Higher incidence of screening-related cancers in the employed population

Higher incidence of screening-related cancers in the employed population Abstract Background Employment may confound the risk of a cancer diagnosis in both directions. We hypothesized that a higher baseline rate of employment among cancer patients may explain the lack of association between a cancer diagnosis and later unemployment in many studies. Aims To assess the unemployment rate among cancer patients before diagnosis compared with a matched cancer-free control group. Methods Using data from the Israeli National Central Bureau of Statistics 1995 census (persons aged between 15 and 60 years old), the Israeli Tax Authority database and the Israel Cancer Registry, cancer patients (diagnosed between the years 2000 and 2007 and alive at 2011) were compared with matched cancer-free controls. Results There were 8797 cancer patients and 26166 cancer-free controls. We found that, in general, cancer was not associated with unemployment 2 years before diagnosis (adjusted odds ratio [OR] = 0.96, 95% confidence interval [CI] 0.90–1.009, P = NS) after adjustment for age, gender, ethnicity, educational years and residential socioeconomic position. However, the diagnoses associated with screening (breast, prostate, colorectal and cervix cancers) were inversely associated with unemployment 2 years before diagnosis (adjusted OR = 0.90, 95% CI 0.84–0.97, P < 0.01). Conclusions The results from the current study suggest that a higher baseline rate of employment among cancer patients, mainly those who were diagnosed with screening-associated cancers, explains false negative results in previous studies assessing cancer survivors’ work issues. Cancer, case–control study, employment Introduction Working is an indicator of a sense of normality, improved quality of life and is considered helpful when coping with cancer experiences [1,2]. On the other hand, unemployment or a delay in returning to work affects one’s salary, health and welfare, the family budget and relationships at work places [3,4]. Improvement in long-term survival has led to a growing population of cancer survivors [5,6], and about half of the 3.45 million new cancer cases reported in Europe in 2012 involved people of working age [7]. Data about the unemployment of people living with a history of cancer are highly variable among studies. Different unemployment rates may reflect variances in social security systems, health care insurance, variation in laws that protect individuals from discrimination and the implementation of these laws [8] and the prevailing unemployment rate in a specific country or region [9]. Additional variables such as age [1], gender [8] and education [10,11] may also affect the rates of unemployment following a cancer diagnosis. A meta-analysis [9] found that cancer survivors are more likely to be unemployed than healthy control participants (34 versus 15%, relative risk = 1.37; 95% confidence interval [CI] 1.21–1.55). Yet, after adjustment for the diagnosis, age and prevailing unemployment rate, this association weakened and lost significance [9]. People who are employed have distinct characteristics that may affect the risk of a cancer diagnosis. These include a better socioeconomic status [12], but also an increased risk for occupational exposures [13,14]. In addition, the employed may be in better health than the general population [15], since relatively healthy individuals are likely to gain employment and remain employed; while, chronically disabled and severely ill individuals are frequently not employed [15]. Since cancer screening is aimed at healthy persons [16] and is strongly associated with socioeconomic status [16], workers may have an increased risk for being diagnosed with a screening-associated cancer. We hypothesized that the baseline rate of unemployment is lower among cancer patients compared with cancer-free persons. Furthermore, we hypothesized that this difference is greater among patients who were diagnosed with screening-associated cancers. The difference in baseline rates of employment may explain the lack of association between a cancer diagnosis and unemployment in many studies. Taskila et al. [10] found that the employment rate among cancer patients before diagnosis (78%) was the same as in the cancer-free population. However, stratification according to cancer types was not available. Thus, the aim of this study was to (i) assess the unemployment rate among cancer patients before diagnosis compared with a matched cancer-free control group, (ii) assess the unemployment rate among persons diagnosed with a cancer associated with comprehensive screening programmes compared with a matched cancer-free control group and (iii) examine the association between a cancer diagnosis and baseline unemployment rates after controlling for potential confounders. The use of large national databases allowed us to adjust the results for potential confounders and reduce selection and information biases. Methods This was a case–control study, which included baseline measurements from the Israeli Central Bureau of Statistics 1995 census. The study included a representative sample of the whole population who completed a comprehensive interview (1113420 persons, which represents 20% of the population in Israel). Data on cancer incidence were obtained from the Israel National Cancer Registry. By law, the Israel Cancer Registry has been notified of cancer incidences since 1982. The registry receives compulsory notification from numerous data sources, including pathology reports, discharge summaries and death certificates. Completeness of the registry was about 95% for solid tumours [17]. Melanoma cases were excluded from the analyses since melanoma may be a direct effect of being a worker. Persons aged 15 up to age 60 years old at the time of interview (1995) were included in the current study. Thus, the minimal age of the patients at the time of cancer diagnosis (between the years 2000 and 2007) was 20 years old. Matched groups in a 3:1 ratio according to 5-year age groups, gender and ethnicity (Jewish versus non-Jewish) were sampled from the population in the census who completed the comprehensive interview. The stratification according to ethnicity was done due to lower socioeconomic status and more negative health outcomes of the non-Jewish community compared with the Jewish population [18]. Persons who were diagnosed with cancer were not included in the cancer-free group, and participants who died during the study period (until the end of 2011) were excluded from the current study. Employment status was classified as employed or unemployed. The latter group included unemployed persons, homemakers, students, military conscripts, those on disability pension, retired individuals or unknown. Employed was defined as any reported annual income (>$0). Variables assessed in relation to being unemployed before a cancer diagnosis included age (at the time of cancer diagnosis or in the case of non-cancer healthy control participants, the age at the time of diagnosis of his/her matched cancer patient), gender, number of education years in 1995 (continuous variable), residential socioeconomic position (based on the town/city of residence, according to a national classification of 10 clusters by geographical units) [19] and cancer. Non-specific symptoms that appear in the year prior to the cancer diagnosis may increase the risk of being unemployed. Thus, employment status at baseline was defined 2 years prior to the diagnosis of cancer (or at the same year of the matched cancer patient for non-cancer controls) using the Israeli Tax Authority database. The data were derived from individuals’ tax returns filed with the Tax Authority, which receives yearly compulsory reports from each individual. Data since 1998 were available for the current study. Thus, we included only patients who were diagnosed with a cancer after 2000. We further stratified the study’s population into (i) patients who were diagnosed with a screening-associated cancer (breast, prostate, colorectal and cervix); (ii) patients who were diagnosed with a cancer not associated with screening and (iii) a healthy population (reference group). Lung cancer was not included in the group of cancers associated with comprehensive screening programmes since screening was not common during the study period (2000–07). The current study was approved by the Committee on Human Research at the Hadassah-Hebrew University Medical Center. For analysis of the study population’s characteristics, continuous variables were compared by analysis of variance test and categorical variables were compared with the Wald test and the likelihood-ratio test statistics. A binary logistic regression analysis was constructed to predict the risk of not being employed at baseline using the likelihood-ratio test statistics. We used adjusted analyses to determine the predictive values of the independent variables (age, gender, ethnicity, education, residential socioeconomic position and cancer) and the dependent variable (not being employed at baseline). In addition, we compared the stages at cancer diagnoses between the employed and unemployed populations, who were diagnosed with breast or colorectal cancers. All statistical tests were two-sided, with P <0.05 considered statistically significant. The SPSS programme (18th version; Chicago, IL) was used for the statistical analysis. Results After excluding participants who were diagnosed with melanoma, the study included 26166 match controls and 8797 cancer patients. The most frequent cancers were breast (n = 2365), prostate (n = 938), colorectal (n = 881), cervix (n = 565), bladder (n = 504), thyroid (n = 480) and lymphoma (n = 447). Cancers that are associated with comprehensive screening programmes (breast, prostate, colorectal and cervix) were reported in more than half (n = 4749, 54%) of the cancer patients. The baseline characteristics of the study population are presented in Table 1. Overall, the absolute differences between persons without a history of cancer and cancer patients were minor for all baselines variables. In contrast, patients who were diagnosed with a screening-associated cancer differed from individuals without a history of cancer. The former group was often female and Jewish and was more literate with higher resident socioeconomic positions. In cancer patients, the unemployment rate 2 years before the diagnosis was 34% (n = 3002) compared with 37% (n = 9550) among persons without a history of cancer (P < 0.001). Rates of 34% were found for patients who were diagnosed with both screening-associated (n = 1621) and non-screening-associated (n = 1381) cancers. Table 1. Characteristics of participants in the comprehensive interview of the Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–07 Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) aOrdinal variable based on the town/city of residence, according to a national classification of 10 clusters by geographical units. bBreast, prostate, colon and cervix cancers. View Large Table 1. Characteristics of participants in the comprehensive interview of the Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–07 Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) aOrdinal variable based on the town/city of residence, according to a national classification of 10 clusters by geographical units. bBreast, prostate, colon and cervix cancers. View Large Unemployment rates were lower for patients with all screening-associated cancers when compared with the matched cancer-free population (Figure 1). Yet, significant changes in unemployment rates were only found for patients who were diagnosed with breast cancer (36 versus 40% for the matched cancer-free population, P < 0.01) and prostate cancer (24 versus 26% for the matched cancer-free population, P < 0.05). Among patients who were diagnosed with non-screening-associated cancers, differences between cancer patients and matched cancer-free population were small and uniformly non-significant. In the adjusted models (Table 2), unemployment 2 years before diagnosis was associated with age (increased risk), gender (lower risk in men), ethnicity (lower risk in Jewish individuals), educational years (inverse) and residential socioeconomic position (inverse). In the first model, cancer was not associated with unemployment 2 years before diagnosis (OR = 0.96, 95% CI 0.90–1.009, NS). In the second model, cancer patients were divided into those with (i) screening-associated cancer and (ii) non-screening-associated cancers. Only screening-associated cancer had an inverse relationship with unemployment 2 years before diagnosis (OR = 0.90, 95% CI 0.84–0.97, P < 0.01); while, no association was seen for non-screening-associated cancers (OR = 1.02, 95% CI 0.95–1.10, NS). Among patients who were diagnosed with breast or colorectal cancers, the percentages of the various stages at diagnosis were similar among the employed and unemployed populations (Table S1, available as Supplementary data at Occupational Medicine Online). Figure 1. View largeDownload slide Unemployment rates according to cancer sites (screening-associated cancers are presented in the left) among patients who were diagnosed with cancer (black) compared to match persons free cancer population (grey). Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–2007. Figure 1. View largeDownload slide Unemployment rates according to cancer sites (screening-associated cancers are presented in the left) among patients who were diagnosed with cancer (black) compared to match persons free cancer population (grey). Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–2007. Table 2. Associations between baseline unemployment a and age, gender, ethnicity, education, residential socioeconomic and cancer, Israel, 2000–07 Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS NS, not significant. aIncludes students, soldiers, housewives and unemployed. bBased on the town/city of residence, according to a national classification of 10 clusters by geographical units. cBreast, prostate, colon and cervix cancers. View Large Table 2. Associations between baseline unemployment a and age, gender, ethnicity, education, residential socioeconomic and cancer, Israel, 2000–07 Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS NS, not significant. aIncludes students, soldiers, housewives and unemployed. bBased on the town/city of residence, according to a national classification of 10 clusters by geographical units. cBreast, prostate, colon and cervix cancers. View Large Discussion The study’s hypotheses were partially proven. The unemployment rate among patients diagnosed with cancer was lower than the matched cancer-free population (34 versus 37%, P < 0.001). In addition, in the adjusted model, patients who were diagnosed with screening-associated cancers had a decreased risk for unemployment 2 years before diagnosis (OR = 0.90, 95% CI 0.84–0.97, P < 0.01). In contrast, unemployment 2 years before diagnosis was not associated with the whole cancer population or with patients diagnosed with non-screening-associated cancers. The reduced risk for unemployment before a cancer diagnosis was modest among patients with screening-associated cancers. Yet, this effect was calculated after adjustment for various confounders, including socioeconomic variables. In addition, unemployment after a cancer diagnosis is associated with both being unemployed 2 years before diagnosis and being diagnosed with cancer [1,20]. The magnitude of the risk for being unemployed 2 years before diagnosis is currently much higher than being diagnosed with cancer (OR >6.5 versus <1.5) [1,20]. Furthermore, non-differential baseline misclassification bias can be substantial for working cessation, which have a low incidence and high prevalence [21]. The current study emphasizes the need for corrected baseline employment statuses, mainly among patients diagnosed with screening-associated cancers, in studies that assess cancer survivors’ work issues. A better health status among workers may explain the current results [15]. Workers are healthy enough to be hired initially and those who remain at work stayed healthy enough to maintain employment; whereas, general populations include people unfit for work because of impaired health [15]. Cancer screening programmes are intended for healthy persons [16], and the use of such programmes may be higher among employees. Furthermore, this population may undergo more screenings due to health promotion programmes at their workplace [22]. In addition, lower rates of unemployment may serve as a marker for better socioeconomic status. Numerous publications suggest that socioeconomic status is positively associated with access to health care [23,24], even in Israel despite the universal health care system [25]. People with low socioeconomic status may have less access to medical services such as screening [26]; while, a higher socioeconomic status may lead to improved access and utilization of diagnostic technology [23]. Indeed, patients who were diagnosed with screening-associated cancers in the current study had higher years of education and residential socioeconomic positions. Yet, lower rates of unemployment among this population were significant after adjusting for these variables. Furthermore, among patients diagnosed with breast or colorectal cancers, the distribution of stages was similar among the employed and unemployed subgroups (Table S1, available as Supplementary data at Occupational Medicine Online). Thus, the current results are not explained merely by socioeconomic status. However, residual confounding by socioeconomic status could have influenced our findings. On the other hand, being employed may be a marker for an increased risk of breast cancer among employed women. Employed women are frequently older at the time of birth of their first child or nulliparous, and both these characteristics are known risk factors for breast cancer [27]. The current study has several strengths. The study design is not associated with recall bias, which is common in classical case–control formats. In addition, using a high-quality data set and linkage to highly validated databases (Israeli Central Bureau of Statistics Israel Cancer Registry [17] and the Israeli Tax Authority database) supports the internal validity of the current study; while, using a population-based data supports the external validity. Furthermore, using these databases eliminates non-response bias [28], which is quite frequent in occupational studies [29]. The external validity of the current study to other populations depends on the health care system and insurance. Socially determined variables play a central role in labour force participation because they are the result of the interaction of the labour market, job search behaviour, economic incentives and health insurance [9]. Thus, the study’s results may vary from one country to another. Yet, Israel is considered a western democracy with regards to health and social parameters. It has a highly developed health care system [30] and a comprehensive social security system. Consequently, we assume that our results highlight the possibility of false negative results in studies evaluating the employment status of cancer survivors in other modern western countries. In conclusion, the results of the current study suggest that a higher baseline rate of employment among cancer patients, mainly among patients diagnosed with screening-associated cancers, may explain false negative results in studies assessing cancer survivors’ work issues. These findings provide solid support for the importance of controlling the baseline rate of employment in studies that compare working-related issues between cancer patients and individuals without a cancer history. Understanding the complex mechanisms whereby employment may affect a cancer diagnosis remains an important research priority. Key points This study used population-based data from over 34000 participants to investigate the association between cancer diagnosis and baseline rate of employment. We found that cancer was not associated with unemployment 2 years before diagnosis. However, the diagnoses associated with screening (breast, prostate, colorectal and cervix cancers) were inversely associated with a decreased risk of unemployment 2 years before diagnosis. A higher baseline rate of employment among cancer patients, mainly those who were diagnosed with screening-associated cancers, may explains false negative results in previous studies assessing cancer survivors’ work issues. Funding This work was supported by the Israel Cancer Association Grant – The Ethel Cohen Memorial Fund (2017). Competing interests None declared. References 1. Rottenberg Y , Ratzon NZ , Jacobs JM , Cohen M , Peretz T , de Boer AG . Unemployment risk and income change after testicular cancer diagnosis: a population-based study . Urol Oncol 2016 ; 34 : 5.e27 – 5.e33 . Google Scholar CrossRef Search ADS 2. Vrkljan BH , Miller-Polgar J . Meaning of occupational engagement in life-threatening illness: a qualitative pilot project . Can J Occup Ther 2001 ; 68 : 237 – 246 . Google Scholar CrossRef Search ADS PubMed 3. Hoffman B . Cancer survivors’ employment and insurance rights: a primer for oncologists . Oncology (Williston Park) 1999 ; 13 : 841 – 846 ; discussion 846, 849, 852. Google Scholar PubMed 4. Sanchez KM , Richardson JL , Mason HR . The return to work experiences of colorectal cancer survivors . AAOHN J 2004 ; 52 : 500 – 510 . Google Scholar PubMed 5. Luker K , Campbell M , Amir Z , Davies L . A UK survey of the impact of cancer on employment . Occup Med (Lond) 2013 ; 63 : 494 – 500 . Google Scholar CrossRef Search ADS PubMed 6. Verdecchia A , Francisci S , Brenner H et al. ; EUROCARE-4 Working Group . Recent cancer survival in Europe: a 2000–02 period analysis of EUROCARE-4 data . Lancet Oncol 2007 ; 8 : 784 – 796 . Google Scholar CrossRef Search ADS PubMed 7. van Muijen P , Duijts SF , Kornet-van der Aa DA , van der Beek AJ , Anema JR . Work disability assessment of cancer survivors: insurance physicians’ perspectives . Occup Med (Lond) 2015 ; 65 : 558 – 563 . Google Scholar CrossRef Search ADS PubMed 8. Brown RF , Owens M , Bradley C . Employee to employer communication skills: balancing cancer treatment and employment . Psychooncology 2013 ; 22 : 426 – 433 . Google Scholar PubMed 9. de Boer AG , Taskila T , Ojajärvi A , van Dijk FJ , Verbeek JH . Cancer survivors and unemployment: a meta-analysis and meta-regression . J Am Med Assoc 2009 ; 301 : 753 – 762 . Google Scholar CrossRef Search ADS 10. Taskila-Brandt T , Martikainen R , Virtanen SV , Pukkala E , Hietanen P , Lindbohm ML . The impact of education and occupation on the employment status of cancer survivors . Eur J Cancer 2004 ; 40 : 2488 – 2493 . Google Scholar CrossRef Search ADS PubMed 11. Mehnert A , de Boer A , Feuerstein M . Employment challenges for cancer survivors . Cancer 2013 ; 119 ( Suppl. 11 ): 2151 – 2159 . Google Scholar CrossRef Search ADS PubMed 12. Schuring M , Robroek SJ , Otten FW , Arts CH , Burdorf A . The effect of ill health and socioeconomic status on labor force exit and re-employment: a prospective study with ten years follow-up in the Netherlands . Scand J Work Environ Health 2013 ; 39 : 134 – 143 . Google Scholar CrossRef Search ADS PubMed 13. Richardson DB , Cardis E , Daniels RD et al. Risk of cancer from occupational exposure to ionising radiation: retrospective cohort study of workers in France, the United Kingdom, and the United States (INWORKS) . Br Med J 2015 ; 351 : h5359 . Google Scholar CrossRef Search ADS 14. Ollier M , Chamoux A , Naughton G , Pereira B , Dutheil F . Chest CT scan screening for lung cancer in asbestos occupational exposure: a systematic review and meta-analysis . Chest 2014 ; 145 : 1339 – 1346 . Google Scholar CrossRef Search ADS PubMed 15. Thygesen LC , Hvidtfeldt UA , Mikkelsen S , Brønnum-Hansen H . Quantification of the healthy worker effect: a nationwide cohort study among electricians in Denmark . BMC Public Health 2011 ; 11 : 571 . Google Scholar CrossRef Search ADS PubMed 16. Smith RA , Manassaram-Baptiste D , Brooks D et al. Cancer screening in the United States, 2015: a review of current American cancer society guidelines and current issues in cancer screening . CA Cancer J Clin 2015 ; 65 : 30 – 54 . Google Scholar CrossRef Search ADS PubMed 17. Barchana M , Liphshitz I , Rozen P . Trends in colorectal cancer incidence and mortality in the Israeli Jewish ethnic populations . Fam Cancer 2004 ; 3 : 207 – 214 . Google Scholar CrossRef Search ADS PubMed 18. Baron-Epel O , Garty N , Green MS . Inequalities in use of health services among Jews and Arabs in Israel . Health Serv Res 2007 ; 42 : 1008 – 1019 . Google Scholar CrossRef Search ADS PubMed 19. Levine H , Afek A , Shamiss A et al. Risk of germ cell testicular cancer according to origin: a migrant cohort study in 1,100,000 Israeli men . Int J Cancer 2013 ; 132 : 1878 – 1885 . Google Scholar CrossRef Search ADS PubMed 20. Ratzon NZ , Uziely B , de Boer AG , Rottenberg Y . Unemployment risk and decreased income two and four years after thyroid cancer diagnosis: a population-based study . Thyroid 2016 ; 26 : 1251 – 1258 . Google Scholar CrossRef Search ADS PubMed 21. Jurek AM , Greenland S , Maldonado G . How far from non-differential does exposure or disease misclassification have to be to bias measures of association away from the null ? Int J Epidemiol 2008 ; 37 : 382 – 385 . Google Scholar CrossRef Search ADS PubMed 22. Allweiss P , Brown DR , Chosewood LC et al. ; Cancer Prevention in the Workplace Writing Group . Cancer prevention and worksite health promotion: time to join forces . Prev Chronic Dis 2014 ; 11 : E128 . Google Scholar PubMed 23. Li N , Du XL , Reitzel LR , Xu L , Sturgis EM . Impact of enhanced detection on the increase in thyroid cancer incidence in the United States: review of incidence trends by socioeconomic status within the surveillance, epidemiology, and end results registry, 1980–2008 . Thyroid 2013 ; 23 : 103 – 110 . Google Scholar CrossRef Search ADS PubMed 24. Addo J , Ayerbe L , Mohan KM et al. Socioeconomic status and stroke: an updated review . Stroke 2012 ; 43 : 1186 – 1191 . Google Scholar CrossRef Search ADS PubMed 25. Filc D , Davidovich N , Novack L , Balicer RD . Is socioeconomic status associated with utilization of health care services in a single-payer universal health care system ? Int J Equity Health 2014 ; 13 : 115 . Google Scholar CrossRef Search ADS PubMed 26. Damiani G , Federico B , Basso D et al. Socioeconomic disparities in the uptake of breast and cervical cancer screening in Italy: a cross sectional study . BMC Public Health 2012 ; 12 : 99 . Google Scholar CrossRef Search ADS PubMed 27. Howlader N , Altekruse SF , Li CI et al. US incidence of breast cancer subtypes defined by joint hormone receptor and HER2 status . J Natl Cancer Inst 2014 ; 106 (5). pii: dju055. doi:10.1093/jnci/dju055. 28. de Winter AF , Oldehinkel AJ , Veenstra R , Brunnekreef JA , Verhulst FC , Ormel J . Evaluation of non-response bias in mental health determinants and outcomes in a large sample of pre-adolescents . Eur J Epidemiol 2005 ; 20 : 173 – 181 . Google Scholar CrossRef Search ADS PubMed 29. Martikainen P , Laaksonen M , Piha K , Lallukka T . Does survey non-response bias the association between occupational social class and health ? Scand J Public Health 2007 ; 35 : 212 – 215 . Google Scholar CrossRef Search ADS PubMed 30. Chernichovsky D . Not ‘socialized medicine’—an Israeli view of health care reform . N Engl J Med 2009 ; 361 : e46 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Occupational Medicine. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Occupational Medicine Oxford University Press

Higher incidence of screening-related cancers in the employed population

Loading next page...
 
/lp/ou_press/higher-incidence-of-screening-related-cancers-in-the-employed-W4T76kzETb
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Occupational Medicine. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0962-7480
eISSN
1471-8405
D.O.I.
10.1093/occmed/kqy055
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Employment may confound the risk of a cancer diagnosis in both directions. We hypothesized that a higher baseline rate of employment among cancer patients may explain the lack of association between a cancer diagnosis and later unemployment in many studies. Aims To assess the unemployment rate among cancer patients before diagnosis compared with a matched cancer-free control group. Methods Using data from the Israeli National Central Bureau of Statistics 1995 census (persons aged between 15 and 60 years old), the Israeli Tax Authority database and the Israel Cancer Registry, cancer patients (diagnosed between the years 2000 and 2007 and alive at 2011) were compared with matched cancer-free controls. Results There were 8797 cancer patients and 26166 cancer-free controls. We found that, in general, cancer was not associated with unemployment 2 years before diagnosis (adjusted odds ratio [OR] = 0.96, 95% confidence interval [CI] 0.90–1.009, P = NS) after adjustment for age, gender, ethnicity, educational years and residential socioeconomic position. However, the diagnoses associated with screening (breast, prostate, colorectal and cervix cancers) were inversely associated with unemployment 2 years before diagnosis (adjusted OR = 0.90, 95% CI 0.84–0.97, P < 0.01). Conclusions The results from the current study suggest that a higher baseline rate of employment among cancer patients, mainly those who were diagnosed with screening-associated cancers, explains false negative results in previous studies assessing cancer survivors’ work issues. Cancer, case–control study, employment Introduction Working is an indicator of a sense of normality, improved quality of life and is considered helpful when coping with cancer experiences [1,2]. On the other hand, unemployment or a delay in returning to work affects one’s salary, health and welfare, the family budget and relationships at work places [3,4]. Improvement in long-term survival has led to a growing population of cancer survivors [5,6], and about half of the 3.45 million new cancer cases reported in Europe in 2012 involved people of working age [7]. Data about the unemployment of people living with a history of cancer are highly variable among studies. Different unemployment rates may reflect variances in social security systems, health care insurance, variation in laws that protect individuals from discrimination and the implementation of these laws [8] and the prevailing unemployment rate in a specific country or region [9]. Additional variables such as age [1], gender [8] and education [10,11] may also affect the rates of unemployment following a cancer diagnosis. A meta-analysis [9] found that cancer survivors are more likely to be unemployed than healthy control participants (34 versus 15%, relative risk = 1.37; 95% confidence interval [CI] 1.21–1.55). Yet, after adjustment for the diagnosis, age and prevailing unemployment rate, this association weakened and lost significance [9]. People who are employed have distinct characteristics that may affect the risk of a cancer diagnosis. These include a better socioeconomic status [12], but also an increased risk for occupational exposures [13,14]. In addition, the employed may be in better health than the general population [15], since relatively healthy individuals are likely to gain employment and remain employed; while, chronically disabled and severely ill individuals are frequently not employed [15]. Since cancer screening is aimed at healthy persons [16] and is strongly associated with socioeconomic status [16], workers may have an increased risk for being diagnosed with a screening-associated cancer. We hypothesized that the baseline rate of unemployment is lower among cancer patients compared with cancer-free persons. Furthermore, we hypothesized that this difference is greater among patients who were diagnosed with screening-associated cancers. The difference in baseline rates of employment may explain the lack of association between a cancer diagnosis and unemployment in many studies. Taskila et al. [10] found that the employment rate among cancer patients before diagnosis (78%) was the same as in the cancer-free population. However, stratification according to cancer types was not available. Thus, the aim of this study was to (i) assess the unemployment rate among cancer patients before diagnosis compared with a matched cancer-free control group, (ii) assess the unemployment rate among persons diagnosed with a cancer associated with comprehensive screening programmes compared with a matched cancer-free control group and (iii) examine the association between a cancer diagnosis and baseline unemployment rates after controlling for potential confounders. The use of large national databases allowed us to adjust the results for potential confounders and reduce selection and information biases. Methods This was a case–control study, which included baseline measurements from the Israeli Central Bureau of Statistics 1995 census. The study included a representative sample of the whole population who completed a comprehensive interview (1113420 persons, which represents 20% of the population in Israel). Data on cancer incidence were obtained from the Israel National Cancer Registry. By law, the Israel Cancer Registry has been notified of cancer incidences since 1982. The registry receives compulsory notification from numerous data sources, including pathology reports, discharge summaries and death certificates. Completeness of the registry was about 95% for solid tumours [17]. Melanoma cases were excluded from the analyses since melanoma may be a direct effect of being a worker. Persons aged 15 up to age 60 years old at the time of interview (1995) were included in the current study. Thus, the minimal age of the patients at the time of cancer diagnosis (between the years 2000 and 2007) was 20 years old. Matched groups in a 3:1 ratio according to 5-year age groups, gender and ethnicity (Jewish versus non-Jewish) were sampled from the population in the census who completed the comprehensive interview. The stratification according to ethnicity was done due to lower socioeconomic status and more negative health outcomes of the non-Jewish community compared with the Jewish population [18]. Persons who were diagnosed with cancer were not included in the cancer-free group, and participants who died during the study period (until the end of 2011) were excluded from the current study. Employment status was classified as employed or unemployed. The latter group included unemployed persons, homemakers, students, military conscripts, those on disability pension, retired individuals or unknown. Employed was defined as any reported annual income (>$0). Variables assessed in relation to being unemployed before a cancer diagnosis included age (at the time of cancer diagnosis or in the case of non-cancer healthy control participants, the age at the time of diagnosis of his/her matched cancer patient), gender, number of education years in 1995 (continuous variable), residential socioeconomic position (based on the town/city of residence, according to a national classification of 10 clusters by geographical units) [19] and cancer. Non-specific symptoms that appear in the year prior to the cancer diagnosis may increase the risk of being unemployed. Thus, employment status at baseline was defined 2 years prior to the diagnosis of cancer (or at the same year of the matched cancer patient for non-cancer controls) using the Israeli Tax Authority database. The data were derived from individuals’ tax returns filed with the Tax Authority, which receives yearly compulsory reports from each individual. Data since 1998 were available for the current study. Thus, we included only patients who were diagnosed with a cancer after 2000. We further stratified the study’s population into (i) patients who were diagnosed with a screening-associated cancer (breast, prostate, colorectal and cervix); (ii) patients who were diagnosed with a cancer not associated with screening and (iii) a healthy population (reference group). Lung cancer was not included in the group of cancers associated with comprehensive screening programmes since screening was not common during the study period (2000–07). The current study was approved by the Committee on Human Research at the Hadassah-Hebrew University Medical Center. For analysis of the study population’s characteristics, continuous variables were compared by analysis of variance test and categorical variables were compared with the Wald test and the likelihood-ratio test statistics. A binary logistic regression analysis was constructed to predict the risk of not being employed at baseline using the likelihood-ratio test statistics. We used adjusted analyses to determine the predictive values of the independent variables (age, gender, ethnicity, education, residential socioeconomic position and cancer) and the dependent variable (not being employed at baseline). In addition, we compared the stages at cancer diagnoses between the employed and unemployed populations, who were diagnosed with breast or colorectal cancers. All statistical tests were two-sided, with P <0.05 considered statistically significant. The SPSS programme (18th version; Chicago, IL) was used for the statistical analysis. Results After excluding participants who were diagnosed with melanoma, the study included 26166 match controls and 8797 cancer patients. The most frequent cancers were breast (n = 2365), prostate (n = 938), colorectal (n = 881), cervix (n = 565), bladder (n = 504), thyroid (n = 480) and lymphoma (n = 447). Cancers that are associated with comprehensive screening programmes (breast, prostate, colorectal and cervix) were reported in more than half (n = 4749, 54%) of the cancer patients. The baseline characteristics of the study population are presented in Table 1. Overall, the absolute differences between persons without a history of cancer and cancer patients were minor for all baselines variables. In contrast, patients who were diagnosed with a screening-associated cancer differed from individuals without a history of cancer. The former group was often female and Jewish and was more literate with higher resident socioeconomic positions. In cancer patients, the unemployment rate 2 years before the diagnosis was 34% (n = 3002) compared with 37% (n = 9550) among persons without a history of cancer (P < 0.001). Rates of 34% were found for patients who were diagnosed with both screening-associated (n = 1621) and non-screening-associated (n = 1381) cancers. Table 1. Characteristics of participants in the comprehensive interview of the Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–07 Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) aOrdinal variable based on the town/city of residence, according to a national classification of 10 clusters by geographical units. bBreast, prostate, colon and cervix cancers. View Large Table 1. Characteristics of participants in the comprehensive interview of the Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–07 Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) Variables Cancer patients compared with cancer-free group Screening-associated cancer patients compared with non-screening-associated cancers and cancer-free group Positive history of cancer (n = 8797), n (%) No history of cancer (n = 26166), n (%) Screening- associated cancerb (n = 4749), n (%) Non-screening- associated cancer (n = 4048), n (%) No history of cancer (n = 26166), n (%) Age, mean (SD) 52.9 (SD 12) 53.3 (SD 11) 54.0 (SD 11) 51.7 (SD 12) 53.3 (SD 11) Gender, males 3327 (37) 10368 (39) 1398 (29) 1930 (47) 10368 (39) Non-Jewish 939 (10) 2920 (11) 432 (9) 507 (12) 3859 (11) Educational years, mean (SD) 12.1 (SD 4) 11.8 (SD 4) 12.4 (SD 4) 11.9 (SD 4) 11.8 (SD 4) Unemployment 2 years before diagnosis 3002 (34) 9550 (36) 1621 (34) 1381 (34) 9550 (36) Residential socioeconomic positiona, mean (SD) 6.2 (SD 2) 6.1 (SD 2) 6.4 (SD 2) 6.1 (SD 2) 6.1 (SD 2) aOrdinal variable based on the town/city of residence, according to a national classification of 10 clusters by geographical units. bBreast, prostate, colon and cervix cancers. View Large Unemployment rates were lower for patients with all screening-associated cancers when compared with the matched cancer-free population (Figure 1). Yet, significant changes in unemployment rates were only found for patients who were diagnosed with breast cancer (36 versus 40% for the matched cancer-free population, P < 0.01) and prostate cancer (24 versus 26% for the matched cancer-free population, P < 0.05). Among patients who were diagnosed with non-screening-associated cancers, differences between cancer patients and matched cancer-free population were small and uniformly non-significant. In the adjusted models (Table 2), unemployment 2 years before diagnosis was associated with age (increased risk), gender (lower risk in men), ethnicity (lower risk in Jewish individuals), educational years (inverse) and residential socioeconomic position (inverse). In the first model, cancer was not associated with unemployment 2 years before diagnosis (OR = 0.96, 95% CI 0.90–1.009, NS). In the second model, cancer patients were divided into those with (i) screening-associated cancer and (ii) non-screening-associated cancers. Only screening-associated cancer had an inverse relationship with unemployment 2 years before diagnosis (OR = 0.90, 95% CI 0.84–0.97, P < 0.01); while, no association was seen for non-screening-associated cancers (OR = 1.02, 95% CI 0.95–1.10, NS). Among patients who were diagnosed with breast or colorectal cancers, the percentages of the various stages at diagnosis were similar among the employed and unemployed populations (Table S1, available as Supplementary data at Occupational Medicine Online). Figure 1. View largeDownload slide Unemployment rates according to cancer sites (screening-associated cancers are presented in the left) among patients who were diagnosed with cancer (black) compared to match persons free cancer population (grey). Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–2007. Figure 1. View largeDownload slide Unemployment rates according to cancer sites (screening-associated cancers are presented in the left) among patients who were diagnosed with cancer (black) compared to match persons free cancer population (grey). Israeli National Central Bureau of Statistics 1995 census who developed cancer and matched cancer-free group, Israel, 2000–2007. Table 2. Associations between baseline unemployment a and age, gender, ethnicity, education, residential socioeconomic and cancer, Israel, 2000–07 Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS NS, not significant. aIncludes students, soldiers, housewives and unemployed. bBased on the town/city of residence, according to a national classification of 10 clusters by geographical units. cBreast, prostate, colon and cervix cancers. View Large Table 2. Associations between baseline unemployment a and age, gender, ethnicity, education, residential socioeconomic and cancer, Israel, 2000–07 Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS Model controlling for cancer Model controlling also for screening-associated cancer Odds ratio P value Odds ratio P value Age (per year) 1.06 (1.05–1.06) <0.001 1.06 (1.05–1.06) <0.001 Gender (males versus females) 0.49 (0.47–0.52) <0.001 0.49 (0.47–0.52) <0.001 Ethnicity (Jewish versus non-Jewish) 0.41 (0.38–0.45) <0.001 0.41 (0.41–0.38) <0.001 Educational year (per year) 0.88 (0.88–0.89) <0.001 0.88 (0.88–0.89) <0.001 Residential socioeconomic positionb 0.93 (0.92–0.95) <0.001 0.94 (0.92–0.95) <0.001 No cancer (n = 26166) 1 (reference) 1 (reference) Cancer diagnosis (n = 8797) 0.96 (0.90–1.00) NS Screening-associated cancerc (n = 4749) 0.90 (0.84–0.97) <0.05 Non-screening-associated cancer (n = 4048) 1.02 (0.9–1.1) NS NS, not significant. aIncludes students, soldiers, housewives and unemployed. bBased on the town/city of residence, according to a national classification of 10 clusters by geographical units. cBreast, prostate, colon and cervix cancers. View Large Discussion The study’s hypotheses were partially proven. The unemployment rate among patients diagnosed with cancer was lower than the matched cancer-free population (34 versus 37%, P < 0.001). In addition, in the adjusted model, patients who were diagnosed with screening-associated cancers had a decreased risk for unemployment 2 years before diagnosis (OR = 0.90, 95% CI 0.84–0.97, P < 0.01). In contrast, unemployment 2 years before diagnosis was not associated with the whole cancer population or with patients diagnosed with non-screening-associated cancers. The reduced risk for unemployment before a cancer diagnosis was modest among patients with screening-associated cancers. Yet, this effect was calculated after adjustment for various confounders, including socioeconomic variables. In addition, unemployment after a cancer diagnosis is associated with both being unemployed 2 years before diagnosis and being diagnosed with cancer [1,20]. The magnitude of the risk for being unemployed 2 years before diagnosis is currently much higher than being diagnosed with cancer (OR >6.5 versus <1.5) [1,20]. Furthermore, non-differential baseline misclassification bias can be substantial for working cessation, which have a low incidence and high prevalence [21]. The current study emphasizes the need for corrected baseline employment statuses, mainly among patients diagnosed with screening-associated cancers, in studies that assess cancer survivors’ work issues. A better health status among workers may explain the current results [15]. Workers are healthy enough to be hired initially and those who remain at work stayed healthy enough to maintain employment; whereas, general populations include people unfit for work because of impaired health [15]. Cancer screening programmes are intended for healthy persons [16], and the use of such programmes may be higher among employees. Furthermore, this population may undergo more screenings due to health promotion programmes at their workplace [22]. In addition, lower rates of unemployment may serve as a marker for better socioeconomic status. Numerous publications suggest that socioeconomic status is positively associated with access to health care [23,24], even in Israel despite the universal health care system [25]. People with low socioeconomic status may have less access to medical services such as screening [26]; while, a higher socioeconomic status may lead to improved access and utilization of diagnostic technology [23]. Indeed, patients who were diagnosed with screening-associated cancers in the current study had higher years of education and residential socioeconomic positions. Yet, lower rates of unemployment among this population were significant after adjusting for these variables. Furthermore, among patients diagnosed with breast or colorectal cancers, the distribution of stages was similar among the employed and unemployed subgroups (Table S1, available as Supplementary data at Occupational Medicine Online). Thus, the current results are not explained merely by socioeconomic status. However, residual confounding by socioeconomic status could have influenced our findings. On the other hand, being employed may be a marker for an increased risk of breast cancer among employed women. Employed women are frequently older at the time of birth of their first child or nulliparous, and both these characteristics are known risk factors for breast cancer [27]. The current study has several strengths. The study design is not associated with recall bias, which is common in classical case–control formats. In addition, using a high-quality data set and linkage to highly validated databases (Israeli Central Bureau of Statistics Israel Cancer Registry [17] and the Israeli Tax Authority database) supports the internal validity of the current study; while, using a population-based data supports the external validity. Furthermore, using these databases eliminates non-response bias [28], which is quite frequent in occupational studies [29]. The external validity of the current study to other populations depends on the health care system and insurance. Socially determined variables play a central role in labour force participation because they are the result of the interaction of the labour market, job search behaviour, economic incentives and health insurance [9]. Thus, the study’s results may vary from one country to another. Yet, Israel is considered a western democracy with regards to health and social parameters. It has a highly developed health care system [30] and a comprehensive social security system. Consequently, we assume that our results highlight the possibility of false negative results in studies evaluating the employment status of cancer survivors in other modern western countries. In conclusion, the results of the current study suggest that a higher baseline rate of employment among cancer patients, mainly among patients diagnosed with screening-associated cancers, may explain false negative results in studies assessing cancer survivors’ work issues. These findings provide solid support for the importance of controlling the baseline rate of employment in studies that compare working-related issues between cancer patients and individuals without a cancer history. Understanding the complex mechanisms whereby employment may affect a cancer diagnosis remains an important research priority. Key points This study used population-based data from over 34000 participants to investigate the association between cancer diagnosis and baseline rate of employment. We found that cancer was not associated with unemployment 2 years before diagnosis. However, the diagnoses associated with screening (breast, prostate, colorectal and cervix cancers) were inversely associated with a decreased risk of unemployment 2 years before diagnosis. A higher baseline rate of employment among cancer patients, mainly those who were diagnosed with screening-associated cancers, may explains false negative results in previous studies assessing cancer survivors’ work issues. Funding This work was supported by the Israel Cancer Association Grant – The Ethel Cohen Memorial Fund (2017). Competing interests None declared. References 1. Rottenberg Y , Ratzon NZ , Jacobs JM , Cohen M , Peretz T , de Boer AG . Unemployment risk and income change after testicular cancer diagnosis: a population-based study . Urol Oncol 2016 ; 34 : 5.e27 – 5.e33 . Google Scholar CrossRef Search ADS 2. Vrkljan BH , Miller-Polgar J . Meaning of occupational engagement in life-threatening illness: a qualitative pilot project . Can J Occup Ther 2001 ; 68 : 237 – 246 . Google Scholar CrossRef Search ADS PubMed 3. Hoffman B . Cancer survivors’ employment and insurance rights: a primer for oncologists . Oncology (Williston Park) 1999 ; 13 : 841 – 846 ; discussion 846, 849, 852. Google Scholar PubMed 4. Sanchez KM , Richardson JL , Mason HR . The return to work experiences of colorectal cancer survivors . AAOHN J 2004 ; 52 : 500 – 510 . Google Scholar PubMed 5. Luker K , Campbell M , Amir Z , Davies L . A UK survey of the impact of cancer on employment . Occup Med (Lond) 2013 ; 63 : 494 – 500 . Google Scholar CrossRef Search ADS PubMed 6. Verdecchia A , Francisci S , Brenner H et al. ; EUROCARE-4 Working Group . Recent cancer survival in Europe: a 2000–02 period analysis of EUROCARE-4 data . Lancet Oncol 2007 ; 8 : 784 – 796 . Google Scholar CrossRef Search ADS PubMed 7. van Muijen P , Duijts SF , Kornet-van der Aa DA , van der Beek AJ , Anema JR . Work disability assessment of cancer survivors: insurance physicians’ perspectives . Occup Med (Lond) 2015 ; 65 : 558 – 563 . Google Scholar CrossRef Search ADS PubMed 8. Brown RF , Owens M , Bradley C . Employee to employer communication skills: balancing cancer treatment and employment . Psychooncology 2013 ; 22 : 426 – 433 . Google Scholar PubMed 9. de Boer AG , Taskila T , Ojajärvi A , van Dijk FJ , Verbeek JH . Cancer survivors and unemployment: a meta-analysis and meta-regression . J Am Med Assoc 2009 ; 301 : 753 – 762 . Google Scholar CrossRef Search ADS 10. Taskila-Brandt T , Martikainen R , Virtanen SV , Pukkala E , Hietanen P , Lindbohm ML . The impact of education and occupation on the employment status of cancer survivors . Eur J Cancer 2004 ; 40 : 2488 – 2493 . Google Scholar CrossRef Search ADS PubMed 11. Mehnert A , de Boer A , Feuerstein M . Employment challenges for cancer survivors . Cancer 2013 ; 119 ( Suppl. 11 ): 2151 – 2159 . Google Scholar CrossRef Search ADS PubMed 12. Schuring M , Robroek SJ , Otten FW , Arts CH , Burdorf A . The effect of ill health and socioeconomic status on labor force exit and re-employment: a prospective study with ten years follow-up in the Netherlands . Scand J Work Environ Health 2013 ; 39 : 134 – 143 . Google Scholar CrossRef Search ADS PubMed 13. Richardson DB , Cardis E , Daniels RD et al. Risk of cancer from occupational exposure to ionising radiation: retrospective cohort study of workers in France, the United Kingdom, and the United States (INWORKS) . Br Med J 2015 ; 351 : h5359 . Google Scholar CrossRef Search ADS 14. Ollier M , Chamoux A , Naughton G , Pereira B , Dutheil F . Chest CT scan screening for lung cancer in asbestos occupational exposure: a systematic review and meta-analysis . Chest 2014 ; 145 : 1339 – 1346 . Google Scholar CrossRef Search ADS PubMed 15. Thygesen LC , Hvidtfeldt UA , Mikkelsen S , Brønnum-Hansen H . Quantification of the healthy worker effect: a nationwide cohort study among electricians in Denmark . BMC Public Health 2011 ; 11 : 571 . Google Scholar CrossRef Search ADS PubMed 16. Smith RA , Manassaram-Baptiste D , Brooks D et al. Cancer screening in the United States, 2015: a review of current American cancer society guidelines and current issues in cancer screening . CA Cancer J Clin 2015 ; 65 : 30 – 54 . Google Scholar CrossRef Search ADS PubMed 17. Barchana M , Liphshitz I , Rozen P . Trends in colorectal cancer incidence and mortality in the Israeli Jewish ethnic populations . Fam Cancer 2004 ; 3 : 207 – 214 . Google Scholar CrossRef Search ADS PubMed 18. Baron-Epel O , Garty N , Green MS . Inequalities in use of health services among Jews and Arabs in Israel . Health Serv Res 2007 ; 42 : 1008 – 1019 . Google Scholar CrossRef Search ADS PubMed 19. Levine H , Afek A , Shamiss A et al. Risk of germ cell testicular cancer according to origin: a migrant cohort study in 1,100,000 Israeli men . Int J Cancer 2013 ; 132 : 1878 – 1885 . Google Scholar CrossRef Search ADS PubMed 20. Ratzon NZ , Uziely B , de Boer AG , Rottenberg Y . Unemployment risk and decreased income two and four years after thyroid cancer diagnosis: a population-based study . Thyroid 2016 ; 26 : 1251 – 1258 . Google Scholar CrossRef Search ADS PubMed 21. Jurek AM , Greenland S , Maldonado G . How far from non-differential does exposure or disease misclassification have to be to bias measures of association away from the null ? Int J Epidemiol 2008 ; 37 : 382 – 385 . Google Scholar CrossRef Search ADS PubMed 22. Allweiss P , Brown DR , Chosewood LC et al. ; Cancer Prevention in the Workplace Writing Group . Cancer prevention and worksite health promotion: time to join forces . Prev Chronic Dis 2014 ; 11 : E128 . Google Scholar PubMed 23. Li N , Du XL , Reitzel LR , Xu L , Sturgis EM . Impact of enhanced detection on the increase in thyroid cancer incidence in the United States: review of incidence trends by socioeconomic status within the surveillance, epidemiology, and end results registry, 1980–2008 . Thyroid 2013 ; 23 : 103 – 110 . Google Scholar CrossRef Search ADS PubMed 24. Addo J , Ayerbe L , Mohan KM et al. Socioeconomic status and stroke: an updated review . Stroke 2012 ; 43 : 1186 – 1191 . Google Scholar CrossRef Search ADS PubMed 25. Filc D , Davidovich N , Novack L , Balicer RD . Is socioeconomic status associated with utilization of health care services in a single-payer universal health care system ? Int J Equity Health 2014 ; 13 : 115 . Google Scholar CrossRef Search ADS PubMed 26. Damiani G , Federico B , Basso D et al. Socioeconomic disparities in the uptake of breast and cervical cancer screening in Italy: a cross sectional study . BMC Public Health 2012 ; 12 : 99 . Google Scholar CrossRef Search ADS PubMed 27. Howlader N , Altekruse SF , Li CI et al. US incidence of breast cancer subtypes defined by joint hormone receptor and HER2 status . J Natl Cancer Inst 2014 ; 106 (5). pii: dju055. doi:10.1093/jnci/dju055. 28. de Winter AF , Oldehinkel AJ , Veenstra R , Brunnekreef JA , Verhulst FC , Ormel J . Evaluation of non-response bias in mental health determinants and outcomes in a large sample of pre-adolescents . Eur J Epidemiol 2005 ; 20 : 173 – 181 . Google Scholar CrossRef Search ADS PubMed 29. Martikainen P , Laaksonen M , Piha K , Lallukka T . Does survey non-response bias the association between occupational social class and health ? Scand J Public Health 2007 ; 35 : 212 – 215 . Google Scholar CrossRef Search ADS PubMed 30. Chernichovsky D . Not ‘socialized medicine’—an Israeli view of health care reform . N Engl J Med 2009 ; 361 : e46 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Occupational Medicine. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Occupational MedicineOxford University Press

Published: Apr 4, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off