Unemployment and health-related quality of life in melanoma patients during the COVID-19 pandemic: A web-based cross-sectional study

Background The outbreak of coronavirus disease-2019 (COVID-19) ineluctably caused social distancing and unemployment, which may bring additional health risks for patients with cancer. To investigate the association of the pandemic-related impacts with the health-related quality of life (HRQoL) among patients with melanoma during the COVID-19 pandemic, we conducted a cross-sectional study among Chinese patients with melanoma.


Results
A total of 70 patients with melanoma completed the study. The mean age of the patients was 55.2 ± 14.8 years, 45.7% (32/70) were male, and 21.4% (15/70) were unemployed since the epidemic. Unemployment of the patients and their family members and income loss were signi cantly associated with a lower FACT-G score, while the MS score was associated with the unemployment of the patients' family members.

Conclusions
Our ndings suggested that unemployment is associated with impaired HRQoL in melanoma patients during the COVID-19 epidemic.

Background
The outbreak and pandemic of the coronavirus disease 2019 (COVID-19) brought profound impacts on the entire society and the individual's life. Besides the con rmed and suspected cases in hospitals or health facilities, the majority of people started the self-isolation at home voluntarily. However, the lockdown resulted in income loss and unemployment in some people. Under the circumstances, patients with cancers or chronic conditions may face higher risks of job loss and more mental and physical stress.
Melanoma is the most serious type of skin cancers with a highly aggressive ability. The 5-year survival rate can be low as 5% for advanced melanoma [1], and the median overall survival time is about 8 months [2]. Such poor survival outcomes result in heavy mental stress and impaired quality of life.
Approximately one-third of patients with melanoma reported psychological distress to some extent, and symptoms of anxiety were more prevalent than depression [3]. Neoadjuvant therapy such as checkpoint therapy and target therapy has been rapidly developed in recent years, and signi cantly bene ted melanoma patients. However, high medical costs imposed heavy nancial burdens on these patients. It was estimated that over 25,000 US dollars were cost per person annually for patients in late stages [4]. As a result, unemployment and income loss may not only result in emotional problems but also lead to unaffordability during the advanced treatment for melanoma.
In the current study, we investigated the association of unemployment, income loss and other epidemicrelated impacts with the health-related quality of life (HRQoL) in melanoma patients in China, based on an online questionnaire survey.

Methods
A cross-sectional study among Chinese melanoma patients was conducted between 4 Apr, 2020 and 11 Apr, 2020. The online survey link to facilitate the collection of questionnaires was distributed on social media (WeChat groups and teledermatology platforms). Every single IP address was allowed only one entry and submission in order to avoid repeated submissions by individual patients. The nal submission required the patients to complete all the questions. The study was reviewed and approved by the institutional research ethics boards of Xiangya Hospital, Central South University (approval number: 202002024). Electronic informed consent was collected from all participants before the survey.
The exposure variables (employment status, income change and outdoor activity restriction) and covariates (gender, age, educational level, annual income, marital status, clinical stage of melanoma, course of melanoma, site of melanoma, current status of disease, comorbidities, adherence to treatment, and healthcare utilization) were collected. Functional Assessment of Cancer Therapy-Melanoma (FACT-M), an outcome measurement system, was use to evaluate the quality of life in melanoma patients (www.facit.org). FACT-M is composed of a core questionnaire, the FACT-General (FACT-G), and a diseasespeci c module including melanoma subscale (MS), and melanoma surgery scale (MSS) [5]. In our study, the primary outcome was the FACT-G score (the sum of physical well-being (PWB), social/family wellbeing (SWB), emotional well-being (EWB) and functional well-being (FWB)), and the secondary outcome was the MS score. The possible score ranges of the FACT-G and MS were 0-108 and 0-64, respectively, and higher scores indicate better HRQoL. The MSS was not analyzed since not all patients received surgery. The use of the FACT-M was authorized, and the Chinese version was obtained from the copyright holder before the survey.

Statistical Analyses
There was no missing data in the questionnaire survey. The data were exported from the online survey system and analyzed with R version 3.5.2. Continuous variables with normal distribution were expressed as mean ± standard deviation (SD) and compared with analysis of variance (ANOVA). Hotelling's T2 test was used for the multivariate comparison of multiple subscale scores. Continuous data with skewed distribution were presented as median (interquartile range, IQR) and compared with Wilcoxon rank sum test. Categorical variables were summarized as counts (percentages) and compared using the chi-square test or Fisher's exact test. Stepwise linear regression with the corrected Akaike information criteria was used to identify variables that were associated with the outcomes. The effect size of the association was presented as regression coe cient and 95% con dence interval. P-value less than 0.05 was considered statistically signi cant. Reporting of the results followed the STROBE guideline.

Results
From April 4, 2020 to Apr 11, 2020, a total of 70 valid questionnaires was collected and analyzed. None reported con rmed infection with COVID-19. The mean age of the patients was 55.2 ± 14.8 years, and 32 (45.7%) were male. The characteristics of participants by employment status were shown in Table 1.
Educational level, marital status, history of hypertension, income change since the epidemic, and employment status of family members were signi cantly different across the groups. The intracluster correlation coe cients of the subscales of FACT-M varied from 0.82 to 0.92, indicating good internal reliability. The mean FACT-G and MS scores were 72.1 ± 17.1 and 52.5 ± 7.6, respectively. The distribution of the FACT-G and MS scores by employment status are shown in Fig. 1a, with a clear positive correlation between the two scores. According to the Hotelling's Trace for multivariate test, unemployment was signi cantly associated with the PWB, SWB, and FWB subscale scores, but was not associated with the EWB and MS subscale scores (Fig. 1b). The FACT-G score was the lowest in patients who were unemployed since the epidemic (P = 0.003); in contrast, the difference was not signi cant between the patients who were unaffected and those who were unemployed or retired before the epidemic (Fig. 1c). Income loss and unemployment of family members were also signi cantly associated with lower FACT-G but not MS score, while adherence to treatment and isolation status were not associated with both outcomes (Fig. 2). It is noteworthy that the patients who reported income loss had a lower proportion of immune therapy at present (9%) compared to those whose income were unaffected (17%).
The stepwise linear regression identi ed lower educational level, unemployment since the epidemic, and progression of melanoma as risk factors for impaired HRQoL measured as the FACT-G ( Table 2). The progression of melanoma and unemployment of family members were signi cantly associated with the MS score.

Discussion
In this study, we investigated the pandemic-related impacts on the HRQoL in melanoma patients through an online survey. The impaired HRQoL of melanoma was associated with the unemployment of the patients and their family members, as well as income loss. These ndings provide new insights into the health inequity issue arisen in the particular period.
It is well established that the health status and psychological characteristics of patients are associated with HRQoL. Previous studies found that patients with melanoma perceived persistent worries about developing new or metastatic cancers [6]. A pervasive sense of uncertainty increases psychological distress, and impairs quality of life [7]. Investigators found that systemic therapies can decrease HRQoL in a short time; however, the long-term HRQoL of melanoma survivors is comparable with the general population [8,9]. In the current study, we found that over 20% of the melanoma patients were unemployed since the epidemic of COVID-19 and their HRQoL were decreased notably. Interestingly, unemployment and a history of melanoma were previously identi ed to be associated with greater cancer worry, which indicated the importance of both melanoma and unemployment on psychological status [10]. Our ndings further emphasize the need to track the mental health of cancer patients under unemployment.
It has been estimated that over 100,000 people will be diagnosed with melanoma this year in the United States according to American Society of Clinical Oncology (ASCO). The incidence rate of melanoma in China was estimated at 0.9 per 100,000, with a 110.3% rise compared to the 1990s [11]. The increasing incidence of melanoma makes it a public health concern and brings the global burden of disease [12]. The disability-adjusted life-years (DALYs) of melanoma have been increasing in China during the last decade, while the DALYs of common cancers such as esophageal cancer and stomach cancer have decreased signi cantly [11]. In recent years, the development of immune therapy bene ts cancer patients substantially, but also leads to a heavy economic burden for them [13]. A review indicates that higher spending on cancer is consistently associated with lower mortality [14]. In other words, loss of income and work-related bene ts experienced by the unemployed may lead to adverse health outcomes [15]. Unemployment of the family members further impairs the social support and nancial source, resulting in the unaffordability of the advanced therapies and, the progression of cancer and mental stress. This is supported by our nding that unemployment of family member was an independent risk factor for lower MS score. Additionally, a lower proportion of the use of immune therapy during the epidemic period reported by the melanoma patients with income loss indicates that nancial status is an essential consideration in the selection of treatment strategy. These ndings indicate that health disparities in cancer could result in a vicious circle. Social setting, incorporating poverty, culture, and social justice were found to play a part in disease outcome [16]. Some poverty-related barriers, such as income, education, and health insurance could in uence melanoma outcomes [17]. This makes social justice quite essential to diminish the disparities in melanoma.

Conclusion
To the best of our knowledge, this is the rst empirical study of pandemic-related impacts on the HRQoL of melanoma patients. Our nding indicates that early and timely mental health intervention, telemedicine, and health education are needed for melanoma patients. More importantly, health decisionmakers should consider the reimbursement policy of advanced treatment methods for cancers with high fatality rate.

Limitations
There are limitations to our study. First, we conducted an online survey instead of the in-person interview because of the high infectivity of the COVID-19. Therefore, the representativeness of our sample might be limited owing to selection bias. Second, a total of 70 patients participated in the study despite our efforts to recruit as more patients as we could. The sample size is relatively small which may lead to insu cient power of test to identify differences. Third, the exposure and outcome variables were self-reported, and recall bias might be introduced. Last, the survey was conducted among Chinese patients, which may not fully represent melanoma patients beyond China owing to the differences in culture, reimbursement policy, and social system. Electronic informed consent was collected from all participants.
Consent for publication: Electronic informed consent for publication was obtained from all participants.
Availability of data and materials: The datasets generated and analysed during the current study are not publicly available but are available from the corresponding author on reasonable request. Violin-boxplot for the distribution of the FACT-G and MS scores by income loss, unemployment of family members, isolation status, and adherence to treatment.