The Effect of COVID-19 Pandemic on Adolescent and Young Adult Cancer Patients and Survivors

Coronavirus disease 2019 (COVID-19) specific psychological stress

Assessed per responses to the 12 questions pertaining to COVID-19 specific psychological stress within the adolescent and young adults (AYA) Cancer Survivor COVID-19 Survey section titled, "COVID-19 Related Distress (Emotional and Physical Reactions) and Health Behaviors.'' This survey includes both a 5-level Likert scale for the respondent's current level of concern (Not at all, A little, Neutral, A lot, Very Much), plus an ordinal 3-level scale for the respondent to rate the perceived level of change compared to before COVID-19 (Less, Same, More). Responses to individual questions will be summarized at each time point as means (for the Likert scale) and percentages (for discrete levels of change), together with 95% confidence intervals. For each question, will also summarize the percentages of patients in each group checking one of the 3 levels (Less, Same, More) indicating whether they perceived a change in that question since before COVID-19.

Survey responses

Will be summarized by group and time point. Associations between endpoints and demographic, treatment-related and resilience variables, as well as differences among groups will be assessed by t-test, analysis of variance or Chi-square test. Non-parametric tests (Wilcoxon rank sum, Kruskal-Wallis, Fisher's exact) will be employed when appropriate. Regression models (e.g., linear, logistic etc.) will also be employed. Change from baseline to subsequent time points in Likert scores will be modeled by mixed-effect models, while blocking on patient to control for repeated measures. Models will include baseline demographic, treatment-related, and resilience factor variables as covariates.

Patient reported outcomes

Will be summarized by group and time point. Associations between endpoints and demographic, treatment-related and resilience variables, as well as differences among groups will be assessed by t-test, analysis of variance or Chi-square test. Non-parametric tests (Wilcoxon rank sum, Kruskal-Wallis, Fisher's exact) will be employed when appropriate. Regression models (e.g., linear, logistic etc.) will also be employed. Change from baseline to subsequent time points in Likert scores will be modeled by mixed-effect models, while blocking on patient to control for repeated measures. Models will include baseline demographic, treatment-related, and resilience factor variables as covariates.

Changes of survey responses

Will be summarized by group and time point. Associations between endpoints and demographic, treatment-related and resilience variables, as well as differences among groups will be assessed by t-test, analysis of variance or Chi-square test. Non-parametric tests (Wilcoxon rank sum, Kruskal-Wallis, Fisher's exact) will be employed when appropriate. Regression models (e.g., linear, logistic etc.) will also be employed. Change from baseline to subsequent time points in Likert scores will be modeled by mixed-effect models, while blocking on patient to control for repeated measures. Models will include baseline demographic, treatment-related, and resilience factor variables as covariates.

Changes in discrete responses

Will be summarized by group and time point. Associations between endpoints and demographic, treatment-related and resilience variables, as well as differences among groups will be assessed by t-test, analysis of variance or Chi-square test. Non-parametric tests (Wilcoxon rank sum, Kruskal-Wallis, Fisher's exact) will be employed when appropriate. Regression models (e.g., linear, logistic etc.) will also be employed. Change from baseline to subsequent time points in Likert scores will be modeled by mixed-effect models, while blocking on patient to control for repeated measures. Models will include baseline demographic, treatment-related, and resilience factor variables as covariates.