Merit Pay Programs

Predictors of Sustained Return to Work After Work-Related Injury or Disease: Insights from Workers’ Compensation Claims Records

Janneke Berecki-Gisolf • Fiona J. Clay •

Alex Collie • Roderick J. McClure

Published online: 6 December 2011

� Springer Science+Business Media, LLC 2011

Abstract Aim After work-related injury or disease, mul-

tiple spells of work absences and unsuccessful return to work

(RTW) are common. The purpose of this study was to

identify predictors of sustained RTW and work disability

recurrences. Methods Australian WorkSafe Victoria claims

containing income compensation payments starting between

January 1st, 2001 and December 31st, 2004 (n = 59,526)

were analysed over a 2-year observation window. Time until

first RTW and final RTW, and ‘recurrences’ (cessations of

payments of [7 days), were derived from claims payments data. Regression models were used relating demographic,

occupational, workplace and injury characteristics to RTW

outcomes. Results Although 94% of claimants had at least

one RTW, only 79% achieved sustained RTW during fol-

low-up. Median time until first RTW was 50 days; median

time until final RTW was 91 days. Independent predictors of

delayed final RTW were older age, afflictions involving the

neck or multiple locations, and working in manufacturing.

Of those who returned to work, 37% had at least one

recurrence: risk factors were ages 35–55, female sex,

working as a labourer, working in manufacturing, traumatic

joint/ligament or muscle/tendon injury and musculoskeletal

and connective tissue diseases, and afflictions involving the

neck or multiple locations. Conclusions Work disability

recurrences are common and have considerable impact on

sustained RTW outcomes. A policy focus on education

about secondary prevention may help improve long-term

RTW outcomes, particularly for persons with musculo-

skeletal disorders and those working in manufacturing.

Keywords Work disability � Recurrence � Compensable injury � Lost time claims

Introduction

Of the 10.8 million Australians who worked at some time

between June 2005 and June 2006, 690,000 (6.4%) expe-

rienced at least one work-related injury or illness [1]. More

than half of work-related injuries involved time away from

work: 21% of injured workers missed 1–4 days of work,

28% missed five or more days and a very small group (2%)

did not return to work (RTW) after the illness or injury

occurred [1].

There are a range of RTW outcome measures, which

generally serve several functions, as listed by Krause et al.

[2]: as indicator of the overall burden of occupational

injury on society; to determine the burden on specific

groups such as injured workers, their families, employers,

or industries; to evaluate the effectiveness and efficacy of

intervention programs and policies to help injured workers

RTW; and to measure the impact of interventions on

societal costs of injuries and illnesses [2]. Work outcome

measures are continually evolving to meet the needs of

employees, employers and other stakeholders, and should

continue to do so [3].

‘Time to first return’ has been found to seriously

underestimate the total duration of work disability when

compared to other outcome measures [2]. In a study among

workers claiming income compensation for lower back

pain in California, eleven outcome measures categorised

J. Berecki-Gisolf (&) � R. J. McClure Monash Injury Research Institute, Monash University,

Building 70 Clayton Campus, Melbourne, VIC 3800, Australia

e-mail: [email protected]

F. J. Clay � A. Collie Institute for Safety, Compensation and Recovery Research,

Monash University, Melbourne, VIC, Australia

123

J Occup Rehabil (2012) 22:283–291

DOI 10.1007/s10926-011-9344-y

under 1. ‘calendar time event’ outcomes; 2. ‘cumulative

time’ outcomes and 3. ‘point prevalence’ outcomes were

compared. Allowing gaps of 7 or less days in the estimate

of ‘time to first return’ did little to reduce the underesti-

mation: more than 20% of claimants experienced multiple

temporary disability episodes during 1–3.5 years of follow-

up [2]. A study among Ontario workers with permanent

partial impairments due to injuries reported that as much as

60% of those who returned to work had one or more sub-

sequent injury-related work absences [4].

There appears to be a substantial difference between first

RTW and sustained RTW. Multiple spells of work absen-

ces and unsuccessful RTW are common, especially among

workers with chronic conditions. Identifying characteristics

that influence post-injury employment patterns will assist

the development of disability management strategies [4].

The purpose of this study was to identify predictors of

sustained RTW, using a set of RTW outcome measures

derived from workers’ compensation claims data. The

particular focus was on the differences between first RTW

and final RTW, and the reason for the discrepancy between

the two: relapses. Identification of risk factors for relapses

is an important step in removing barriers to successful

RTW.

Methods

Sample

WorkSafe Victoria provides workers’ compensation

insurance for employers. Following a workplace injury or

disease, claimants may be eligible to receive income

replacement, and medical and rehabilitation expenses,

regardless of who was at fault. WorkSafe Victoria pays

weekly compensation until the injured or sick worker can

return to pre-injury work. For the first 13 weeks, the

weekly entitlement is 95% of pre-injury average weekly

earnings (PIAWE); from 14 to 130 weeks, 80% of PIAWE.

After 130 weeks, a weekly entitlement of 80% of PIAWE

is only paid if the claimant is still not capable of working,

as a consequence of the injury.

For this study, de-identified WorkSafe claims data and

compensation payment data were analysed to identify fac-

tors associated with successful and unsuccessful RTW. Non-

fatality claims including income compensation payments

(i.e. lost time claims) with a first weekly payment dating

between January 1st, 2001 and December 31st, 2004 were

selected (n = 70,761 claims). These claims were made by

66,081 claimants: after removing claims made by claimants

with multiple claims within the data collection and follow-

up period, 59,526 claims remained (Fig. 1). Claims with less

than 10 days of compensation are usually covered by the

employer; this analysis is only relevant to absenteeism of

10 days or more. The number of days of compensation paid

by the employer preceding the claim was included in the

analysis; for 95% of claimants, this amounted to 10 days.

These claims were followed for 2 years, i.e. payments

referring to this claim dating more than 2 years after the first

income compensation payment were deleted, and the claim

was marked as censored. Censoring at 2 years avoids the

steep drop in payments expected around 130 weeks, which

can refer to either RTW, or to a cease in payments due to

compensation scheme policy. A schematic representation of

data preparation is given in Fig. 1.

Defining Return to Work Outcomes

Cessation of income compensation payments was assumed

to reflect a RTW; this was, however, not verified. Although

this will generally be the case, some payment cessations

could be due to activities other than RTW, such as study

commencement, or unemployment benefits.

First full RTW was defined as the first gap in payments of

[7 days; time until the first RTW was the number of days from the start of the first weekly income compensation until

the start of the first [7 day gap in payments; this is an adaptation of the method described by Oleinick et al. [5].

Partial RTW was not taken into account; the focus of this

analysis was on full RTW. Final full RTW was defined as the

All claims data 1,832,814

entries

All payment data 27,323,661

entries Only payments

referring to compensation

are kept 9,768,822

entries

Payments started 2yrs+

after first payment are

deleted 5,934,811

entries Claims by those with more than 1

claim are removed 59,526

claimants

Claimants with more than one

claim 6555 claimants

One entry per claim_ID

187,953 claims

Fatalities (n=30) are removed

70,761 claims

Claims and payments merged by claim ID

Claims starting before 1-1-1 and after 31-12-4 are

deleted 70,791 claims

One entry per claimant_ID

66081 claimants

Fig. 1 Schematic representation of the data preparation

284 J Occup Rehabil (2012) 22:283–291

123

last income compensation payment. Claimants who received

any income compensation payment in the last 6 months of

the 2-year follow-up were considered not to have achieved a

final full RTW and the claim was marked as censored at the

date of the last payment. Failed attempts at RTW were called

relapses, and were defined as gaps in payment of [7 days, followed by a resumption of payments.

Terminology: Relapse Versus Recurrence Versus

Failed Attempt at RTW

In this paper the use of the term ‘relapses’ requires some

clarification. A cessation of income compensation payments

of [7 days followed by a re-starting of payments under the same claim ID is interpreted as a ‘relapse’; however, we

cannot certify whether this is due to a recurrence of the

injury or disease, recurrence of pain, or other reasons.

‘Relapse’ as reported here therefore refers to a work dis-

ability recurrence within the same claim, without specifi-

cation of the underlying reason. Although we have used this

synonymously with ‘failed RTW’ throughout this paper,

this is a matter of interpretation. Even if we could be certain

that every gap in payments exceeding 7 days corresponded

with a RTW, a starting-and-stopping pattern of RTW is not

necessarily a ‘failure’, unless it is perceived by the worker

as such, and discourages future RTW attempts. Alterna-

tively, a number of ‘short trials’ of returning to work may be

a process gradually leading to sustained RTW without any

negative connotations for the worker, employer or insurer.

Variables

Occupational, workplace and injury/disease as well as

demographic details were derived from the claims data.

Age refers to claimant age at the time of injury/disease.

Pre-injury gross weekly income refers to the total weekly

earnings prior to illness or injury. Those categorised as

‘zero’ pre-injury gross weekly income were working under

an arrangement other than regular hours per week: their

actual estimated weekly earnings were $591 [pp 25–75:

$480–751]. Occupation is categorised into nine occupation

group codes using the Australian and New Zealand Stan-

dard Classification of Occupations (ANZSCO) [6]. Indus-

try is categorised according to the Australian and New

Zealand Standard Industrial Classification (ANZSIC) [7].

Bodily location of the injury/disease refers to the location

of the first affliction (i.e. head, neck, trunk, etc.). Nature of

affliction refers to the type of injury or disease, in 18 cat-

egories (7 for injury, 10 for disease and 1 for other). Both

bodily location and affliction nature are categorised

according to the National Occupational Health & Safety

Commission’s Type of Occurrence Classification System

(TOOCS) version 3 [8].

Statistical Analysis

The data analysis for this paper was generated using SAS

software, Version 9.2 [9].

First RTW and Final RTW

General survival functions describing first and final RTW,

stratified by injury versus disease, were computed using the

product limit method (SAS LIFETEST procedure); a gap in

payments was defined as an event. To assess the association

between potential predictors and time to final RTW, a Cox

model was constructed (SAS PHREG procedure). This

model contained all listed covariates and was stratified by

sex. Time dependent variables were introduced in the model

and were statistically significant (and clinically meaning-

ful). We therefore decided to split the follow-up period into

2 segments: B13 weeks, and [13 weeks–2 years, based on the drop in payments from 95 to 80% of pre-injury income

at 13 weeks. Separate Cox models were constructed for the

two time periods (left-censoring, not truncation, was

applied in the model predicting final RTW after 13 weeks).

Relapses

Among those who attempted a RTW, a histogram of the

number of relapses was computed (SAS UNIVARIATE

procedure). The association between potential predictors

and failed attempts at RTW was then assessed using binary

logistic regression modelling (SAS LOGISTIC procedure).

Fully adjusted models of having at least one relapse during

the follow up period were constructed; only claimants who

attempted to RTW were included.

Repeat Claimants

Repeat claimants were excluded from the RTW analysis. In

analysing the RTW process of a worker with multiple

claims, a cessation in compensation payments may be due

to RTW or due to a new claim starting. In other words, for

repeat claimers, compensation payments cannot be ana-

lysed for the full duration of the follow-up. Claimants with

repeat claims have been reported to differ from those with

single claims [10]; the RTW results presented here are

limited to claimants with a single claim.

Results

Of the 59,526 claims included in the analysis, 34,817

(58%) were due to injury and 24,708 (42%) were due to

disease (and 1 unknown). Men were in the majority with

39,644 (67%) men and 19,881 (33%) women.

J Occup Rehabil (2012) 22:283–291 285

123

First RTW

The median time until first RTW was 50 days. Of the

injured claimants, 1,925/34,817 (5.5%) did not attempt a

full RTW during the 2-year follow-up; median time until a

first full RTW was 46 days [95% CI 45–47]. Of the

claimants with a disease, 1,760/24,708 (7.1%) did not

attempt a full RTW; median time until RTW was 56 days

[95% CI 55–57]. A Kaplan–Meier plot for the duration of

income compensation until a cessation in compensation

payments of [7 days is shown in Fig. 2.

Final RTW

The median time until final RTW was 91 days. Of the

injured claimants, 6,979/34,817 (20%) did not fully RTW

during the 2-year follow-up; the median time until a full

RTW was 77 days [95% CI 75–78]. Of the claimants with

a disease, 5,767/24,708 (23%) did not fully RTW; median

time until full RTW was 117 days [95% CI 114–121].

Figure 2 shows a Kaplan–Meier plot for the duration of

income compensation until the end of the last payment,

which was interpreted as a final RTW; those with any

income compensation payment in the last 6 months of data

collection were censored. Fully adjusted Cox model results

of time until final RTW are shown in Table 1.

During the first 13 weeks, the time until a final RTW

was increased for those aged [35 years. Time to final RTW did not clearly differ by pre-injury income, except

for the lowest income range ($1–$250 per week) which

was associated with a relatively rapid final RTW. Of the

various occupational groups, labourers, elementary cleri-

cal, sales and service workers and intermediate production

and transport workers had a relatively delayed final RTW.

The industries construction and manufacturing were asso-

ciated with a delayed final RTW. Final RTW differed a

great deal per affliction nature: spinal cord injuries and

intracranial injuries were associated with the most delayed

final RTW, whereas burns and digestive system diseases

were associated with rapid final RTW. Of the bodily

locations, a relatively delayed final RTW was seen in those

with afflictions involving the neck or multiple locations;

afflictions involving the upper and lower limbs were

associated with a rapid final RTW.

Between 13 weeks and 2 years, predictors of final RTW

were largely similar to predictors of final RTW in the first

Fig. 2 Kaplan-Meier plots for the duration of income

compensation until a cessation

in compensation payments of

[7 days, interpreted as a first return to work (top); or until a final payments stop, interpreted

as a final RTW (bottom). Data were stratified by nature of the

affliction, grouped as injury

versus disease (left) or by sex (right)

286 J Occup Rehabil (2012) 22:283–291

123

Table 1 Columns 3 and 4 a : Potential predictors of time to sustained

return to work: results of separate Cox models for the first 13 weeks

(early effects) and [13 weeks–2 years (late effects). Column 5b:

Work disability recurrences (‘relapses’). Results of logistic regression

modelling of having at least one relapse during the follow-up period,

among those who returned to work (N = 19,650/52,577)

Variable Final return to work Relapses

Early effects

First 13 weeks

Late effects

14 weeks to 2 years

C1 Relapses

N (%) HR [95% CI] HR [95% CI] OR [95% CI]

Age

\25 7,099 (12) 1.34 [1.29–1.39] 1.30 [1.23–1.38] 0.76 [0.71–0.82] 25–34 12,888 (22) 1 REF 1 REF 1 REF

35-44 16,008 (27) 0.87 [0.84–0.90] 0.82 [0.78–0.86] 1.18 [1.12–1.24]

45–54 15,991 (27) 0.79 [0.77–0.82] 0.78 [0.74–0.81] 1.22 [1.16–1.29]

C55 7,528 (13) 0.77 [0.74–0.80] 0.83 [0.79–0.87] 1.09 [1.02–1.16]

Gender

Male 39,644 (67) 0.65 [0.62–0.69]

Female 19,881 (33) 1 REF

Gross weekly income (pre-injury)

0 c

11,185 (19) 0.97 [0.94–1.00] 1.12 [1.08–1.17] 0.94 [0.89–0.99]

1–250 2,256 (4) 1.21 [1.14–1.28] 1.16 [1.07–1.27] 0.69 [0.62–0.76]

251–500 11,694 (20) 1.03 [1.00–1.07] 1.02 [0.97–1.06] 0.90 [0.86–0.95]

501–750 21,298 (36) 1 REF 1 REF 1 REF

751–1,000 8,782 (15) 0.99 [0.95–1.03] 1.00 [0.95–1.05] 1.03 [0.97–1.09]

1,001? 4,299 (7) 1.04 [0.99–1.09] 0.96 [0.89–1.03] 0.88 [0.81–0.96]

Occupation group

Managers and administrators 1,551 (3) 0.95 [0.87–1.04] 1.09 [0.98–1.21] 0.97 [0.85–1.11]

Professionals 6,135 (10) 1.02 [0.96–1.08] 1.09 [1.02–1.17] 0.96 [0.89–1.05]

Associate professionals 4,835 (8) 1.05 [0.99–1.12] 1.01 [0.93–1.09] 0.88 [0.80–0.96]

Tradespersons and related workers 12,451 (21) 1.01 [0.96–1.07] 0.96 [0.89–1.03] 0.95 [0.88–1.04]

Advanced clerical and service workers 692 (1) 1.07 [0.95–1.22] 1.20 [1.05–1.38] 0.80 [0.67–0.95]

Intermediate clerical, sales and service

workers

5,643 (9) 1 REF 1 REF 1 REF

Intermediate production and transport

workers

11,387 (19) 0.89 [0.84–0.94] 0.98 [0.92–1.06] 1.05 [0.97–1.15]

Elementary clerical, sales and service

workers

2,532 (4) 0.92 [0.86–0.99] 1.03 [0.94–1.12] 1.05 [0.94–1.17]

Labourers and related workers 14,288 (24) 0.85 [0.81–0.90] 0.92 [0.86–0.98] 1.12 [1.04–1.22]

Industry group (ANZSIC)

A—Agriculture, forestry and fishing 1,812 (3) 1.17 [1.09–1.26] 1.02 [0.91–1.13] 0.76 [0.67–0.87]

B—Mining 144 (0.3) 1.12 [0.89–1.40] 0.63 [0.43–0.93] 0.96 [0.64–1.43]

C—Manufacturing 13,878 (24) 1.01 [0.97–1.06] 0.86 [0.80–0.91] 1.50 [1.39–1.62]

D—Electricity, gas, water and waste

services

513 (1) 1.14 [1.00–1.29 1.01 [0.86–1.20] 1.41 [1.15–1.72]

E—Construction 5,766 (10) 1 REF 1 REF 1 REF

F—Wholesale trade 3,461 (6) 1.14 [1.07–1.21] 0.92 [0.85–1.01] 1.36 [1.23–1.50]

G—Retail trade 3,151 (5) 1.18 [1.11–1.26] 0.99 [0.90–1.08] 1.03 [0.93–1.15]

H—Accommodation and food services 2,056 (4) 1.09 [1.01–1.18] 1.03 [0.93–1.14] 1.09 [0.97–1.24]

I—Transport, postal and warehousing 4,011 (7) 1.19 [1.12–1.26] 1.04 [0.96–1.13] 1.09 [0.98–1.20]

J—Information media and

telecommunications

316 (1) 1.06 [0.89–1.27] 1.11 [0.92–1.35] 1.47 [1.15–1.89]

J Occup Rehabil (2012) 22:283–291 287

123

Table 1 continued

Variable Final return to work Relapses

Early effects

First 13 weeks

Late effects

14 weeks to 2 years

C1 Relapses

N (%) HR [95% CI] HR [95% CI] OR [95% CI]

K—Financial and insurance services 291 (1) 1.19 [0.98–1.43] 1.09 [0.89–1.34] 0.95 [0.72–1.24]

L—Rental, hiring and real estate services 535 (1) 1.13 [1.00–1.29] 0.76 [0.64–0.91] 1.35 [1.11–1.64]

M—Professional, scientific and technical

services

1,129 (2) 1.16 [1.05–1.27] 1.03 [0.92–1.17] 1.16 [1.00–1.34]

N—Administrative and support services 4,516 (8) 1.36 [1.29–1.43] 0.95 [0.87–1.03] 0.94 [0.85–1.03]

O—Public admin. and safety 2,842 (5) 1.13 [1.05–1.22] 0.93 [0.85–1.03] 1.42 [1.27–1.59]

P—Education and training 2,822 (5) 1.42 [1.32–1.53] 1.08 [0.98–1.19] 0.96 [0.85–1.08]

Q—Health care and social assist. 7,329 (13) 1.33 [1.25–1.41] 1.14 [1.05–1.24] 1.18 [1.08–1.30]

R—Arts and recreation services 1,210 (2) 1.31 [1.21–1.43] 1.06 [0.94–1.21] 0.97 [0.84–1.13]

S—Other services 1,687 (3) 1.09 [1.01–1.18] 0.96 [0.87–1.07] 1.11 [0.98–1.27]

Affliction(TOOCSv3)

Injury: Intracranial injuries 221 (0.4) 0.77 [0.61–0.96] 0.94 [0.70–1.27] 0.87 [0.63–1.21]

Injury: Fracture 6,190 (10) 1.18 [1.14–1.23] 1.33 [1.25–1.40] 0.61 [0.57–0.65]

Injury: Wounds, lacerations, amputations

and internal organ damage

8,279 (14) 1.81 [1.75–1.88] 1.22 [1.15–1.29] 0.48 [0.45–0.51]

Injury: Burn 490 (1) 2.74 [2.47–3.04] 1.31 [0.98–1.75] 0.29 [0.22–0.38]

Injury: Injury to nerves and spinal cord 44 (0.07) 0.38 [0.19–0.75] 0.74 [0.44–1.26] 0.94 [0.48–1.86]

Injury: Traumatic joint/ligament and

muscle/tendon injury

18,701 (31) 1 REF 1 REF 1 REF

Injury: Other injuries 886 (1) 1.51 [1.36–1.66] 0.95 [0.81–1.10] 0.65 [0.55–0.77]

Disease: Musculoskeletal and connective

tissue diseases

14,839 (25) 0.96 [0.93–0.99] 1.11 [1.07–1.16] 1.00 [0.96–1.05]

Disease: Mental diseases 5,292 (9) 0.91 [0.62–1.31] 0.84 [0.55–1.27] 0.61 [0.33–1.14]

Disease: Digestive system diseases 2,505 (4) 2.66 [2.52–2.81] 3.25 [2.90–3.64] 0.19 [0.16–0.21]

Disease: Skin and subcutaneous tissue

diseases

283 (0.48) 1.52 [1.30–1.78] 1.27 [0.99–1.62] 0.68 [0.52–0.89]

Disease: Nervous system and sense organ

diseases

1,146 (2) 0.92 [0.84–1.01] 1.34 [1.21–1.48] 0.84 [0.73–0.95]

Disease: Respiratory system diseases 183 (0.3) 0.91 [0.69–1.20] 0.81 [0.58–1.13] 0.79 [0.54–1.18]

Disease: Circulatory system diseases 213 (0.4) 0.85 [0.68–1.07] 1.02 [0.78–1.34] 0.71 [0.51–0.99]

Disease: Infectious and parasitic diseases 130 (0.2) 1.27 [0.95–1.71] 0.77 [0.50–1.20] 0.64 [0.41–1.00]

Disease: Neoplasms 11 (0.02) 0.57 [0.18–1.78] 0.38 [0.09–1.50] 1.57 [0.41–5.93]

Disease: Other diseases 67 (0.1) 1.20 [0.84–1.73] 0.89 [0.52–1.51] 0.71 [0.41–1.25]

Disease: Other claims 34 (0.1) 1.33 [0.80–2.19] 0.78 [0.36–1.68] 0.94 [0.43–2.06]

Bodily location of injury/disease

Head 1,449 (2) 1.07 [0.99–1.15] 0.63 [0.55–0.73] 1.22 [1.06–1.41]

Lower limbs 11,228 (19) 1.08 [1.04–1.11] 1.08 [1.03–1.13] 0.86 [0.81–0.90]

Mental 5,308 (9) 0.68 [0.47–0.99] 1.26 [0.83–1.90] 1.18 [0.63–2.19]

Multiple locations 801 (1) 0.57 [0.50–0.64] 0.73 [0.64–0.84] 1.49 [1.27–1.76]

Neck 2,366 (4) 0.75 [0.70–0.81] 0.77 [0.71–0.84] 1.32 [1.20–1.45]

Systemic 311 (1) 0.83 [0.65–1.05] 0.99 [0.74–1.34] 1.04 [0.73–1.47]

Trunk 18,642 (31) 0.86 [0.83–0.89] 0.71 [0.68–0.74] 1.21 [1.15–1.27]

Unspecified 35 (0) 0.56 [0.33–0.97] 0.73 [0.37–1.44] 1.20 [0.55–2.60]

Upper limbs 19,374 (33) 1 REF 1 REF 1 REF

288 J Occup Rehabil (2012) 22:283–291

123

13 weeks, with several noteworthy exceptions. The occu-

pational groups ‘professionals’ and ‘advanced clerical and

service workers’ were statistically significantly associated

with a rapid RTW after, but not before, 13 weeks. The

industries administrative and support services; education

and training; and arts and recreation services, which were

associated with rapid final RTW in the first 13 weeks, were

no longer statistically significantly different from the ref-

erence category (construction) after 13 weeks. Manufac-

turing did significantly worse than construction in terms of

RTW after, but not before 13 weeks.

Pre-Injury Income

Pre-injury gross weekly income as an explanatory variable

in the models described above led to results that merit

further exploration; particularly for the categories ‘0’ and

‘1–250’. Those in the ‘zero’ pre-injury gross weekly

income category were workers without recorded pre-injury

weekly working hours; their actual estimated average

weekly income was $591 [pp 25–75: 580–751]. There were

no striking differences between ‘zero’ income workers and

other income groups, except for having a higher proportion

of injury as opposed to disease (63% injury vs. 58% injury

among other claimants; difference = 5%[95% CI 4–6]).

The ‘1–250’ weekly income group was relatively small and

differed from other claimants in several respects: claimants

in this group were more likely to be female (59 vs. 32%),

under 30 years of age (37 vs. 21%), and to work in

administrative and support services (19 vs. 7%). Afflictions

were more likely to be an injury (68 vs. 58%) and to

involve the upper limbs (40 vs. 32%).

Failed RTW

Relapses were common with 20,934 (37%) claimants

having at least one failed attempt at RTW during follow-

up. A histogram of the distribution of the number of

relapses among injury- and disease related claims is shown

in Fig. 3. Of those with a relapse, the median duration of

the first attempted RTW (as determined from gaps of more

than 7 days in compensation payments) was 26 days [pp

25–75: 14–53]. The results of fully adjusted logistic

regression models of relapse (one or more vs. none) are

shown in Table 1. Most at risk for relapses were workers

aged between 35 and 55; women; workers in the middle

income range; labourers and related workers; persons

working in the industry groups manufacturing, electricity,

gas, water and waste services; information media and

Number of relapses 0 1 2 3 4 5 6 7 8 9 10

% o

f c

la im

s

0

10

20

30

40

50

60

70

Injury Disease

Fig. 3 Histogram of the distribution of the number of work disability recurrences among injury- and disease related claims

Table 1 continued

Variable Final return to work Relapses

Early effects

First 13 weeks

Late effects

14 weeks to 2 years

C1 Relapses

N (%) HR [95% CI] HR [95% CI] OR [95% CI]

First return to work

Early (B13 weeks) 40,541 (68) 0.50 [0.48–0.52]

Late ([13 weeks) 15,300 (26) 1 REF No attempted RTW 3,685 (6)

Of note, in the Cox models, HRs exceeding unity imply a rapid final RTW compared to the reference group (thus a favourable outcome), whereas

in the logistic regression model an OR exceeding unity implies more relapsing compared to the reference group (an unfavourable outcome)

HR hazards ratio, CI confidence interval, OR odds ratio, ANZSIC Australian and New Zealand standard industrial classification [7], TOOCS v3 National Occupational health and safety commission’s type of occurrence classification system version 3 [8] a

Models are adjusted for all listed variables and stratified by gender b

Final models are adjusted for all listed variables c

Claimants categorised as having ‘zero’ pre-injury gross weekly income were working under an arrangement other than regular hours per week:

their actual estimated weekly earnings were $591 [pp 25–75: 480–751]

J Occup Rehabil (2012) 22:283–291 289

123

telecommunications; and public administration and safety;

workers with traumatic joint/ligament and muscle/tendon

injury or musculoskeletal and connective tissue disease;

and those with afflictions affecting the neck, or multiple

locations. Late ([13 weeks) first RTW was also a risk factor for relapses. Least at risk for relapses were persons

under 25 years of age; low income earners (weekly income

of $1–250); advanced clerical and service workers; persons

working in agriculture, forestry and fishing; and workers

with fractures, wounds, burns or digestive system diseases.

Discussion

In this study we found that time to sustained RTW was 1.8

times longer than the time until first RTW: work disability

recurrences were common with 37% of those attempted a full

RTW having at least one relapse. Factors associated with

‘sustained’ RTW were generally similar in the early recovery

stage (B13 weeks) and in the later stages ([13 weeks– 2 years), but there were subtle differences that need to be

taken into account; particularly the effect of the industry on

RTW was not proportional throughout the 2-year follow-up

period. Overall, older age, afflictions involving the neck or

multiple locations, and working in manufacturing were

independent predictors of delays in final RTW. Risk factors

for relapses were delayed final RTW, ages 35–55, female

sex, working as a labourer, working in manufacturing,

musculoskeletal disorders, and afflictions involving the neck

or multiple locations. The results of this study confirm pre-

vious reports on frequent relapsing after a first RTW [4, 11],

and extend the current knowledge with an analysis of factors

associated with work disability recurrences, across a full

range of work-related injuries and diseases.

The rate of work disability recurrence reported in pre-

vious studies varies depending on the study population and

methodology: in an administrative study of US workers

with work-related lower back pain (LBP), 11% of the

cohort experienced recurrent disability because of work-

related LBP [12]. Although this is a substantially lower

recurrence rate compared to our results, this may be due to

a shorter follow-up period (12 months vs. 24 months in our

study) and more rigorous definitions: first RTW was

defined as working for 15 consecutive days (vs. [7 day in our study) and recurrences were defined as resumption of at

least 15 consecutive days of disability payments (vs.

C1 day in our study). Another study of LBP recurrence

reported work disability recurrence rates of 17, 14 and 11%

after minimum breaks in indemnity payments of 1, 3 and

8 days, respectively [13]; no minimum period of time off

work was specified in this study (vs. 10 workdays in our

study). To verify if the discrepancy between these reports

and our results were due to the nature of the injury, we

repeated our analyses selecting only lower back pain

claims (specific location of the first affliction: ‘lower back’;

affliction: ‘traumatic joint/ligament and muscle/tendon

injury’ or ‘musculoskeletal and connective tissue dis-

eases’): among workers with LBP, 48% had at least one

relapse. The vast difference between previously reported

recurrence rates and the findings in our study may be

related to differences in claims handling: for example,

differences in the tendency to start a new claim rather than

re-open an existing claim with a new episode of an old

injury/disease. In an Alberta study of workers’ compensa-

tion claimants with lower back pain, 18% of claimants

resumed benefits after initial suspension, 14% had their

claim re-opened or filed a new back-related claim, and with

overlap between these groups, the total recurrence rate was

25% [14]. Differences between compensation systems and

claims handling and how this impacts work disability

recurrence rates are ground for future research.

Female sex [4, 13], blue collar occupation [13] and age

[13] have previously been reported as risk factors for

recurrence of work disability, as has treatment by a phys-

ical therapist or physician service versus chiropractic ser-

vice [12]. Among workers with LBP, longer time until first

RTW is reported to be the most powerful predictor of

recurrence, and this is thought to be a proxy for the severity

of the LBP episode [13]. This is confirmed in our study.

Working in manufacturing was associated with relapsing in

our study: manufacturing was also associated with delayed

final RTW, particularly after 13 weeks. Manufacturing is

the largest industry group in the lost time claims data, and

this group may be amenable to policy change: in addition

to the focus on timely RTW, emphasis on education and

secondary prevention may help improve long-term sus-

tained RTW outcomes. Similar recommendations apply to

workers with musculoskeletal disorders.

Study Limitations

This study has several limitations that need to be addres-

sed. The main limitation of this study relates to the use of

administrative data and can best be discussed by consid-

ering the present study in the light of the conceptual

framework described by Young et al. of RTW as an

evolving process comprising four phases: ‘off work’, ‘work

re-entry’, ‘retention’ and ‘advancement’ [15, 16]. Admin-

istrative data does not provide the information needed to

fully place RTW outcomes in this context: for example,

cessation of income replacement benefits in the present

study is considered to be a proxy for RTW (i.e. ‘work re-

entry’) but could actually reflect time off work to gain

qualifications needed for career advancement, or a range of

other situations. Cessation of benefits have previously been

reported to be poor measure of RTW: in a study among low

290 J Occup Rehabil (2012) 22:283–291

123

back injury claimants in California, the amount of lost

work time as reported by injured workers was significantly

underestimated by administrative measures based on the

duration of wage replacement benefits[17]. The results

presented here are RTW outcomes solely from a payer

perspective [18]: no distinction is made between cessation

of payments (a ‘good’ outcome) due to RTW, unemploy-

ment benefits, retirement, study, caring duties or other.

The modelling in the current study relied entirely on

information from the claims records, and a range of impor-

tant variables, such as psychosocial factors, were missing.

There was also no information available relating to severity

of the affliction or pain. Within the diagnostic categories,

differences in recovery patterns and pain levels are likely to

impact the time until first RTW and the cumulative duration

on benefits, as was demonstrated among a cohort of workers

with back pain [19]. How pain and recovery patterns impact

the RTW process is therefore out of the scope of this study. A

further study limitation is the potential selection bias intro-

duced by removing claimants with multiple claims within the

data collection and follow-up period. The RTW results

presented here apply to claimants with a single claim:

untangling patterns of repeat claiming and RTW is a topic for

future research. Finally, this study does not cover all

absenteeism but only periods off work of[2 weeks, because absenteeism of less than 2 weeks is usually covered by the

employer.

Conclusions

Work disability recurrences as determined from income

compensation data were much more common than reported

in two previous studies using similar methodology. How

recurrence rates are affected by the compensation system

and claims handling procedures is ground for future

research. Working in manufacturing and having musculo-

skeletal disorders were associated with delayed final RTW

and frequent relapses: a policy focus on education about

secondary prevention may help improve long-term RTW

outcomes in this group.

Acknowledgments This study is based on administrative claims data from the Compensation Research Database held at the Institute

for Safety, Compensation and Recovery Research (Victoria, Austra-

lia). The research was supported by funding from WorkSafe Victoria

and the Transport Accident Commission (TAC).

Conflict of interest There was no conflict of interests.

References

1. Australian Bureau of Statistics. Work-related injuries 2007.

2. Krause N, Dasinger LK, Deegan LJ, Brand RJ, Rudolph L.

Alternative approaches for measuring duration of work disability

after low back injury based on administrative workers’ com-

pensation data. Am J Ind Med. 1999;35(6):604–18.

3. Amick BC, III, Lerner D, Rogers WH, Rooney T, Katz JN. A

review of health-related work outcome measures and their uses,

and recommended measures. Spine (Phila Pa 1976). 2000;25(24):

3152–60.

4. Butler RJ, Johnson WG, Baldwin ML. Managing work disabil-

ity—why 1st return to work is not a measure of success. Ind

Labor Relat Rev. 1995;48(3):452–69.

5. Oleinick A, Gluck JV, Guire K. Factors affecting first return to

work following a compensable occupational back injury. Am J

Ind Med. 1996;30(5):540–55.

6. Australian Bureau of Statistics and Statistics New Zealand.

Australian and New Zealand standard classification of occupa-

tions (ANZSCO). 1st ed. 2006. Available from: http://www.abs.

gov.au/ausstats/abs@nsf/mf/1220.0

7. Australian Bureau of Statistics and Statistics New Zealand. Aus-

tralian and New Zealand standard industrial classification (ANZ-

SIC). 2006 [16 June 2011]. Available from: http://www.ausstats.

abs.gov.au/Ausstats/subscriber.nsf/0/19C21C5659BCAE73CA2

574C8001474E4/$File/12920_2006%20(revision%201).pdf.

8. Australian Safety and Compensation Council. Type of occurrence

classification system 3rd ed. Revision 1. Canberra2008 [16 June

2011]. Available from: http://www.safeworkaustralia.gov.au/Abo

utSafeWorkAustralia/WhatWeDo/Publications/Documents/207/

TypeOfOccurrenceClassificationSystem(TOOCS)3rdEditionRevi

sion1.pdf.

9. SAS software, Version 9.2. Copyright (c) 2002–2008 by SAS

Institute Inc., Cary, NC, USA.

10. Ruseckaite R, Collie A. Repeat workers’ compensation claims:

risk factors, costs and work disability. BMC Public Health. 2011;

11:492.

11. Baldwin ML, Johnson WG, Butler RJ. The error of using returns-

to-work to measure the outcomes of health care. Am J Ind Med.

1996;29(6):632–41.

12. Cifuentes M, Willetts J, Wasiak R. Health maintenance care in

work-related low back pain and its association with disability

recurrence. J Occup Environ Med. 2011;53(4):396–404.

13. Wasiak R, Verma S, Pransky G, Webster B. Risk factors for

recurrent episodes of care and work disability: case of low back

pain. J Occup Environ Med. 2004;46(1):68–76.

14. Gross DP, Battie MC. Functional capacity evaluation perfor-

mance does not predict sustained return to work in claimants with

chronic back pain. J Occup Rehabil. 2005;15(3):285–94.

15. Young AE, Roessler RT, Wasiak R, McPherson KM, van Poppel

MNM, Anema JR. A developmental conceptualization of return

to work. J Occup Rehabil. 2005;15(4):557–68.

16. Wasiak R, Young AE, Roessler RT, McPherson KM, van Poppel

MNM, Anema JR. Measuring return to work. J Occup Rehabil.

2007;17(4):766–81.

17. Dasinger LK, Krause N, Deegan LJ, Brand RJ, Rudolph L.

Duration of work disability after low back injury: a comparison

of administrative and self-reported outcomes. Am J Ind Med.

1999;35(6):619–31.

18. Young AE, Wasiak R, Roessler RT, McPherson KM, Anema JR,

van Poppel MNM. Return-to-work outcomes following work

disability: stakeholder motivations, interests and concerns.

J Occup Rehabil. 2005;15(4):543–56.

19. Chen C, Hogg-Johnson S, Smith P. The recovery patterns of back

pain among workers with compensated occupational back inju-

ries. Occup Environ Med. 2007;64(8):534–40.

J Occup Rehabil (2012) 22:283–291 291

123

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.