Sickness certification of workers compensation claimants by general practitioners in Victoria, 2003–2010

The past decade has seen substantial health policy reform aimed at increasing labour force participation by injured and ill workers. Underpinning these reforms is a growing body of evidence showing that safe work is good for health1,2 and that return to work (RTW) after injury or illness can promote recovery.3,4 General practitioners play a key role in the RTW process, being the first point of contact with the health care system for many injured workers, and the main “gatekeepers” to workers compensation and disability benefits.

Understanding the behaviour and attitudes of GPs towards injured and ill workers is an area of growing interest among primary care and work injury researchers. To date, no Australian data on this have been published. Overseas research has shown that, despite most GPs believing that work is generally beneficial for health, there are divergent views among GPs regarding their role in RTW processes.5 Some believe that their focus should be on the medical recovery of the worker, but others acknowledge their role in RTW processes.6 Many GPs do not consider RTW processes and the work–health nexus to be part of their role,7 and some would prefer to have no role in sickness certification.8

Evidence that GP sickness certification practices vary according to worker characteristics is growing. A common finding is that GPs are more likely to certify workers with mental health conditions (MHCs) as unfit for work than those with physical conditions.9,10 Risk factors for longer duration of work incapacity include older age, social deprivation, and presence of mild and severe MHCs.11 Other factors which influence GP attitudes about managing RTW include weak doctor–patient relationships, pressure on consultation time, fears for personal safety, limited knowledge about the patient’s workplace, and the administrative burden of dealing with the compensation system.8,12-14

Much of the research on sickness certification practices has been conducted in discrete clusters of primary care practices in Europe. Certification data can also be obtained from jurisdiction-level databases, which are held by some workers compensation authorities. These have the advantages of population-based data capture and a method for verifying the work-relatedness of workers’ conditions (because only claims for work-related injury and illness are accepted).

We examined patterns of sickness certification of injured and ill workers by GPs in the state of Victoria, Australia, over an 8-year period.

Victoria had a working population of about 2.8 million in June 2011.15 Employers are required to maintain workers compensation insurance through the Victorian WorkCover Authority (VWA) unless they are able to self-insure or obtain insurance through the national workers compensation scheme, or if they are sole traders. All injuries and illnesses that exceed the pecuniary threshold for health care expenses or have required more than 10 days work absence are required to be lodged with the VWA via one of six private insurers.

The Victorian workers compensation system requires production of a medical certificate to accept a compensation claim. Certificates can be submitted by GPs or by hospital-based medical practitioners; they include the practitioner’s recommendation regarding fitness to work, and start and end dates. Initial certificates are handwritten and provided to the injured worker’s employer, who submits the certificate to the insurer. Subsequent certificates can be provided directly to the insurer and the employer. The insurer enters certificate data into a standardised database. This information is transferred to the VWA with other claim information, enabling statewide capture of medical certification data.

There are statutory limits for the duration of certificates which state that a patient is unfit for work; these are defined in Victorian workers compensation regulations. An initial medical certificate for a workers compensation claim can be for up to 14 days. Subsequent certificates can be for up to 28 days, unless special conditions exist.

We accessed the Compensation Research Database, which was established at the Institute for Safety, Compensation and Recovery Research, Monash University. It contains case-level administrative data received from the VWA, including data on industry, occupation, employer, workplace, demographics, injury, health care services, payments and medical certificates.16 Ethics approval was obtained from the Monash University Human Research Ethics Committee.

Of 207 949 accepted compensation claims that we extracted, 83 607 were excluded; 78 086 were for health care expenses only, 5439 had an initial medical certificate not written by a GP, and 82 contained errors in logic. Over the study period, initial medical certificates were issued by 9750 GPs to 124 424 workers.

The numbers of initial medical certificates issued and rates of certification, by certificate type and injury and illness category, are shown in Box 1. The largest proportion of certificates was for musculoskeletal injuries and diseases (40.2%), followed by back pain and strains (16.4%) and other traumatic injuries (15.6%). The majority of certificates (74.1%) recommended that workers were unfit for work, 22.8% recommended alternative duties and 3.1% recommended that workers were fit for work. Men received more certificates in all injury and illness categories except the MHCs category, in which more than half of certificates were issued to women. Comparing mean ages, workers categorised as having other traumatic injuries were slightly younger, and those with MHCs and other diseases were slightly older.

From 2003 to 2010 in Victoria, more than 70% of initial medical certificates issued by GPs for injured and ill workers were unfit-for-work certificates, and fewer than a quarter recommended alternative duties. This is an important finding in the context of growing evidence regarding the health benefits of work,1,18 the importance of workforce participation for social and economic development,2 and the potential health and economic benefits of early RTW.3

The rate of sickness certification per 1000 workers varied substantially between injury and illness categories. The highest proportion of unfit-for-work certificates were written for workers with a MHC, followed by fractures, other traumatic injuries, back pain and strains, musculoskeletal injuries and diseases, and other diseases. Recent studies in the United Kingdom have also shown that GPs are more likely to issue unfit-for-work certificates for patients with MHCs than other conditions.9,10 Conversely, GPs in our study were more likely to certify workers with other diseases as suitable for alternative duties than those in other categories. Our data also show that the median duration of unfit-for-work certificates issued by GPs to patients with MHCs is longer than in patients with musculoskeletal injuries and diseases, back pain and other traumatic injuries.

There is a stigma associated with mental illness that may influence pursuit of workers compensation claims among workers with MHCs.19 Health professionals are more likely to perceive people with mental illness as having poorer health outcomes than they really have.20 There is often confusion about when a worker with a MHC should resume work, and workplace accommodations are often not in place to facilitate the RTW process.21 The ability of a workplace to accommodate an injured or ill worker may influence GP certification behaviour.

We also found substantial sex differences. Women with MHCs were more likely to receive certificates than men, but men received a greater proportion of certificates in all other categories. This is consistent with previous findings showing that women have a proportionately greater burden of sickness absence due to mental illness22 and that women are more susceptible to depression as a result of job stress than men.23

Our data were obtained from a population-based database of injured and ill workers across a diverse range of industries and occupations, and included GPs operating in a diverse range of health care practices. The six categories of injury and illness that we used were derived from a national standard occupational injury classification system.17 To maintain maximum data retention across types of injury and illness, it was necessary to group diverse conditions in some categories, while other categories represented more homogenous clinical populations. Within the more heterogenous categories, there may be wide ranges of expected recovery and RTW trajectories for different conditions. Also, workers compensation data do not include cases where an injured or ill worker does not file a compensation claim. Such cases account for about 23% of work-related conditions for which patients attend Australian GP practices,19 so this study does not represent a complete picture of GP certification practices.

Our study is the first to assess sickness certification of injured and ill workers in Australia, and the first to assess certification by GPs in a population as large as the Victorian population of full-time workers. It provides a foundation for comparison with other jurisdictions and health care systems, and a benchmark against which future public health initiatives aimed at modifying GP certification behaviour can be evaluated.

We identified substantial variation in GP practice depending on worker condition. Future research is needed to determine why rates of alternative-duties and fit-for-work certificates among workers with MHCs are less than those for workers with other conditions. It is possible that differences in certification behaviour are due to severity of injury, workplace conditions (eg, lack of appropriate modified duties), challenges in dealing with mental health claims and other factors.

2 Duration of initial medical certificates for workers compensation claims, by certificate type and injury and illness category, Victoria, 2003–2010*

* Horizontal lines represent median values; lower and upper limits of the vertical bars represent 1st and 3rd quartiles, respectively; and error bars represent 95% CIs.