A Systematic Review of the Effect of Cancer Treatment on Work Productivity of Patients and Caregivers

Abstract

BACKGROUND:

Cancer is a leading cause of death with substantial financial costs. While significant data exist on the economic burden of care, less is known about the indirect costs of treatment and, specifically, the effect on work productivity of patients and their caregivers. To examine the full effect of cancer and the potential value of new therapies, all aspects of care, including indirect costs and patient-reported outcomes, should be evaluated.

OBJECTIVE:

To perform a systematic review of the literature examining the effect of cancer treatment on work productivity in patients and their caregivers.

METHODS:

Articles, abstracts, and bibliographies were searched in MEDLINE, Cochrane, Scopus, CINAHL, and conference lists from the American Society of Clinical Oncology, International Society for Pharmacoeconomics and Outcomes Research, and Academy of Managed Care Pharmacy up to January 2016. The PRISMA guidelines were used. Controlled search terminology included individual pharmacologic therapies for cancer and terms related to patient and caregiver work productivity. Citations were included if they evaluated the effect of cancer treatment on work productivity, used and described productivity assessments and instruments, and were written in English. Studies that reported only clinical outcomes or assessed only nonpharmacological treatments were excluded. Identified studies were screened and extracted for study inclusion by 2 independent reviewers, with adjudication by 2 secondary reviewers during the final eligibility phase.

RESULTS:

Of 978 potential citations, 62 articles or abstracts were included. Forty-six studies (74.2%) evaluated patient-related productivity; 10 studies (16.1%) focused on caregivers, and 6 studies (9.7%) were a combination. Sixteen countries contributed literature, including 26 studies (41.2%) conducted in the United States. The most commonly studied cancer was breast cancer (53.2%). Nearly 22% of the studies were conducted on multiple types of cancer. The significant diversity of study methodologies and measurements rendered a single unifying conclusion difficult. A variety of metrics were used to quantify productivity (hours lost, return to work, change of status, and activity impairment). The Work Productivity and Activity Impairment questionnaire was the most commonly used standardized tool (n = 9; 14.5%). Factors found to be associated with impairment in productivity included disease- and treatment-related effects, such as disease progression and severity, cognitive and neurological impairments, poor physical and psychological status, receipt of chemotherapy, and time and expenses required to receive therapy.

CONCLUSIONS:

This review highlights the considerable variety of studies that have assessed work productivity for cancer treatment and the multifaceted reasons affecting patients and caregivers. With increasing emphasis being given to understanding the value that patients assign to various aspects of cancer treatment, more streamlined information on productivity may be important to patients as they play a greater role in selecting treatment goals through shared decision making with their providers.

What is already known about this subject

• The cost of cancer care in the United States has been steadily increasing because of improvements in survival and emerging treatments.

• Cancer as a disease results in significant productivity loss and economic burden to society.

• Several value and benefit frameworks have been recently proposed in response to high prices of cancer treatment and are intended to assist payers and health systems in their coverage and payment decisions.

What this study adds

• This study provides a review of the diversity of existing studies on the effect of cancer treatment on work productivity in patients and their caregivers and recommends additional research using standardized measures and quantification of indirect costs to better understand the full economic burden of cancer and value of treatments.

• Most patient-focused studies found an overall negative effect of cancer treatment on patient productivity because of treatment adverse effects that led to missed work, cognitive impairments, and challenges for lifting heavy loads and keeping pace with others at work.

• For caregivers, factors associated with productivity loss were intensive cancer treatment, travel for treatment, costs associated with treatment, and change of work hours during treatment.

In 2014, approximately 14.5 million people were living with cancer in the United States, with a projected increase to 19 million by 2024.¹ In 2016, over 1.6 million were estimated to be diagnosed.¹ Because of steady improvements in treatment, patient survival has been significantly improved. In a longitudinal registry analysis of over 1 million patients diagnosed with 7 major cancer types, survival rates for patients diagnosed between 2005 and 2009 were largely increased across all cancers, compared with those diagnosed between 1990 and 1994, with the sole exception of ovarian cancer.²

In addition to clinical burden, cancer imposes a substantial economic burden on society. Projections for the cost of cancer care among survivors (assuming constant costs, incidence, and survival) have been estimated at nearly $158 billion by the year 2020 as a result of growth and aging in the U.S. population.³ Continuing care costs for breast and prostate cancers from increased survival were found to be particularly important drivers. Medicare Part B spending on drugs (primarily resulting from cancer treatment) increased 267% from 1997 to 2004, compared with a rise in overall Medicare spending of 47% during the same period.⁴ Additionally, annual costs as a result of lost productivity because of cancer mortality have been projected to reach $147.6 billion in 2020.⁵ Inclusion of the value of caregiving and household activity, which is often overlooked in estimates, increases these estimated losses to $308 billion.⁵

Increasing survival of patients with cancer means that many patients continue to work or return to work after treatment. Data from the 2008-2011 Medical Expenditure Panel Survey found that 42.1% (95% confidence interval [CI] = 37.9-46.2) of cancer survivors reported changes in work (e.g., extended paid time off, unpaid time off, change in hours, duties, and employment status) as a result of their diagnoses.⁶ Additionally, cancer was reported to interfere with physical (25.1%; 95% CI = 20.9-29.2) and mental tasks (14.4%; 95% CI = 11.4-17.3) required by work. The importance of continued employment and work productivity to patients with cancer is largely unrecognized and is likely to have a significant effect on quality of life and choices made regarding their treatment. Understanding and recognizing these effects could stimulate more treatment innovations that are designed to keep people on the job and productive. New oral targeted therapies for cancer represent an excellent example, where administration occurs largely in a home setting, which prevents time lost from work for administration and allows patients to avoid infusion clinics and experience a more “normal” life overall.

Recently, a number of frameworks have been proposed by various agencies and organizations for assessing the value of oncology treatments, including the American Society of Clinical Oncology (ASCO), the European Society of Medical Oncology (ESMO), Memorial Sloan Kettering (MSK) Cancer Center, the National Comprehensive Cancer Network (NCCN), and the Institute for Clinical and Economic Review (ICER).^7-11 For the most part, these resources assess clinical benefits, risk of toxicity, quality of evidence, and financial aspects, but they have come under scrutiny for their purported lack of consideration of the patient perspective or those elements of care that may contribute the most value to patients.^12,13 Accordingly, ASCO issued an update to its value framework in May 2016,¹⁴ which now includes consideration of quality of life (QoL); this represents a first key step for the inclusion of patient-reported outcomes among these frameworks. Patients routinely report QoL, continued independence, and ability to perform normal activities as highly important during cancer treatment,^15,16 which can all be associated with continued productivity. Others acknowledge a need to remain employed to retain medical insurance as a key consideration during their treatment.¹⁷ Work productivity during treatment for cancer represents a specific aspect of patient perspective that could be considered for incorporation into existing value frameworks in concert with QoL measurements.

The objectives of this study were to (a) perform a systematic review of the literature that examines the effect of cancer treatment on work productivity in patients and their caregivers and (b) identify gaps in knowledge and provide guidance for future research in this area.

Methods

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁸

Literature Search Strategy

MEDLINE (PubMed), Scopus, Cochrane, CINAHL (Cumulative Index to Nursing and Allied Health Literature), and conference abstracts from ASCO, International Society for Pharmacoeconomics and Outcomes Research, and the Academy of Managed Care Pharmacy were systematically searched for all the relevant publications between the start dates of each respective database through January 2016. A primary search was conducted in all the databases and conference abstracts using controlled search terminology that combined individual pharmacologic therapies for cancer (169 unique therapies) and terms related to patient work productivity and caregiver-related terms (Table 1). The lists of individual pharmacological therapies for cancer and productivity and caregiver terms were compiled through an investigator literature search. For example, the search term bevacizumab AND absenteeism reflects the use of a therapy search term combined with a patient productivity search term. After this initial comprehensive search, a secondary search was conducted only in MEDLINE (PubMed), using a list of general oncology terms combined with the productivity and caregiver terms (Table 1). Bibliographies of identified articles were also checked for additional studies related to research objectives.

TABLE 1.

List of Search Terms Used Related to Patient and Caregiver Productivity and Cancer

Work Productivity Terms	Caregiver Burden Terms	Oncology Terms
Absenteeism	Caregiver(s)	Oncology
Presenteeism	Carers	Cancer
Work productivity	Caregiver burden	Neoplasm
Loss of work productivity	Caregiver stress	Tumor
Work impairment	Caregiver support	Malignant tumor
Work	Caregiver need	Benign tumor
Indirect cost	Caregiver role strain	Carcinomas
Societal cost	Family caregiver burden	Metastatic
Total work impairment	Family caregivers	Metastases
Sick leave		Cancroid
Missed work		Polyp
Time away from work		Sarcoma
Work ability
Short-term disability
Use of ancillary services
Workforce dropout
Employment
Long-term disability

Inclusion and Exclusion Criteria

Articles and conference abstracts were included in this systematic review if they met the following criteria: (a) evaluated the effect of any pharmacological cancer treatment on work productivity in patients or their caregivers relevant to their treatment phase; (b) described cancer clinical trials that used productivity assessments; (c) contained a description of productivity instruments used; (d) performed economic evaluations of treatments; and (e) were written in the English language. Randomized clinical trials that reported only clinical outcomes, articles that assessed only nonpharmacological treatments, economic evaluations of cancer, QoL, or other quality improvement studies that did not address work productivity were excluded from the review, as were theses and dissertations, commentaries, editorials, and summary reports.

Data Screening and Extraction

Under supervision, 5 graduate student reviewers searched each of the electronic databases to identify potential records. Titles and abstracts deemed to be generally relevant were selected, and records were excluded that did not focus on productivity or examined disease processes other than cancer. Two additional reviewers (authors Dashputre and Ghosh) executed a screening phase, which removed additional records after a more in-depth assessment of the abstracts and relevance to the goal of this study. The broad types of records excluded at this stage were editorials, commentaries, reviews, randomized controlled trials with only clinical outcomes, and studies that evaluated the overall cost of cancer or included only nonphar-macological therapies. These reviewers then obtained full-text articles of records and completed a final screening based on additional information revealed within the full articles. A detailed data extraction log of these articles were created that included desired information relevant to this review, such as objective of the study, year of publication, type of cancer studied, country where study was conducted, target population studied, outcomes, measures to assess productivity losses, and factors affecting the loss in productivity. Two senior reviewers (authors Kamal and Covvey) cleaned and reviewed extracted data against the original articles. This process was used to ensure that the included studies met the criteria, and correct information was extracted.

Results

The review process was based on the PRISMA guidelines as shown in Figure 1.¹⁸ The initial search identified 978 potentially relevant publications, of which 889 were further excluded during in-depth review of the abstracts. Of 89 publications that underwent further eligibility assessment, 27 were excluded. As a result, 62 relevant studies (56 full-text articles and 6 abstracts) met criteria and were included for evaluation in the final review.

FIGURE 1.

Flowchart of Publications Selection Process

Overall Study Characteristics

A summary of the publications included in this review is presented in Table 2 (for specific cancer types) and Table 3 (for multiple and/or unspecified cancer types). The most common cancer studied was breast cancer, which accounted for 53.2% (n = 33) of the included studies.^19-41 Other types of cancers studied were colorectal (n = 5),^42-46 prostate (n = 3),^47-49 ovarian (n = 2),^50,51 and lung (n = 2).^52,53 Other less commonly studied cancers (each with only 1 study) were brain tumor/glioblastoma,⁵⁴ leukemia,⁵⁵ head and neck,⁵⁶ oropharyngeal,⁵⁷ pancreatic,⁵⁸ renal cell,⁵⁹ and thyroid.⁶⁰ Nearly 22% of the studies (n = 14) were conducted on multiple types of cancer.^61-74 Sixteen countries contributed literature, with more than half of the studies conducted outside of the United States and 41.2% of the studies (n = 26) conducted in the United States.^{21,34,35,37,39-42,46,62,63,69,71-76} Other contributing countries included Canada,^5,9,14,64 the United Kingdom,^32,47,61,65 Sweden,^{27-30,56,58,67,77} France,^{23,24,52,53,60} Italy, ^{33,38,45,51-53} the Netherlands,^20,30 Singapore,³⁶ Australia,^43,54 Ireland,⁴⁴ Brazil,⁵⁵ Germany,^52,53 Japan,⁷⁰ South Korea,⁶⁴ Norway, ⁶⁶ and Spain.^52,53

TABLE 2.

Description of Studies Assessing Specific Cancer Types

Author, Year Country	Target Population; Publication Type	Study Design	Treatment Type	Study Objective	Patient Population/Setting	Measures/Methods Used to Assess Productivity	Main Findings
Breast cancer (n = 24)
Hilton, 2000 Canada26	Caregiver; full-text	Prospective qualitative interview	Chemotherapy	To study men’s perspectives on their experiences and how their wives’ breast cancer and chemotherapy affected them and their families	10 male spouses of women with breast cancer and at least 1 child at home	Identification of themes from semistructured interviews	Spouses noted increased need to be involved in household and child care activities as a result of their spouses’ treatment. Help from others outside the home was common, and effect on work life balance was apparent. Limitations: Small sample size, over-representation of Caucasians, single interview
Maunsell, 2004 Canada31	Patient; full-text	Retrospective, population-based comparative cohort	Chemotherapy, hormone therapy, radiation, surgery	To evaluate the presence of evidence of discrimination at work, defined as negative or involuntary changes in employment situation (including changes in position wages, and other conditions), associated with breast cancer	646 survivors (3 years postdiagnosis), and 890 comparators selected from the Quebec Tumour Registry	Stopping work, changes in working conditions/situation	21% of survivors and 15% of comparators were not working at 3 years (ARR = 1.29; 95% CI = 1.05-1.59) Limitation: Subject to selection bias
Drolet, 2005 Canada22	Patient; full-text	Prospective telephone interview	Chemotherapy, radiation, surgery	To examine work absences of 4 weeks or more among women who had breast cancer during the 3 years after diagnosis and compared their absences with those of women who had never had cancer	567 disease-free survivors, 79 survivors with new breast cancer (3 years postdiagnosis) from Quebec Tumour Registry, and 890 comparators selected from provincial health care files	Missed work	85% of disease-free survivors and 18% of comparators had missed at least 4 weeks of work at 1-year postdiagnosis (mean duration: 5.6 vs. 1.7 months, P < 0.001). Adjuvant chemotherapy prolonged absence duration (9.5 vs. 5.4 months, P < 0.0001).
Balak, 2008 Netherlands20	Patient; full-text	Retrospective records analysis	Surgery plus either (a) chemotherapy, (b) radiation, (c) multimodal treatment, or (d) no adjuvant treatment	To evaluate (a) the return to work rate following early stage breast cancer and (b) the effect of type of treatment and cancer-related symptoms on return to work	72 women from the Occupational Health Department register of Groningen, Friesland, Drenthe, and Overijssel of the Netherlands	Missed work	35% of patients were absent longer than 1 year, and 6% did not return to work within 2 years postdiagnosis. Chemotherapy (HR = 0.31; 95% CI = 0.12-0.81) or multimodal treatment (HR: 0.24; 95% CI: 0.10-0.54) increased absence duration. Limitations: Validity, self-reported cancer-related symptoms
Munir, 2010 United Kingdom32	Patient; full-text	Prospective qualitative focus group	Chemotherapy plus hormone therapy, radiation, or surgery	To investigate women’s awareness of chemotherapy-induced cognitive changes and their perception of cognitive limitations in carrying out daily tasks and subsequent return to work	13 women recruited via cancer charity support groups affiliated with NHS oncology clinics	Identification of themes from semistructured interviews	Treatment was reported to cause cognitive impairments that influence working life (short-term memory, verbal abilities, speed of processing information, executive functioning). Appraisal of return to work was based on side effects from chemotherapy, awareness of cognitive failures, and impact on confidence. Limitations: Small sample size, generalizability
Fantoni, 2010 France23	Patient; full-text	Prospective survey questionnaire	Chemotherapy, hormone therapy, radiation, surgery	To evaluate the objective and subjective factors that influence when and if women with breast cancer should return to work	379 women working at diagnosis selected from national medical computerized program, which covers all hospital admissions and discharges at a reference cancer treatment center in the northern district of France	Return to work, duration of sick leave	At 12 months post-treatment, mean duration of sick leave was 10.8 months, and 54.3% returned to work. At 3 years posttreatment, 82.1% returned to work. Medical factors negatively influencing return to work were chemotherapy, radiotherapy, and lymphedema. Limitations: Generalizability; support and barrier questions were dependent on respondent perceptions
Mujahid, 2010 United States76	Patient; full-text	Prospective survey questionnaire	Chemotherapy, radiation, surgery	To examine impact of sociodemographic, treatment, and work support on missed work after breast cancer diagnosis	589 women working at diagnosis with nonmetastatic disease selected from Los Angeles County SEER registry	Missed work, stopping work	After initial treatment, 44% of women missed ≤ 1 month of work; 24% missed > 1 month of work; and 32% of women stopped working completely. Amount of missed work was positively associated with number of comorbidities, mastectomy, and receipt of chemotherapy. Limitations: Lack of control group, unknown consequences of missed work on patient and family life
Hedayati, 2013 Sweden77	Patient; full-text	Prospective survey questionnaire with follow-up	Chemotherapy, hormone therapy, no medical treatment	To identify any associations between cognitive, psychosocial, somatic, and treatment factors with time to return to work among women treated for breast cancer	44 women working before diagnosis who returned to work (n = 29) or were on sick leave (n = 15) recruited from mammography screening program at Stockholm South General Hospital	Return to work, Cognitive Stability Index; EORTC QLQ, Breast Cancer module	Factors associated with not working at 11 months postdiagnosis were advanced disease, lymph node involvement, and HER 2+ tumor. Chemotherapy was associated with longer periods of sick leave. Cognitive functions did not predict return to work. Limitation: Small sample size
Jagsi, 2011 United States40	Patient; abstract	Prospective survey questionnaire	Chemotherapy, radiation, surgery	To study the consequences of treatment on work of breast cancer survivors	482 women with nonmetastatic disease from Los Angeles and Detroit SEER registries	Stopping work	Work loss was associated with initial treatment with chemotherapy (51% vs. 32%, P = 0.001), which held association in a multivariate model (OR = 2.6, P = 0.006). Limitation: Generalizability
Johnsson, 2011 Sweden27	Patient; full-text	Prospective exploratory cohort	Surgery plus chemotherapy, hormone therapy, radiation	To generate new knowledge about factors predicting return to work among women treated for early-stage breast cancer and about changes in life satisfaction, and coping, and to examine the association between these concepts and returning to work	102 women with early-stage disease working at diagnosis from Department of Oncology at Karolinska University Hospital, Sodersjukhuset in Stockholm	Return to work, General Coping Questionnaire and the Life Satisfaction Questionnaire	28% of women treated with chemotherapy, and 85% of women not treated with chemotherapy returned to work 6 months postsurgery. Chemotherapy, > 30 days of sick leave in past year, low satisfaction with ADLs, and not being born in Sweden were predictive of no return to work at 6 months in a multivariate model. Limitation: Small sample size
de Lima Lopes, 2011 Singapore36	Patient; full-text	Simulation model	Trastuzumab adjuvant to chemotherapy	To determine the societal cost and benefit of trastuzumab used for the treatment of HER2+ breast cancer	2 private and 2 public treatment centers in Singapore comprising 60%-70% of all patients with breast cancer	Incremental cost analysis, with productivity losses derived from gross national income per capita	Incremental costs were $26,971.05 USD. Average cost per QALY was $19,174.59. Costs (benefits) to society ranged from a cost of $79.42 to a benefit of $9,263.06, depending on the model used. Limitations: The economic model is based on clinical data from North America and Europe rather than from Singapore. Also, the study does not consider if a patient is employed and productive.
Stroppa, 2011 Italy33	Patient; abstract	Prospective longitudinal interview	Chemotherapy, hormone therapy, radiation, surgery	To study return to work after diagnosis of breast cancer	76 patients with early-stage disease admitted to outpatient medical oncology department in an Italian hospital	Stopping work, return to work	45% of patients stopped work 1 month postdiagnosis; 55.3% stopped work at 3-9 months. Increased age, chemotherapy, and physical/psychological status affected return to work. Limitation: Small sample size
Kaufman, 2012 United States37	Patient; abstract	Prospective cohort	Chemotherapy, hormone therapy	To study the baseline characteristics and the correlations between health-related quality of life, symptoms, activities of daily living, and work productivity	277 patients with locally advanced or metastatic breast cancer	WPAI, MDASI, and Activity Level Scale from the Rotterdam Symptom Checklist	5 most severe symptoms (fatigue, decreased sexual interest, disturbed sleep, drowsiness, and emotional distress) were most severe in those receiving chemotherapy and correlated with work productivity indices.
Bradley, 2013 United States21	Caregiver; full-text	Prospective cohort	Chemotherapy, radiation, surgery	To evaluate the effect of cancer and its treatment on employment and weekly hours worked for employed men whose wives were newly diagnosed with cancer	373 married, employed men compared with noncancer control sample (n = 451 and n = 328 for 2- and 9-month surveys, respectively) from the Current Population Survey in Virginia	Changes in work hours	At 2- and 9-month interviews, men in cancer sample decreased weekly work hours at a greater rate than controls. Men in cancer sample were 2.6 times more likely to stop working at 2 months. At 2-month posttreatment, men in cancer sample were 1.45 times and 1.7 times (if on treatment) more likely to reduce number of weekly hours. Limitations: Reporting bias, erroneous propensity scoring for matching controls at baseline
Frederix, 2013 Netherlands, Sweden30	Patient; full-text	Prospective health utility study	Not specified	To estimate the utility values in laypeople and productivity loss for women with breast cancer	213 women currently or previously treated for metastatic disease; 200 men/women as general public sample	WPAI-GH, health state ratings	Those currently receiving treatment reported productivity reductions of 69% (Netherlands) and 72% (Sweden); those who had recently completed therapy reported reductions of 41% (Netherlands) and 40% (Sweden). Limitation: Utilities based on convenience sample
Ganem, 2013 France24	Patient and physician; abstract	Prospective survey questionnaire	Chemotherapy, hormone therapy, radiation, surgery, targeted therapy	To study the impact of chemotherapy on work in breast cancer patients who were working at the time of diagnosis	216 patients working at diagnosis, 97 oncologists	Job satisfaction, stopping work, missed work	31% of patients continued working without interruption; 66% of patients feared that continuing to work would make them face at least 1 type of medical difficulty; 92% of oncologists shared same fear.
Lilliehorn, 2013 Sweden28	Patient; full-text	Prospective longitudinal interview	Surgery plus chemotherapy or hormonal therapy	To describe the sick-leave pattern of a group of Swedish women with primary breast cancer and to explore their ideas about what motivates and discourages their return to work	56 newly diagnosed women from the radiation department of a Swedish university hospital	Length of sick leave, stopping work	The average sick leave over 18-24 months was 410 days (range 0-942 days). Patients treated with chemotherapy had twice as long sick leave as those who did not receive chemotherapy. Limitation: Only included ideas about work rather than factors such as side effects
Lundh, 2013 Sweden29	Patient; full-text	Prospective longitudinal questionnaire	Antibody therapy, chemotherapy, hormone therapy, radiation	To investigate whether longitudinal changes in HRQoL among breast cancer patients vary by prediagnosis occupational status or postdiagnosis changes in working time	841 women recruited from breast cancer quality register of the Uppsala– Örebro Region in Central Sweden	Changes in work hours, EORTC QLQ-C30, EORTC QLQ-BR23	20% of employed patients stopped work at 16 months; 12% decreased working time; and 5% had increased. Employed women had better symptoms and functioning compared with those on sick leave, disability, or retirement. Limitations: Questionnaires were completed within time intervals and not at a certain point in time, which may have influenced the variation in HRQoL at baseline
Vacante, 2013 Italy38	Patient; full-text	Prospective clinical study	Capecitabine adjuvant to surgery	To evaluate the effect of chemotherapy with capecitabine in patients with breast cancer on work productivity and daily activity	34 patients with metastatic disease who desired to keep working	WPAI	Nonsignificant increase in absenteeism, presenteeism, work productivity loss, and daily activity impairment after 1 and 6 cycles compared with baseline. Limitations: Small sample size, no comparator
Wan, 2013 United States39	Patient and caregiver; full-text	Retrospective matched cohort claims analysis	Chemotherapy, hormone therapy, radiation, surgery	To study and compare the indirect costs of productivity loss between metastatic breast cancer and early-stage breast cancer patients, as well as their respective family members	139 patients with metastatic disease, 432 patients with early disease, and 820 controls eligible for sick leave; 432 metastatic, 1,552 early, and 4,682 controls eligible for STD. Patients from the MarketScan CCEA and HPM databases	Sick leave, STD, costs	Patients not receiving chemotherapy during follow-up were associated with lower STD costs. Metastatic patients’ family members had higher leave days and associated costs than matched patients. Limitations: Reason for sick leave is unknown; workday absence only calculated for personal reasons or FMLA; not all caregivers and areas of productivity loss were captured.
Cleeland, 2014 United States35	Patient; full-text	Cross-sectional patient-reported outcomes	Chemotherapy, hormone therapy, targeted therapy	To study work productivity in patients receiving first-line hormonal therapy or chemotherapy and/or targeted therapy	152 patients with locally recurrent or metastatic disease from the VIRGO study (NCT00726661)	WPAI-Specific Health Program, Activity Level Scale, HRQoL, MDASI	At baseline, 38.1% of patients were employed, with impairment for 20% (work time missed), 30% (impairment while working due to problem), and 40% (overall work impairment), with a mean 7.3 hours missed per week. Fatigue and decreased sexual interest had higher impairment of daily activities and work productivity. Limitations: Analyses restricted to baseline data, selection bias, small sample size
Azarkish, 2015 Iran19	Patient; full-text	Prospective questionnaire	Chemotherapy, surgery	To identify factors related to the return to work in women under breast cancer treatment	175 women currently working recruited from 4 hospitals (referral centers of oncology) in eastern regions of Iran	Return to work	At the time of interview, 80% had returned to work after diagnosis, and mean duration of sick leave was 104.35 ± 99.23 days. Women who had no pain or surgery scar as well as those who had no lymphedema after treatment were more likely to return to work. Limitation: Generalizability
Wan, 2015 United States41	Patient; full-text	Retrospective database cohort analysis	Eribulin, gemcitabine, capecitabine, vinorelbine	To study indirect costs among metastatic breast cancer patients receiving eribulin vs. other commonly used chemotherapies in the treatment of metastatic breast cancer	43 patients received eribulin, 99 gemcitabine, 55 vinorelbine, and 303 capecitabine. Patients from the MarketScan CCEA and HPM databases	Sick leave, STD, costs	Patients receiving eribulin had either numerically lower or similar STD days/patient/month compared with other therapies. Eribulin treatment was less likely to result in STD leave than gemcitabine. Limitation: Reason for patient sick leave is unknown. Presenteeism, transportation costs, and time spent seeking cancer treatment and care was not included.
Yin, 2015 United States34	Patient; abstract	Retrospective cohort	Not specified		19,496 women with breast cancer at different stages in a commercial health claims database linked to employer records	Missed work	Women treated for breast cancer missed an average of 79 hours per quarter, and disease progression was associated with greater work hours missed. Limitations: Women whose cancer progressed were more likely to be in the employment-based dataset, making estimated productivity effects likely conservative.
Brain tumor/glioblastoma (n = 1)
Gzell, 2014 Australia54	Patient; full-text	Prospective cohort/case series	Radiation, temozolomide	To study effect of glioblastoma progression and treatment on work	112 patients selected from Northern Sydney Cancer Center	Return to work	28% returned to work by 6 months, and 27% returned to work by 12 months. Return to work was related to presence of neurological deficit before starting radiotherapy. Limitations: Most patients were professionals, and few were manual laborers; work sustainability or retention was not assessed.
Colorectal cancer (n = 5)
Bradley, 2011 United States42	Patient; full-text	Simulation model	Chemotherapy	To estimate and project productivity costs of colorectal cancer and to model savings from 4 approaches: risk factor reduction, improved screening, improved treatment, and mix of all 3	Hypothetical population using projections from the NCI Cancer Intervention and Surveillance Modeling Network	Productivity costs	The cumulative cost of productivity loss attributable to cancer decreased from $339 to $331 billion with increased use of adjuvant chemotherapy, but prevention and control strategies were more effective to reduce productivity losses. Limitations: Years of life lost and benefits of risk reduction may be underestimated; not all risk factors were included; model does not provide confidence intervals; surgery and radiation were not included.
Hanly, 2013 Ireland44	Caregiver; full-text	Prospective postal survey questionnaire, cost analysis	Chemotherapy, radiation, surgery	To estimate time allocated to care by informal caregivers of colorectal cancer survivors and effect of using different cost evaluation methods to this time	154 caregivers recruited from National Cancer Registry in Ireland	Time allocated to caring, costs	Caregivers spent mean 42.5 hours/week during diagnosis and initial treatment, and 16.9 hours/week in domestic-related caring activities in ongoing care. Time spent for hospital activities correlated with later stage disease. Limitations: Selection bias, overrepresentation of rectal cancer survivors, and influence of other caregivers may underestimate total costs.
Malaguarnera, 2013 Italy45	Patient; full-text	Prospective clinical study	Capecitabine	To evaluate the effect of chemotherapy with capecitabine in patients with colorectal cancer on work productivity and daily activity	30 patients with stage III disease	WPAI	Significant increase in absenteeism at 1 and 6 cycles, and after 3-month follow-up. Limitations: Small sample size, lack of comparator
Gordon, 2014 Australia43	Patient; full-text	Prospective matched cohort with telephone/postal survey	Chemotherapy, radiation, surgery	To describe transitions in, and key factors influencing, work participation during the 12 months following a diagnosis of colorectal cancer	239 patients at cancer diagnosis and 717 noncancer comparator cohort selected from the Household, Income and Labour Dynamics in Australia (HILDA) Survey	Stopping work, missed work, sick leave	More participants with colorectal cancer had stopped working at 12 months than controls (27% vs. 8%). Treatment resulted in a mean time off of 91 days. Factors related to delayed work re-entry included having or had chemotherapy. Limitation: Low response rate, did not assess if participants had cancer recurrence during the study or their motivation to return to work
Veenstra, 2015 United States46	Patient; abstract	Prospective survey questionnaire	Not specified	To study the impact of paid sick leave on job status of patients with colorectal cancer	1,469 patients working with stage III disease recruited from Georgia and Detroit SEER registries	Work retention, sick leave	During treatment, 46% of patients did not retain their jobs, and 32% reported high financial burden. Paid sick leave was associated with job retention. Limitation: Generalizability
Chronic myeloid leukemia (n = 1)
Hamerschlak, 2014 Brazil55	Patient; full-text	Prospective telephone interview	Imatinib, other therapies	To evaluate the QoL for patients receiving treatment by the public health system in Brazil	443 patients on imatinib and 220 control patients selected from Brazilian Association of Lymphoma and Leukemia (ABRALE)	Functional Assessment of Chronic Illness Therapy tool	Imatinib was associated with a better ability to work. Hydroxyurea and chemotherapy was inversely associated with work capability, while imatinib was associated with better ability to work. Limitations: Reporting bias, small sample size
Head and neck cancer (n = 1)
Isaksson, 2016 Sweden56	Patient; full-text	Prospective longitudinal interview	Chemotherapy, radiation, surgery	To (a) investigate employment status at diagnosis, sick leave, and returning to work patterns in correlation to QoL, anxiety, and depression in patients treated for head and neck cancer and (b) explore patients’ experiences of the process of returning to work	66 patients (treatment with radiation [32%], surgery [2%], surgery/radiation [17%], and chemotherapy, radiation and/or surgery [17%]) recruited from a university hospital	Return to work, EORTC QLQ-C30, EORTC QLQ-H&N35, and Hospital Anxiety and Depression Scale	53% of patients returned to work at 24 months after treatment. QoL parameters were worse in patients who did not return to work. Limitations: Interviews were not tape recorded with risk of missing data; did not address potential differences in age and gender.
Lung cancer (n = 2)
Stanisic, 2010 France, Germany, Italy, Spain53	Patient; full-text	Simulation model	Bevacizumab and platinum-based chemotherapy	To estimate potential economic benefits from a societal perspective in patients returning to work when treated with bevacizumab-based combination therapy	Simulation model for nonsquamous metastatic non-small cell disease estimating 1,481 patients from France, 1,492 from Germany, 820 from Italy, and 745 from Spain identified via epidemiological estimates and 2 phase III clinical trials (E4599 and AVAiL)	Cost benefit analysis using human capital approach	Mean cost savings due to reduction in productivity losses resulted in €21,667 at year 1 and €39,001 at year 1.5 in France; €21,171 and €38,107 in Germany; €7,578 and €31,640 in Italy; and €12,401 and €22,322 in Spain. Limitation: Return to work expected in 20% progression-free patients (sourced from clinical experts)
Lister, 2012 France, Germany, Italy, Spain52	Patient and caregiver; full-text	Simulation model	Bevacizumab or standard chemotherapy	To assess the savings accrued using bevacizumab-based treatment for non-small cell lung cancer	Simulation model for non-small cell disease via publicly available databases or published literature	Cost analysis using Markov model	Gains in productivity and associated reductions in societal expenditure are higher in the bevacizumab treatment arm than in the chemotherapy arm. Limitations: Exclusion of private expenditures
Oropharyngeal cancer (n=1)
Baxi, 2015 United States57	Patient; abstract	Cross-sectional survey	Chemotherapy and radiation	To assess the impact of definitive chemotherapy/radiation on employment and QOL in patients with human papillomavirus+ oropharynx cancer patients	129 patients with advanced, human papillomavirus+ disease recruited from Memorial Sloan Kettering Cancer Center in New York City	Stopping work, missed work, change in responsibilities	During chemotherapy/radiation, 9% of patients stopped working and did not return; 69% took time off; 19% reduced responsibilities; and 3% continued working. There was no significant difference in time off by chemotherapy type.
Ovarian cancer (n = 2)
Calhoun, 2001 United States50	Patient and caregiver; full-text	Prospective cost analysis	Chemotherapy	To study the total costs of chemotherapy-induced toxicity incurred by patients and caregivers	83 patients with neurotoxicity, neutropenia, or thrombocytopenia from Northwestern University Medical School	Patient estimation of resource costs	Indirect costs for hematologic was $2,422 for neutropenia and $2,122 for thrombocytopenia, while neurotoxicity was associated with $2,837 in caregiver work loss. Indirect costs were evident for caregivers as well. Limitations: Small sample size, self-reporting bias.
Angioli, 2015 Italy51	Caregiver; full-text	Prospective cost analysis	Chemotherapy, surgery	To study economic impact of firstline treatment of advanced ovarian cancer on caregiver	172 caregivers for advanced disease recruited from Division of Gynecologic Oncology, University Campus Bio-Medico of Rome	WPAI (adapted from Crohn’s Disease Caregiver), Medical Outcomes Social Support Survey, and Profile of Mood States	Caregivers reported 8.7% loss in work time, 12.8% impairment while working due to caregiving, and overall work productivity loss of 20.67%. Limitations: Only direct nonmedical costs considered, estimation of costs not well discussed.
Pancreatic cancer (n = 1)
Tingstedt, 2011 Sweden58	Patient; full-text	Retrospective registry cohort	Gemcitabine, surgery	To study health care costs and loss to productivity of costs of patients with pancreatic cancer	83 adults selected from Lund University Hospital	Cost analysis using human capital approach	Younger patients received more chemotherapy and CT scans than elderly patients. Approximately 4-5 productive years were lost per patient in addition to the loss of production due to absenteeism while still alive. Productivity losses were estimated at €287,420 per patient less than aged 65 years. Limitations: Small sample size
Prostate cancer (n = 3)
Wilson, 2008 United States49	Patients; abstract	Retrospective registry cohort, cost analysis	Cryotherapy, hormone therapy, surgery	To compare long-term patterns of work loss and costs over 10 years in newly diagnosed patients with prostate cancer	5,414 newly diagnosed patients recruited from CaPSURE, a national disease registry at 31 academic and community urology practices	Missed work, changes in work status, and lost wages	Average weekly work reduction was 15.7 hours in first 6 months, with 88% work status remaining the same over next 6 months. Over 10 years, cumulative wage loss was $146,500, which was greater for patients with androgen therapy, and less for cryotherapy and radical prostatectomy.
Grunfeld, 2013 United Kingdom47	Patient; full-text	Prospective qualitative interview with follow-up	Brachytherapy, chemotherapy, radiation, surgery	To qualitatively study the meaning of work among prostate cancer survivors and to describe the link between masculinity and work following cancer treatment	50 men recruited from outpatient departments of hospitals in 3 U.K. health care trusts	Identification of themes from semistructured interviews	Side effects and worry about future were perceived to have an effect on work. Social stigma and disclosure about cancer was associated with work. Limitations: Small sample size.
Li, 2013 United States48	Caregiver; full-text	Retrospective analysis of survey data, cost analysis	Hormone therapy	To study lost productivity and costs associated with informal caregiving among partner caregivers of localized prostate cancer patients within 1 year after diagnosis	88 caregivers of patients identified from the Family and Cancer Therapy Selection study	Changes in hours, changes in responsibilities, time on caregiving, and costs	44.3% of caregivers were working at baseline, compared with 39.8% at follow-up. Mean working hours decreased from 14.0 hours/week to 10.9 hours/week. Annual economic burden for caregivers was estimated at $6,063 for full-time workers and $2,530 among women working 10-14 hours/week. Limitations: Convenience sample, small sample size, and no measurements on psychological effects of caregiving
Renal cell carcinoma (n = 1)
Yang, 2009 United States59	Caregiver; abstract	Secondary analysis of randomized clinical trial	Temsirolimus or interferon alfa	To analyze the impact of informal caregiving on caregiver’s workplace productivity	416 patients with advanced disease from a phase III trial	WPAI-CG, Caregiver Reaction Assessment Instrument	Absenteeism of 11 hours/week and a 27% reduction in productivity at work was reported. Caring for patients treated with temsirolimus was associated with lower absenteeism (22% vs. 40%, P = 0.0339); lower overall work productivity loss (34% vs. 49%, P = 0.0178); and lower overall impairment in regular activity (29% vs. 38%, P = 0.0305) than interferon.
Thyroid cancer (n = 1)
Borget, 2007 France60	Patient; full-text	Prospective clinical study with survey questionnaire, cost analysis	Recombinant human thyroid-stimulating hormone	To compare the frequency, duration, and cost of sick leave for follow-up control between recombinant human thyroid-stimulating hormone and withdrawal	292 patients recruited from 3 French hospitals	Sick leave, costs	Patients on thyroid-stimulating hormone were less likely to require sick leave compared with withdrawal (11% vs. 33%; P < 0.001). The valuation of indirect costs of absenteeism by the method of friction cost was €1,537 ± 2,899 for active patients treated by withdrawal vs. €454 ± 1,673 for active patients treated by recombinant human thyroid-stimulating hormone. Limitation: Use of prospective and retrospective recruitment

ADLs = activities of daily living; ARR = adjusted relative risk; CCEA = MarketScan Commercial Claims and Encounters database; CI = confidence interval; EORTC QLQ = European Organization for Research and Treatment of Cancer Quality of Life Questionnaire; FMLA = Family Medical Leave Act; HPM = MarketScan Health and Productivity Management database; HR=hazard ratio; HRQoL = health-related quality of life; MDASI = MD Anderson Symptom Inventory; NCI = National Cancer Institute; NHS = National Health Service; OR = odds ratio; QALY = quality-adjusted life-year; QoL = quality of life; SEER = Surveillance, Epidemiology, and End Results Program; STD = short-term disability; USD = U.S. dollars; WPAI = Work Productivity Activity Impairment questionnaire.

TABLE 3.

Description of Studies Assessing Multiple or Undefined Cancer Types

Author, Year Country	Target Population; Publication Type	Study Design; Quality Assessment	Treatment Type	Study Objective	Patient Population/Setting	Measures/Methods Used to Assess Productivity	Main Findings
Multiple cancer types (n = 14)
Bradley, 2002 United States62	Patient; full-text	Prospective interview	Not specified	To examine employment patterns of long-term cancer survivors; to better understand the economic late effects of treatment (any treatment for the mentioned cancers) by examining employment decisions and outcomes	253 long-term survivors (58 lung, 53 colorectal, 73 breast, and 69 prostate cancer) selected from Metropolitan Detroit Cancer Surveillance System (MDCSS)	Changes in work status, modified questions from Health and Retirement Study and the Current Population Survey	54.5% reduced work schedule at least once while undergoing treatment. Limitations from cancer and its treatment were most likely for lifting heavy loads (26%) and keeping pace with others (22%). Limitations: Examined long-term survivors and not total cancer population, recall bias.
Passik, 2005 United States71	Patient and caregiver; full-text	Prospective interview	Chemotherapy	To study the impact of chemotherapy-induced cancer patients’ fatigue on their spouses’ quality of life, including their own levels of fatigue, depressive symptoms, activity levels, work absenteeism, and marital adjustment	40 couples with 1 partner with chemotherapy-induced fatigue or anemia; 25 completed assessment at 1 month (most common cancer: 13 breast, 4 ovarian, and 3 prostate)	Changes in responsibilities, changes in work hours, effectiveness at work, missed work; Caregiver-Fatigue Symptom Inventory, Caregiver Strain Index, Zung Self-Rating Depression Scale, Dyadic Adjustment Scale, Activity Level Rating Scale	28% caregivers handling fewer responsibilities at work; 32% reduced their work hours; and 32% felt less effective overall at work. Caregiver strain related to fatigue at baseline, lower likelihood to engage in social activity, and work at 1-month follow-up. $630,000 in total costs due to chemotherapy-related toxicity, with over half of these costs in lost wages. Limitations: Small sample size, short follow-up, greater fatigue among dropouts
Bradley, 2006 United States63	Patient; full-text	Prospective interview	Chemotherapy, hormone therapy, radiation, surgery	To examine the number of days employed patients undergoing treatment for either breast or prostate cancer were absent from their jobs	239 breast cancer and 206 prostate cancer patients selected from Metropolitan Detroit Cancer Surveillance System (MDCSS)	Missed work	Women who received only surgery missed an average of 26.5 days; women who received surgery in combination with radiation or chemotherapy missed 19.0 and 61.6 days. Men missed an average of 27 days overall, and median days missed were 25 for men treated surgically without hormone or radiation therapy. Limitations: Generalizability, recall bias.
Longo, 2006 Canada68	Patient and caregiver; full-text	Prospective survey questionnaire	Chemotherapy, radiation	To determine financial and family resources burden associated with treatment of cancer	282 patients (74 breast, 70 colorectal, 68 lung, and 70 prostate) recruited from 5 cancer clinics in Ontario	Out-of-pocket costs, financial burden, and missed work	Mean monthly out-of-pocket cost was $213, plus $372 related to travel costs. 20% of patients perceived financial burden to be significant or unmanageable. Limitations: Self-reporting bias; data on patients’ lost time from work not verified due to privacy issues
Amir, 2007 United Kingdom61	Patient; full-text	Prospective postal survey questionnaire	Chemotherapy, hormone therapy, radiation, surgery	To explore the association of rate at which cancer patients returned to paid employment after 18 months of diagnosis, with demographic, clinical, and work perception	267 patients (breast 48%, colorectal 14%, prostate 9%) employed and recruited via North Western Cancer Intelligence Service in Manchester	Return to work, sick leave, job satisfaction, and prospects	82% returned to work; 25% worked during their treatment; and 20% not able to return to work. Length of sick leave (OR = 1.68; 95% CI = 1.23-2.28) and absence of surgery (OR = 0.28; 95% CI = 0.08-0.94) were associated with return to work. Those who returned reported work deterioration due to cancer. Limitation: Reliance on general practitioners to include/exclude respondents
Choi, 2007 South Korea64	Patient; full-text	Prospective longitudinal survey questionnaire	Chemotherapy, radiation, surgery	To study job loss and reemployment in Korean patients with cancer	305 men (stomach 32%, liver 38%, and colorectal 20%) employed and recruited via National Cancer Center Korea	Job loss, reemployment	53% lost job during 24-month follow-up (liver cancer most likely at 63%). At 24 months, 23% had been reemployed, which was more likely for colorectal (31%) and stomach (30%) cancers. Limitations: Generalizability, male patients only
Ohguri, 2009 Japan70	Patient; full-text	Retrospective records analysis	Chemotherapy, radiation, surgery	To evaluate work limitations and attendance rates after employees diagnosed with cancer returned to work from sick leave and to identify related factors for limitations and attendance rates	129 men and 4 women (most common cancer: stomach 24% and colorectal 17%) from employee medical records from a manufacturing company	Limitation in work, changes in hours, and sick leave	Work limitations in 59% of employees upon return, including 27% with alteration of work, 23% with prohibition of shift work, and 41% with prohibition of overtime. Chemotherapy use was correlated with limitations. Decreased attendance related to disease- and treatment-related factors. Limitations: Not all employees were examined before their return to work; small sample size, generalizability
Schwartzberg, 2009 United States72	Patient and staff; full-text	Prospective time motion observational study	Cetuximab or rituximab	To study the impact of infusion reactions on staff time and costs among patients receiving an initial infusion of cetuximab and rituximab	161 patients (71 cetuximab, 90 rituximab) from 27 geographically representative community oncology sites	Time and costs	Total human resource cost accrued through 10 days postadministration was $53.9 (no reaction), $79.9 (mild/moderate), and $210.4 (severe) for cetuximab and $98.7 (no reaction), $110.7 (mild/moderate), and $155.1 (severe) for rituximab. Limitations: No random assignment, time and costs assessed for monoclonal antibody infusions
Song, 2009 United States73	Patient; full-text	Retrospective matched cohort study	Chemotherapy	To examine the impact of chemotherapy-induced neutropenic complications on STD among cancer patients receiving chemotherapy	280 patients with chemotherapy-induced neutropenic complications (breast 29.3%, lung 10.4%, non-Hodgkin’s lymphoma 2.9%, other 57.5%) and 2,154 controls. Patients from the MarketScan CCEA and HPM databases	STD, costs	Patients with complications experienced significantly greater STD leave and indirect costs (3.2 vs. 2.3 days, P = 0.046 and $549 vs. 394, P = 0.050) per month than patients with no neutropenic complications from chemotherapy. Limitations: Claims data availability, indirect cost calculation for caregivers not examined
Tevaarwerk, 2010 United States74	Patient; abstract	Secondary analysis of randomized clinical trial with survey questionnaire	Not specified	To study effect of receiving an increasing number of treatment modalities (e.g., radiotherapy, chemotherapy) as a risk factor for change in employment	1,976 patients (breast cancer 59%) recruited from 7 academic and 32 community medical oncology clinics within trial E2Z02 (ECOG’s SOAPP study)	Change in work status and level	Number of treatment modalities was not a statistically significant predictor for change in employment status. Metastatic disease, age > 45 years, ongoing therapy, and performance status 2-4 was associated with change in employment status and level.
Gudbergsson, 2011 Norway66	Patient; full-text	Prospective postal survey questionnaire	Chemotherapy, hormone therapy, radiation, surgery, other	To conduct a post primary treatment comparative study of cancer patients with short- or long-term leaves	840 patients after primary treatment for the 10 most common invasive types of cancer (breast or gynecological cancer 76%) recruited from 4 hospitals	Sick leave, Work Ability Index	Longer sick leave after primary cancer treatment was associated with lower health status, use of chemotherapy, hormonal therapy and multimodal treatment, and changes in employment due to cancer. Limitations: Sick leave was studied in absolute rather than relative terms.
Mazanec, 2011 United States69	Caregiver; full-text	Cross-sectional correlational study	Not specified	To study health promotion behaviors and work productivity loss in informal caregivers of individuals with advanced stage cancer	70 caregivers (to patients with stages II to IV pancreatic cancer or stages III or IV lung, gastrointestinal, or gynecologic cancer) from Midwestern National Cancer Institute-designated Comprehensive Cancer Center	Caregiver health promotion behaviors, WPAI (adapted from Crohn’s Disease Caregiver), Caregiver Reaction Assessment, and Medical Outcomes Social Support Survey	For absenteeism and presenteeism, the mean percentage of overall work productivity loss due to caregiving was 22.88%. Overall work productivity loss was not significantly correlated with receiving treatment or any other characteristic of the care recipient. Limitations: Convenience sample, generalizability
Cooper, 2013 United Kingdom65	Patient; full-text	Prospective longitudinal survey questionnaire	Chemotherapy, radiation, surgery	To examine the role of clinical, sociodemographic, work and psychological factors in return to work following treatment for breast, gynecological, head and neck, and urological cancers	290 patients (breast 89, gynecological 56, urological 88, head and neck 47) recruited from outpatient department of hospital	Return to work, work values, Illness Perceptions Questionnaire, EORTC QLC-30, Hospital Anxiety and Depression Scale, Fear of Recurrence Scale	Return to work in breast cancer was a median of 30.0 weeks, head and neck cancer (18.3 weeks), gynecological cancer (17.9 weeks), and urological cancer (5.0 weeks). For breast cancer, patients with greater control over the effect of their cancer at work were more likely to return to work sooner. Limitations: Small sample size, lack of gender differences in head and neck patient sample
Hoven, 2013 Sweden67	Caregiver; full-text	Prospective longitudinal telephone interviews	Not specified	To study socioeconomic conditions of parents of children with cancer by means of a longitudinal assessment of work situation, sick leave, and household income	277 caregivers (mothers 139, fathers 138) and 149 children (leukemia 36.2%, central nervous system tumor 20.8%, and other solid tumors 43.0%) recruited from 4 pediatric oncology centers in Sweden	Changes in work status, sick leave, and household income	The greatest period of sick leave was during the child’s treatment phase. Predictors of sick leave longitudinally included being female or a child with more intensive treatment. Limitations: Lack of individual data and comparison group
Cancer type not specified (n = 5)
Ihbe-Heffinger, 2004 Germany79	Patient; full-text	Prospective, multicenter, cost of illness	Chemotherapy	To assess the impact of health care resource utilization of delayed CINV and to estimate costs imputable to CINV	244 patients enrolled from 3 hospitals and 3 office-based facilities in Germany	Direct and indirect costs	53.9% were on sick leave during or after chemotherapy; 33% used health care resources due to CINV. Characteristics associated with high attributable costs imputable to CINV were cisplatin-containing regimen; experience of emesis; and presence of delayed CINV. Limitations: External validity, nonsocietal perspective
Henry, 2008 United States78	Patient and caregiver; full-text	Cross-sectional online/telephone survey questionnaire	Chemotherapy, radiation	To examine the prevalence of chemotherapy or radiotherapy-associated side effects and related treatment burden and correlates of fatigue and missed work days among cancer patients	814 patients receiving chemotherapy and/or radiotherapy derived from Harris Support Care Cancer Interactive’s online chronic illness panel and from a telephone list of physician-diagnosed cancer patients	Missed work, burden of treatment and adverse effects, attitudes toward work (modified work productivity questions from National Health Interview Survey), Functional Assessment of Chronic Illness Therapy	28% reported a clinic visit to treat side effects of this treatment. 43% of patients were employed with an average loss of 26 days due to therapy, and 18 days due to side-effect treatment. Missed work was associated with higher side-effect burden in the regression model. Limitations: Recall bias, self-reporting of side-effect burden
Miedema, 2008 Canada80	Caregiver; full-text	Prospective qualitative interview	Bone marrow transplant, chemotherapy, radiation	To study economic impact on families caring for a child with cancer	28 families in Newfoundland and Labrador, recruited with the assistance of Candle lighters Canada–Newfoundland and Labrador Division (childhood cancer support foundation)	Identification of themes from semistructured interviews	4 major themes emerged: travel expenses for care, loss of income due to reduction/termination, out-of-pocket expenses, and inability to use assistance programs. 61% fathers and 86% mothers changed work hours during treatment. Limitations: Small sample size, questions may not address all concepts associated with work.
Haiderali, 2011 United States25	Patient; full-text	Prospective observational study	Chemotherapy	To prospectively assess the prevalence and burden of CINV	178 patients recruited from 32 oncology specialty care settings in 19 different states	WPAI-Nausea and Vomiting, Functional Living Index-Emesis questionnaire, costs	61.2% reported experiencing CINV. 37.2% of all patients reported reduced daily functioning. Total costs due to CINV were $778.58 per patient from the day of administration through the 5 days following the first cycle of chemotherapy; patients with more severe symptoms typically had higher costs. Limitations: Small size, resource costs extracted from external sources
Blackmon, 2016 United States75	Caregiver; abstract	Retrospective analysis of survey data	Not specified	To study the impact of caregiving on caregivers’ work life and jobs	6,310 respondents from the LIVESTRONG Survey of People Affected by Cancer	Changes in work status and paid time off	41% of survivors reported that caregivers made a work modification: 57% took paid time off, 41% took unpaid time off, 4% switched from full time to part time, and 3% took early retirement.

CINV = chemotherapy-induced nausea and vomiting; EORTC QLQ = European Organization for Research and Treatment of Cancer Quality of Life Questionnaire; STD = short-term disability; WPAI = Work Productivity Activity Impairment.

Types of cancer treatments studied were surgery, chemotherapy, radiation, hormonal, antibodies, and targeted therapies; these treatments were generally studied as multimodal regimens rather than individualized therapies, as would be expected for most cancer treatment. Only 12 (19.4%) studies evaluated specific therapeutic agents, including trastuzumab,³⁶ capecitabine,³⁸ eribulin,⁴¹ gemcitabine,⁴¹ vinorelbine,⁴¹ temozolomide,⁵⁴ imatinib,⁵⁵ bevacizumab,⁵³ temsirolimus,⁵⁹ interferon-alfa,⁵⁹ recombinant human TSH,⁶⁰ cetuximab,⁷² or rituximab.⁷² Various metrics were used to quantify changes in work productivity, most frequently including missed work or hours lost,^{21,22,34,44,49,63,76,78} return to work,20,25,27,31,33,43,54,56,61,65 sick leave and disability,19,23,25,28,66,67,73,79 and activity impairment.^{32,35,37,55,62,70} Ten studies focused on quantification of direct,^68,72 indirect,^{39,42,50,52,53,60,68} or total costs^25,50 as their main study outcome. Fourteen studies (22.6%) involved use of instruments that were designed by investigators to assess productivity. The most common standardized instrument used was the Work Productivity Activity Impairment (WPAI) questionnaire, which was used in 9 studies (14.5%). 25,30,35,37,38,45,51,59,69

Effect of Cancer Treatment on Patient Work Productivity

Overall, 52 studies (83.9%) evaluated effect of work productivity specifically among patients. Forty-four studies (71.0%) evaluated only patient-related productivity, and 6 studies (9.7%) evaluated patient and caregiver-related productivity.^{39,50,52,68,71,78} One study evaluated patient and physician-related productivity, ²⁴ while the final study evaluated patient and staff-related productivity.⁷²

Impairment in Patient Work Productivity.

Most studies conducted on patients noted impairment in patient productivity from cancer treatment. Factors found to be associated with negative productivity were disease progression or severity, ^34,35,39,63 treatment-related side effects that led to missed work, ^{20,22,23,31,35,39,41,43,46,47,54,66,70,77,79} treatment-associated cognitive impairments,³² and lifting heavy loads and keeping pace with others at work.⁶² Chemotherapy was found to be associated with greater limitations after return to work,⁷⁰ longer periods of sick leave or delayed return,^{20,22,23,28,43,66,77} and greater probability of not returning to work.^27,33,40,76 Cumulative effects of multimodal treatments appeared to have greatly affected patients as well. In Bradley et al. (2006),⁶³ women with breast cancer who received a combination of surgery, radiation, and chemotherapy missed an average of 68.6 days from work, compared with only surgery (missed 26.5 days), surgery and radiation (missed 19 days), and surgery and chemotherapy (missed 61.6 days).⁶³ Similarly, Mujahid et al. (2010) found that women receiving mastectomy and chemotherapy were more likely to stop working than women who did not; this relationship was independent of any sociodemographic or other treatment-related factors.⁷⁶ In a qualitative assessment, perceptions of women’s appraisal of return to work were based on side effects from chemotherapy, awareness of cognitive failures, and effect on confidence.³² It was reported that chemotherapy caused cognitive impairments that influenced working life (e.g., shortterm memory, verbal abilities, speed of processing information, and executive functioning).³² A study conducted by Hedayati et al. (2013) also found that cognitive functioning was not associated with return to work; however, this study was limited by a small sample size.⁷⁷

Calhoun et al. (2001) quantified indirect costs associated with chemotherapy-induced toxicity using a simulation model.⁵⁰ They found that indirect costs for patient were $4,220, $3,834, and $4,282 for neurotoxicity, neutropenia, and thrombocytopenia, respectively. Indirect costs for hematologic toxicity were associated with patient work loss ($2,422 for neutropenia [63% of the total indirect costs] and $2,122 for thrombocytopenia [50% of the total costs]). Although it is not clear from the Calhoun et al. article which year the cost was adjusted to (appears to be 1999), it was one of the first studies to highlight the effect of treatment-related toxicity on indirect costs, thereby emphasizing the need for comprehensive assessment of toxicity.

Improvement in Patient Work Productivity.

A few studies described positive effects of treatment on patient productivity, such as return to work during or after treatment. Amir et al. (2007) explored the rate at which cancer patients returned to paid employment after 18 months of diagnosis and found that 82% returned to work, 25% worked during their treatment, and 20% were not able to return to work.⁶¹ The length of sick leave and absence of surgery were associated with return to work.⁶¹ Azarkish et al. (2015) reported that 80% of women in an Iranian study returned to work after breast cancer treatment, with no pain or lymphedema and good supervisor support being the positive predictors of return.¹⁹ Frederix et al. (2013) assessed the effect of breast cancer on work productivity using the WPAI questionnaire in the Netherlands and Sweden; those who had completed therapy had improved work productivity compared with those currently undergoing treatment.³⁰ A simulation model study estimated a decrease in the cumulative cost of productivity loss attributable to increased use of adjuvant chemotherapy in colorectal cancer, although the reduction in productivity losses was more in preventative and control strategies.⁴²

Among specific therapies, the use of imatinib was associated with a better ability to work for patients with chronic myeloid leukemia compared with chemotherapy, which was found to be inversely associated with work capability.⁵⁵ Another study assessing the savings accrued using bevacizumab-based treatment for non-small cell lung cancer was conducted in France, Germany, Spain, and Italy by the same research group.⁵² This study found gains in productivity and reduction in societal expenditure from the bevacizumab-based treatment as compared with standard chemotherapy.⁵² An abstract describing patients with metastatic breast cancer identified a lower need for short-term disability and indirect costs for patients who received eribulin, compared with those who received gemcitabine.⁴¹ Finally, economic estimates of adjuvant trastuzumab for early HER2-positive breast cancer demonstrated an average incremental societal cost benefit of $4,375.89,³⁶ while bevacizumab-based treatment demonstrated societal cost savings compared with standard chemotherapy for non-small cell lung cancer.⁵²

Effect of Cancer Treatments on Caregiver Work Productivity

Overall, 16 studies (25.8%) focused on evaluation of caregiver work productivity. Ten studies (16.1%) focused on caregivers,^{21,26,44,48,51,59,67,69,75,80} and 6 studies (9.7%) focused on caregivers and patients.^{39,50,52,68,71,78}

Caregiver Work Productivity Studies.

Of the studies evaluating only caregiver-related productivity, most found productivity impairment from treatment, with 1 contributing factor being intensive cancer treatment.⁶⁷ Other factors reported were travel and treatment-related expenses and change of work hours during treatment, which affected caregiver work productivity because of increased involvement in household activities and time spent on hospital-related activities, such as treatment, appointments, and adverse effects.^{21,26,44,51,75,80} Estimates for overall caregiver work productivity loss as a function of absenteeism/presenteeism were estimated at 21%,⁵¹ 23%,⁶⁹ and 27%⁵⁹ across 3 different studies. A qualitative study of how men coped after their wives were diagnosed with breast cancer revealed themes such as adaptation in work schedules to accommodate family needs, but also the use of work as an escape from stress at home.²⁶ In a Canadian study of the effect of childhood cancer on the family, loss of income, out-of-pocket expenses, and lack of financial assistance programs collectively contributed to caregiver strain.⁸⁰ An analysis comparing the effect of temsirolimus and interferon-alfa on caregiver productivity found that caregivers caring for temsirolimus patients reported significantly lower absenteeism, lower overall work productivity loss, and lower overall impairment in regular activity as compared with interferon-alfa.⁵⁹

Patient and Caregiver Work Productivity Studies.

Treatment was found to impair work productivity in most studies that were conducted on patients and caregivers.^{39,50,68,71,78} Factors associated with negative effect included time and financial burden associated with treatments side effects, time spent to visit providers for treatment and follow-up, time spent in other patient-related household activities, and fatigue. A study conducted by Henry et al. (2008) assessed the prevalence of chemotherapy or radiotherapy-associated side effects and related treatment burden, and correlates of fatigue and missed work days among cancer patients.⁷⁸ Henry et al. found that 28% of the patients reported a clinic visit to treat side effects of the treatment; 46% reported that the location for these visits was outside their local community; and 43% of patients were employed and had an average annual productivity loss of 26 days due to therapy and 18 days due to treatment side effect.

Discussion

The cost of cancer care in the United States has been steadily increasing as a result of improvements in survival and treatments. With the growing cost of cancer treatments, frameworks by organizations such as ASCO,^7,14 ESMO,⁸¹ MSK Cancer Center,⁸² NCCN,¹² and ICER¹³ assess the true value of these treatments. These frameworks use multiple criteria such as clinical efficacy, toxicity, cost, innovation, burden of illness, patient and societal burden, and ethical and equity considerations to assess the value of oncology treatments, although no single model encompasses the components needed for a full societal assessment.⁸³ Accordingly, there is a gap in meeting the diverse needs of different health care stakeholders. This review uses work productivity outcomes as an example of an often-overlooked potential value assessment for oncology treatment. This systematic review included 62 studies that examined the effect of cancer treatment on work productivity in patients and their caregivers. The overall trend in results indicated a negative effect of cancer treatment on patient and caregiver productivity. Factors associated with negative productivity were many and differed across patients (disease progression and severity, cognitive and neurological impairments, poor physical and psychological status, and receipt of chemotherapy) and caregivers (intensive cancer treatment, travel for treatment of patient, time due to treatment adverse effects, and change of work hours during treatment). There were, however, a few studies that provided evidence for a positive effect of chemotherapy and certain treatments (imatinib, bevacizumab, and temsirolimus) on work productivity, measured against comparators.^52,53,55,59 Notably, these results were found among the limited number of studies detailing specific agents, were for newer therapies, and the benefits were identified in direct comparison with older or standard treatments. These studies rendered the most useful overall results in the review, giving specific indications as to therapy preferences from the productivity perspective.

This review identified a variety of quantitative (hours of work lost and time to return to work); qualitative (change of job status, degree of activity impairment, and patient-reported experiences); and economic (loss of earnings and cost of caregiving and sick leave) measures to quantify work productivity. While providing a breadth of information, estimation of an overall effect is difficult. The most common standardized instrument used across studies to measure work productivity was the WPAI questionnaire, and this was used in only 13.4% of studies. The WPAI is a 7-item assessment that measures the degree of presenteeism, absenteeism, and activity impairment caused by general health status or specific health problems.⁸⁴ The instrument evaluates current employment status, hours worked, hours of work missed, degree of productivity effect, and degree of activity effect. It is a frequently used instrument because of its strong psychometric properties and ease of administration and has been used not only in cancer studies, but also across a host of other chronic diseases. Additionally, it has been adapted for use with caregivers,⁸⁵ which was also employed in studies within this review. Interestingly, there is only 1 cancer-specific WPAI available, for castration-resistant prostate cancer, and it is adapted for patients and for caregivers. However, none of the reviewed studies reported use of this instrument.^47-49 While the WPAI does have significant benefits as a validated measure of work productivity, a major limitation is a patient recall period of only 7 days (to assess work effect), as used for formal validation studies.

An important consideration for variability in productivity could be attributable to differences in populations and, importantly, in countries where these studies were conducted. Since more than half of the reviewed studies were conducted outside of United States, it is conceivable that variations across studies are related to health care systems and social norms of these other countries, both of which may have a greater role in influencing health care decisions, resource allocation, and even work philosophy. The mixed private and public structure of the U.S. health care system may promote less incentive to adopt the societal perspective for assessments of therapeutic agents, compared with single-payer or socialized health systems.⁸¹ There is some evidence in other disease areas, such as rheumatoid arthritis, that shows that differences in a country’s social policy could influence work-disability rates.⁸² However, as the cost of cancer therapy increases, and patients continue to work (either to maintain insurance or pay for out-of-pocket costs or posttreatment as survivors), the intersection of these concepts will become increasingly important to consider.

Recommended Areas for Future Research

The oncology market is changing rapidly, and the advent of innovative treatments has further put increasing drug prices under scrutiny. Thus, areas of future research that should be explored include bringing together clinical, economic, and humanistic outcomes, including patient and caregiver factors, in value assessments of new oncology drugs. Patient-reported outcomes in several forms (e.g., QoL, health status, functioning, and productivity) should be considered for inclusion in clinical trials for new oncology therapies, which has been recommended by ASCO in its updated framework.¹⁴ The development of cancer-specific productivity assessment instruments, particularly given the increasing introduction of oral oncolytic therapies, would be another useful contribution in this area, as well as efforts to better understand patient preferences for meaningful productivity endpoints.

Limitations

This study has a few limitations that should be considered. First, as previously mentioned, the nonspecificity and heterogeneity of studies (among populations, treatments, measurements, and outcomes) included in this review provide limited information on the specific effects of work productivity. Accordingly, a quantitative assessment (via meta-analysis) was not deemed appropriate for this review, nor was a standard assessment of quality for each study able to be performed for this analysis. Many studies were conducted retrospectively, making it difficult to truly distinguish the effect of the cancer disease process versus cancer treatment on work productivity or to eliminate the potential for patient recall bias in the data. Finally, results of U.S. studies may not be directly comparable with international studies, given the differences in disease prevalence, population characteristics, and social and health care policies.

Conclusions

This systematic review details several effects of cancer treatment on patient and caregiver work productivity, including adverse effects, cognitive impairments, difficulty meeting regular job requirements, and time and money needed to undergo treatment. Because of the varied assessments and significant implications of cancer treatment found in this study, we recommend that additional research using standardized measures and quantification of indirect costs is needed to better understand the full economic burden of cancer and value of treatments. This information will be of greater importance as patients become more engaged in their care through shared decision making, where patients take a more active role in choosing the care that best fits with their values and treatment goals.