METHODS

Study Design and Participants
This cross-sectional study was conducted by a consortium of researchers from different countries aiming to measure the patterns of cancer care during the COVID-19 pandemic and quantify the impact of the pandemic on various components of cancer care delivery. The survey included questions to assess the performance of oncology centers in different countries in response to the COVID-19 pandemic. The tool evaluated cancer services management during the crisis and the perception of oncologists about the potential harm to patients.
The survey was disseminated electronically using the SurveyMonkey platform (SurveyMonkey, San Mateo, CA) and targeted a convenience sample of individuals identified as key informants in their institutions who were aware of the management and updates about oncology practices and services during the pandemic. SurveyMonkey was chosen by the research team for its advanced design capabilities.
The key informants were senior oncologists who were active in the global oncology community and had a good network of contacts. We targeted oncologists who were involved in clinical care and aware of their center information. They were selected from different geographical regions and continents and acted as regional coordinators. The regional coordinators invited country coordinators to disseminate the survey to the targeted participants in their respective areas. Patients were not involved in this study.

Procedures and Study Instrument
The data collection instrument (online survey) consisted of 51 English-language questions developed by the research team, which was composed of oncologists and research methodologists who have monitored cancer care and tracked key cancer care indicators across countries during the COVID-19 pandemic. Content validity of the instrument was assessed by presenting it to a group of eight experts (six oncologists, an epidemiologist, and a nurse manager). The group evaluated the relevance and appropriateness of each item of the scale, and a content validity ratio (0.79) and content validity index (0.87) were generated. All changes were corrected as suggested by the panel of experts. The obtained final version of the tool was further piloted on a sample of physicians who worked with patients with cancer in different institutions (n = 20) to ensure that all the items used were clear and understandable.
The online survey included characteristics of the surveyed centers (location of the center, type of services offered, and number of new patients served annually), interruption of cancer care (reasons for interrupting usual care and reducing or discontinuing services and access to products and medications), potential harm to patients as a result of interrupting services (missing treatment, cancellation of clinic visits, seeking care elsewhere, and type of harm that may have affected the patients), diagnosis and management of COVID-19 (diagnosed cases in the city, diagnosed patients without cancer, diagnosed patients with cancer, diagnosed cancer center staff, availability of PPE, and availability of practice guidelines for COVID-19 diagnosis and management), and virtual management of patients with cancer and remote care (availability of tumor boards, availability and management of virtual tumor boards, availability and management of virtual clinics, possibility of persistence of these services after the pandemic subsides, and dissemination of medication to patients with cancer).
The questions for this study were adapted from a literature review of the topic being measured, and respondents familiar with the concept of cancer care management were interviewed to help to generate instrument items. Content validity of the constructed instrument was ensured by the judgment of experts on the relevancy and clarity of the items. SurveyMonkey was used to disseminate the survey and compile the responses.

Data Analysis
Responses were collected anonymously and recorded. Upon completion of the data collection process, data were imported from SurveyMonkey into Stata 14.0 software (StataCorp, College Station, TX) for statistical analysis. Descriptive analyses to examine the characteristics of the sample were performed. Means, standard deviations, and frequencies with their corresponding 95% CIs were reported for every surveyed cancer institution.
Ethical approval was obtained from the institutional review board at King Abdullah International Medical Research Center at the King Saud bin Abdulaziz University for Health Sciences. Comparative analysis was conducted to compare the severity of the pandemic impact on the centers based on the World Bank income classification of the responding countries, such as interruption of care, access to PPE and medications, use of virtual technology to run clinic visits or tumor boards, and delivery of medications to patients’ homes.