Medical oxygen and respiratory support requirements for patients hospitalised with COVID-19 in 23 low-income and middle-income countries: a prospective, observational cohort study

These datasets are derived from the WHO O2CoV2 REDCap project exports (screening, enrolment/follow-up, and facility forms), harmonised across languages, relabelled using the REDCap data dictionary, and further cleaned using documented manual correction files (e.g., day numbering and oxygen flow corrections). The objects below correspond to analysis-ready tables. The deposit is organised by dataset: for each table we provide (i) a machine-readable CSV file containing the analysis-ready data and (ii) a companion HTML file that documents variable distributions, missingness, and coding, using the original REDCap labels where available. data_fup_d1 — Daily follow-up, Days 1–7 (D1 form) Patient-day level dataset containing daily clinical status and respiratory support information for the first week after enrolment. Records were cleaned to remove duplicates and non-standard extra follow-up days, and day numbering was harmonised using manual correction files. Derived variables include harmonised oxygen modality (daily_o2_type), grouped modality (daily_o2_type_2), estimated FiO2 when missing, and computed daily oxygen consumption (daily_o2, daily_o2_lmin), plus saturation-based indicators (e.g., daily_o2sat_90). data_fup_d2 — Daily follow-up, Days 8–30 (D2 form) Patient-day level dataset covering days 8–30 of follow-up. Respiratory support modality is reconstructed from the advanced support checklist fields (adv_supptype___*) into harmonised categorical variables (daily_o2_type, daily_o2_type_2). A consistent follow-up day rank (daynum_2) is generated to align with the day numbering used in D1. data_patient — Patient-level analytical dataset One-row-per-patient dataset combining enrolment forms (baseline demographics, comorbidities, vaccination status, etc.) with Day 1 follow-up information (data_day1), follow-up length, and derived summaries (e.g., highest oxygen support during hospitalisation). Additional recodes/derived variables include comorbidity counts and categories (n_como, n_como_2), age and BMI categories, harmonised sex/pregnancy fields, and region/facility identifiers. data_facility — Facility-level analytical dataset One-row-per-facility dataset describing facility characteristics, oxygen supply systems, electricity sources, bed capacity by care level, staffing, and availability of respiratory support devices. The dataset includes documented manual corrections for beds/devices/staff and derives standardised indicators such as ICU bed counts and percentages, and device/staff rates per 100 beds (e.g., fac_hfnc_100_bed, fac_cpap_100_bed, fac_bme_100_bed).