Nearly half of individuals randomised to WBRT reduced their employment from full time to part time at 12 months, and for those who were unemployed at baseline, the proportion who were unemployed rose from one half to two thirds at 12 months. There was a small but generally downward shift in income bracket for both groups. Negative employment change and household income at 12-months were not significantly different between the two groups.
The median age was 61 years in both groups, which is over the minimum age that an individual can access their superannuation in Australia (55 to 60, depending on their date of birth) [18]. The Australian Bureau of Statistics (ABS) 2016-17 Survey found that the average age of retirement in those who had retired in the previous 5 years was 62.9 years [19]. There is a trend for Australians to retire later in life and this is reflected by the Australian pension eligibility age which is currently 66 years and is expected to increase to 67 years by 2023. The ABS found that among the men and women whose last job was less than 20 years ago, the most commonly reported reason for ceasing their last job was due to reaching “retirement age/eligible for superannuation/pension” and the second main reason was due to “own sickness, injury or disability”. As the median age of our study population was close to the average retirement age, it is possible that a number of our trial participants stopped working due to retirement rather than illness.
A 2003 Survey of Disability, Aging and Carers found that nearly half of older Australians with cancer (aged between 45 and 64 years) were not in the labour force. This was also reflected in a study by Bastes et al. which suggested that people with cancer were less likely to be in the labour workforce (46%), resulting in a reduction of $1.7 billion to the Australian gross domestic product. Munir et al suggested that the type of cancer and the intensity of the treatment regimen involved will differentially affect an individual’s ability to work [11]. Paul et al conducted a study in two hospitals in Australia and found that among 255 patients who had a cancer diagnosis, there was a 67% change in employment, which in 16% was job loss [12]. In another study by Zajacova et al. it was found that individuals are 10% less likely to be employed after diagnosis, with annual working time decreased by up to 200 hours in the first year after diagnosis [13].
In the current randomised study of adjuvant WBRT versus observation following treatment for melanoma brain metastases, we found that individuals in the WBRT group incurred a reduction in their employment status that was similar to the reduction that occurred in the observation group. This could be because the trial did not show a statistically significant difference in oncological outcomes between the WBRT and observation groups.
There was no significant difference between the WBRT group and the observation group for the changes to household income during the first year of treatment. This could be due to other family members increasing work hours or starting work, leave entitlements (e.g, sick leave, annual and long service leave), superannuation in those who had reached the retirement age, and government assistance schemes (e.g. the pension/allowance) which may have assisted in maintaining the same household income as was achieved prior to the diagnosis of brain metastasis.
Previously, a number of studies have had the limitation that they examined the effects of a cancer diagnosis and treatments using the general population as a control group. The systematic review performed by Munir et al. highlighted this limitation, and reported that the majority of the studies that they reviewed compared cancer survivors with the healthy general population [11]. In our study, the control group was patients with SRS/surgery only, not the general population, providing a more appropriate comparison of cancer diagnosis and treatments on financial status.
Our sub-study nevertheless had several limitations. The sample size of individuals with advanced melanoma was too small for precise estimates. Due to disease progression, a number of our patients were not able to complete the scheduled follow-up visits, which resulted in missing data. The missed visits were attributed mainly to death due to melanoma, neurocognitive impairment, and disease progression. These patients were therefore most likely not in the workforce. We did not capture financial assistance where individuals used specific sources of financial assistance, or reasons for not assessing financial assistance. Reporting errors from respondents might also distort the true effects of cancer treatment on individuals’ income and employment status. Thus our findings may have underestimated or overestimated the true effects of WBRT compared with observation. Furthermore, income is an important yet challenging concept to measure in surveys as it is a sensitive topic and some people do not like to divulge personal or household weekly income. Self-reporting of data relies on individual interpretation of the question. A number of individuals did not know their exact income so we had to create categories to facilitate responses. This could be improved in future by linking to government databases such as the Australian Taxation Office, although privacy issues would make this difficult to achieve.