Validation of the European Oncology Quality of Life Toolkit

Several patient-reported questionnaires have been developed and validated to measure quality of life (QoL) of cancer patients. However, a comprehensive tool, developed in close collaboration with people who have experienced cancer, widely applicable across Europe and tailored to the health status of the individual patient, is lacking. Such a tool will be important for incorporating the patient's perspective into the evaluation of policies and programs addressing cancer at the European level.

The European Oncology Quality of Life Toolkit (EUonQoL-Kit) is a unified patient-centred tool for the assessment of QoL based on preferences of cancer patients and survivors. It is developed from a patient perspective, administered digitally, available in the languages of the European Union (EU) 27 and several associated countries, and applicable in future surveys to contribute to the EU's mission on cancer. It includes three static questionnaires specifically developed for different disease phases (patients in active treatment, survivors, and patients in palliative care), and three dynamic versions of the same questionnaires based on Item Response Theory (IRT) and Computer Adaptive Testing (CAT).

This is an observational study aimed to perform the psychometric validation of the EUonQoL-Kit through its first large scale application in a pan European pilot survey.

The EUonQoL-Kit will be tested in a sample of cancer patients and survivors from 45 oncological centres located in 33 European countries. The sample will include patients in active treatment (group A), survivors (group B), and patients in Palliative Care (group C). Each centre will recruit a sample of 100 participants (40 from group A, 30 from group B, 30 from group C), stratified for primary diagnosis: lung cancer, breast cancer, colorectal cancer, haematological malignancies, prostate cancer, others.

In each centre, data collection will be performed in pre-identified outpatient clinics and inpatient wards. Participation in the study will be offered to eligible patients until the pre-defined sample size is reached.

All patients in the sample will fill in the EUonQoL-Kit. Additionally, three subgroups of patients will be proposed to fill in the following additional questionnaires:

• FACT-G and EQ-5D-5L, to evaluate concurrent validity ("concurrent validity" sub-sample: 10% of the overall sample, stratified for the three disease conditions: A, B, and C).
• Live CAT version of the EUonQoL-Kit, to validate the static and dynamic versions against each other ("live CAT" sub-sample: 10% of the overall sample, stratified for the three disease conditions).
• EUonQoL-Kit, a second time, (at least 1 hour after the first administration), to assess test-retest reliability ("test-retest" sub-sample: 10% of the overall sample, stratified for the three disease conditions).

The development of the EUonQoL-Kit has involved multiple stakeholder groups, also including patients, through different steps:

• Systematic search, collection, and analysis of existing validated QoL tools, metrics, item banks in databases to identify QoL dimensions not adequately covered by existing tools.
• A mixed-method study including patient interviews, a Delphi survey with patients and healthcare providers, aimed at collecting patient priorities and preferences on QoL dimensions, followed by a usability testing of a preliminary version of the EUonQoL-Kit.
• Development of the EUonQoL-Kit, including three static questionnaires specifically developed for patients in active treatment, survivors and in palliative care, and three dynamic versions of the same questionnaires. The final version of the six questionnaires will be submitted as an amendment to the current protocol when available.
• Translation and cultural adaptation of the EUonQoL-Kit across European countries.

The CAT version of the EUonQoL-Kit requires an item bank containing IRT-calibrated items (questions) and an algorithm for selecting the most relevant item to ask, based on the previous answers. Such dynamic system ensures that each patient is asked the most relevant and informative items.

CAT/IRT-based technology will be implemented as:

• Short forms within the EUonQoL-Kit static version. Three static versions will be designed to secure the optimal fit between items and patient groups.
• Live CAT, within the EUonQoL-Kit dynamic version. Each item is selected based on the previous answers, while the patient completes the questionnaire.

The final EUonQoL-kit will include both traditional and CAT/IRT-based items.

The overall sample size is planned to be 4,500 patients. Expecting a lower recruitment rate of 10-15%, at least 4,000 patients will be enrolled, a number suitable for data analyses.

Scientifically sound recommendations on statistical power/sample size in validation studies are lacking and minimum rule-of-thumb requirements are provided. The highest number of participants recommended for exploratory and confirmatory factor analyses (EFA \& CFA) is 1,000; thus, the sample size by subgroup mentioned above is appropriate to evaluate construct validity.

Analyses of the primary aim will include:

• EFA \& CFA to assess structural validity. The sample will be divided into two random sub-samples, stratified by disease stage and country. The first sub-sample will be used to perform EFA, and the model obtained will be confirmed by CFA in the second sub-sample. Goodness-of-fit will be measured by the Root Mean Square Error of Approximation, Comparative Fit Index, and Tucker-Lewis Index.
• Distribution of scale scores, examined by calculating the observed range, floor and ceiling effects, and statistics of central tendency and dispersion.
• Reliability in terms of internal consistency (estimated with Cronbach's alpha coefficient of multi-item scales) and reproducibility (agreement between the two administrations of EUonQoL-Kit estimated with the Intra-class Correlation Coefficient).
• Concurrent validity, assessed by the multi-trait multimethod matrix between the EUonQoL-Kit and the FACT-G/EQ-5D-5L questionnaires.
• Construct validity based on hypotheses-testing, assessed through the patterns of EUonQoL-Kit scores across known groups defined by variables such as disease stage, treatment, and Eastern Cooperative Oncology Group performance status. Mean differences among groups will be tested with ANOVA, and the magnitude of the difference between them will be estimated by Effect Size coefficient: > 0.8 high, 0.5 moderate, and 0.2 low.
• Test of Differential Item Functioning, used to assess item equivalence.

Analyses of the secondary aims will include:

• Assessment of acceptability and respondent burden, based on the response rate and percentage of missing items, as well as the time needed to complete the questionnaire.
• Validation of the static and dynamic versions of the EUonQoL-Kit against each other, performed by 1) comparing scores obtained with the static and dynamic versions to test if they produce interchangeable results; and 2) assessing whether the items selected for the static versions should be adjusted to obtain optimal assessment.
• To analyse QoL inequalities across clusters of populations, countries, and healthcare systems, multi-level regression models will be applied across three levels: individual, cancer centre and country. It will also be explored whether the magnitude of subgroup differences (e.g., for age or education) varies across countries.
• Analysis of QoL disparities across different patient groups will include gender, ethnicity, religion, geographical area, psychological and socio-economic factors, and education level.

Data collected include personal information belonging to special categories, like health-related data, origins, lifestyles etc. Legal basis to collect and process information is the data subject consent, which is necessary to take part in this study. A data protection expert will assess the impact on data protection throughout the duration of the study. Patient registration and CRF data collection will be centralized through the CRF.net platform, developed and owned by Istituto Nazionale Tumori, IRCCS - Fondazione G. Pascale, Napoli (INT-NA) and released with the GPL v3 license. Data will be stored by INT-NA on cloud services provided by Telecom Italia S.p.A. and located in Acilia (Italy).

Patient-reported data associated to QR codes will be collected by an ad hoc mobile app developed by Clinical Research Technology (CRT), Salerno (Italy).

The CAT version of EUonQoL-Kit will be administered through the integration of a CAT engine provided by the European Organization for Research and Treatment of Cancer (EORTC), sited in Bruxelles (Belgium).

Data processing is managed by:

• Istituto Nazionale Tumori, IRCCS - Milan (INT): INT serves as the data controller.
• Istituto Nazionale Tumori, IRCCS - Fondazione G. Pascale, Napoli (INT-NA). INT-NA will be appointed as data processor under article 28 GDPR and will be responsible for patient enrolment, questionnaire assignment through the generation of a QR code, clinician reported data collection through the electronic Case Report Form (eCRF), management of the interface with the CAT engine and with the mobile app for the administration of EUonQoL-kit items.
• CRT: based on a QR code provided by CRF.net platform, the mobile app administers specific QoL questionnaires for specific subgroup of patients. For each item administered, the app will directly transfer the answer to the platform, without recording/storing any content. Such a procedure avoids any possibility to link contents to patients' IDs.
• EORTC: in case of the dynamic EUonQoL-Kit administration, EORTC CAT will be appointed as data processor under article 28 GDPR and will get access to each pseudonymized answer as provided by the CRF.net platform. The EORTC CAT will delete any content once the questionnaire has been submitted including scores sent to the CRF.net platform.

The CRF.net platform is accessible online by investigators and data managers through username and password and is protected through encrypted data certified by SSL certificates and HTTPS protocols. For each registered patient, the CRF.net platform generates a Quick Response (QR) code that will be scanned with the tablet to allow the patient to complete the questionnaires.

The interface of the platforms allows data review and cleaning in the central archive directly (with a limited and controlled access of data-managers and investigators). Each operation done in the archive database is registered through track-change.

Each participating centre will receive tablet devices from the sponsor to get access to the CRF.net platform and collect data directly from patients. Results will be shared as open data in a completely anonymous and aggregated form. After 5 years since data collection, all personal information related to the study, including informed consents, will be anonymized/destroyed according to the delay stated by the Ethics Code on data processing for statistics or scientific research purposes issued by the Italian Data Protection Authority under article 20, par. 4, LD 101/2018.

Patients could exercise their rights by submitting requests to the staff identified in the private information. All centres have appointed a data protection officer under the General Data Protection Regulation (GDPR) and an internal data breach policy.

This study was designed and shall be implemented and reported in accordance with national and European legal and ethical requirements. Moreover, the survey will follow the ethical principles laid down in the Declaration of Helsinki and the ethical principles of observational research on potentially fragile patients.

The protocol, Informative Sheet, and Informed Consent Form must be reviewed and approved by ethics committees of each centre involved. No study procedure can be performed before the written informed consent has been provided.