Depression, anxiety, & loneliness among cancer survivors during the COVID-19 pandemic

Abstract Purpose To assess the impact of the COVID-19 pandemic on depression, anxiety, and loneliness between those with and without a history of cancer. Design This prospective observational study used a quantitative approach. Participants Adult members of the Kaiser Permanente Research Bank (N = 104,640). Methods Participants completed a series of surveys from May to December 2020. The difference in score of depression, anxiety, and loneliness were estimated using linear mixed regression. Findings Among cancer survivors, 21% and 19% met the thresholds for increased risk of depression and anxiety. Among cancer survivors, younger age groups and females reported increased depression, anxiety, and loneliness scores. Conclusions This study highlights the continued necessity of addressing mental health needs and social support in cancer survivors during and after a public health emergency. Implications for psychosocial providers Cancer survivors may need particular resources after cancer treatment to strengthen resilience and improve quality of life.

Recently published longitudinal and cross-sectional studies have reported that cancer patients may be at greater risk for COVID-19 morbidity and the various social and emotional impacts associated with the pandemic compared to those with other chronic conditions. 6,7Furthermore, cancer patients reported increased emotional distress during lockdown, especially recently diagnosed individuals or those who started treatment during the initial months of the pandemic. 8,9Although several studies have evaluated the mental health impacts in those recently diagnosed with cancer during the pandemic, few studies have focused on cancer survivors who are past their primary cancer treatment.
The objective of this study was to assess the mental health impacts of the COVID-19 pandemic between those with and without a history of cancer.Furthermore, we sought to identify those who needed mental health services referral and evaluation and to examine demographic or clinical characteristics which differentially affected cancer survivors.Using survey and electronic health record (EHR) data from participants of the Kaiser Permanente Research Bank (KPRB), we conducted a cross-sectional study of the association between history of cancer and validated measures for depression, anxiety, and loneliness during the first year of the COVID-19 pandemic.

Study sample
The study sample were members of the KPRB that includes EHR information, lifestyle surveys, and biospecimens (saliva or blood) from adult Kaiser Permanente (KP) health plan members (https://researchbank.kaiserpermanente.org/). 10,11Beginning recruitment in September 2015, the biorepository builds on the KP Research Program on Genes, Environment, and Health (RPGEH) cohort, and uses email, direct mail, and in-person outreach to invite all adult KP members to join. 10 The KPRB includes members from all eight KP regions nationwide, including: Colorado, Georgia, Hawaii, Mid-Atlantic States (District of Columbia, Maryland, and Virginia), Northern California, Southern California, Oregon, and Washington.In addition to general recruitment of all adult KP members, the KPRB conducts enhanced recruitment for recently diagnosed cancer patients and cancer survivors. 11All participants in the study sample provided written informed consent.This study was reviewed and approved by the Institutional Review Board (IRB) at KP Mid-Atlantic States, which is the IRB of record for the KPRB.
We included KPRB participants who completed the COVID-19 initial survey and at least one survey during the follow-up period where depression, anxiety, and loneliness were assessed (N = 129,483).Participants with a history of cancer were diagnosed at least 12 months prior to initial COVID-19 survey completion.To ensure ascertainment of adjudicated cancer history information, we excluded those from KP regions with unavailable tumor registry data, including those from KP Georgia and KP Southern California (N = 20,220), those with <12 months of prior KP health plan enrollment (N = 1189), those who had a prior non-breast in-situ or benign/borderline cancer diagnosis (N = 3176) or an invasive cancer diagnosis within the 12 months prior to initial COVID-19 survey completion (N = 145), or those who had missing age, sex, or survey completion date (N = 113).

Measures
After consent, KPRB members were invited to complete a self-administered electronic or paper-based survey that covered demographic factors, including race and ethnicity, and social determinants of health.In addition to the intake survey, the KPRB sent a series of electronic surveys via email starting in April 2020 to assess the impact of the COVID-19 pandemic on KPRB members.These surveys were sent to all KPRB participants who were also KP health plan members as of April 1, 2020.The initial COVID-19 survey collected information on COVID-19 infection, household characteristics, and potential COVID-related risk factors, including cigarette smoking history, using a standardized instrument based on the Behavioral Risk Factor Surveillance System (BRFSS) assessment of nicotine use. 12Among those who completed the initial COVID-19 survey, the KPRB sent up to 11 follow-up surveys between May 2020 through December 2020 to evaluate the impact of the COVID-19 pandemic on KPRB members.][15] All KP regions capture EHR and claims data for research using a standardized, structured common data model, known as the Virtual Data Warehouse (VDW).The VDW is a well-established data model employed by health systems within the Health Care Systems Research Network.Because the health systems included in our analyses are providers of both healthcare and insurance, the VDW allows for nearly complete capture of healthcare visits, diagnoses, and procedures; it also includes comprehensive data on patient characteristics and medication use dating back at least two decades. 16,17For most KP regions, the VDW also includes tumor registry data.These tumor registries employ North American Association of Central Cancer Registries (NAACCR) protocols to identify, confirm, and abstract common data elements for each cancer case occurring within the health system. 18We used the VDW and associated tumor registry data to collect data on cancer history (cancer diagnoses, time since diagnosis, and stage at diagnosis).We also used the VDW to collect information on comorbidities occurring within the 12 months prior to the COVID-19 initial survey and as an additional source of information on demographic characteristics including age, sex, race, and ethnicity.
The three main outcomes for this study were depression, anxiety, and loneliness assessed at any point in time between May 2020 and December 2020.Depression was measured using the Patient Health Questionnaire-2 item (PHQ-2), which asks if an individual has experienced depression over the past two weeks with a score ranging from 0 to 6. 13 Individuals who screen with a score of 3 or higher are at increased risk for a depressive disorder. 13Anxiety was measured using the Generalized Anxiety Disorder 2-item (GAD-2), which asks if an individual has experienced anxiety over the last two weeks with a score ranging from 0 to 6. 14 Individuals who screen with a score of 3 or higher are at increased risk for generalized anxiety disorder. 14Loneliness was measured using the Three-Item Loneliness Scale, with a score ranging from 0 to 9. 15 For each of the measures used, individuals with a higher score are experiencing greater depression, anxiety, or loneliness. 15

Statistical analysis
The difference in score of PHQ-2, GAD-2, and loneliness in those with and without a history of cancer were estimated using linear mixed regression.Participants had to respond to at least one survey in the follow-up period assessing PHQ-2, GAD-2, and loneliness to be included in the study sample.For participants who responded to more than one follow-up survey, statistical models accounted for within-subject correlation.All models adjusted for age group, sex, race/ethnicity, KP region, smoking status (never, former, current), and Charlson comorbidity score (a score to predict risk of death within one year of hospitalization for those with 16 specific comorbid conditions). 19he difference in score of PHQ-2, GAD-2, and loneliness in cancer survivors in relation to demographic and cancer characteristics also adjusted for time since diagnosis and summary stage at diagnosis.We reported mean score, standard deviation, linear mixed regression coefficients, 95% confidence intervals (CIs), and p values for all models.Analysis was completed using SAS 9.4 (SAS Institute Inc, Cary, NC).

Participants
Approximately 48% of those invited completed the initial COVID-19 survey and 53% completed at least one survey assessing depression, anxiety, and loneliness during the follow-up period.Analyses included 16,231 individuals with a history of cancer and 88,409 without a history of cancer (Table 1).Most of the population was between the ages of 65 and 79 (50%), female (63%) and non-Hispanic (NH) White (82%).Fifty-one percent of individuals in this study were diagnosed with cancer more than 10 years ago, while 22% were diagnosed two to five years ago.Sixty-three percent never smoked and 48% had a Charlson comorbidity score of 0. Compared to those without a history of cancer, cancer survivors had a higher proportion of former smokers and those with three or more comorbid conditions (Table 1).

Comparison of PHQ-2, GAD-2, and loneliness
Among those without a history of cancer, 24% met the established thresholds for increased risk of both depression and anxiety.Compared to those without a history of cancer, cancer survivors had a lower proportion of participants who met these thresholds (21% and 19%, respectively).Both groups reported decreases in mean depression, anxiety, and loneliness scores between May and July 2020, however, mean scores increased from August to December 2020 (Figure 1).Compared to those without a history of cancer, cancer survivors had decreased PHQ-2, GAD-2, and loneliness scores throughout the survey period (Figure 1, Table 2).

Conclusions
After accounting for demographic and cancer characteristics, cancer survivors had decreased PHQ-2, GAD-2, and loneliness scores compared to those without a history of cancer from May 2020 to December 2020 during the COVID-19 pandemic.About 21% and 19% of KPRB cancer survivors met the criteria for further referral and screening for depression and anxiety in our study, respectively.Among cancer survivors, there were statistically significant differences between depression, anxiety, and loneliness by age group, sex, race/ethnicity, smoking status, and comorbidity burden.We hypothesized that cancer survivors would be disproportionally impacted by depression, anxiety, and loneliness during the COVID-19 pandemic compared to those without a history of cancer; however, we did not observe this in our study.Our findings conflict with studies on mental health in cancer populations during the pandemic.In a prospective cohort study of 1051 cancer patients and survivors, 48% reported moderate to severe loneliness and deterioration in emotional function during the pandemic. 9However, in a longitudinal COVID-19 survey of 355 cancer patients, there were no clinically meaningful differences in patient-reported mental health outcomes. 20Systematic reviews in this topic area cite considerable increase in emotional distress among cancer patients, primarily because of fear of COVID-19 infection due to immunocompromised status and disruption of healthcare access. 21,22Our findings may differ because 51% of survivors in our study population were diagnosed greater than 10 years prior to the pandemic and most of these were unlikely to be undergoing active treatment.However, our findings could also be an example of resiliency in cancer survivorship.Resilience is a person's ability to maintain or restore psychological and physical functioning when confronted with external stressors. 23In a survey administered to 409 Italian breast cancer patients and survivors, high levels of resilience acted as a protective factor against psychological distress, regardless of COVID-19 exposure. 24][27] Among cancer survivors in our study, both younger age groups and females reported increased depression, anxiety, and loneliness scores compared to older age groups and males during the follow-up period.9][30] A prospective cohort study of 401 cancer patients in Spain reported higher levels of emotional distress in younger women during the pandemic. 28Furthermore, eleven pooled longitudinal cohort studies of 49,993 participants in the United Kingdom reported higher levels of psychological distress during the initial year of the pandemic in women aged 25-44. 5Between 2020 and 2021, the COVID-19 pandemic impacted nearly every activity of daily living, including employment, education, and childcare.It may be that these challenges caused considerable mental health strain in younger, female populations. 29,30mong cancer survivors in our study, NH Black and Asian participants reported decreased depression, anxiety, and loneliness scores compared to NH White participants during the follow-up period.Other studies in both cancer and non-cancer populations during the pandemic have reported mixed or conflicting results regarding race/ethnicity.Several studies among non-cancer populations, including the Women's Health Initiative cohort study and the COVID-19 Southern Cities study, reported a decrease in anxiety and loneliness among Black men and women. 31,32Furthermore, a series of Centers for Disease Control and Prevention (CDC) surveys in 2020 of 9896 people stated Asian respondents reported lower prevalence of anxiety or depressive symptoms. 3In contrast, a prospective cohort study in 890 African American cancer survivors evaluated the impact of the pandemic on patient-reported health outcomes and 40% of participants reported feeling anxious, depressed, and/or isolated. 33Lastly, several studies evaluating mental health during the pandemic in both cancer and non-cancer populations found no differences by race/ethnicity. 5,24

Strengths and limitations
This study includes a large sample of KPRB participants from six different health plan regions in the United States over the first year of the COVID-19 pandemic.This study used validated survey instruments to assess depression, anxiety, and loneliness.Despite these strengths, our study has several important limitations to note.The KPRB includes an insured population, so the results of this study may not be generalizable to populations without insurance.We did not have prior baseline assessments of depression, anxiety, or loneliness, so there was no way to determine the mental health status of participants pre-pandemic.Furthermore, cancer history is derived from tumor registry data which has a lag of one to two years.Therefore, misclassification may have occurred for those newly diagnosed with cancer.Lastly, there may be bias in the outcomes of interest due to self-reporting and the stigma attached to mental health assessment.

Implications for psychosocial providers
Cancer survivors reported decreased PHQ-2, GAD-2, and loneliness scores compared to those without a history of cancer during the initial year of the COVID-19 pandemic.Among cancer survivors, depression, anxiety, and loneliness were higher in younger women.The results of this study highlight the continued necessity of thoroughly assessing mental health needs and social support in cancer survivors, especially during and after a public health emergency.Younger and female cancer survivors may need particular attention after cancer treatment has ended to determine their specific life-course related needs, support their resilience, and help improve their quality of life.

Ethical approval
This study was reviewed and approved under Institutional Review Board (IRB) # 182281-1 by the Kaiser Permanente Mid-Atlantic States IRB, which is the IRB of record for the Kaiser Permanente Research Bank.

Consent form
Informed consent was obtained from all individual participants included in the study.

Figure 1 .
Figure 1.adjusted model of mean PhQ-2 and gaD-2 scores in relation to cancer history.a a adjusted model includes age group, gender, race/ethnicity, Kaiser Permanente region, smoking status, and charlson comorbidity score.

Figure 2 .
Figure 2. adjusted model of mean PhQ-2 and gaD-2 scores in those with a history of cancer by age group.a a adjusted model includes age group, gender, race/ethnicity, Kaiser Permanente region, smoking status, charlson comorbidity score, time since diagnosis, and summary stage at diagnosis.

Figure 3 .
Figure 3. adjusted model of mean PhQ-2 and gaD-2 scores in those with a history of cancer by sex. a a adjusted model includes age group, gender, race/ethnicity, Kaiser Permanente region, smoking status, charlson comorbidity score, time since diagnosis, and summary stage at diagnosis.
Note: Boldface indicates statistical significance (p < .05). a adjusted model includes age group, sex, race/ethnicity, Kaiser Permanente region, smoking status, and charlson comorbidity score.

Table 3 .
linear mixed regression results for PhQ-2, gaD-2, and 3-item loneliness scale among those with a history of cancer in relation to demographic and cancer characteristics (May through December 2020, N = 16,231).