Patient‐reported outcomes in advanced NSCLC before and during the COVID‐19 pandemic: Real‐world data from the German prospective CRISP Registry (AIO‐TRK‐0315)

Patients with lung cancer under treatment have been associated with a high risk of COVID‐19 infection and potentially worse outcome, but real‐world data on patient‐reported outcomes (PROs) are rare. We assess patients' characteristics and PROs before and during the COVID‐19 pandemic in an advanced non‐small cell lung cancer (NSCLC) cohort in Germany. Patients with locally advanced or metastatic NSCLC from the prospective, multicentre, observational CRISP Registry (NCT02622581) were categorised as pre‐pandemic (March 2019 to Feb 2020, n = 1621) and pandemic (March 2020 to Feb 2021, n = 1317). From baseline to month 15, patients' health‐related quality of life (HRQoL) was assessed by FACT‐L, anxiety and depression by PHQ‐4. Association of pandemic status with time to deterioration (TTD) in QoL scales adjusted for potential covariates was estimated using Cox modelling. PROs were documented for 1166 patients (72%) in the pre‐pandemic, 979 (74%) in the pandemic group. Almost 60% of patients were male, median age was 66 years, comorbidities occurred in 85%. Regarding HRQoL, mean‐change‐from‐baseline plots hardly differed between both samples. Approximately 15%–21% of patients reported anxiety, about 19%–27% signs of depression. For the pandemic group, TTD was slightly, but statistically significantly, worse for the physical well‐being‐FACT‐G subscale (HR 1.15 [95%CI 1.02–1.30]) and the anxiety‐GAD‐2 subscale (HR 1.14 [95%CI 1.01–1.29]). These prospectively collected real‐world data provide valuable insights into PROs before and during the COVID‐19 pandemic in advanced NSCLC. For the patients, the pandemic seemed to be less of a burden than the disease itself, as there was a considerable proportion of patients with anxiety and depression in both groups.

the pandemic group.Almost 60% of patients were male, median age was 66 years, comorbidities occurred in 85%.Regarding HRQoL, mean-change-from-baseline plots hardly differed between both samples.Approximately 15%-21% of patients reported anxiety, about 19%-27% signs of depression.For the pandemic group, TTD was slightly, but statistically significantly, worse for the physical well-being-FACT-G subscale (HR 1.15 [95%CI 1.02-1.30])and the anxiety-GAD-2 subscale (HR 1.14 [95%CI 1.01-1.29]).These prospectively collected real-world data provide valuable insights into PROs before and during the COVID-19 pandemic in advanced NSCLC.For the patients, the pandemic seemed to be less of a burden than the disease itself, as there was a considerable proportion of patients with anxiety and depression in both groups.
The COVID-19 pandemic has had far-reaching impacts on public health and disease management, and patients with lung cancer are at higher risk of COVID-19 infection.Here, the authors examined how COVID-19 affected patient-reported outcomes for people with advanced nonsmall cell lung cancer (NSCLC).Using data from the CRISP Registry in Germany, the authors assessed patients' quality of life, including anxiety and depression, before and during the pandemic.Although they found considerable anxiety and depression both before and after the onset of the pandemic, the pandemic itself did not seem to pose a particular burden.

| INTRODUCTION
On March 11, 2020, the World Health Organization declared the coronavirus disease 2019 (COVID-19) outbreak a pandemic, referring to the over 118,000 cases of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections around the world. 1 Healthcare systems worldwide were overwhelmed, with priority given to COVID-19 treatment and many extraordinary emergency measures being required to prevent and control spread of the virus. 2,3The COVID-19 pandemic had important implications for the management of other diseases, including malignant diseases 4 such as lung cancer which remains the leading cause of cancer death worldwide. 5tients with lung cancer under treatment have been associated with a high risk of SARS-CoV-2 infection and potentially worse outcome. 6,7][10] Psychological symptoms such as depression together with physical symptoms such as weight loss at the time of lung cancer diagnosis are prognostically relevant and predict worse survival. 11,12While there are some studies on the impact of the COVID-19 pandemic on patient-reported outcomes (PROs) in lung cancer, those analyses either included rather small patient cohorts, 4,[13][14][15] more selective patients from randomised controlled clinical trials (RCTs) 12 or assessed PROs only at one time point. 12,16Pandemic-related longitudinal PRO data from real-world patients with lung cancer, however, are scarce.

| Study design and cohort definition
The CRISP Registry is an open, non-interventional, prospective, multicentre, interdisciplinary clinical research platform collecting data on treatment, patient and tumour characteristics, biomarker testing and outcome of patients with NSCLC or SCLC.The first patient was recruited on December 17, 2015.Currently, more than 9000 patients have been enrolled into CRISP by over 170 certified lung cancer centres, comprehensive cancer centres, hospitals and office-based oncology practices across Germany.[19] For this analysis, eligible patients are aged ≥18 years with histo-

| Questionnaires and statistical analysis
Descriptive statistics were used to summarise patient and tumour characteristics by pandemic status (before vs during the COVID-19 pandemic).
To assess patients' general HRQoL and levels of anxiety and depression over time, patients were asked to fill in questionnaires at time of recruitment (baseline), every 2 months until month 12 and thereafter every 3 months for up to 36 months.Here, we evaluated questionnaires until month 15.Scoring of the questionnaires was performed according to the respective manuals.
HRQoL was assessed by the 36-item questionnaire FACT-L, version 4, consisting of the 27-item FACT-General (FACT-G) covering the four domains physical, social/family, emotional and functional well-being and the 9-item lung cancer subscale (LCS). 20Items are rated on a 5-level Likert-type scale.The FACT-L has a possible total score range of 0-136.Change from baseline was calculated as mean value of the individual difference between baseline and the respective follow-up time point (2, 4, 6, 8, 10, 12 and 15 months).
Anxiety and depression were assessed by the ultra-brief 4-item PHQ-4 including the subscales PHQ-2 for depression and the Generalized Anxiety Disorder Scale-2 (GAD-2). 21,22The PHQ-4 has a possible score range of 0-12.Probable cases of depression (PHQ-2) or anxiety (GAD-2) were defined as a score ≥3 on the respective scale at the respective time point.

Association of pandemic status with time to deterioration (TTD)
in QoL scales (FACT-G, depression, anxiety) adjusted for potential covariates was estimated using Cox modelling.First, TTD for all subscale scores was calculated.Deterioration of QoL was defined as a minimum 10 point worsening on a 100 point scale compared to baseline score.Cox proportional hazards models were then used to assess the hazard ratio (HR) of patients in the pandemic group compared to those in the pre-pandemic group.Deterioration was adjusted for common covariates such as age, sex, BMI, ECOG performance status, CCI, smoking status, histology, presence of metastases (M status) and firstline treatment strategy.HRs and 95% confidence intervals (CI) of pandemic status for all subscales were displayed as forest plot.Time to events were estimated using the Kaplan-Meier method. 23Progression-free survival (PFS) was defined as the interval between start of first-line treatment and the date of progression or death.Patients without such an event before start of second-line treatment were censored at start of second-line treatment or at last contact.Overall survival (OS) was defined as the interval between start of first-line treatment and the date of death from any cause.documented for 1166 patients (72%) in the pre-pandemic and for 979 (74%) in the pandemic sample.
Patient and tumour characteristics for all patients with documented PRO data are summarised in Table 1.In both samples, almost 60% of patients were male.The median age at start of first-line treatment was 66.6 years in patients from the pre-pandemic and 65.2 years for patients included in the pandemic group.The proportion of patients with ECOG 0 was similar in both samples (29% prepandemic; 32% pandemic).At least one concomitant disease was present in 85% of patients in both groups.Bones (28% pre-pandemic; 30% pandemic), brain (21% pre-pandemic; 23% pandemic) and lung (21% pre-pandemic; 24% pandemic), respectively, were the most frequently affected metastatic sites.
Of all patients with documented PRO data from the pandemic sample (n = 979), a positive SARS-CoV-2 test result was documented for 4% (n = 34).Approximately half of those patients with a positive test result (n = 15) presented with severe/critical SARS-CoV-2 symptoms and had to be hospitalised (data not shown).

| Patient-reported outcome
Nearly the proportion of PRO questionnaires were sent back in both groups at baseline (72% pre-pandemic; 74% pandemic).Especially at early time points, the return rate was slightly higher in the pandemic than the pre-pandemic sample: 59% vs 52% (after 2 months) and 52% vs 45% (after 4 months; data not shown).
The mean baseline scores for the FACT-G total scale as well as the proportion of patients with depression (by PHQ-2) and/or anxiety (by GAD-2) at baseline in relation to the monthly numbers of patients recruited are shown in Figure 2.There were no substantial differences between pre-pandemic and pandemic group.
Mean-change-from-baseline plots (mean ± 95% CI) for the FACT-G subscales (physical, social/family, emotional and functional well-being), the LCS and the FACT-L total scale are displayed in Figure 3, together with the proportion of patients reporting depression and/or anxiety.There were neither substantial differences in the mean-change-from-baseline plots between pre-pandemic and pandemic group (Figure 3A-F), nor in the subscale scores of the PHQ-2 and the GAD-2, which were almost identical between both groups (data not shown).In total, the proportion of probable cases of anxiety ranged from 15% to 21%, those of depression from 19% to 27% (Figure 3G, H).
Figure 4 shows the association of the pandemic status with TTD in QoL scales adjusted for potential covariates.HRs and their respective 95% CIs of pandemic status (reference: pre-pandemic group) were close to 1 for all subscales indicating no substantial influence of the pandemic status on TTD.However, TTD was slightly, but statistically significantly, worse for the physical well-being FACT-G subscale To the best of our knowledge, this is the first study on the impact of the COVID-19 pandemic on HRQoL, depression and anxiety over time using prospectively collected real-world data on a large patient cohort of advanced NSCLC.In our patient population, the pandemic had no substantial influence on PROs.For the pandemic group, small but statistically significant deteriorations were observed in physical well-being and anxiety.In both samples, a considerable proportion of patients reported levels of anxiety and depression.
Our results demonstrating that PROs of patients with advanced NSCLC were only slightly affected by the COVID-19 pandemic are consistent with those obtained from other studies.In a crosssectional analysis of data from patients with advanced lung cancer (NSCLC, small cell or mesothelioma) who had been included in two RCTs in the United states, the COVID-19 pandemic did not further decrease HRQoL or increase depression symptoms. 12A crosssectional study on patients with lung, breast or colorectal cancer treated in a day hospital in France found that patients' HRQoL was preserved and levels of anxiety remained low during the COVID-19 pandemic, but the region studied was only slightly affected by the pandemic. 13In a further cross-sectional study from Germany, patients with thoracic malignancies who were treated in a lung cancer centre and surveyed in October 2020 using self-designed questionnaires, reported their HRQoL as mostly unaffected. 16In contrast to our work, however, in all these studies, PROs of patients were only assessed at one time point as a snapshot and not longitudinally.
An Israeli study aimed to longitudinally evaluate PROs in 164 patients with cancer including 41 with lung cancer.A major limitation, though, is that this study does not represent a change over time but rather a longitudinal comparison of separate factors, that is, initial COVID-distress at the beginning of the first wave compared to levels of depression and anxiety 1 month later. 14ile PROs of patients' hardly differed before and during the pandemic, those of healthy individuals did as shown by the literature: whereas normative samples showed mean PHQ-4 scores of 1.92 for women and 1.56 for men, 22 scores of participants surveyed during the first nationwide lockdown in Germany were 3.91 for women and 3.21 for men 24 and thus largely comparable to the mean baseline PHQ-4 scores observed for patients included in this study (3.6 for pre-pandemic, 3.5 for pandemic).However, this comparison should be interpreted with caution, since both cohorts differ and the mean age of the sample including healthy participants (40 years) 24 is much lower than that of our cohort (66 years).Levels of anxiety and depression prior to and during the pandemic (about 15%-21% and 19%-27%, respectively) are almost in line with those reported for the general population related to the COVID-19 pandemic: in a systematic review and meta-analysis, the pooled prevalences of anxiety and depression of the general public across different countries were 32% and 27%, respectively. 25Regarding HRQoL, the mean baseline FACT-G total score of 71 for both the pre-pandemic and the pandemic group is within the range of 68-79 observed in other patient cohorts with advanced lung cancer 12,26,27 and in a sample of the US general population recruited during the COVID-19 pandemic. 28However, HRQoL F I G U R E 3 Change in HRQoL and reported anxiety/depression from baseline to month 15.Mean-change-from-baseline plots (mean ± 95% CI) for the FACT-G subscales physical well-being (A), social well-being (B), emotional well-being (C), functional well-being (D) and the lung cancer subscale LCS (E), as well as for the FACT-L total scale (F).For one of the 979 patients with documented PRO data from the pandemic group, data were not reported for all subscales and the patient was therefore excluded from this analysis.For better readability, a line graph format was chosen; however, individual data points are shown and not a change over time.Proportion of probable cases of depression (G) and anxiety (H) for patients from the pre-pandemic and pandemic sample.Change from baseline is calculated as mean of the individual difference between baseline and respective time point for all patients with data available at baseline and the respective time points.Probable cases of depression (PHQ-2) or anxiety (GAD-2) are defined as a score ≥3 on the respective scale at the respective timepoint.Since the vast majority of patients included in our study were of stage IV, however, they were assigned high priority for visits for the continuation of (outpatient) treatment even during the COVID-19 era according to adapted guidelines by the European Society of Medical Oncology (ESMO), 16,31 the German Society for Hematology and Medical Oncology (DGHO) 32,33 and other societies. 34,35 summary, our results indicate no major additive effect of the existential threat posed by the COVID-19 pandemic beyond cancer as one might suspect. 12,15Most patients from the CRISP cohort appeared to be already highly burdened prior to the pandemic, so that outcomes were only minimally affected or worsened by the pandemic-in contrast to healthy individuals. 24Nevertheless, one should not neglect the small but statistically significant deteriorations in physical well-being and anxiety observed for patients during the pandemic.Worsening in physical well-being may be the result of the nationwide lockdowns and the related stay-at-home orders.In addition, services such as physical therapy and sports facilities were sometimes not available during the height of the pandemic.Exercise has been shown to benefit patients with cancer even in an advanced stage, for example, by increasing control of treatment-related symptoms and improving patients' QoL. 16,36r some later time points, the mean changes from baseline, such as for social well-being, differed between the pre-pandemic and pandemic groups.However, this difference was less than one score point which is below the clinically meaningful change score of two to three points. 37sides the small deterioration in anxiety by the pandemic, the considerable proportion of patients reporting anxiety and depression both before and during the pandemic is of particular concern.Lower levels of anxiety and depression have been shown to be positively associated with treatment adherence. 2,38Furthermore, as the mean baseline FACT-G total score has been proven as a predictor of survival of advanced lung cancer, individuals with lower baseline HRQoL have a shorter survival than those with higher baseline HRQoL scores. 26[41] Assessing PROs and developing strategies to minimise levels of anxiety and depression while maintaining long-term HRQoL should therefore be a priority in the management of lung cancer, not only during times of crises but by default.

| Limitations
Major strengths of this study are the prospective design and the analysis of data from a large real-world cohort.Another strength is the longitudinal assessment of PROs taking account several COVID-19 waves and associated restrictions, which, to our knowledge, has not been done so far.Since both recruitment periods (pre-pandemic, pandemic) covered 1 year, possible seasonal effects were also considered.We are aware that some patients in the pre-pandemic group-depending on their date of inclusion-partially reported on their PROs during the pandemic, especially at later time points if they were still alive.After performing a sensitivity analysis between patients included in the first half-year (March to August) and those included in the second half-year (September to February), however, we observed similar findings to those from the primary analysis.
Although patient and treatment characteristics were largely similar between both groups, this comparison is based on descriptive statistics hampering the interpretation of results.As patients were enrolled at the time of their diagnosis, there is no information on when the first symptoms occurred prior to diagnosis and whether this period has thus been extended by the pandemic.Although the COVID-19 pandemic had no substantial impact on PROs according to our results, we cannot comment on the extent of disruptions in routine practice and on patients' accessibility to outpatient care, as this was not subject of this article.Our findings reflect the specific pandemic situation in Germany and might not correspond to the outcomes in other regions with different health care systems and other strategies to cope with the outbreak.

| CONCLUSIONS
These prospectively collected real-world data on a large patient cohort from the CRISP Registry provide valuable insight into the characteristics and long-term PROs of patients with advanced NSCLC prior to and during the COVID-19 pandemic in Germany.For the patients, the pandemic seemed to be less of a burden than the disease itself, as PROs only slightly differed, but a considerable proportion of patients reported high anxiety and depression scores in both groups.
We used the prospective, German clinical research platform CRISP to show characteristics, PROs and clinical outcome in a large cohort of patients with locally advanced, inoperable or metastatic non-small cell lung cancer (NSCLC, stage IIIB/C, IV) before and during the COVID-19 pandemic.In Germany, the first cases of SARS-CoV-2 infections were confirmed at the end of January 2020.In March 2020, the German government announced the first nationwide lockdown.For this analysis, eligible patients with start of first-line treatment between March 1, 2019 and February 29, 2020 were included into the pre-pandemic group, those beginning first-line treatment between March 1, 2020 and February 28, 2021 into the pandemic group.We analysed PROs over time in relation to pandemic status, assessing health-related quality of life (HRQoL) with the Functional Assessment of Cancer Therapy-Lung (FACT-L) questionnaire and symptoms of anxiety and depression using the Patient Health Questionnaire 4 (PHQ-4).
logically confirmed diagnosis of metastatic or locally advanced NSCLC, if ineligible for curative surgery and/or radiochemotherapy (stage IV or stage IIIB/C according to UICC 8; here, collectively referred to as 'advanced').Patients with start of first-line treatment between March 1, 2019 and February 29, 2020 were included into the pre-pandemic group, those with start of first-line treatment between March 1, 2020 and February 28, 2021 into the pandemic group.Data cut for this analysis was June 30, 2022.
Patients alive or lost to follow-up at data cut (June 30, 2022) were censored at last contact.All analyses were calculated using R software, version 4.0.5 (2021-03-31), and SAS software, version 9.4 of the SAS System for Windows.Copyright 2002-2012 SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc., Cary, NC, USA.

3 | RESULTS 3 . 1 |F
Patient and tumour characteristicsFrom 7774 evaluable patients with advanced stage IIIB/C and IV NSCLC recruited into CRISP until data cut for this interim analysis (flow chart see Figure1), a total of 1621 patients were included in the pre-pandemic group (start of first-line treatment between March 1, 2019 and February 29, 2020), 1317 in the pandemic group (start of first-line treatment between March 1, 2020 and February 28, 2021).Thus, recruitment was slightly higher in the pre-pandemic than in the pandemic period (see also Figure2).PRO data have been I G U R E 1 Flow chart.Patient flow chart of all patients with advanced stage IIIB/C and IV NSCLC included in this analysis with start of 1st-line treatment between March 1, 2019 and February 29, 2020 (pre-pandemic sample) or between March 1, 2020 and February 28, 2021 (pandemic sample), starting from the total number of patients recruited into the CRISP Registry from December 2015 until June 2022.NSCLC, non-small cell lung cancer; PRO, patientreported outcome.

First-line treatment
characteristics and clinical outcome of patients are summarised in Table2.The majority of patients were treated with a checkpoint inhibitor (CPI) in first-line, either as monotherapy (17% pre-pandemic; 15% pandemic), or in combination with chemotherapy of patients in risk group[%]   Probable case of anxiety Probable case of depression(B)F I G U R E 2 Recruitment and QoL, depression and anxiety at baseline during pre-pandemic and pandemic period.Number of recruited patients and mean baseline scores for FACT-G total scale (A) and proportion of depression (by PHQ-2) and anxiety (by GAD-2) at baseline (B) during prepandemic (March 2019 to February 2020) and pandemic (March 2020 to February 2021) periods.For better readability, a line graph format was chosen; however, individual data points are shown and not a change over time.FACT-G TS, Functional Assessment of Cancer Therapy-General total scale; GAD-2, Generalized Anxiety Disorder Scale-2; PHQ-2, Patient Health Questionnaire-2.T A B L E 1 Patient and tumour characteristics of patients with documented PRO data at baseline.
HR 1.15 [95% CI 1.02-1.30])andthe anxiety GAD-2 subscale (HR 1.14 [95% CI 1.01-1.29])forthe pandemic group.T A B L E 2 a Unless otherwise indicated.bThere are no specifications as to the timing, frequency or criteria of tumour assessment, thus registry response data should be considered as the best clinical approximation and might not be identical to the response determined in clinical trials.
29me to deterioration.Association of pandemic status with TTD in scales of the FACT-G and with anxiety (GAD-2) and depression (PHQ-2) questionnaires.Displayed are the HRs of being in the pandemic group for deterioration of 10 points on a 100 point scale, adjusted for age, sex, BMI, ECOG performance status, CCI, smoking status, histology, presence of metastases (M status) and first-line treatment strategy.BMI, body mass index; CCI, Charlson comorbidity index; ECOG, Eastern Cooperative Oncology Group; EWB, emotional well-being; FACT-G, Functional Assessment of Cancer Therapy-General; GAD-2, Generalized Anxiety Disorder Scale-2; FWB, functional well-being; HR, hazard ratio; PHQ-2, Patient Health Questionnaire-2; PWB, physical well-being; TS, total scale; SWB, social well-being; TTD, time to deterioration.is lower compared to a normative sample of the general population in Austria reporting a mean FACT-G total score of 86.5.29Similar to the PRO results, the clinical outcomes observed in this study were also comparable between the pre-and the pandemic group (median PFS of 7.8 vs 8.0 months, median OS of 16.7 vs 16.6 months).This contrasts with an Italian work showing a worse