https://orcid.org/0000-0002-9719-9110
Abstract
Background
Being able to work during and after breast cancer treatments is important for patients to have a sense of normalcy, financial security, and improved quality of life, and for society due to the economic burden of sick leave. Factors influencing the length of sick leave can be sociodemographic factors, workplace adaptations, recurrences, symptoms, and type of treatment. The aim of this study is to analyse factors associated with prolonged sick leave after adjuvant breast cancer treatments.
Methods
The population of this registry study consists of 1333 early breast cancer patients diagnosed and treated in Helsinki University Hospital between 2016 and 2018. Data on patient demographics, disease characteristics, treatment, and healthcare resource utilization were obtained from Helsinki University Hospital and data on income level and sick leave were obtained from Kela sickness benefits registry. Prolonged sick leave was determined as the patient accumulating 30 or more reimbursed sick leave days during a 60-day follow-up period after the end of active oncological treatment. Univariate analysis and multivariate analysis were conducted.
Results
A total of 26% of the patients in this study were on sick leave for 30 or more days after the active treatments ended. Study findings show that chemotherapy, triple-negative breast cancer, reconstructive surgery, amount of outpatient visits, and income are associated with prolonged sick leave. Independent predictors of prolonged sick leave were treatment line, number of outpatient contacts, reconstruction, and triple-negative breast cancer.
Conclusions
Our study shows that prolonged sick leave affects a substantial number of working-age women with early breast cancer. Independent predictors for prolonged sick leave were all treatment-related. Targeted support for treatment-related side-effects already during the treatment period could lead to better recovery and earlier return to work.
Background
Breast cancer is the most common cancer in women world-wide accounting for 28% of the total cancer cases in Europe [Citation1]. Most of the women diagnosed belong to the working age group with about 70% of new breast cancer cases globally occurring in women of working age [Citation2,Citation3].
Being able to work during and after breast cancer treatments is important for having a sense of normalcy [Citation4], financial security [Citation5], and improved quality of life [Citation6]. Being able to work is also important for society due to the economic burden of sick leave [Citation6–8].
The number of sick leave days for breast cancer patients in Europe varies. The median sick leave days have been identified as 155 days in France [Citation9] and 216 days for breast cancer patients in the Netherlands [Citation10]. In this study, modified from Drolet et al. (2005) [Citation11], a prolonged sick leave is determined as lasting over 30 days after the last treatment visit (radiation therapy or chemotherapy).
Several studies have investigated issues related to employment and returning to work among breast cancer patients. For example, sociodemographic factors, workplace adaptations, treatment, recurrences, fatigue, and symptoms [Citation10,Citation12,Citation13], and time to return to work [Citation14] have been studied.
The type of treatment received, especially chemotherapy, seems to have had the largest individual/direct impact on working capacity [Citation9,Citation12]. Drolet et al. (2005) [Citation11] found that receiving adjuvant chemotherapy prolonged absence duration by up to 4 months compared to women not receiving chemotherapy. Furthermore, higher age and lower income and education [Citation15], the lack of support from employers [Citation16], as well as psychological factors such as depression and emotional distress [Citation15] increase the likelihood of longer sick leave/not returning to work.
The aim of the study was to analyse factors associated with prolonged sick leave after adjuvant breast cancer treatments based on registry data in a Finnish cohort of breast cancer patients diagnosed 2016–2018. In Finland in 2014, breast cancer had the highest healthcare costs of all cancer types, at 186 million Euros, accounting for 20% of the total cancer costs with eleven per cent of the costs of breast cancer caused by sick leave [Citation17].
Methods
Study population and design
The study is retrospective and based on data from patients with breast cancer from the registries of Helsinki University Hospital (HUS) and the Social Insurance Institution of Finland (Kela). The study population consists of breast cancer patients diagnosed and treated in HUS between 2016 and 2018. HUS is responsible for breast cancer treatments of the whole population of the Uusimaa region (1.7 million). Patients were included if they were of working age (18 to minimum retirement age), had a municipality of residence in the HUS area, and had breast cancer (ICD-10 codes C50 or D01.5) recorded as the main diagnosis on a first-time visit between 2016–2018, and had their last recorded active treatment visit no later than 30.10.2018. Leaving a 60-day follow-up period from the end of active treatment. The upper age limit of the patients varied between 63–64.25 years as the minimum retirement age depends on the person’s year of birth.
To capture only patients with early-stage breast cancer patients had to meet the following inclusion criteria, (1) undergone surgery for breast cancer (Nordic Classification of Surgical Procedures HAC, HAB, HAF), and (2) received either chemotherapy or radiotherapy or both. Patients were excluded if they, 1) died during active treatment or during 12 months after the end of active treatment (surgery, adjuvant chemotherapy or radiotherapy were considered as active treatment), (2) had 10 or fewer radiotherapy visits, or (3) had excessive periods of chemotherapy. Excessive periods of chemotherapy were defined as exceeding 8 cycles per treatment period, with treatment periods defined as series treatments with less than 60 days between consecutive cycles. Treatments including paclitaxel (max. 12 treatments/period), trastuzumab or pertuzumab (max. 22 treatments/period) were exceptions to the definition of excessive periods of chemotherapy.
Patients were classified into three groups based on the number of sick leave days accumulated after active oncological treatments, “no sick leave”, “expected sick leave” and “prolonged sick leave”. In this study, prolonged sick leave was defined as the patient accumulating 30 or more days reimbursed sick leave days during a 60-day follow-up period after the end of active oncological treatment (Supplement material Figure 1). The definition was modified by Drollet et al. (2005) [Citation11] and is in line with the clinical experience from HUS that acute treatment-related side-effects usually ease in less than 30 days. If the patient needs more sick leave the situation is unusual and needs to be evaluated carefully.
Data
Data on patient demographics, disease characteristics, treatment, and healthcare resource utilization were obtained from HUS electronic medical record (EMR) and data on income level and sick leave was obtained from Kela sickness benefits registry. All information across different data sources was linked to the patient based on their unique personal identity code. The structured EMR data included data on outpatient visits and other contacts, radiotherapy visits, inpatient stays, surgical procedures, diagnostics, prescriptions, at-hospital administered drugs as well as gender, age, and municipality of residence.
In all patient characteristics, except for income level, were defined at the time of diagnosis utilizing structured EMR data. Income information was calculated from Kela’s sickness benefits data and therefore only covers patients with sickness allowance paid by Kela. The sickness allowance is determined on the basis of an annual income. The annual income is calculated for a reference period of 12 calendar months prior to the calendar month that precedes the start of the work disability (https://www.kela.fi/sickness-allowance-amount-and-payment).
Table 1. Characteristics of study participants.
Each patients’ annual income level was defined as the highest income level recorded between 2016-2019. Histological classification was derived from the ICD-10 classification, C50.X1 ductal carcinoma; C50.X2 lobular carcinoma; and C50.X0, C50.X9 and D01.5 other. Oestrogen (ER) and progesterone (PR) receptor, and HER2 statuses were retrieved from pathological reports.
Curative surgeries (treatment code for mastectomies HAC, breast-conserving surgeries HAB, and other breast cancer surgery HAF), chemotherapy administration days, and radiotherapy visits were defined as active treatment events. An active treatment episode was defined as a series of active treatment events, with a maximum of 60 days between each treatment event. If there were more than 60 days between treatment events, the treatments were assigned to separate treatment episodes (Supplement material Figure 1). In addition to active treatment, patients could receive other treatment during their active treatment episode or during the 60-day follow-up period after the end of active treatment. These treatments included reconstruction surgery (treatment codes HAE, HAD), axillary evacuation (treatment codes PJA, PJD), endocrine treatment (defined as drug prescriptions with ATC classes L02), or other medications (defined as drug prescriptions with ATC classes M01 or N02 for pain killers, and A04 for antiemetics).
Healthcare resource utilization including both outpatient visits and inpatient stays was considered only during active treatment episodes, and they include all events regardless of the specialty and primary diagnosis associated with the event. Outpatient visits were classified into emergency department visits and other outpatient visits based on service transaction codes. Contacts that occurred on the same days with active treatment events were excluded.
Sick leave data included data from the registries of Kela - The Social Insurance Institution in Finland that pays a sickness allowance and disability pensions for people of working age who cannot perform their regular work due to an illness. This data includes information on all sickness benefits, their duration, diagnoses, and the amount paid for by Kela. However, for the first 10 days of sick leave the employee is paid salary by their employer. As the data only had information on benefits paid for by Kela, information on leave of 10 or fewer days was not included. Patients with disability pension were classified as being on sick leave after the date of the first granted pension.
Statistical analysis
Logistic regression analysis including calculations of odds ratio (OR) and 95% confidence interval (95%CI) was used to examine the impact of patient characteristics on the likelihood of prolonged sick leave. Variables significantly associated with prolonged sick leave in univariate analysis were entered into multivariate analysis with backward stepwise selection. Two multivariate models were formed as yearly income was obtained only for those with reported sick leave. Thus, Model 1 was formed for the whole cohort, and yearly income was not entered into the analysis. Model 2 was formed for the cohort consisting only of those with reported sick leave, and yearly income was entered into the analysis. VIF values of the variables in the final models were less than 2.0 indicating that multicollinearity did not occur. Prior to analysis, data were divided into training (80%) and test data sets (20%). Models were built using training data, and areas under the receiver operating characteristic (AUROC) curves were reported for a test data set. StepAIC method was used for feature selection. All analyses were performed with RStudio (RStudio Team (2020). RStudio, Integrated Development for R. RStudio, PBC, Boston, MA URL http://www.rstudio.com/). P-value<.05 was considered as statistically significant.
Results
Patient characteristics
Characteristics of study participants are presented in for No sick leave, Expected sick leave and Prolonged sick leave groups as well as for the entire cohort. A total of 1333 patients were included in the study. From 4769 breast cancer patients diagnosed and treated in HUS oncology between 2016 and 2018, a total of 3436 patients were excluded (246 municipality of residence outside the HUS area, 1267 did not receive active treatment [as defined in methods], 453 had last active treatment later than 30.10.2018, 44 received excessive periods of chemotherapy or had less than 10 radiotherapy visits, 33 died during or 12 months after active treatment, and 1393 were older than defined upper age limit at the time of last active treatment). No sick leave group included 147 (11%) patients, Expected sick leave 845 (63%) patients and Prolonged sick leave 341 (26%) patients (). The mean age of patients was 52.3 (SD 7.7) years old, and almost all of them were women (n < 5 males in cohort) (). The most common treatment line included surgery, chemotherapy, and radiotherapy (54%). The mean active treatment duration was 5.1 (SD 3.4) months, and included 8.2 (SD 6.2) outpatient visits, 0.4 (SD 0.9) ED visits and 0.3 (SD 0.6) inpatient stays (). 1259 (94%) of the patients had at least one outpatient visit, and 322 (24%) of the patients visited ED at least once during their active treatment.
Sick leave during and after treatment
Sick leave during treatment
Sick leave during active treatment was reported for 1105 (83%) of the patients. For most of the patients, sick leave was prescribed in one sick leave episode (n, 949, 71%) and only for a few patients in several episodes (n, 156, 12%). On average patients were on sick leave 56% (SD 40) of the length of their treatment. Sick leave lasted 43% (SD 38) of the length of treatment for patients treated with surgery and radiotherapy, 68% (SD 38) for patients treated with surgery, chemotherapy, and radiotherapy, and 64% (SD 40) for patients treated with surgery and chemotherapy. Sick leave covered over 90% of the length of treatment for 451 (34%) of patients. During the treatment, the patients had on average 16.8 (11.9) sick leave days per month of active treatment. Patients in Prolonged sick leave group had 27.1 (5.9) sick leave days, and patients in the Expected sick leave group had 15.6 (10.9) sick leave days per month of active treatment ().
Table 2. Sick leaves of the patients before, during, and after active breast cancer treatment.
Sick leave after treatment
During the first two months (days 0-60) after active treatments ended, the patients had on average 16.5 (23.1) sick leave days. A total of 341 (26%) patients had over 30 sick leave days, and these patients were thus grouped into the Prolonged sick leave group. Others with at least one reported sick leave during 2016–2018 were grouped to Expected sick leave group, and the rest of the patients (without any reported sick leave) to No sick leave group. Patients in Prolonged sick leave group had 53.4 (10.0) sick leave days, and patients in Expected sick leave group had 4.5 (7.8) sick leave days during the first two months (days 0-60) after active treatment.
During the first month (days 0-30) after the active treatment, patients had on average 10.2 (12.9) sick leave days (). Patients in Prolonged sick leave group had 29.3 (3.0) sick leave days, and patients in Expected sick leave group 4.3 (7.6) sick leave days (). During the second month (days 31-60) after active treatment, the patients had on average 6.3 (11.6) sick leave days, patients in Prolonged sick leave group had 24.1 (9.6) sick leave days, and patients in Expected sick leave group had 0.2 (1.8) sick leave days ().
Both during the first and second month after active treatment, sick leave was mainly due to breast cancer diagnosis (87.4% and 78.8% of all the sick leave days, respectively). Mental and behavioural diseases (F-group diagnosis) covered 4.3% and 7.1%, and diseases of the musculoskeletal system and connective tissue (M-group diagnosis) 1.2% and 2.2% of all the sick leave days during the first and second month after active treatment, respectively.
During the first year (days 0-360) after active treatment, patients had on average 3.8 (7.1) sick leave days per month. Patients in Prolonged sick leave group had 12.7 (8.9) sick leave days per month, and patients in Expected sick leave group 0.9 (2.3) sick leave days per month.
Univariate and multivariate analyses
Univariate analyses
The demographic factors significantly associated with an increased likelihood of prolonged sick leave in univariate analysis were younger age, municipality of residence and lower yearly income (). With respect to disease and treatment characteristics, ductal histology, triple negative type of cancer, treatment line that included chemotherapy, performed mastectomy or reconstruction, or prescriptions of symptom medications were associated with an increased likelihood of prolonged sick leave, while ER and PR receptor expression, as well as BCS were associated with a decreased likelihood ().
Table 3. Univariate and multivariate logistic regression analysis for prolonged sick leave after active breast cancer treatment.
Multivariate analyses
The remaining independent predictors of prolonged sick leave in multivariate analysis were treatment line, high number of outpatient visits/contacts during active treatment, reconstruction, and triple-negative breast cancer (Model 1 and 2, ). In addition, the municipality of residence was associated with an increased likelihood of prolonged sick leave (Model 1), but the association was no longer significant after considering yearly income in the sick leave cohort (Model 2, ). The VIF values of all the variables in the models were less than 2.0. AUROC of Model 1 (for the test data set) was 0.764 (95%CI 0.699–0.829) and of AUROC of Model 2 (for test data set) 0.740 (95%CI 0.670–-0.812).
Discussion
The objective of this study was to investigate the sick leave patterns and predictive factors for prolonged sick leave after the active treatment period of breast cancer patients who have undergone adjuvant therapies. Breast cancer treatments have several acute side-effects and the majority of the patients are on sick leave during and straight after their treatments. It is important to detect the patients who are at risk and do not return to work and possible modifiable factors for supportive interventions for these patients. Better supportive care with symptoms and/or side effects due to aggressive treatment may be important for patients to return to work. Targeted support during the treatment period could lead to better recovery and earlier return to work. Our main finding is that as many as 26% of the patients were on sick leave for more than 30 days after the active treatments ended. Factors increasing the risk of prolonged sick leave, are in line with previous studies which found that treatment-related factors [Citation15], particularly patients receiving chemotherapy have more often prolonged sick leave [Citation11]. We showed that treatment-related factors, chemotherapy, triple-negative breast cancer and reconstructive surgery predicted longer (more than 30 days after treatment) sick leave days. Furthermore, the higher amount of outpatient visits during the treatment period independently predicted prolonged sick leave post-treatment. The association between more aggressive treatment and sick leave is logical due to treatment-related side-effects. The association between the higher amount of outpatient visits and sick leave could be related also to the side-effects of the treatments since the visits during the active treatment days were excluded from the analyses. Triple negativity is associated with a worse prognosis and the need for chemotherapy [Citation18] which can cause both psychological and physiological burdens to patients. This probably explains the association between triple negativity and prolonged sick leave.
Socio-demographic factors such as education and ethnicity [Citation15] have in previous studies been identified as influencing factors related to the return to work of breast cancer patients. In our study, we could not identify the level of education or ethnicity of the patients in the cohort. However, results from our univariate analysis indicated that younger age, municipality of residence, and lower yearly income are associated with prolonged sick leave. Age has previously been identified as an influencing factor for prolonged sick leave of breast cancer patients [Citation15]. However, interestingly previous findings showed that older women tend to have longer sick leave [Citation16] than younger. The contradictory results related to age might be explained by contextual factors, such as differences between countries e.g., in relation to social security and pensions. The results of our multivariate analysis indicate that the level of income was the only demographic factor associated with longer sick leave. The association between income level and prolonged sick leave could be due to the nature of work and the level of physical requirements in the employment. Employment related factors, particularly employer support [Citation16,Citation19] has previously been identified to be associate to the return to work of breast cancer patients. To better support women at risk of prolonged sick leave, further research could be conducted on employment-related factors, such as organisational culture of the employer, maturity and size of the employer, contract type, as well as nature and physical recruitments of the work.
The strength of this study is a fairly high amount of early breast cancer patients treated according to the international guidelines with detailed information about their sick leave days. We have also some limitations in our study. We were not able to collect detailed information about the nature of employment of the patients nor psychosocial factors (e.g., quality of life, anxiety, or depression) and analyse further these indicators as they are not collected to the registries that we used in this study. Further research is needed to investigate the association between psychological factors such as depression and emotional distress [Citation15] to longer sick leave and not returning to work. Future research could also include and focus on integrating patient reported, psychosocial measure such as QoL; anxiety and depression in the analysis which would help in targeting the interventions for patients.
Conclusions
The aim of this study was to analyse factors associated with prolonged sick leave after adjuvant breast cancer treatments in a Finnish cohort of breast cancer patients diagnosed 2016–2018. The study findings indicate that up to 26% of breast cancer patients have prolonged sick leave, indicating a decrease in the patients’ quality of life. Our findings also show that chemotherapy, triple negative breast cancer, reconstructive surgery, amount of outpatient visits and income are associate with prolonged sick leave of early breast-cancer patients. Thus, there is potential to identify the patients at risk, and to offer these patients support early on.
Author contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by [Paula Pennanen], [Mari Lahelma]. All authors participated in the writing of the manuscript. All authors have read and approved the final manuscript.”
Ethics approval
This received a study permit from the Helsinki University hospital (10.12.2020) HUS/175/2020.
| Abbreviations | ||
| AUROC | = | areas under the receiver operating characteristic |
| BCS | = | Breast conserving surgery |
| CI | = | confidence interval |
| ED | = | Emergency department |
| ER | = | oestrogen |
| EMR | = | Electronic medical record |
| HCRU | = | health care resource utilization |
| HER2 | = | human epidermal growth factor receptor 2 |
| HUS | = | Helsinki University Hospital |
| IQR | = | interquartile range |
| Kela | = | Kansaneläkelaitos – The Social Insurance institution of Finland |
| M | = | month |
| n | = | number |
| NCSP | = | Nordic Classification of Surgical Procedures |
| OR | = | odds ratio |
| PR | = | progesterone |
| QoL | = | Quality of Life |
| SD | = | standard deviation |
| VIF | = | Variance inflation factor |
Supplemental Material
Download MS Word (13.1 KB)Supplemental Material
Download JPEG Image (1 MB)
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The data that support the findings of this study are available on request from the corresponding author PPS. The data are not publicly available due to restrictions e.g., their containing information that could compromise the privacy of research participants.
Additional information
Funding
References
- “WHO | Breast cancer, prevention and control,” WHO. [cited 2018 Oct 5]. Available from: http://www.who.int/cancer/detection/breastcancer/en/.
- Global Burden of Disease Study 2015 (GBD 2015). Results United States, Institute for Health Metrics and Evaluation (IHME). [cited 2017 Jan 7]. http://ghdx.healthdata.org/gbd-results-tool. 2016.
- Hauglann B, Benth JŠ, Fosså SD, et al. A cohort study of permanently reduced work ability in breast cancer patients. J Cancer Surviv. 2012;6(3):345–356. doi: 10.1007/s11764-012-0215-0.
- Blinder VS, Murphy MM, Vahdat LT, et al. Employment after a breast cancer diagnosis, a qualitative study of ethnically diverse urban women. J Community Health. 2012;37(4):763–772. doi: 10.1007/s10900-011-9509-9.
- Mackenzie CR. ‘It is hard for mums to put themselves first’, how mothers diagnosed with breast cancer manage the sociological boundaries between paid work, family and caring for the self. Soc Sci Med. 2014;117:96–106. doi: 10.1016/j.socscimed.2014.07.043.
- Kennedy F, Haslam C, Munir F, et al. Returning to work following cancer, a qualitative exploratory study into the experience of returning to work following cancer. Eur J Cancer Care (Engl). 2007;16(1):17–25. doi: 10.1111/j.1365-2354.2007.00729.x.
- Peteet JR. Cancer and the meaning of work. Gen Hosp Psychiatry. 2000; 22(3):200–205. doi: 10.1016/s0163-8343(00)00076-1.
- Rasmussen DM, Elverdam B. The meaning of work and working life after cancer, an interview study. Psychooncology. 2008; 17(12):1232–1238. doi: 10.1002/pon.1354.
- Arfi A, Baffert S, Soilly AL, et al. Determinants of return at work of breast cancer patients, results from the OPTISOINS01 french prospective study. BMJ Open. 2018;8(5):e020276. doi: 10.1136/bmjopen-2017-020276.
- De Boer AG, Verbeek JH, Spelten ER, et al. Work ability and return-to-work in cancer patients. Br J Cancer. 2008;98(8):1342–1347. doi: 10.1038/sj.bjc.6604302.
- Drolet M, Maunsell E, Mondor M, et al. Work absence after breast cancer diagnosis, a population-based study. CMAJ. 2005;173(7):765–771. doi: 10.1503/cmaj.050178.
- De Boer AG, Taskila T, Ojajärvi A, et al. Cancer survivors and unemployment, a meta-analysis and meta-regression. Jama. 2009;301(7):753–762. doi: 10.1001/jama.2009.187.
- Carlsen K, Oksbjerg Dalton S, Frederiksen K, et al. Cancer and the risk for taking early retirement pension, a danish cohort study. Scand J Public Health. 2008;36(2):117–125. doi: 10.1177/1403494807085192.
- Balak F, Roelen CA, Koopmans PC, et al. Return to work after early-stage breast cancer, a cohort study into the effects of treatment and cancer-related symptoms. J Occup Rehabil. 2008;18(3):267–272. doi: 10.1007/s10926-008-9146-z.
- Islam T, Dahlui M, Majid HA, et al. Factors associated with return to work of breast cancer survivors, a systematic review. BMC Public Health. 2014;14(Suppl 3):S8. doi: 10.1186/1471-2458-14-S3-S8.
- Roelen CA, Koopmans PC, De Graaf JH, et al. Sickness absence and return to work rates in women with breast cancer. Int Arch Occup Environ Health. 2009;82(4):543–546. doi: 10.1007/s00420-008-0359-4.
- Torkki P, Leskelä RL, Linna M, et al. Cancer costs and outcomes for common cancer sites in the Finnish population between 2009–2014. Acta Oncol. 2018;57(7):983–988. doi: 10.1080/0284186X.2018.1438656.
- Won KA, Spruck C. Triple‑negative breast cancer therapy, current and future perspectives. Int J Oncol. 2020;57(6):1245–1261. doi: 10.3892/ijo.2020.5135.
- Lidgren M, Wilking N, Jönsson B. Cost of breast cancer in Sweden in 2002. Eur J Health Econ. 2007;8(1):5–15. doi: 10.1007/s10198-006-0003-8.