Life expectancy estimations and determinants of return to work among cancer survivors over a 7-year period

Due to advances in medical science and technology, the number of cancer survivors continues to increase. The workplace needs and employment difficulties cancer survivors face after treatment need to be addressed to protect these individuals’ right to work and to maintain the overall labor force of the country. We conducted a retrospective cohort study with a follow-up period from 2004 to 2010. All data analyzed in the study were obtained from the Labor Insurance Database, the Taiwan Cancer Registry of the Ministry of Health and Welfare, and the National Health Insurance Research Database. The relationships between risk factors and the presence of returning to work were analyzed by a Cox proportional hazard model. The survival rates of patients with different cancer stages were evaluated using Kaplan–Meier survival analysis. Among the employees with an initial diagnosis of cancer, 70.4% remained employed through 1 year after the diagnosis, accounting for 83.4% of all cancer survivors; only 51.1% remained employed through 5 years after the diagnosis, accounting for 78.7% of all cancer survivors, a notable decrease. Age, gender, salary, treatment method, company size, and cancer stage were the factors that affected whether employees could return to work or not. The long-term survival of people diagnosed with cancer depends on their chances of returning to work. Strengthening existing return-to-work policies and assisting cancer survivors with returning to work after the treatment should be priorities for protecting these individuals’ right to work and for maintaining the overall labor force.


Methods
The period investigated in this study was from 2004 to 2010, a total of 7 years, and the Labor Insurance Database was used as the main source of information. Labor insurance established by Taiwan government is a compulsory program for workers above 15 years and below 65 years of age intended to protect their rights and interests. First, we identified all employees covered by labor insurance using "Insured File" and "Enterprise File" in the Labor Insurance Database. All selected employee data included salary, company size, county and city where the company is located, changes in insurance, effective date of insurance, employment category and other related information. Combined with the Taiwan Cancer Registry, we identified employees with an initial diagnosis of cancer, thus establishing the cohort for this study. There were 136,342 eligible participants enrolled in this cohort study, including 69,619 patients returning to work and 66,723 patients not returning to work. In addition, we used the National Health Insurance Research Database to obtain inpatient and outpatient records, comorbidities, and cancer-related treatments (including surgery, radiation therapy, chemotherapy, and hormone therapy) as well as prediction of death for each case. This study was reviewed and approved by the Institutional Review Board of Tri-Service General Hospital.
Covariables. Relevant variables were collected from the databases for the purpose of the study. Personal characteristics included age, gender, monthly salary (divided into ≤ 28,800, 28,800-38,200, and > 38,200 New Taiwan Dollars), medical care accessibility (divided into North, Middle, South, and East according to health insurance subdivision) and employment category. Health status included a medical history of major chronic diseases in the year prior to the cancer diagnosis (according to the International Classification of Diseases 9th edition, ICD-9). The clinical comorbidities included disorders of lipid metabolism (ICD-9-CM codes, 272), alcohol abuse (ICD-9-CM codes, 265. 2 Primary outcome. The main outcome of this study was RTW after an initial diagnosis of cancer. Full RTW was defined as the time in calendar days of sick leave until complete work resumption 13 . RTW was confirmed based on employment data from the Labor Insurance database. Unemployment was defined as the employee withdrawing from insurance and not being insured again for 5 years after the initial diagnosis of cancer. The secondary endpoint was the all-cause mortality after RTW within the follow-up period for workers with cancer. Statistical analysis. The SAS statistical software package (version 9.3, SAS Institute Inc., Cary, North Carolina) was used to analyze the descriptive statistics. Continuous variables are expressed as the mean and standard deviation, and categorical variables are expressed as frequencies and percentages. The independent samples t-test, chi-square test, Pearson product difference correlation, ANOVA and logistic regression were used for inferential statistics analyses. Univariate and multivariate adjustments for the Cox proportional hazard model were used to determine the risk factors that can significantly predict the presence or absence of returning to work. The fully adjusted model includes age, treatment, income range, industrial classification, company size, and cancer stage. The hazard ratio (HR) and 95% confidence interval (CI) of the HR are provided to indicate the degree of risk. Finally, the survival rates of survivors with different cancer stages were evaluated using Kaplan-Meier survival analysis. In this study, the statistical significance level α was defined as 0.05. A p-value lower than the significance level indicated statistical significance.
Ethical approval. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent. Informed consent was obtained from all individual participants included in the study. Table 1 provides a summary of returning to work at the fifth year for patients from a fixed cohort with an initial diagnosis of cancer. The average age reemployed patients was 47.6 ± 9.4 years old, and the majority were women (66.7%). The average age of those who did not return to work was 52.1 ± 9.7 years old, and male patients accounted for the majority (60.1%). In terms of personal disease factors, participants who returned to work had fewer comorbidities than those who did not return to work. In terms of disease treatment, the largest proportion of workers received operation (returning to work: 76.5%; non-returning to work: 47.7%) after cancer diagnosis. In terms of pathological grades, the majority of cancer survivors who returned to work had stage 2-4 disease. Female breast cancer accounted for the highest proportion (24.1%) of cancer survivors, followed by cervical (15.9%) and colorectal and anal (10.1%) cancer survivors.

Results
In Supplementary Fig. 1, the OR of RTW rapidly increased after the 2nd year in all cancer survivors. A rapid decline was noted after the 4th year. This implied that cancer survivors might return to work within the first 2-6 years after diagnosis with cancer. Table 2 shows the number of deaths, survivors, employed individuals, and the number of people who left the workplace in the 1st to 5th years after the initial diagnosis of cancer and the 10 most common cancers. A total of 70.4% of cancer survivors remained employed through the first year after the initial diagnosis of cancer, accounting for 83.4% of the cancer survivors. After the fifth year, 51.1% of the cancer survivors remained employed, accounting for 78.7% of the cancer survivors. In the 1st year after the initial diagnosis of cancer, patients with cervical cancer (86.1%), female breast cancer (83.6%) and thyroid cancer (82.6%) represented the highest proportion of cancer survivors who remained employed; in the 5th year, patients with cervical cancer (80.6%), thyroid cancer (79.0%), and female breast cancer (72.0%) represented the highest proportion among those who remained employed. Figures 1 and 2 show the results of the univariate analysis of factors that affect returning to work in the 2nd and 5th years after the initial diagnosis. The results showed that age (older) and gender (male) were negatively www.nature.com/scientificreports/ associated with returning to work. Alcohol abuse, hypertension, myocardial infarction, heart failure, peripheral arterial disease, cerebrovascular disease, dementia, chronic pulmonary diseases, peptic ulcer, mild liver disease, hemiparesis, kidney disease, moderate-severe liver diseases and mental illness were significantly negatively associated with RTW in the 2nd and 5th years, and dyslipidemia and depression were significantly negatively associated with returning to work only in the fifth year. Surgical treatment showed a significant positive relationship with RTW in the 2nd and 5th years. Radiation therapy and chemotherapy were significantly negatively associated with returning to work in the 2nd and 5th year. Based on histopathological staging, compared to the highest  www.nature.com/scientificreports/ stage (stage 4), lower stages were significantly positively associated with returning to work in the 2nd and 5th year. In terms of cancer types, we used oral cancer, which is prevalent in Taiwan and Southeast Asia, as the reference. The results showed that patients with salivary gland cancer, nasopharyngeal carcinoma, osteocarcinoma, chondrocarcinoma, sarcoma, skin cancer, female breast cancer, female genital cancer, testicular cancer, bladder cancer, renal cancer and thyroid cancer were significantly associated with increased rates of returning to work in the 5th years. In contrast, esophageal cancer, gastric cancer, hepatic cancer and intrahepatic cholangiocarcinoma, gallbladder cancer and extrahepatic cholangiocarcinoma, pancreatic cancer, thoracic cancer, brain cancer, and leukemia were significantly associated with reduced returning to work in the 5th year. Figure 3 presents the relationship between the different variables and the occurrence of returning to work in the 2nd and 5th years. The results showed that the HR of returning to work in the 2nd and 5th years was decreased for patients with the following characteristics: older age, male sex, difference in salary level greater than 38,200 New Taiwan Dollars, receiving chemotherapy, working in medium/small scale industry, and advanced cancer stage. In contrast, for patients who underwent surgery or radiotherapy, the HR of returning to work in the 2nd and 5th years was still increased. The most notable change was for radiation therapy. In the univariate analysis, radiation therapy was a negative factor for returning to work, but it was a positive factor in the multivariate analysis. Figure 4 displays the survival rates of survivors with different cancer stages (A: all stages; B: stage 0; C: stage 1; D: stage 2; E: stage 3; F: stage 4) evaluated using Kaplan-Meier survival analysis. The survival rates of all cancer stages were significantly higher in the returning to work group than in the non-returning to work group (p < 0.001). The returning to work group had significantly higher survival rates than the non-returning to work group for survivors with stage 1, 2, 3, and 4 diseases (p < 0.001). Table 3 shows the association between returning to work and all-cause mortality. Patents who returned to work had significantly reduced risk of all-cause mortality with an HR of 0.46 (95%CI: 0.44-0.48, p < 0.001) after fully adjusting for the variables.

Discussion
The aim of this study was to analyze the return to work of cancer survivors in Taiwan by combining Taiwan's Labor Insurance Database, National Health Insurance Research Database and Taiwan Cancer Registry. This study used a retrospective cohort study design to establish a cohort of cancer survivors with an initial diagnosis of cancer from 2004 to 2010. From the beginning to the end of the observation, there were no newly added research cases, and no study cases remained. This study found that among the patients with an initial diagnosis, 70.4% of the workers remained employed through the first year, accounting for 83.4% of the cancer survivors. In the fifth year after the diagnosis, 51.1% of the patients remained employed, accounting for 78.7% of cancer survivors, a decrease from the first year. The results of our study showed that the important factors affecting whether patients with an initial diagnosis of cancer returned to work include age, gender, salary level, treatment method, enterprise size and cancer stage and whether returning to work was the key factor affecting the future survival of patients.
The relationship between cancer and returning to work has been presented in numerous studies. Fantoni et al. demonstrated that a high proportion of employed patients with breast cancer returned to work in 36 months after treatment 14 . In a recent study, the risk factors affecting returning to work of cancer survivors included   16 . Among the cancer and symptom factors, in our study, the chances of returning to work were higher for the patients who received surgical treatment. However, if patients received chemotherapy and radiotherapy, the chances of returning to work were relatively low. The reasons for this may be that the patients who received only surgical treatment had their cancer detected at an early stage. Patients with relatively advanced stage disease may require chemotherapy and radiotherapy. In addition, chemotherapy and radiotherapy require 3-6 months, and patients need to visit the hospital or outpatient clinic for several courses of treatment and are more prone to complications or discomfort. Consistent with our findings, a study on breast cancer survivors with a median monthly follow-up of 36 months found that chemotherapy or radiotherapy restricted or postponed returning to work, possibly due to end-stage cancer patients requiring multiple treatment strategies 14 . The multifactor analysis indicated that changes in receiving radiotherapy were a positive factor for returning to work; however, the possible reason for this result was that our study did not group survivors according to multiple treatment strategies. For cancer staging, compared to patients with stage 4 cancer, patients with lower-stage cancer (e.g., stage 1) had a greater chance of returning to work. Studies from other counties also showed that patients with terminal cancer or palliative treatments were associated with a lower return to work rate 17,18 . In terms of cancer types, previous studies have shown that liver cancer, pulmonary cancer, brain cancer, blood cancer, gastrointestinal cancer, pancreatic cancer, head and neck cancer, and gynecological cancer are all significantly correlated with unemployment or losing jobs [18][19][20][21][22] . Another study found that male and female genital cancer, skin cancer and breast cancer had the highest return to work rate 2 years after a cancer diagnosis 23 . Cervical cancer and female breast cancer survivors had a higher return to work rate, possibly due to the continued promotion of Pap smear screening and breast cancer screening by the Taiwan Health Promotion Administration 24 . Women over the age of 30 can receive an annual free Pap smear examination and women aged 45-69 years old or 40-44 years old with a family history of breast cancer in firstor second-degree relatives can receive free mammography once every 2 years. These screening programs can greatly improve the diagnostic rate of early cervical cancer and breast cancer, thereby reducing disability and work-related issues in working women 25,26 .
For demographic factors, based on age, the older an individual is, the lower the chance of the patient returning to work, which may also be related to the retirement age in Taiwan (according to the Ministry of Labor's statistics, Taiwan's actual retirement age from 2010 to 2015 was 63.3 years for men and 60.6 years for women. In this study, the average age of patients in the fifth year after the initial diagnosis was 49.8 years old, the average age of patients who returned to work was 47.6 years old, and the average age of patients who did not return to work was 52.1 years old). Because the age at diagnosis was close to the retirement age, patient motivation to return to work may be decreased. In a Danish study, compared with younger patients, older patients (50-60 years) had a higher unemployment rate 27 . In addition, studies have shown that demographic factors that affect return to work included gender and low socioeconomic levels 18 . Marino's study addressing gender and return to work showed www.nature.com/scientificreports/  www.nature.com/scientificreports/ that among patients who were still alive 2 years after a cancer diagnosis, older men returned to work later than did older women, but married men returned to work earlier than did married women 28 . In our cohort, women were more likely to return to work than men, probably due to the high proportion of women with cervical cancer and breast cancer, which are mainly identified at early stages. Oral cancer is more common in Taiwanese men and is mostly caused by chewing betel nuts and smoking. In the fifth year, returning to work of men with oral cancer was only half that of women with cervical cancer and breast cancer. Survival prognosis is an important indicator for cancer survivors. After confirming prognostic factors that affect survival, health education, rehabilitation and treatment are performed to improve the survival and the quality of life of patients. In our study, returning to work significantly affected the prognosis of patients, especially for patients with stage 4 cancer. Daily physical performance, for example, can be measured with the US Eastern Cooperative Oncology Group and Karnofsky performance scales In a study of pancreatic cancer, poor daily physical performance was significantly correlated with a poor prognosis 29 . In patients with breast cancer that metastasized to the brain, physical performance in the presence of metastasis can predict survival 30 . Good physical performance is especially important for patients with advanced stage cancer, indicating that return to work can effectively predict the survival of these patients 31 .
An advantage of this study is the analysis of big data from a labor insurance and health insurance database in Taiwan. At the end of 2016, there were a total of 10,165,434 people were enrolled in labor insurance. The 2 databases were merged, and information related to the diagnoses and treatment of cancer survivors and changes in employment over 11 years were tracked. A limitation of this study is that the database did not include other important factors for determining the return to work of cancer survivors, such as education level, family support, personal physical performance status, work effort level, etc. As a result, there are limitations related to the inferences that can be made.
In this study, important factors for returning to work RTW included chronic diseases, gender, age, salary, and cancer stage. In particular, we also found that returning to work was significantly associated with the survival of the patients. For the country and society, the return to work of patients with cancer can boost the social labor force and promote the national economy. For working-age patients, a cancer diagnosis and subsequent return to work after treatment is an important transition milestone from being a cancer patient to being a cancer survivor. Returning to work indicates recovery. The path to the return to work of cancer survivors requires not only interdisciplinary professional intervention but also the assistance of the state, society, and employers. Strengthening existing returning to work policies and assisting high-risk populations to return to work after treatment should be a priority to protect these individuals' right to work and to maintain the overall labor force.