The goal of this pilot clinical trial is to compare standard of care, low-flow oxygen, and high-flow nasal canula oxygen in pediatric patients aged 1-59 months with pneumonia and an oxygen saturation of 90-93% in Malawi. The main question it aims to answer is:

• Does the protocol for the randomized control trial work well?
• Can the researchers safely conduct the protocol for the trial?

Participants will be randomly assigned to one of the three groups (normal care without oxygen, low-flow oxygen, and high-flow nasal cannula oxygen) and treated with that therapy in the hospital. Researchers will look at the ability to safely conduct each part of the study.

Pneumonia is the leading infectious cause of under 5-year-old deaths globally and responsible for >50% of deaths in Africa. The World Health Organization (WHO) defines low blood oxygen saturation (SpO2) levels (hypoxemia) as 90%. Hypoxemia is identified in 31% of child pneumonia cases in Africa and is a key marker of elevated mortality risk. When children are hypoxemic, the WHO recommends oxygen treatment. Importantly, the WHO threshold of 90% for hypoxemia was based on concerns over limited oxygen supply and hospital over-crowding in low- and middle-income countries (LMICs), rather than quality evidence. In most LMICs, low oxygen flow is the mainstay of oxygen delivery. Recently, in high-income settings high-flow nasal cannula (HFNC) oxygen has emerged as a safe and effective alternative. HFNC oxygen delivers higher flow warmed, humidified gas via nasal prongs to reverse hypoxemia, and potentially improve outcomes.

Recent evidence challenges whether the WHO & 90% hypoxemia threshold is optimal for identifying all children at higher risk of mortality in LMICs. One meta-analysis from 13 LMICs reported 3.66-fold-higher odds of death (95% confidence interval (CI), 1.42, 9.47) for children with a SpO2 93%. The investigators research from Malawi and Bangladesh established children with pneumonia and SpO2 between 90-93% (moderate hypoxemia) is common, and, compared to higher SpO2 levels, conveys higher mortality risk. To date, African children with a SpO2 90-93% are not recommended for oxygen treatment. Observational data from Malawi found children with moderate hypoxemia and treated with oxygen had higher survival than those referred with a SpO2 90%. Currently, no randomized trials have determined whether low flow oxygen or HFNC oxygen treatment reduces the mortality of children with moderate hypoxemia (SpO2 90-93%) in African LMICs.

Aim 1: Conduct a pilot open label, three armed, parallel, randomized controlled trial (RCT) comparing standard care, low-flow oxygen, and HFNC oxygen for children with clinical pneumonia and a SpO2 90-93% to determine feasibility of a larger trial. The investigators hypothesize it will be feasible to recruit, randomize, treat, and safely follow-up all participants. Children with SpO2 90-93% will be randomized 1:1:1 to standard care without oxygen (controls), low flow oxygen (intervention #1), or HFNC oxygen (intervention #2). The primary outcome will be feasibility, defined as the proportion of enrolled children with 2 protocol violations. Secondary outcomes include consent refusal, intervention efficacy, participant attrition, and safety.

Aim 2: Determine the prevalence of young Malawian children with a SpO2 90-93% at the designated study hospital. The investigators hypothesize a SpO2 90-93% will be common among children presenting to the trial hospital. The investigators will measure the SpO2 of all children under-five years old (not limited to pneumonia cases) presenting to the hospital 1 week per month over 12-months. Conservatively assuming an average volume of 30 children per day, based on prior data, the investigators will generate 1,400 SpO2 measurements.

Inclusion Criteria:

• 1-59 months of age
• Pneumonia (as defined by the World Health Organization)
• Oxygen saturation 90-93% without oxygen

Exclusion Criteria:

• Emergency signs (signs of severe illness as defined by the World Health Organization) including:
• absent or obstructed breathing,
• severe respiratory distress,
• shock,
• decreased mental status,
• convulsions, or
• severe dehydration

The research proposed is a pilot randomized, open-label trial of acute lower respiratory infection with moderate hypoxemia oxygen use in Malawi. The pilot trial will enroll 21 participants with a primary outcome of study feasibility. The datasets will report de-identified patient outcomes including mortality, hospital length of stay, and treatment failure as well as trial feasibility and acceptability data. The shared format of data will be .csv files.

Data will be provided at the time of the primary outcome publication(s) for a period of at least 10 years.

Publicly shared data will be deposited in an open access repository that meets the specifications required by the NIH. The selected repository will provide metadata, unique identifiers, and data access for at least 10 years. Use of keywords will allow for findability using searches. Access to data will be controlled and limited to qualified investigators who sign a data use agreement and provide a reasonable research question for application to the data. Data requests will require at least a description of the research, objective, design, analysis plan that includes data safeguards, publicly available research use statement, documentation of institutional Review Board (IRB) review and approval. The open access repository will review the request and has final authority for deciding on access to the data. If the study is active, then