scans that occur at acute stages of injury. Similar trends have
been reported in recent studies where the frequency of physiotherapy intervention did not decrease in lowrisk groups (Foster 2011).
Figure 4 Servicing patterns for workers certified as not fully fit at various intervals
Services to workers with back injuries not fully cleared for work
100.0
89.9
90.0
80.0
84.6
82.4
80.1
77.5
67.4
70.0
Percentage (%)
92.2
89.8
57.0
60.0
50.0
50.0
43.0
40.0
42.2
41.2
37.2
36.7
34.7
32.6
30.0
26.2
22.5 23.3
20.0
13.6
10.0
6.7
0.2
3.6
1.0
0.0
Not fully fit 1 month
Not fully fit 2 months
Not fully fit 3 months
Not fully fit 6 months
Not fully fit 12 months
services at different times of claim
% receiving allied health
% chemist prescript
% scans-xray/MRI/CT
% Voc rehab
% psychology
Findings on accuracy of early clinical assessment:
During the acute/subacute phase of the injury, information was sought from the worker’s treating medical
practitioner who provided an assessment on:

the worker’s functional limitations and pain

treatment and referrals
Page 14

potential co-morbidities

prediction of return to work outcomes

how much the worker is depressed.
Table 2 summarises the results from the clinicians.
Table 2 Summary of clinical assessment data from treating practitioners
Information
sought from
treating doctor
Results
Comment
Previous episodes
40% reported knowing about a previous
episode
60% didn’t notice or reported no previous
episode
Knowledge was quite variable and lacked
consistency
This may possibly indicate:

a lack of knowledge

lack of interest, or

insufficient time to pursue the worker’s
relevant medical history
Prediction of RTW
13% reported high risk of chronic pain however
equal number incorrectly identified risk
Predicting a high proportion of chronicity which
is less than actual level
Other conditions
impacting on RTW
20% reported other conditions
Most common conditions were:
(i)
other injuries e.g shoulder, neck
(ii)
mental disorder
(iii)
general musculoskeletal problem
(iv)
co-morbidities such as obesity
(v)
pregnancy
Treatment and
investigations
Nearly 60% prescribed rest
75% referred to physiotherapy
More than one third requested x-rays and
other imaging investigations
Passive treatment such as rest does not
accord with the WCSA Back Guidelines
The WCSA Imaging Guidelines do not
recommend x-rays etc at this stage
Participating clinicians when asked to assess the psychosocial risk in their patients were unable to correctly
identify more than 5% of those with a ’non-organic’ presentation at the subacute stages of injury and only
13% accurately identified a ’moderate risk of chronic pain’ with equal numbers incorrectly identifying as high
risk those that did not continue to become chronic.
Thus these study findings suggest that clinical judgment if used alone to inform medical treatment will need
to be better informed by the significantly more accurate psychosocial assessment provided by the tools
validated in this study. This is important as a more specific management is recommended in the relevant
guidelines (National Health and Medical Research Council (Australia) 1999) for managing those ‘at risk’.
Accuracy of comprehensive assessment tools
The validity of the brief screening tools has been previously demonstrated in the PICS study (described in
Appendix A). Therefore, the main object of this study focussed on data from the comprehensive assessment
tools to measure the accuracy of these tools and their effectiveness in helping to identify specific risk issues.
Data analysis has identified a small number of discrete risk factors present at acute stages of injury which
accurately identified two thirds of the ten percent of injured workers who subsequently developed chronic
pain related work disability.
The scores on the comprehensive assessment tools in each risk domain differentiated between injured
workers with chronic outcomes as defined by ‘any income maintenance paid beyond 12 months post-injury’
from those that had ‘ceased income maintenance before three months post-injury’.
Page 15
The six risk categories were as follows:
1. high pain intensity and quality
2. high perceived disability which was impacting on most home, work, recreational and social
activities
3. excessive pain focus and fear of aggravating pain through normal daily activity
4. beliefs about having a negative future prognosis and belief that work activities were unsafe
5. poor pain related self-efficacy, poor belief in own self-management of pain
6. prominent psychological distress including anxiety and depressive symptoms which further
adversely affected their ability to cope.
The relevant instruments in each risk category accurately predicted compensation and return-to-work
outcomes at acute stages of injury as shown in Table 3. The scores for the group with ‘no income
maintenance at any stage’ (not shown in the Table 3) have lower scores that the two groups shown. This
additional group is likely to represent the background levels of pain, disability and recurrence in the
community as those claimants had only a minor exposure to the compensation environment.
Table 3 shows that workers with chronic outcomes have significantly higher risk scores at acute stages
except the PSEQ where a reduction in scores also indicates poor self-efficacy. Figure 5 provides graphical
comparison of the trends for the disability scores observed between workers who have chronic outcomes to
those that do not.
Further comparison to data from studies of chronic pain patients indicates that the ‘at risk’ workers in the
current study had as a group, risk scores on the psychosocial assessment tools at levels equivalent to
scores reported in chronic pain patients seeking treatment (Nicholas, Asghari et al. 2008). These scores
decreased marginally over two months post-injury gradually increasing with ongoing disability between one
to two years post-injury. By contrast, those workers who did not follow a chronic trajectory, while having
elevated scores on relevant instruments, these were significantly below the scores obtained by those
workers with chronic outcomes. Furthermore the scores of those not at risk decreased more steeply than
those identified as ‘at risk’.
Table 3 Scores from groups with: (i) no income after 14 weeks, (ii) income for more than 12 months on the key
psychometric instruments
Tests
1.
Income
maintenance status
Initial
Subsequent
6
months
12
months
<15 weeks
4.6
2.3
1.4
2.0
>12mths
6.5
5.6
5.2
5.9
Perceived disability
(PDI)
<15 weeks
27.3
15.3
11.9
8.6
>12mths
43.8
38.8
37
38
Pain focussing (PCS)
<15 weeks
17.9
15.2
14.2
10.3
>12mths
24.7
23.8
21.7
21.3
Beliefs about work,
prognosis (FABQ_W)
<15 weeks
20.2
17.7
N.A.
N.A.
>12mths
25.4
26.8
N.A.
N.A.
Poor self-efficacy
(PSEQ)
<15 weeks
41.6
46.6
45.7
47.1
>12mths
22.1
26.8
21.9
23.8
Distress DASS-S
<15 weeks
12.4
9.3
7.8
8.3
>12mths
19.9
19.3
20.1
19.8
Pain -
(Numerical Rating Score)
2.
3.
4.
5.
6.
Note PSEQ is reverse scored – higher the score, better coping skills, lower the score, poorer the self-efficacy
Page 16
Figure 5 Disability scores at various stages after injury for different cohorts
Pain Disability Index
50
40
30
20
10
IM >12 mths
all claims
no IM
0
initial
subsequent
after 6 months
after 12 months
claim duration
This finding supports the view that psychological factors are present early and continue over time (Dunn
2010) reinforcing the need to identify those at risk early and act to address these issues as recommended in
relevant guidelines. Most of these risk categories include issues which can be managed if identified early
and practitioners are provided with relevant information and management tools.
Page 17
Discussion
The objective of this study was to improve the early identification of specific psychosocial risk factors that
could be targeted by evidence-based medical and vocational management. Psychometric and other data
were used to identify relevant risk domains and evaluate the accuracy of the current clinical risk identification
in informing clinical practice.
To date, very few relevant studies have been conducted in the Australian workers’ compensation context.
The study extended the findings of a 2001 WorkCover PICS study that demonstrated the validity of early
risk-screening using the Orëbro low-back questionnaire (Linton and Hallden 1998) at acute stages of a workrelated back injury to detect general risks of long term disability.
The current study found strong evidence for the accuracy of early risk-assessment which identified two thirds
of the ten percent of injured workers who subsequently developed a chronic pain-related work disability. The
comprehensive instruments used were able to accurately measure six related but separate domains of risk
which included: high pain intensity; perceived disability; pain focus and avoidance of feared activities of daily
living; beliefs about having a negative future work prognosis and belief that work activities were unsafe; poor
pain related self-efficacy; and psychological distress. Those identified as ‘at risk’ had risk scores on the
psychosocial assessment tools at levels equivalent to scores reported in chronic pain patients seeking
treatment and these scores decreased only marginally with time.
Also found was that current clinical assessment did not accurately identify those at risk, particularly in
distinguishing different domains of risk - pain, psychosocial ‘non-organic’ issues, the fears of work tasks and
those with prominent psychological distress. The accuracy rates were at 10-20% with an equal number of
false positives (those identified as at risk when they are not at risk). These findings support the view that
psychological factors are present early (Dunn 2010) and reinforce the need to identify those at risk early and
to address these issues as recommended in relevant guidelines. In addition to various educational
strategies, any intervention will need to use general and more comprehensive risk-assessment to better
inform clinicians about those at risk and what specific domains of risk need to be targeted in treatment. Few
studies (Linton 2001; Boersma and Linton 2005), have used prospective designs or were conducted in the
compensation arena (Schultz, Crook et al. 2005; Nicholas 2010). One study at the Concord Hospital in NSW
has trialled an early risk-assessment and management approach. That study, conducted within one large
workplace (with further studies planned), is not directly applicable to the WorkCoverSA scheme that has
many and varied employers and different treatment providers.
Any intervention needs to be based on accurate, timely, specific and cost-effective screening. The study
findings indicate that the most cost-effective and optimal screening could be done from four to eight weeks
post-injury using a two-staged process. The first stage would involve an initial brief screening to identify any
general risks. Those identified as at risk at the first stage can have a general psychosocial intervention or
undergo a more comprehensive assessment to confirm the risk and identify the specific risk domains that
require targeted intervention.
While both brief and comprehensive screening instruments were shown to be accurate, key challenges
remain in implementing risk-screening prompting the following questions: Who should conduct the screening
and management and at what stage of injury? How would it be done, what instruments to use? Should this
be a one or two-staged process? How to integrate risk assessment with improving current treatment
strategies?’
Page 18
Who should do the assessment and when?
The study findings indicate that risk screening is feasible at each stage of injury, however the most costeffective and optimal screening could be done at early stages from four to eight weeks post-injury. While
earlier assessment has been shown to be accurate (eg Concord study), given the nature of the claims
management process, size of employers and current medical management it is unlikely that an externally
triggered strategy is feasible under three weeks post injury.
Given the above limitations and the need to early assessment and intervention a number of potential
professionals may be considered to provide screening, assessment and or intervention.
A number of screening approaches have been trialled including case-manager administered questionnaires
(Hazard, Haugh et al. 1996), using primary medical clinicians, OHS nurse practitioners and physiotherapists
(Foster, Hill et al. 2011). Each approach has advantages and disadvantages and is specific to each
compensation scheme. Key issues are accessibility, trust and ability to refer and treat early. The following
section discusses the potential professionals who could be involved in this process.
Case manager
The relationship with the worker is remote with insufficient involvement in the early stages. Is there sufficient
trust in the case manager to answer personal questions? The worker might be wary of the impact on their
claim.
General practitioner
Doctors, often first in the injury management process, are endorsed by the system, society and the worker to
diagnose, determine interventions and treatments. Screening by questionnaire or interview techniques would
be an acceptable part of the clinical process. The doctors could be trained to interpret the results and provide
feedback to the worker. The probability that doctors would systematically undertake that role without
assistance is low, however, with sufficient education and support this may change. Currently, some doctors
do screen for risk-factors such as psychological distress in mental health domains using K10 (but not for
musculoskeletal disorders). These practices should be extended to the routine assessment of
musculoskeletal injuries. An informal survey of doctors who participated in this study indicated a preference
for having clinic staff handing out and collecting completed questionnaires for posting to an assessment
facility off site which can provide relevant feedback similar to the process for assessing blood samples.
Workplace rehabilitation provider or vocational rehabilitation provider
This group possesses general skills but is not always involved early and may not have be sufficient clinical
skills and confidence to interpret the results and provide feedback to the worker. They are able to
recommend the need for specialist treatment, but not refer.
Physiotherapist
This clinical group is often an early participant and continues to be involved at all stages of injury at the
highest levels of servicing. They have some awareness of psychosocial issues but workers with
psychological conditions are outside their skill domain. However, the worker might be comfortable being
screened by a physiotherapist which can be helpful to provide interventions for fear and avoidance beliefs.
Some physiotherapists use Orëbro or Oswestry to screen for high risk factors but generally do not use
psychological screeners generally and do not target psychological issues in a structured way.
Page 19
Occupational therapist
This clinical group is not often involved in the early stages but may help with screening; only some are
licensed under Medicare arrangements as suitable to provide range of psychological treatments.
Rehabilitation & return to work coordinator
This group is based at the workplace (or worksite) therefore the worker may not have sufficient confidence in
their ability to take on “medical type” tasks of assessment and interpretation of pain and psychosocial issues.
As a group they lack the specialist skills and confidence to interpret the results and may not be able to
communicate the assessment findings to medical or other specialists. However with appropriate training this
group can be important in assisting in the process for an early screening and management of psychological
issues.
Psychologist
A group rarely involved until psychological issues are obvious and disability behaviours are well entrenched.
An intervention for those identified with high risk factors needs to done early in the claim. It is probably not
effective to have routine screening by psychologists; a psychologist with relevant training and experience in
clinical and organisational psychology will be needed, and there could be waiting times for those who are
skilled in this arena. As there are no clinicians experienced in the early intervention a specific training would
need to be provided to selected clinicians.
Worker’s supervisor
Similar issues to the RTW coordinators.
Nurse
If based at the workplace or doctor’s practice, a nurse might have the trust of the worker to interview,
administer a questionnaire and obtain compliance from the worker.
Which psychometric instruments to use?
An extensive array of psychometric instruments has been tested in this study. Some instruments are strongly
correlated with each other and have been able to corroborate the results. The instruments that measure the
six key domains that best predict psychosocial risk are discussed below. The results for recommended
instruments are summarised in Table 4 and disability scores for different cohorts are presented in Figure 4.
The order of the risk domains relates to the impact that domain has on the outcomes. The most significant
predictor of future disability is the current disability; however the other dimensions are also key determinants
of risk. Comprehensive assessment should include the following elements:
1. Pain Severity
Recommendation: Pain severity scale of current and weekly pain (NRS) from the existing tools
such as the Orëbro or Oswestry (for low back pain only)
2. Perception of function or self-perceived disability
Both the Pain Disability Index (PDI) and the Oswestry Disability Index are highly correlated and
predictive of future outcomes. Even though the Oswestry is commonly used by physiotherapists
and doctors, it is specific to back injuries. While either can be used, the advantage of the PDI in
the WorkCover context is broader applicability across injuries and a potential to compare across
a range of body locations injuries to assess progress and the impact of management or
Page 20
treatmentstrategies.
Recommendation: Oswestry for low back pain only, PDI for all body locations
3. Pain focus
The Pain Catastrophising Scale (PCS) provides a measure of catastrophising personality traits.
People high in this dimension can be successfully treated with CBT (Main, Sullivan et al. 2008).
Recommendation: PCS
4. Beliefs about work and prognosis
The fear of work activities and a belief that a person will not be able to return to work duties is
one of the key predictors of future disability (Shaw, van der Windt et al. 2009).The Orebro has a
brief scale that assesses this dimension.
Recommendation: abbreviated Orebro (Linton, Nicholas et al. 2011) with supplementary
workplace relevant questions from the occupational disability screener (Shaw, Pransky et al.
2005) which was validated for use in the WorkCover context as a part of this study.
5. Coping skills with regard to pain
The Pain Self Efficacy Questionnaire (PSEQ) measures the confidence to self manage the
impact of pain. The PSEQ is one of the few instruments that provides a measure of positive
future focused approach. The PSEQ and PCS are moderately correlated, therefore the use of
both instruments might not be warranted.
Recommendation: the PCS and PSEQ
6. Distress and depression
Three measures of distress have been tested, Depression Anxiety and Stress Scale (DASS)
(Lovibond and Lovibond 1995), Kesslers (K10)(Kessler, Andrews et al. 2002), Positive and
Negative Affective Scales (PANAS)(Watson, Clark et al. 1988). The DASS is a more clinically
specific instrument extensively validated in the Australian and UK context (Lovibond and
Lovibond 1995; Crawford and Henry 2003) and also with pain patients(Nicholas, Asghari et al.
2008). The K10 is a more general instrument with less sensitivity but is simpler to use. Most
GP’s use this test routinely and are familiar with and the meaning of the K-10 scores. PANAS
measures stable affective states and can detect longer-term personal predisposition to
neuroticism (Crawford and Henry 2004) which has been associated with dysfunctional coping in
a range of “psychosomatic” conditions and particularly the impact on work performance (Kaplan,
Bradley et al. 2009). Our study indicated a possible use of PANAS in research on general
personality predictors and cross-correlation with other instruments measuring related concepts,
however the direct clinical utility in the WorkCover context is being evaluated further as a part of
the longitudinal component of this study.
Recommendation: Use K10 by GP’s in the acute phase to detect general distress in primary
medical practice. If the K10 score is above 30 it can also be used in the more chronic phases.
The DASS is useful if interpreted by an experienced clinician such as a psychologist. It can
inform the treating practitioner about more specific aspects of the worker’s distress e.g. anxiety
and depression. Currently the clinical utility of PANAS in routine screening is uncertain.
Page 21
Brief screening instrument: applicability of Orëbro
There is widespread interest in using the original Orëbro tool and a range of recent abbreviated versions.
Following our experience of using this tool in combination with other tools, we offer the following
observations about its applicability in this workers’ compensation environment:
- It is a composite tool mainly validated with acute and subacute populations.
- As it provides a “global risk score”, it is not generally used to triage risk groups and it does not assess
the impact of work related barriers therefore stream to appropriate risk intervention. A second step is
probably needed to identify the main risk factors.
- Consequently, additional screening is needed to obtain more detail on the specific risk issues.
- Some major deficiencies of this tool are the lack of questions about the workplace (relationships,
supports, barriers) and that it in not related to return to work in any way. Therefore a number of more
recently developed tools now exist (see Shaw (Shaw, Pransky et al. 2005), StarT tool – (Hill, Dunn et
al. 2010) for use in primary care).
Overall it is a valuable addition to the risk screening toolkit. It is a potentially valuable screening instrument if
used as a first step in combination with five item questions relevant to workplace from the occupational
disability questionnaire (Shaw, Pransky et al. 2005). The second step would involve more comprehensive
risk assessment process, other tests or assessments are needed to identify specific areas requiring
intervention / management.
Guidance on potential risks and potential interventions
A summary of the broader risks, relevance and potential interventions for those workers whose current
status is ‘High impact on activities to daily living and more than two weeks not cleared for work’ is provided in
Table 5. The risks align to the most significant findings (see results section) and are also colour-coded using
the ‘flags’ model to identify the source of the psychosocial risk; ‘yellow’ indicates personal psychosocial,
‘blue’ indicates workplace psychosocial and ‘red’ indicates potential medical co-morbidities.
Page 22
Table 5 Guidance on the key psychosocial issues (personal, work and health) for workers who report high pain levels
Current status
Relevance
Risks
Possible intervention
1
High impact on activities to
daily living and more than
two weeks not cleared for
work
Away from workplace
Deconditioning and likely reaggravation
Coordinated supervised RTW activities
Program by physiotherapist and
psychologist
2
High fear and avoidance
beliefs about work
The avoidance behaviour relates to work
activities not general activity
These beliefs prevent the worker being
active with a high probability of reaggravation
Beliefs addressed and graduated
return to feared activities as 2nd step of
physiotherapy and psychology
program
3
A high pain focus and fear
of negative outcomes
A catastrophising personality that is very
pain-focussed
Fear of pain, symptoms or activity
If the worker exhibits the behaviours 1
to 3, is a high probability of poor
recovery and low RTW rates
Same as for 2
4
High pain self-efficacy and
optimism
Coping ability
Based on a global measure of optimism
about RTW
If high - chance of a good outcome
If less than average, coping will be a
barrier to RTW and is a strong
indicator of worker’s predicted
outcome
Clarify reasons, address potential
RTW issues continue with graduated
program
5
Psychological distress
Overall distress or more specifically the
clinical symptoms of depression or anxiety
(not applicable to injuries with a perceived
traumatic element)
May impact on some outcomes.
Depression early in the claim is likely
to be a pre-existing condition
Clarify extent of problems using GP/
psychologist/psychiatrist; address with
medication or CBT*
6
Work-related issues
blue flags
(i) Supervisor support
(ii) Availability of duties
(iii) Co-worker attitudes
(iv) Work is heavy and monotonous
Sustained RTW is not assured if (i) to
(iii) are negative
Work intervention with help from
RRTW* coordinator or workplace
rehabilitation provider; focus on
suitable duties and support as part of a
structured program to address any
fears identified in 2
7
Health risks factors
Red flags
Potential co-morbidities:
Smoking
Obesity
Lack of exercise
Longer claim duration for those
reporting poor self-reported health or
presence of multiple risk factors
Return to long-term management with
help from GP
* Key: CBT Cognitive behaviour therapy
RRTW rehabilitation and return to work coordinator
Page 23
Conclusion and recommendations
The study again confirmed that most injured workers returned to work and ceased extensive treatment within
three months of injury with only 10-20% who had longer periods of disability. The current evidence-based
guides specify intervention within 10 weeks of injury to prevent poor outcomes. However treatment and
occupational intervention requires additional resources, training and costs which to be cost effective need an
early and accurate identification of those needing this intervention. The study indicated that an early and
brief psychometric screening was feasible and more accurate than routine medical or physiotherapy
assessment at identifying those at risk. Furthermore the study also found that by using comprehensive
assessment tools more specific domains of risk could be identified which can then be targeted with
appropriate treatment.
Therefore, a process is proposed comprised of early risk assessment tools to assist and inform clinical
judgement to significantly improve compensation outcomes as summarised in Figure 2 (page 4). Two
intervention strategies proposed are:
1) Conduct a trial with primary treating clinicians (general practitioners and physiotherapists) at early stages
of injury to improve the current injury management model using a brief screener such as the Orebro.
2) Develop risk-screening involving a two-stage process to help identify subgroups needing specific
interventions. This should involve in addition to the procedures described in 1) a triage process undertaken
by a nominated and trained psychologist to further confirm if the person identified through the brief screening
is at risk and likely to benefit from a targeted treatment. This process will identify the key risk issues in each
domain and provide or facilitate relevant short-term treatment in coordination with the treating doctor and
physiotherapist. This strategy should involve training 10 psychologists, physiotherapists and medical
practitioners in the proposed process and in the standardised treatment. This training would be provided by
leading experts in the field including Clinical Psychologist Dr Michael Nicholas and other relevant
professionals (eg physiotherapists and medical practitioners) with expertise in this field.
© WorkCoverSA, 20
Page 24
Appendix A Psychosocial studies conducted in WorkCoverSA
Psychosocial indicators of chronicity study
In 2001 the project, the Psychosocial indicators of chronicity study PICS) was conducted to assess the
usefulness of initial risk screening on injured workers. The study was a prospective longitudinal cohort of 400
workers with musculoskeletal claims. This study was the first validation study in Australasia using two
instruments: (i) Acute Low Back Pain Questionnaire (ALBPQ)-later renamed Orëbro against workers’
compensation return to work (RTW) data; (ii) the Brief Multi-dimensional Pain Inventory Instrument2 and a
Chronic Pain Grade Instrument (Dixon, Pollard et al. 2007). The results showed that high risk claims could
be identified; however the Orebro tool while able to differentiate general risk was not able to differentiate
between different categories of risk. Therefore additional tools were required to help identify categories of
risk issues likely to be amendable to intervention and management. Additional information required to enable
full risk identification included:

baseline data on risk profiles in relevant risk domains - claim duration, type of claim, work and
compensation issues, individual characteristics, claims history, provider involvement

integrating relevant workplace information

integrating data from the workers’ compensation database with data collected from individuals

including baseline data on factors affecting long term outcomes post injury such as recurrence,
absenteeism, presenteeism
One key recommendation arising from that project was to develop tools and an intervention strategy to help
identify subgroups with shared risk profiles and pilot an intervention to manage key risk areas.
Projects since 2007
Phased approach
Since 2007 we have been taking a phased approach to determine those psychosocial factors that influence
the duration of workers’ compensation back claims within the Australian compensation context. This project
has continued the work started in the PICS project (and incorporating the growing evidence from other
studies in the field) but targeting low back injuries as the first step in the risk identification. As back injuries
are the most prevalent in chronic compensation claims, a number of risk assessment tools were chosen to
make it possible to cross link the strategy with other available initiatives world wide.
The three phase project started with a pilot study to develop and refine the methodology to collect both
quantitative and qualitative data. The second phase, a prospective longitudinal study, is nearing completion
and forms the basis of this report. See the next section for more detail. The third phase discussed in the
“Conclusion and recommendations section of this report will be based on interventions to address the
significant risk factors identified in the longitudinal study.
Prospective study
2
Refer to literature by Turk et al
© WorkCoverSA, 20
Page 25
Objectives
The aims for the prospective longitudinal study were refined during the pilot study and were to:

Assess in more detail those psychosocial factors that influence claim duration in work related back
injuries within the South Australian compensation context.

Identify those risk factors amenable to change.

Determine the feasibility of using a one or two step risk assessment process; that is using brief
clinician administered screening questionnaire or using an additional comprehensive assessment to
further refine a persons specific risk issues.

Validate an existing risk tool(s) or if relevant refine or develop a new tool(s).

Determine the most appropriate person(s) to do the assessment – for example treating clinician,
physiotherapist, case manager, vocational rehabilitation professional or others.

Develop a protocol for early intervention studies targeting key risk areas and assess benefits of risk
intervention.
© WorkCoverSA, 20
Page 26
Appendix B Examples of key conceptual models representing the domains of psychosocial risk.
(a) the “flags” model
Risk domains
Biological
Flag
identifier
Red
Psychiatric
Orange
Psychosocial/
personal
Yellow
Psychosocial/
environmental
Blue
Systemic
Black
Factors
Serious pathology
Co-morbidity
Depression
PTSD
Unhelpful coping strategies
Emotional distress
Passive role in recovery
Overly solicitous carers
Perceived low social support at work
Perceived unpleasant work
Low job satisfaction
Perception of excessive demands
Legislative criteria for compensation
Nature of workplace (eg. heavy work)
Threats to financial security
(b) Simplified representation of the model by Loisel
et al 2005

(c) A matrix of key psychosocial issues incorporating using flags (a) concepts Modified from Loisel et al 2005, Burton and Kendall 2009.
=
© WorkCoverSA, 20
Individual issues
Workplace perception and issues
Coping
Severity of disorder, recurrence, symptoms
Clinical history, comorbidity
Thoughts and beliefs
Mood and emotion
Behaviour and activity
Quality of contact
Availability of suitable duties
Job satisfaction, co-worker support
Management support, culture
Supervisor involvement
Pre-existing performance issues
Health care system- treatment issues.
Systemic issues
Delays in treatment
Non-evidence based treatment
Contradictory treatments
Uncoordinated treatment
Focus exclusively on physical injury
Lack of specialised training in pain related
areas
Compensation and variable legislative
requirements
Financial and legal
Social issues
Financial support, dispute – hardship
Legal issues, proof, claim validity
Reluctance to accept psychosocial issues
Page 27
Appendix C Screening principles
The positive predictive value is the proportion of subjects with a ‘positive’ screening test who do go on to
long-term incapacity - a measure of false negatives. Negative predictive value is the proportion of subjects
with a “negative” screening test who do return to work – a measure of false positive.
Other relevant concepts that impact on the effectiveness of the screening exercise are sensitivity and
specificity.
Sensitivity is the proportion of clients who do go on to long-term incapacity who are correctly predicted by the
screening tool. (It is actual long-term incapacity taking account of false negatives.)
Sensitivity = a / (a + c): in this example, a / (a + c) = 16/20 = 80% sensitivity.
Specificity: the proportion of clients who do not go on to long-term incapacity who are correctly predicted by
the screening tool. (It is actual return to work taking account of false positives).
Specificity = d / (b + d): in this example, d / (b + d) = 56/80=70% specificity.
Economic deadweight: those who receive a work focused intervention but would have returned to work
without the intervention. If a screening tool is used to select clients for the intervention, economic deadweight
is the false positives. Economic deadweight = b / (a + b): in this example, b / (a + b) = 24/40= 60% economic
deadweight.
Actual Outcome
Screening tool
prediction
Long-term
incapacity
Long term incapacity
Return to work
Correct prediction of
incapacity, a
False positive, b
(b=24)
(a=16)
Return to work
False negative, c
(c=4)
© WorkCoverSA, 20
Correct prediction of
return to work, d
(d=56)
Page 28
Appendix D Psychometric instruments used and key domains measured
Screen
Type
Instruments
Dimensions covered
A
Orëbro (various versions)
Functional questions from Orëbro from data from cohort in 2001
collected using full version of Orëbro; also used abbreviated version of
Multi-dimensional Pain Inventory (Turk et al)
A
Shaw Back Disability Risk
Questionnaire
Brief version of Orëbro risk screening tool including questions on
workplace relationships
A
Pain last week - NRS/10
Intensity of pain
A
Current Severity
Self-report “Oswestry q1” on the impact of pain and quality of pain
A
Function-Orëbro /100+
Single item risk screeners from Orëbro on perceived functional
limitations including ‘bothersome pain’, activities of daily living (ADLs)
A
Belief - Pain persists /100
Single items about beliefs
A
Work in 6 months
Expectation of RTW in 6 months
A
Normal Work in 4 weeks/10
Expectation of RTW in 4 weeks
B
Function - Oswestry /100
Questions from Oswestry on perceived functional limitations (ADLS)
B
Pain Disability Index/70
Questions from PDI on perceived functional limitations (ADLs)
B
Tampa
Fear avoidance beliefs about activity and meaning of pain (hurt=harm)
B
Pain Catastrophising Scale
Pain focusing scale (how do you habitually think about pain not just
post-injury) - more a measure of trait
B
Pain Self Efficacy Scale
Confidence in ability to be active despite pain; correlates with PCS
and general function; global expectations of recovery; workplace (blue
flags) and compensation system issues (black flags)
B
Dass- D
Common indicator of depressed mood, not affected by pain related
symptoms
B
Dass-A
Common indicator of anxiety, not affected by pain related symptoms
B
Dass-S
Common indicator of perceived stress (more situational or
environmental); not affected by pain related symptoms
B
K-10
Generic measure of distress, used in population studies and by
general practitioners
A = from a brief screener
B = from a comprehensive measure
© WorkCoverSA, 20
Page 29
Appendix E Who is in the study?
Frequency of claims by age group
Education com pleted
Time with employer
250
35%
35%
30%
30%
200
25%
150
20%
20%
%
Counts of claim
25%
15%
100
15%
10%
10%
50
5%
5%
0%
0
<20
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65+
Primary only
HS to yr 10
HighS y11+
Tafe-4
0%
uni-5
Age groups
w ith Employer < w ith Employer 0
0.5 yrsr yrs
.5-1yr yrs
categories
w ith Employer
2-5 yrs
w ith Employer
5+ yrs
Health indicator - sm oking
Health indicator - exercise
Health indicator BMI
w ith Employer
1-2 yrs
45%
40%
70%
40%
60%
35%
35%
50%
30%
30%
%
20%
15%
25%
40%
20%
30%
%
25%
15%
20%
10%
10%
10%
5%
5%
0%
underweight <20
© WorkCoverSA, 20
healthy 20-25
overweight 25-30
obese 30+
0%
0%
no exercise
rarely
once
2 to 3 w eekly
smokes/day 0
4 or more
w eekly
Page 30
smokes/day <10
smokes/day 1020
smokes/day 20+
Bibliography
Asghari, A. and M. K. Nicholas (2001). "Pain self-efficacy beliefs and pain behaviour. A prospective study." Pain 94(1):
85-100.
Boersma, K. and S. J. Linton (2005). "Screening to identify patients at risk - Profiles of psychological risk factors for early
intervention." Clinical Journal of Pain 21(1): 38-43.
Buchbinder, R. (2008). "Self-management education en masse: effectiveness of the Back Pain: Don't Take It Lying Down
mass media campaign." Medical Journal of Australia 189(10): S29-S32.
Burton, A. K., K. M. Tillotson, et al. (1995). "Psychosocial Predictors of Outcome in Acute and Subchronic Low-Back
Trouble." Spine 20(6): 722-728.
Crawford, J. R. and J. D. Henry (2003). "The depression anxiety stress scales (DASS): Normative data and latent
structure in a large non-clinical sample." British Journal of Clinical Psychology 42: 111-131.
Crawford, J. R. and J. D. Henry (2004). "The Positive and Negative Affect Schedule (PANAS):Construct
validity,measurement properties and normative data in a large non-clinical sample." British Journal of Clinical
Psychology 43(3): 245-265.
Deyo, R. A., M. Battie, et al. (1998). "Outcome measures for low back pain research - A proposal for standardized use."
Spine 23(18): 2003-2013.
Dixon, D., B. Pollard, et al. (2007). "What does the chronic pain grade questionnaire measure?" Pain 130(3): 249-253.
Dunn, K. M. (2010). "Extending conceptual frameworks: life course epidemiology for the study of back pain." Bmc
Musculoskeletal Disorders 11: 11.
Dunn, K. M. and P. R. Croft (2005). "Classification of Low Back Pain in Primary Care: Using "Bothersomeness" to Identify
the Most Severe Cases. [Miscellaneous Article]." Spine 30(16): 1887-1892.
Dworkin, R. H., D. C. Turk, et al. (2005). "Core outcome measures for chronic pain clinical trials: IMMPACT
recommendations." Pain 113(1-2): 9-19.
Fairbank, J. C. T. M. D. and P. B. Pynsent (2000). "The Oswestry Disability Index. ." Spine 25(22): 2940-2953.
Foster, N. E., K. S. Dziedzic, et al. (2009). "Research priorities for non-pharmacological therapies for common
musculoskeletal problems: nationally and internationally agreed recommendations." Bmc Musculoskeletal
Disorders 10: 10.
Foster, N. E., J. C. Hill, et al. (2011). "Subgrouping patients with low back pain in primary care: Are we getting any better
at it?" Manual Therapy 16(1): 3-8.
Furukawa, T. A., R. C. Kessler, et al. (2003). "The performance of the K6 and K10 screening scales for psychological
distress in the Australian National Survey of Mental Health and Well-Being." Psychological Medicine 33(2): 357362.
Gatchel, R. J., Y. B. Peng, et al. (2007). "The biopsychosocial approach to chronic pain: Scientific advances and future
directions." Psychological Bulletin 133(4): 581-624.
Gronblad, M. M. D. P. D., E. M. S. Jarvinen, et al. (1994). "Relationship of the Pain Disability Index (PDI) and the
Oswestry Disability Questionnaire (ODQ) with Three Dynamic Physical Tests in a Group of Patients with Chronic
Low-Back and Leg Pain." Clinical Journal of Pain 10(3): 197-203.
Hazard, R. G., L. D. Haugh, et al. (1996). "Early prediction of chronic disability after occupational low back injury." Spine
21(8): 945-951.
Hill, J. C., K. M. Dunn, et al. (2008). "A primary care back pain screening tool: Identifying patient subgroups for initial
treatment." Arthritis Care & Research 59(5): 632-641.
Hill, J. C., K. M. Dunn, et al. (2010). "Subgrouping low back pain: A comparison of the STarT Back Tool with the Örebro
Musculoskeletal Pain Screening Questionnaire." European Journal of Pain 14(1): 83-89.
Kaplan, S., J. C. Bradley, et al. (2009). "On the role of positive and negative affectivity in job performance: A metaanalytic investigation." Journal of Applied Psychology 94(1): 162-176.
Kendall, N. A. S. (1999). "Psychosocial approaches to the prevention of chronic pain: the low back paradigm." Best
Practice & Research in Clinical Rheumatology 13(3): 545-554.
Kendall, N. A. S., A. K. Burton, et al. (2009). Tacking Musculoskeletal Problems: a guide for clinic and workplace. London,
The Stationary Office.
© WorkCoverSA, 20
Page 31
Kessler, R. C., G. Andrews, et al. (2002). "Short screening scales to monitor population prevalences and trends in nonspecific psychological distress." Psychological Medicine 32(6): 959-976.
Kori, S. H., R. P. Miller, et al. (1990). "Kinesiohobia: A new view of chronic pain behavior." Pain management.(January):
35-43.
Kuijer, W., S. Brouwer, et al. (2006). "Work status and chronic low back pain: exploring the International Classification of
Functioning, Disability and Health." Disability and Rehabilitation 28(6): 379-388.
Lakke, S. E., R. Soer, et al. (2009). "Risk and prognostic factors for non-specific musculoskeletal pain: A synthesis of
evidence from systematic reviews classified into ICF dimensions." Pain 147(1-3): 153-164.
Linton, S. and H. K (1996). "Risk factors and the natural course of acute and recurrent musculoskeletal pain: developing a
screening instrument." 8th World Congress on Pain: 1996.
Linton, S., M. Nicholas, et al. (2011). "Development of a Short Form of the Örebro Musculoskeletal Pain
Screening Questionnaire." Spine DOI: 10.1097/BRS.0b013e3181f8f775 1605-1613.
Linton, S. J. (2001). "Occupational psychological factors increase the risk for back pain: A systematic review." JOURNAL
OF OCCUPATIONAL REHABILITATION 11(1): 53-66.
Linton, S. J., D. Gross, et al. (2005). "Prognosis and the identification of workers risking disability: Research issues and
directions for future research." Journal of Occupational Rehabilitation 15(4): 459-474.
Linton, S. J. and K. Hallden (1998). "Can we screen for problematic back pain? A screening questionnaire for predicting
outcome in acute and subacute back pain." Clinical Journal of Pain 14(3): 209-215.
Loisel, P., R. Buchbinder, et al. (2005). "Prevention of work disability due to musculoskeletal disorders: The challenge of
implementing evidence." Journal of Occupational Rehabilitation 15(4): 507-524.
Loisel, P., M. J. Durand, et al. (2003). "From evidence to community practice in work rehabilitation: The Quebec
experience." Clinical Journal of Pain 19(2): 105-113.
Loisel, P., Q. N. Hong, et al. (2009). "The Work Disability Prevention CIHR Strategic Training Program: Program
Performance After 5 Years of Implementation." JOURNAL OF OCCUPATIONAL REHABILITATION 19(1): 1-7.
Lovibond, P. F. and S. H. Lovibond (1995). "The structure of negative emotional states: comparison of the depression
anxiety stress scales (DASS) with the beck depression and anxiety inventories." Behaviour Research and
Therapy 33: 335-343.
Main, C. J., N. Foster, et al. (2010). "How important are back pain beliefs and expectations for satisfactory recovery from
back pain?" Best Practice & Research in Clinical Rheumatology 24(2): 205-217.
Main, C. J., M. J. L. Sullivan, et al. (2008). Pain management. London, Elsevier.
National Health and Medical Research Council (Australia) (1999). Acute pain management : information for general
practitioners. Canberra, NHMRC.
Nicholas, M. K. (2010). "Obstacles to recovery after an episode of low back pain; the [`]usual suspects' are not always
guilty." Pain 148(3): 363-364.
Nicholas, M. K., A. Asghari, et al. (2008). "What do the numbers mean Normative data in chronic pain measures." Pain
134(1-2): 158-173.
Roe, C., U. Sveen, et al. (2009). "CONSTRUCT DIMENSIONALITY AND PROPERTIES OF THE CATEGORIES IN THE
ICF CORE SET FOR LOW BACK PAIN." Journal of Rehabilitation Medicine 41(6): 429-437.
Schultz, I., J. Crook, et al. (2005). "Predicting return to work after low back injury using the Psychosocial Risk for
Occupational Disability Instrument: a validation study." J Occup Rehabil 15(3): 365 - 376.
Schultz, I., J. Crook, et al. (2008). "A Prospective Study of the Effectiveness of Early Intervention with High-risk Backinjured Workers—A Pilot Study." Journal of Occupational Rehabilitation 18(2): 140-151.
Schultz, I. Z., J. Crook, et al. (2000). "Models of Diagnosis and Rehabilitation in Musculoskeletal Pain-Related
Occupational Disability." Journal of Occupational Rehabilitation 10(4): 271-293.
Shaw, W. S., G. Pransky, et al. (2005). "Early Disability Risk Factors for Low Back Pain Assessed at Outpatient
Occupational Health Clinics." SPINE 30(5): 572-580.
Shaw, W. S., D. A. van der Windt, et al. (2009). "Early Patient Screening and Intervention to Address Individual-Level
Occupational Factors ("Blue Flags") in Back Disability." JOURNAL OF OCCUPATIONAL REHABILITATION
19(1): 64-80.
Sullivan, M., B. SR, et al. (1995). "The Pain Catastrophizing Scale: development and validation." Psychol Assessment(7):
524 - 532.
© WorkCoverSA, 20
Page 32
Turk, D. C., R. H. Dworkin, et al. (2008). "Identifying important outcome domains for chronic pain clinical trials: An
IMMPACT survey of people with pain." Pain 137(2): 276-285.
Turk, D. C. and A. Okifuji (2002). "Psychological Factors in Chronic Pain : Evolution and Revolution." Journal of
Consulting and Clinical Psychology 70(3): 678-690.
Waddell, G., A. Burton, et al. (2003). "Screening to identify people at risk of long-term incapacity for work."
Waddell, G., M. Newton, et al. (1993). "A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance
beliefs in chronic low back pain and disability." Pain 52(2): 157 - 168.
Watson, D., L. A. Clark, et al. (1988). "Development and validation of brief measures of positive and negative affect: The
PANAS scales." Journal of Personality and Social Psychology 54(6): 1063-1070.
© WorkCoverSA, 20
Page 33
Download