Cross-sectional association between working and depression prevalence in cancer survivors: a literature review

Abstract

Objective: To examine the hypotheses that the depression prevalence would be lower in working cancer survivors compared to non-working cancer survivors and that the depression prevalence of working cancer survivors could be as low as that of the cancer-free general population. Methods: We used the PubMed database to search for relevant literature. Out of 299 matches to the used terms, 17 cross-sectional, quantitative studies that compared depression statuses of working cancer survivors with non-working cancer survivors or cancer-free general samples were considered. Results: In some studies, survivors of breast cancer, Hodgkin lymphoma, and hepatocellular carcinoma indicated a significantly lower depression prevalence compared to non-working survivors with the same cancer types. On the other hand, some studies did not demonstrate a significant difference in the association between depression prevalence and working status among breast cancer and head and neck cancer survivors. Working survivors of breast cancer and malignant brain tumor demonstrated significantly worse scores on the depression subscale of the Hospital Anxiety and Depression Scale (HADS-D) compared to the cancer-free general population. However, the depression prevalence was not compared. Conclusions: The existing literature did not consistently support our hypotheses. Working does not function as an absolute prevention method for depression in cancer survivors. Preventive measures against depression are necessary even after cancer survivors return to work. Further research is necessary to determine whether working is effective for preventing depression in cancer survivors.

Introduction

Given the current advances of the 21^st century, cancer is progressively becoming less lethal. In Japan, age-adjusted cancer mortality has been decreasing since the mid-1990s¹⁾. The age-adjusted mortality rates were 226.1 and 108.3 per 100,000 in 1995 for men and women, respectively, while they were 157.5 and 85.0 in 2017¹⁾. The latest 5-year relative survival rate for all types of cancer is estimated at 62.1%¹⁾.

Cancer comorbidities require extensive attention, depression being one of the most common. The pooled prevalence rate of depressive symptoms was estimated as 21% in long-term cancer survivors²⁾, while that of mood disorders was 5.4% in global general populations³⁾. Depression increases all-cause mortality⁴⁾ and the incidence of suicide^5,6) in cancer survivors.

Cancer patients are likely to leave work after getting a cancer diagnosis. They may feel isolated from society, deteriorating their mental health⁷⁾. Returning to work, acting as reinsertion in society, is expected to improve their health. The Japanese government enforced regulations to assist workers in finding balance between work and treatment in the Cancer Control Act¹⁾. Systematic reviews have elucidated the return-to-work rates among cancer survivors ranging from 53.8% to 95.2% in Japan⁸⁾, from 39% to 77% in European countries⁹⁾, and from 24% to 94% (mean: 63.5%) worldwide¹⁰⁾. However, existing evidence is scarce to conclude that working prevents depression in cancer survivors.

A literature review was performed to describe the associations between working and depression in cancer survivors. Our hypotheses were that (i) depression prevalence would be lower in working cancer survivors compared to non-working cancer survivors and (ii) cancer survivors would have depression prevalence as low as that of the cancer-free general population. These hypotheses would be confirmed if returning to and continuing work were effective for preventing depression in cancer survivors.

Materials and Methods

We searched for the existing literature through the PubMed database. The synthesis of search terms was “depression” AND (“survivorship” OR “survivors”) AND (“work” OR “employment”). We searched for papers published between 2005 and 2019. Inclusion criteria for the present review were as follows: (1) cross-sectional, quantitative surveys comparing working cancer survivors with non-working cancer survivors or with the cancer-free general population; (2) depression prevalence or depression measure scores presented as the outcome; (3) working-age cancer survivors (15–65 years) working at the time of the cancer diagnosis. Out of 299 matches to the used terms, duplication occurred for 97 matches. We excluded 185 papers for the following reasons: review and protocol (n=45); qualitative study (n=6); prospective cohort study, randomized controlled trial, case study, and evaluation of screening (n=22); not being written in English or Japanese (n=4); subjects not of working age or their working statuses not identified (n=28); outcome not including depression prevalence or depression measure scores (n=39); and working cancer survivors’ depression not compared to that of non-working cancer survivors or cancer-free general samples (n=41). We ultimately included 17 papers.

For each paper, we extracted the following data: sampling method (hospital-based or cancer registry-based); number, sex, and age of subjects; cancer-related information (site, stage, time since diagnosis/treatment, treatment received); working conditions at the time of survey (employment status, occupation, position); depression-related information (depression prevalence, depression measure scores); and whether a multivariable statistical analysis was employed to examine the associations between working and depression. Depression measures that were adopted in the reviewed papers included the Beck Depression Inventory (BDI)¹¹⁾, the Center for Epidemiologic Studies Depressive Symptoms Scale (CES-D)¹²⁾, the Depression subscale of the Hospital Anxiety and Depression Scale (HADS-D)¹³⁾, and the 9-item Patient Health Questionnaire (PHQ-9)¹⁴⁾. Depression was diagnosed in case of a BDI score of 19 or higher, a CES-D score of 16 or higher, or a HADS-D score of 8 or higher. The review results were summarized separately by cancer site and controls (i.e., non-working cancer survivors and the cancer-free general samples).

Because we only conducted a review of existing literature, it was unnecessary to obtain approval from an ethics review committee per the Ethical Guidelines for Medical and Health Research Involving Human Subjects established by the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare, Japan.

Results

Comparison between working and non-working cancer survivors

We found 13 papers for this comparison (Table 1 and Table 2). The results are presented below by cancer site.

Table 1. Study and subject characteristics: subjects included cancer survivors who were and were not working

Author, year, country	N, sex, and sampling method	Age, years	Cancer				Working conditions at the time of survey
			Site	Stage	Time since Dx/Tx	Tx
Breast cancer
Kim et al. (2008)15) South Korea	1,491 women H	Mean: 47.4 SD: 9.3	Breast: 100%	NR	Years since surgery Mean: 4.6 SD: 2.4	Mastectomy alone: 54.7%	Having a job: 44.9% (Self-employed: 20.7%; Employed: 24.2%)
Ashing-Giwa et al. (2013)16) U.S.	232 women R	Mean: 53 SD: 10.6	Breast: 100%	0: 5% I: 37% II: 41% III: 17%	NR	Lumpectomy: 54% Mastectomy: 41% Radiation therapy: 70% Chemotherapy: 70% Hormonal therapy: 66%	Employed: 39%
Ho et al. (2018)17) Singapore	327 women H	Median (IQR) Employed: 53 (48–58) Unemployed: 57 (50–61) p <0.001	Breast: 100%	in situ or I–II: 86.2%	Median (IQR) years since Dx Employed: 4 (3–8) Unemployed: 5 (3–8) NS	Surgery: 99% Axillary clearance: 59% Radiotherapy: 62% Adjuvant chemotherapy: 66% Hormone therapy: 69%.	Employed: 53%
Puigpinós-Riera et al. (2018)18) Catalonia	2,235 women H	<50: 14.4% 50–65: 45.5% >65: 40.1%	Breast: 100%	in situ: 7.9% Initial phase: 38.6% Locally advanced: 39.8% Metastatic: 1.1%	Years since Dx 5 or fewer: 37.7% 5–10: 42.5% >10: 19.0%	NR	Active worker: 35.3%; Not working: 20.2%; Disabled: 8.5%; Retired: 34.3%
Cervical cancer
Yoo et al. (2013)19) South Korea	858 women H	≥50 years: 66.9%.	Cervix: 100%	Stage: I–IIA: 84.8%.	Years since Tx <5: 44.1%	Chemotherapy: 28.7% Local Tx Surgery only: 72.7% Surgery + radiotherapy: 12.2% Radiotherapy only: 15.1%	Working: 49.4% (Self-employed: 24.1%; Employed: 25.3%)
Head and neck cancer
Verdonck-de Leeuw et al. (2010)20) Netherlands	85 Men: 64% H	Median: 59	Oral cavity/oropharynx: 37% Larynx: 34%	I: 38% II: 17% III: 19% IV: 27%	NR	Radiotherapy: 32% Surgery and radiotherapy: 27% Surgery: 21%	Not working at time of diagnosis: 37.6%; Return to same work: 32.9%; Changed work: 18.8%; not returning to work: 10.6%
Koch et al. (2015)21) Germany	55 Men: 80% H	Mean: 53.0 SD: 5.7	Oropharynx: 27% Larynx: 22% Mouth cavity: 15% Nasopharynx: 13%	I: 18% II: 24% III: 26% IV: 33%	Months since Dx Mean: 66.78 SD: 33.6	The employed and unemployed Surgery: 84%, 100% Neck dissection: 26%, 39% Radiotherapy: 68%, 67% Chemotherapy: 16%, 25% Multimodal therapy: 63%, 67% Rehabilitation: 58%, 81%	Of the employed (34.5%); Blue collar: 42%; White collar: 58% Full-time worker: 100% Hard physical work: 11%; Intermediate physical work: 37%; easy physical work: 16%; sedentary intellectual work: 37%
Hematologic and lymphoid cancer
Magyari et al. (2017)22) Hungary	140 Men: 49.3% H	≥40: 54%	HL: 100%	I–II: 53% III–IV: 47%	Elapsed time ≥10 years: 47%.	ABVD (adriamycin, bleomycin, vinblastine, and dacarbazine): 31%	Occupationally active: 66.4% (Employed: 55.0%; Self-employed: 7.1%; Student: 1.4%; Childbearing: 2.9%)
Mosher et al. (2011)23) U.S.	406 Men: 52% H	Mean: 49.25 SD: 12.82	MM or amyloidosis: 33.7% NHL: 22.4% AML and CML: 11.6%	NR	Months since Tx Mean: 21.23 SD: 6.22	Hematopoietic stem cell transplant: 100%	Employed: 46.6%
Kiserud et al. (2016)24) Norway	281 Men: 60% H	Mean: 52.3 SD: 11.7	Aggressive NHL: 64% HL: 25% Indolent NHL: 11%	I–II: 31% III–IV: 69%	Years since Dx Mean: 12.5 SD: 6.2	High-dose chemotherapy with autologous stem cell transplantation: 100%	Employed: 63.7%
Hepatocellular carcinoma
Mikoshiba et al. (2013)25) Japan	127 Men: 64% H	Mean: 69.0 SD: 8.4	Hepatocellular carcinoma: 100%	NR	Months since Tx Mean: 24.7 SD: 18.5	NR	Employed full-time or part-time: 39.4%
Various cancers
Inhestern et al. (2017)26) Germany	3370 Male: 25.8% R	Mean: 50.1 SD: 6.8	Breast: 52.1% Prostate: 9.8% Female genital organs: 8.2% Skin: 6.6% Digestive organs: 6.4%.	I: 31.5% II: 22.5% III: 12.0%	Months since Dx Mean: 44.4 SD: 23.3	Surgery: 93.6% Radiotherapy: 63.0% Chemotherapy: 56.9%	Employed full-time: 39.0%; Employed part-time: 33.4%; Housewife: 8.8%; Not employed: 18.9%
Dahl et al. (2019)27) Norway	1189 Male: 27% R	Mean: 49.7 SD: 7.8	Breast: 41% NHL: 19% Melanoma: 17% Colorectal: 12% Leukemia: 11%	NR	Years since Dx Mean: 16.8 SD: 6.8	Systemic Tx + others: 50% Minimal surgery: 17% Local Tx: 17% Systemic Tx only: 16%	Of the employed (75.4%); Full-time: 77%; Part-time: 20%; On sick leave: 3%

AML, acute myeloid leukemia; CML, chronic myeloid leukemia; Dx, diagnosis; H, hospital-based sampling; HL, Hodgkin lymphoma; IQR, interquartile range; MM, multiple myeloma; NHL, non-Hodgkin lymphoma; NR, not reported; NS, not significant; R, cancer-registry-based sampling; SD, standard deviation; Tx, treatment.

Table 2. Depression of cancer survivors who were working: compared with those who were not working

Author, year, country	Measurement of depression	Results on depression	Factors adjusted for in the multivariable analysis
Breast cancer
Kim et al. (2008)15) South Korea	BDI	Prevalence Employed: 21.7%; Unemployed: 25.6%; NS	age, monthly income, gastrointestinal and musculoskeletal diseases, pain, dyspnea, insomnia, appetite loss, constipation, arm symptoms
Ashing-Giwa et al. (2013)16) U.S.	CES-D	Prevalence Employed: 53.8%; Unemployed: 53.6%; NS.	NP
Ho et al. (2018)17) Singapore	HADS-D	Prevalence Employed: 16%; Unemployed: 16%; NS	NP
Puigpinós-Riera et al. (2018)18) Catalonia	HADS-D	Prevalence Active worker: 2.7%; Not working: 7.7%; Disabled: 14.7%; Retired: 5.6% AOR (95% CI) for depression (reference: active worker) Not working: 2.38 (1.21–4.67), p=0.012 Disabled: 4.67 (2.27–9.59), p<0.001 Retired: 2.21 (1.01–4.38), p =0.047	Age, stage, time since diagnosis, relapse, social class, cohabitation, and social network and support
Cervical cancer
Yoo et al. (2013)19) South Korea	HADS-D	Least-squares mean score Working: 6.31; Non-working: 7.01; p<0.05	Age, marital status, educational level, monthly income, religion, previous employment status, menopausal status, time since treatment, comorbidities, regular physical activity, and current smoking and drinking
Head and neck cancer
Verdonck-de Leeuw et al. (2010)20) Netherlands	HADS-D	Mean (SD) scores NW: 4.28 (3.69); RSW: 2.89 (3.12); CW: 4.29 (3.36); NRW: 5.00 (3.70); NS Prevalence NW: 22%; RSW: 11%; CW: 21%; NRW: 13%; NS	NP
Koch et al. (2015)21) Germany	PHQ-9	Mean (SD) score of all survivors was 7.4 (4.6). The unemployed indicated higher scores than the employed (p=0.042)	NP
Hematologic and lymphoid cancer
Magyari et al. (2017)22) Hungary	HADS-D	Mean score (SD) Occupationally inactive: 4.9 (3.8); Occupationally active: 2.2 (2.4); p<0.001 Prevalence Occupationally inactive: 22%; Occupationally active: 4%; p=0.005	NP
Mosher et al. (2011)23) U.S.	BDI	Partial regression coefficient between BDI score and currently working: 0.08, NS.	Education, annual income, transplant type (allogenic or autologous), and KPS
Kiserud et al. (2016)24) Norway	HADS-D	Mean (SD) score The employed: 2.8 (3.2); The unemployed: 4.2 (3.5); p<0.001 AOR (95% CI) of HADS-D score (1 point) for current employment: 0.90 (0.76–1.06), p=0.20	Age, sex, and education level for the mean (SD) score. Somatic comorbidity, fatigue, type D personality, health-related quality of life, and workability were additionally adjusted for calculating the AOR.
Hepatocellular carcinoma
Mikoshiba et al. (2013)25) Japan	CES-D	Prevalence Employed: 12.0%; Unemployed: 39.0% AOR (95% CI) for unemployment: 5.18 (1.73–15.54), p=0.003	Age, KPS, and months since treatment
Various cancers
Inhestern et al. (2017)26) Germany	HADS-D	Standardized beta-coefficient of employment with a high HADS-D score (scores ≥11): -0.070, p<0.001	Age, gender, time since diagnosis, cancer site, no. of treatment modalities, physical health status, socioeconomic status, social support, living with a partner, having a child, family functioning,
Dahl et al. (2019)27) Norway	PHQ-9	Mean (SD) scores Employed: 4.4 (4.1); Unemployed: 8.3 (5.3); p<0.001 AOR (95% CI) of PHQ-9 score (1 point) for being unemployed: 1.06 (1.01–1.11), p=0.02	Age, sex, time since diagnosis, treatment type, adverse effect, comorbidity, general health, fatigue, smoking, obesity

AOR, adjusted odds ratio; BDI, Beck Depression Inventory; CES-D, Center for Epidemiologic Studies Depressive Symptoms Scale; CI, confidence interval; CW, changed the work; HADS-D, Depression subscale of the Hospital Anxiety and Depression Scale; KPS, Karnofsky Performance Status; NP, multivariable analysis was not performed; NRW, no return to work; NS, not significant; NW, not working; PHQ-9, 9-item Patient Health Questionnaire; RSW, returned to the same work as before; SD, standard deviation.

Depression was diagnosed in case of a BDI score of 19 or higher, a CES-D score of 16 or higher, or a HADS-D score of 8 or higher.

Breast cancer

We found four studies regarding breast cancer^15,16,17,18), one of which was cancer-registry-based¹⁶⁾, while the remaining three were hospital-based^15,17,18). Although the depression prevalence was high, ranging between 16% and 54%, three studies reported no significant difference in depression prevalence between cancer survivors who were and were not employed^15,16,17). The remaining study reported a significant difference in the prevalence of probable depression, in which prevalence was 2.7%, 7.7%, 14.7%, and 5.6% for the breast cancer survivors who were actively working, not working, disabled, and retired, respectively¹⁸⁾. Those who were not working had about 2.4 times higher risk of depression than those who were actively working and those who were disabled and retired also indicated a higher risk than active workers.

Cervical cancer

A hospital-based study presented the difference in HADS-D scores between cervical cancer workers and non-workers¹⁹⁾. Working survivors showed significantly lower scores than non-working survivors (least-squares mean score: 7.01 vs. 6.31). However, the difference in depression prevalence between them was not reported.

Head and neck cancer

In a hospital-based study²⁰⁾, the survivors were classified into four groups according to working status: those who were not working at the time of the diagnosis, who returned to the same job, who changed jobs, and who did not return to work. Those keeping the same job appeared to indicate low HADS-D scores and a low prevalence of depression. However, the difference between groups was not statistically significant.

Another hospital-based study used the PHQ-9 to screen for depressive disorders²¹⁾. The mean PHQ-9 score of all survivors was 7.4 (standard deviation [SD], 4.6). It was graphically shown that the mean score was significantly higher in unemployed survivors than in employed survivors, but only a univariable analysis was used; neither the detail of score difference nor depression prevalence was exhibited.

Hematologic and lymphoid cancer

A study among Hodgkin-lymphoma survivors²²⁾ reported that both the mean HADS-D score and the depression prevalence were higher in the occupationally inactive survivors than in the occupationally active survivors. However, it was compared only through a univariable analysis.

The BDI scores of hematologic and lymphoid cancer survivors who received a hematopoietic stem cell transplant were not significantly associated with working after adjusting for educational background, annual income, transplant type, and functional status evaluated with the Karnofsky Performance Scale (KPS)²³⁾.

In lymphoma survivors with high-dose chemotherapy and autologous stem cell transplantation, HADS-D scores were significantly higher in those who were employed in comparison to those who were not, even when age, sex, and educational level were controlled²⁴⁾. However, the adjusted odds ratio for a one-point increase of the HADS-D score did not show statistical significance with additional adjustment for somatic comorbidity, fatigue, type D personality, health-related quality of life, and workability.

Hepatocellular carcinoma

In a study on the prevalence of depressive symptoms, these were significantly higher in the unemployed than in the employed (39.0% vs. 12.0%)²⁵⁾. The odds ratio, adjusted for age, KPS scores, and months since treatment, was 5.18.

Various cancers

In a woman-dominant sample (proportion of men: 25.8%)²⁶⁾, the major primary sites were breast, prostate, and female genital organs. A high level of depression (a HADS-D score of 11 or greater) was negatively associated with being employed. The standardized beta-coefficient was only -0.070.

In a different study in a women-dominant sample (proportion of men: 27%)²⁷⁾, breast cancer, non-Hodgkin lymphoma, and melanoma were dominant. Unemployed cancer survivors indicated higher PHQ-9 scores than those who were employed. In a multiple logistic regression analysis, the adjusted odds ratio of the PHQ-9 score for unemployment was 1.06.

Prevalence of depression in working and non-working cancer survivors

As we already pointed out, a total of seven papers reported the prevalence of depression in working and non-working cancer survivors^{15,16,17,18,20,22,25)}, summarized in Figure 1. The HADS-D, CES-D, and BDI were adopted in four studies^17,18,20,22), two studies^16,25), and one study¹⁵⁾, respectively.

Fig. 1.

Prevalence of depression by employment status. Depression was diagnosed in case of a BDI score of 19 or higher, a CES-D score of 16 or higher, or a HADS-D score of 8 or higher. BDI, Beck Depression Inventory; CES-D, Center for Epidemiologic Studies Depressive Symptoms Scale; HADS-D, Depression subscale of the Hospital Anxiety and Depression Scale; HCC, Hepatocellular carcinoma; HL, Hodgkin lymphoma.

Comparison between working cancer survivors and the cancer-free general population

We found four papers for the comparison between working cancer survivors and the cancer-free general population (Table 3 and Table 4). The results are presented below by cancer site.

Table 3. Study and subject characteristics: subjects included working cancer survivors and the cancer-free general population

Author, year, country	Working cancer survivors							Comparison (cancer-free general population)
	N and sex	Age (years)	Cancer				Working conditions at time of survey
			Site	Stage	Time since Dx/Tx	Tx
Breast cancer
Hansen et al. (2008)28) U.S.	100 women	Mean 49.5 SD: 8.5	Breast: 100%	I: 41% II: 41%	≤5 years since Dx: 72.5%	Surgery: 93% Chemotherapy: 79% Radiation: 67%.	Managerial job position: 25.0%	103 healthy working women Age, years Mean: 39.8 SD: 10.8 Managerial job position: 24.3%
Todd et al. (2011)29) U.S.	133 women	Mean: 44.9 SD: 9.11	Breast: 100%	I: 36% II: 47% III: 17%	Years since Tx Mean: 3.07 SD: 2.39	Surgery: 97% Chemotherapy: 83% Radiation: 74% Hormone therapy: 44%	Occupation Clerical/ sales/ service: 19%; Management/ administration: 35%; Professional/ technical/ science: 46%	122 healthy working women Age, years Mean: 39.2 SD: 11.42 Occupation Clerical/ sales/ service: 19%; Management/ administration: 34%; Professional/ technical/ science: 48%
Cheng et al. (2016)30) Hong Kong	30 women	<40: 3.3% 40–49: 40% 50+: 56.7%	Breast: 100%	Early: 42.3% Mid: 30.8% Late: 26.9%	Months since completing primary Tx Mean: 34.6 SD: 38.3	surgery + radiation + chemotherapy: 66.7%	Full-time worker: 53.3%; Part-time worker: 46.7% Occupation Clerical/ sales/ service/ nontechnical worker: 90.0%; Professional/ technical/ science: 10.0%	30 working women with MSC and 30 healthy working females Age, years <40: 13.3%, 46.7% 40–49: 26.7%, 10.0% 50+: 56.7%, 43.4% Full-time worker: 80.0%, 76.7% Part-time worker: 20.0%, 23.3% Occupation Clerical/ sales/ service/ nontechnical worker: 90.0%, 53.3% Management/ administration: 3.3%, 30.0% Professional/ technical/ science: 6.7%, 16.7%
Malignant brain tumor
Calvio et al. (2009)31) U.S.	113 women: 66%.	Not reported	Oligodendro- glioma: 32%; astrocytoma: 30%.	I: 16% II: 45% III; 27% IV: 12%	Years since Dx 5 or less: 75% 6–10: 18% Months since chemotherapy 1–5: 18% 6–11: 33% 12–17: 36% 18–29: 15%.	Combination of surgery, radiation, and chemotherapy: 65% Surgery alone: 26%.	Working full time for more than one year prior to diagnosis: 100%	123 healthy workers Female: 73% Employed for more than one year at the time of survey: 100%

Dx, diagnosis; MSC, musculoskeletal conditions; SD, standard deviation; Tx, treatment.

Table 4. Depression of working cancer survivors: compared with the cancer-free general population

Author, year, and country	Measurement of depression	Results on depression	Factors adjusted for in the multivariable analysis
Breast cancer
Hansen et al. (2008)28) U.S.	HADS-D	Age-adjusted mean score CS: 4.7 NCC 3.3 p<0.01	Age
Todd et al. (2011)29) U.S.	HADS-D	Mean score (Presented only graphically) CS: around 3 NCC: around 2.5 p<0.001	Age, household income, marital status/living arrangement, ethnicity, race, and menopausal status
Cheng et al. (2016)30) Hong Kong	HADS-D	Mean (SD) score CS: 3.03 (2.67) MSC: 2.37 (2.53) HC: 3.10 (2.37) NS	NP
Malignant brain tumors
Calvio et al. (2009)31) U.S.	HADS-D	Mean score (Presented only graphically) CS: around 5 NCC: around 3 p<0.001	Gender, race, education, marital status, health insurance, physical activity, strengthening exercise, enough sleep, job stress

CS, cancer survivors; HADS-D, Depression subscale of the Hospital Anxiety and Depression Scale; HC, healthy comparison; MSC, musculoskeletal conditions; NCC, non-cancer comparison; NP, not performed; NS, not significant; SD, standard deviation.

Breast cancer

We found two studies conducted in the United States on breast cancer^28,29). The subjects were recruited through advertisements in newspapers and websites. The first reported a higher age-adjusted mean HADS-D scores in working cancer survivors than in a cancer-free comparison group (4.7 vs. 3.3)²⁸⁾. No other potential confounders were controlled. The other study reported higher HADS-D scores in working cancer survivors than in the cancer-free group²⁹⁾. The exact scores were not identified, since the result was shown only graphically.

Another study, performed in Hong Kong, found no significant difference in HADS-D scores among cancer survivors, women with musculoskeletal conditions, and healthy women³⁰⁾. Cancer and musculoskeletal patients were recruited in a hospital, while healthy women were recruited using the researchers’ personal network.

Malignant brain tumors

A web-based survey in the United States with a female-dominant sample (66% of cancer survivors and 73% of healthy controls) found that cancer survivors had higher HADS-D scores than the cancer-free group³¹⁾. The exact scores were not identified since the result was shown only graphically.

Discussion

The results of this review do not consistently confirm our hypothesis that the depression prevalence would be lower in working cancer survivors compared to non-working cancer survivors.

Some research did show the hypothesized trend. Working survivors indicated a significantly lower depression prevalence than non-working cancer survivors in studies dealing with breast cancer¹⁸⁾, Hodgkin lymphoma²²⁾, and hepatocellular carcinoma²⁵⁾, as hypothesized. There was a significant association between working and low depression-related questionnaire scores in cervical cancer¹⁹⁾, head and neck cancer²¹⁾, lymphoma²⁴⁾, and various cancer survivors^26,27).

On the other hand, a nonsignificant difference in the depression prevalence was found in studies examining breast cancer^15,16,17) and head and neck cancer survivors²⁰⁾. In those studies, both working and non-working cancer survivors exhibited high depression prevalence, which might have diminished the difference in prevalence by employment. Moreover, one study reported an insignificant correlation between working and BDI scores in hematologic and lymphoid cancer survivors²³⁾.

We further hypothesized that the depression prevalence in working cancer survivors would be as low as that of the cancer-free general population, but again the results did not consistently support this hypothesis. No studies in this condition assessed depression prevalence; instead, the HADS-D scores were reported. Contrary to our hypothesis, the review showed that working breast cancer and malignant brain tumor survivors demonstrated significantly higher HADS-D scores than the cancer-free general population^28,29,31). In another study³⁰⁾, both employed breast cancer survivors and cancer-free healthy women exhibited a low mean HADS-D score. However, they pointed out a potential selection bias due to the small convenience samples as a study limitation.

Differences in the time since receiving the cancer diagnosis/treatment could explain why the depression prevalence differed by study. Depression prevalence in cancer survivors generally decreases as years go by after cancer diagnosis^32,33). This pattern is somewhat present in the current review. In studies dealing with breast cancer survivors, we found a very low depression prevalence; in one study¹⁸⁾, more than 60% of the subjects were diagnosed more than 5 years before the survey in the study. On the other hand, the mean years since surgery were 4.6 in another study, which showed a depression prevalence of more than 20%¹⁵⁾. The median years since diagnosis was 4–5 in a further study showing a depression prevalence of 16%¹⁷⁾.

We reviewed only cross-sectional studies, which might underestimate the depression prevalence in cancer survivors. Some cancer survivors with depression would be expected to have already passed away at the time of the surveys, since depression increases the all-cause mortality in cancer survivors⁴⁾. In addition to the cross-sectional studies, prospective studies, in which researchers track the changes in working cancer survivors’ mental conditions in order to investigate the effect of working on depression in cancer survivors, would be beneficial.

The depression prevalence was high even in working cancer survivors in some studies^{15,16,17,20,25)}. Working cancer survivors showed higher HADS-D scores than the cancer-free general populations^28,29,31). These findings may indicate that working cancer survivors need treatment and preventive measures for depression. Several behavioral and pharmacological approaches are effective in improving the mood of breast cancer patients^33,34). Intervention trials have taken place for preventing depression in breast cancer survivors^{35,36,37,38,39)}. Occupational health workers must be trained to promote good support for working cancer survivors⁴⁰⁾. These achievements should be applied so that working cancer survivors can receive the appropriate treatment and preventive measures while working.

The case-finding ability of the depression measures for cancer patients must be considered in interpreting the depression prevalence. The HADS-D, CES-D, and BDI were adopted to identify the depression prevalence in working and non-working cancer survivors^{15,16,17,18,20,22,25)}. Previous reviews pointed out the moderate case-finding ability of the HADS-D and BDI, although the BDI might have performed better than the HADS-D^41,42). The CES-D has been little examined regarding its case-finding ability of depression in cancer survivors^41,42). Although the BDI, HADS, and PHQ-9 were used as continuous variables^{19,20,21,22,23,24,26,27,28,29,30,31)}, it is not confirmed whether such use holds the case-finding ability or evaluates the severity of depression in cancer survivors. The HADS-D was used to measure depression in the cancer-free general healthy population^28,29,30,31). However, the HADS-D was originally developed to measure depression of non-psychiatric patients¹³⁾. The case-finding ability of the HADS-D might differ when being used outside of the population for which it has been validated.

Our review results had a few limitations. The first is the small number of existing studies. Because of lacking data, it was difficult to assess whether the association between working and depression prevalence differed by cancer site and subjects’ demographic characteristics. The second limitation is that the subjects might not represent the cancer survivor population. Those who feel mentally and physically healthy might tend to participate in the studies more frequently than those who do not. The advertisement-based recruitment used in some studies could exacerbate this tendency. The final limitation is that a multivariable analysis was not performed for the association between working and depression in all studies. The association could be confounded by individual factors (e.g., family history of mood disorders, prior psychiatric history, and personality traits), interpersonal and social risk factors (e.g., working conditions, life events, loneliness, social isolation, low socioeconomic status, social support, and financial problems), and biological factors (e.g., cancer type and stage, comorbid uncontrolled physical symptoms, and treatment side effects)^8,32,33). In fact, elucidating the association between working and depression in cancer survivors was not always the primary purpose in the included studies. The subjects’ working conditions, except for the employment status, were reported insufficiently. Studies that adequately control for such potential confounders are necessary for better evaluation of the associations between depression and working in cancer survivors.

Conclusions

The existing literature did not consistently support our hypotheses that depression prevalence would be lower in working cancer survivors compared to non-working cancer survivors and that the depression prevalence of working cancer survivors could be as low as that of the cancer-free general population. Working may not function as absolute prevention against depression in cancer survivors. Therefore, it seems that preventive measures against depression are necessary for cancer survivors even while working. More studies are needed to further investigate the associations between working and depression in cancer survivors.

Acknowledgments

We thank Editage Group (https://www.editage.jp/) for editing a draft of this manuscript.

Funding information

This work was funded by a grant from the Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (C) (No. 16K09111).

Conflict of interest

The authors declare no conflicts of interest.

References

• 1.  Foundation for Promotion of Cancer Research Japan (FPCRJ). Cancer statistics in Japan 2018. Tokyo: FPCRJ; 2019. https://www.fpcr.or.jp/pdf/p21/cancer_statistics_2018.pdf. Published March 2019. Accessed 18 March, 2020.
• 2.   Brandenbarg  D,  Maass  SWMC,  Geerse  OP, et al. A systematic review on the prevalence of symptoms of depression, anxiety and distress in long-term cancer survivors: implications for primary care. Eur J Cancer Care (Engl). 2019; 28(3): e13086.
• 3.   Steel  Z,  Marnane  C,  Iranpour  C, et al. The global prevalence of common mental disorders: a systematic review and meta-analysis 1980-2013. Int J Epidemiol. 2014; 43(2): 476-493.
• 4.   Satin  JR,  Linden  W,  Phillips  MJ. Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis. Cancer. 2009; 115(22): 5349-5361.
• 5.   Hem  E,  Loge  JH,  Haldorsen  T,  Ekeberg  Ø. Suicide risk in cancer patients from 1960 to 1999. J Clin Oncol. 2004; 22(20): 4209-4216.
• 6.   Misono  S,  Weiss  NS,  Fann  JR,  Redman  M,  Yueh  B. Incidence of suicide in persons with cancer. J Clin Oncol. 2008; 26(29): 4731-4738.
• 7.   Pitman  A,  Suleman  S,  Hyde  N,  Hodgkiss  A. Depression and anxiety in patients with cancer. BMJ. 2018; 361: k1415.
• 8.   Ota  A,  Fujisawa  A,  Kawada  K,  Yatsuya  H. Recent status and methodological quality of return-to-work rates of cancer patients reported in Japan: A systematic review. Int J Environ Res Public Health. 2019; 16(8): 1461.
• 9.   Paltrinieri  S,  Fugazzaro  S,  Bertozzi  L, et al. Return to work in European Cancer survivors: a systematic review. Support Care Cancer. 2018; 26(9): 2983-2994.
• 10.   Mehnert  A. Employment and work-related issues in cancer survivors. Crit Rev Oncol Hematol. 2011; 77(2): 109-130.
• 11.   Beck  AT,  Ward  CH,  Mendelson  M,  Mock  J,  Erbaugh  J. An inventory for measuring depression. Arch Gen Psychiatry. 1961; 4(6): 561-571.
• 12.   Radloff  LS. The CES-D Scale: A self-report depression scale for research in the general population. Appl Psychol Meas. 1977; 1(3): 385-401.
• 13.   Zigmond  AS,  Snaith  RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983; 67(6): 361-370.
• 14.   Cannon  DS,  Tiffany  ST,  Coon  H,  Scholand  MB,  McMahon  WM,  Leppert  MF. The PHQ-9 as a brief assessment of lifetime major depression. Psychol Assess. 2007; 19(2): 247-251.
• 15.   Kim  SH,  Son  BH,  Hwang  SY, et al. Fatigue and depression in disease-free breast cancer survivors: prevalence, correlates, and association with quality of life. J Pain Symptom Manage. 2008; 35(6): 644-655.
• 16.   Ashing-Giwa  K,  Rosales  M,  Lai  L,  Weitzel  J. Depressive symptomatology among Latina breast cancer survivors. Psychooncology. 2013; 22(4): 845-853.
• 17.   Ho  PJ,  Hartman  M,  Gernaat  SAM, et al. Associations between workability and patient-reported physical, psychological and social outcomes in breast cancer survivors: a cross-sectional study. Support Care Cancer. 2018; 26(8): 2815-2824.
• 18.   Puigpinós-Riera  R,  Graells-Sans  A,  Serral  G, et al. Anxiety and depression in women with breast cancer: social and clinical determinants and influence of the social network and social support (DAMA cohort). Cancer Epidemiol. 2018; 55: 123-129.
• 19.   Yoo  SH,  Yun  YH,  Park  S, et al. The correlates of unemployment and its association with quality of life in cervical cancer survivors. J Gynecol Oncol. 2013; 24(4): 367-375.
• 20.   Verdonck-de Leeuw  IM,  van Bleek  WJ,  Leemans  CR,  de Bree  R. Employment and return to work in head and neck cancer survivors. Oral Oncol. 2010; 46(1): 56-60.
• 21.   Koch  R,  Wittekindt  C,  Altendorf-Hofmann  A,  Singer  S,  Guntinas-Lichius  O. Employment pathways and work-related issues in head and neck cancer survivors. Head Neck. 2015; 37(4): 585-593.
• 22.   Magyari  F,  Kósa  K,  Berecz  R, et al. Employment status and health related quality of life among Hodgkin-lymphoma survivors’- results based on data from a major treatment center in Hungary. Health Qual Life Outcomes. 2017; 15(1): 180.
• 23.   Mosher  CE,  DuHamel  KN,  Rini  C,  Corner  G,  Lam  J,  Redd  WH. Quality of life concerns and depression among hematopoietic stem cell transplant survivors. Support Care Cancer. 2011; 19(9): 1357-1365.
• 24.   Kiserud  CE,  Fagerli  UM,  Smeland  KB, et al. Pattern of employment and associated factors in long-term lymphoma survivors 10 years after high-dose chemotherapy with autologous stem cell transplantation. Acta Oncol. 2016; 55(5): 547-553.
• 25.   Mikoshiba  N,  Miyashita  M,  Sakai  T,  Tateishi  R,  Koike  K. Depressive symptoms after treatment in hepatocellular carcinoma survivors: prevalence, determinants, and impact on health-related quality of life. Psychooncology. 2013; 22(10): 2347-2353.
• 26.   Inhestern  L,  Beierlein  V,  Bultmann  JC, et al. Anxiety and depression in working-age cancer survivors: a register-based study. BMC Cancer. 2017; 17(1): 347.
• 27.   Dahl  AA,  Fosså  SD,  Lie  HC, et al. Employment status and work ability in long-term young adult cancer survivors. J Adolesc Young Adult Oncol. 2019; 8(3): 304-311.
• 28.   Hansen  JA,  Feuerstein  M,  Calvio  LC,  Olsen  CH. Breast cancer survivors at work. J Occup Environ Med. 2008; 50(7): 777-784.
• 29.   Todd  BL,  Feuerstein  EL,  Feuerstein  M. When breast cancer survivors report cognitive problems at work. Int J Psychiatry Med. 2011; 42(3): 279-294.
• 30.   Cheng  ASK,  Lau  LOC,  Ma  YNH,  Ngai  RH,  Fong  SSL. Impact of cognitive and psychological symptoms on work productivity and quality of life among breast cancer survivors in Hong Kong. Hong Kong J Occup Ther. 2016; 28(1): 15-23.
• 31.   Calvio  L,  Feuerstein  M,  Hansen  J,  Luff  GM. Cognitive limitations in occupationally active malignant brain tumour survivors. Occup Med (Lond). 2009; 59(6): 406-412.
• 32.   Caruso  R,  Nanni  MG,  Riba  M, et al. Depressive spectrum disorders in cancer: prevalence, risk factors and screening for depression: a critical review. Acta Oncol. 2017; 56(2): 146-155.
• 33.   Yi  JC,  Syrjala  KL. Anxiety and depression in cancer survivors. Med Clin North Am. 2017; 101(6): 1099-1113.
• 34.   Coutiño-Escamilla  L,  Piña-Pozas  M,  Tobías Garces  A,  Gamboa-Loira  B,  López-Carrillo  L. Non-pharmacological therapies for depressive symptoms in breast cancer patients: systematic review and meta-analysis of randomized clinical trials. Breast. 2019; 44: 135-143.
• 35.   Grégoire  C,  Nicolas  H,  Bragard  I, et al. Efficacy of a hypnosis-based intervention to improve well-being during cancer: a comparison between prostate and breast cancer patients. BMC Cancer. 2018; 18(1): 677.
• 36.   Kim  YH,  Choi  KS,  Han  K,  Kim  HW. A psychological intervention programme for patients with breast cancer under chemotherapy and at a high risk of depression: A randomised clinical trial. J Clin Nurs. 2018; 27(3-4): 572-581.
• 37.   Li  Z,  Geng  W,  Yin  J,  Zhang  J. Effect of one comprehensive education course to lower anxiety and depression among Chinese breast cancer patients during the postoperative radiotherapy period - one randomized clinical trial. Radiat Oncol. 2018; 13(1): 111.
• 38.   Blanco  C,  Markowitz  JC,  Hellerstein  DJ, et al. A randomized trial of interpersonal psychotherapy, problem solving therapy, and supportive therapy for major depressive disorder in women with breast cancer. Breast Cancer Res Treat. 2019; 173(2): 353-364.
• 39.   Weihs  KL,  McConnell  MH,  Wiley  JF,  Crespi  CM,  Sauer-Zavala  S,  Stanton  AL. A preventive intervention to modify depression risk targets after breast cancer diagnosis: design and single-arm pilot study. Psychooncology. 2019; 28(4): 880-887.
• 40.   Yoshikawa  E,  Nishikido  N,  Sasaki  M, et al. Development of an occupational health nurse’s guidebook to promote the balance between cancer treatment and work. Environ Occup Health Pract. 2019; 1(2): 31-38.
• 41.   Mitchell  AJ,  Meader  N,  Davies  E, et al. Meta-analysis of screening and case finding tools for depression in cancer: evidence based recommendations for clinical practice on behalf of the Depression in Cancer Care consensus group. J Affect Disord. 2012; 140(2): 149-160.
• 42.   Wakefield  CE,  Butow  PN,  Aaronson  NA,  Hack  TF,  Hulbert-Williams  NJ,  Jacobsen  PB; International Psycho-Oncology Society Research Committee. Patient-reported depression measures in cancer: a meta-review. Lancet Psychiatry. 2015; 2(7): 635-647.