Evaluation of a Multi-country Medical Oxygen Program

Medical oxygen is an essential medicine. Hypoxaemia is deadly, and increases the risk of death by 5-8 fold, requiring prompt recognition and oxygen therapy. Oxygen services are currently inequitable within and between countries, and this has been exacerbated by the COVID19 pandemic. Children are especially vulnerable: in many contexts, fewer than 20% of children admitted to district hospitals with severe hypoxaemia receive oxygen. There are many barriers, at all levels, to ensuring that people who need oxygen will receive it- from delayed care-seeking and referral barriers to facility under-preparedness and over-burdened healthcare workers, to deficiencies in maintenance services, and community misconceptions and fears. Reliable access in rural and remote facilities poses even more challenges.

The Clinton Health Access Initiatives (CHAI), on the back of pre-existing close collaborations with Ministries of Health (MoH) has supported countries in tackling the oxygen access crises of the pandemic. Emerging from this, CHAI and MoH in 9 countries have amalgamated these efforts into a program targeted at improved access to oxygen in each country (the 'MOXY' program). The specific approaches are different in each country but broadly include efforts to: (1) strengthening policies, strategies, and governance of medical oxygen production, distribution, maintenance, and use; (2) building capacity of healthcare workers and technicians to use and maintain oxygen well, and (3) strengthening oxygen-related data acquisition and use for forecasting, budgeting, and monitoring.

MOXY provides the first opportunity to learn from interventions specifically directed at addressing the oxygen problem at large scale, and across different settings (between and within countries). REAL-MOXY is a series of embedded mixed methods studies that aim to better understand the contexts into which oxygen interventions are being introduced; identify and interrogate mechanisms of how these systems work (or not work) to improve health outcomes, and their interaction with different contexts; and synthesise these findings to test and develop theories that can guide policy makers and clinicians in delivering more effective approaches to improve oxygen access.

We will adopt a mixed methods design, with an iterative approach, and co-design to adapt the study methodology to the specific context of each country and facility. We have planned for 5 embedded sub-studies:

The findings of sub-study 1 identifies the facilities that will contribute data (i.e., sequential); and then data collection and integration is concurrent in sub-studies 2-5.

Sub-study 1 aims to identify facilities with high and low functioning oxygen systems, based on current pulse oximetry and oxygen-related clinical practices and facility oxygen readiness. We will use the results of a cross-sectional study already being conducted involving all health facilities in the MOXY catchment areas (part of the MOXY baseline assessment for which ethical approval is already in place). Results will inform facility selection for the subsequent mixed-methods studies.

Sub-study 2 aims to map care pathways (as they are intended) for children (<15 years) with 4 hypothetical clinical scenarios in each participating facility. Data sources include direct observation of patient and equipment flow, and discussions with senior clinicians and managers. Maps to study questions i, ii, and iv.

Sub-study 3 aims to follow patient journeys from arrival through the first 4 hours of care, to understand the sequence of care for acutely unwell children, including how pulse oximetry and oxygen are integrated with other aspects of emergency care. Data sources include direct observation of patients and health workers, patient/caregiver interviews and medical documentation. This sub-study is based in the initial assessment areas of facilities (e.g., emergency or outpatient units). Maps to study questions i, ii, and iv.

Sub-study 4 aims to understand how pulse oximetry and oxygen are used by nurses and medical officers, why, and how this impacts on patient care. Data sources include direct observation of nursing practice, ward rounds, and medical documentation. This sub-study is based in an inpatient unit caring for children. Maps to study questions i, ii and iii.

Sub-study 5 aims to understand the perspectives of a) patients/caregivers, b) healthcare workers, managers and biomedical engineers/technicians. Data sources include focus group discussions and in-depth interviews. Maps to study questions i, ii, iii and iv.

Sub-study 1:

Research Question: What are the baseline pulse oximetry and oxygen practices for children admitted to participating health facilities, and what is the level of institutional readiness for oxygen service delivery? Which facilities, representative of high- and low-performing facilities and different levels of care and facility types, can be selected for additional investigation to understand current functioning of oxygen systems?

Setting and Population: The broader MOXY program baseline cross-sectional assessments involve 9 countries. We have selected 6 MOXY countries for the mixed methods studies outlined in this protocol - Nigeria, Uganda, Liberia, Rwanda, Cambodia and Lao PDR - based on pre-existing research collaborations, and with the aim of representing broadly different geographical contexts.

Baseline assessments are being conducted in all facilities participating in the MOXY program in each of the 6 study countries. From this data set, we will analyse primary and secondary outcomes for wards caring for children <15 years (including neonatal wards where relevant).

Analysis/outcomes: Primary - proportion of admitted children (<15 years, including neonates) screened with pulse oximetry. Secondary - proportion of admitted children with hypoxaemia (<15 years, including neonates) treated with oxygen. For Real-Moxy the secondary outcome will inform identification of facilities for inclusion.

Sub-study 2:

Research question: Where and how are patients managed from arrival to admission and discharge (or transfer), and how does oxygen equipment move within and between clinical areas? How do process maps vary for different clinical scenarios representing different patient groups: i) pneumonia with and without hypoxaemia; ii) severe, acute illness syndrome with WHO emergency signs (e.g., shock, multi-trauma, seizures); iii) surgical condition and iv) neonatal illness (inborn and outborn)?

Setting and Population: This will be conducted in 10 health facilities in each of the 6 countries (Nigeria, Uganda, Liberia, Rwanda, Lao PDR, Cambodia), selected to represent high and low functioning facility oxygen systems and include secondary and tertiary health facilities from government and non-governmental sectors (identified in sub-study 1). We will focus on admitted children (<15 years, including neonates) with (i) pneumonia, (ii) other acute illness with WHO emergency signs, (iii) surgical conditions, and (iv) neonatal illness. We have chosen these conditions to represent diagnoses with a high prevalence of hypoxaemia, and to capture the nuances of how pulse oximetry and oxygen practices are adapted (or not adapted) to clinical scenarios (e.g., having a lower threshold to provide oxygen to a patient in shock; or targeting safe oxygen saturations in neonates).

Analysis/outcomes: Facility maps of patient and equipment flow.

Sub-study 3:

Research question: What is the sequence of emergency care for an unwell child in the first 4 hours, and how are decisions made - particularly relating to oxygen (when to start, stop, how much, what delivery modality, etc.)? What are the points of delays to appropriate care (including pulse oximetry and oxygen) and at what points in time and location could pulse oximetry and oxygen be used better for emergency care of children?

Setting and Population: facilities are same as sub-study 2). Children (<15 years, including neonates) presenting with each of 4 acute illness syndromes: (i) pneumonia, (ii) other acute illness with WHO emergency signs, (iii) surgical conditions, and (iv) neonatal illness.

Analysis/outcomes: Patient journey maps.

Sub-study 4:

Research question: How do healthcare workers use pulse oximetry and oxygen for admitted patients, and how does this change over time and vary between patient groups, facility type, and location? How do practices compare with treatment guidelines and where are the priority areas for improving oxygen care?

Population and setting: Facilities are same as sub-studies 2&3.

Analysis/outcomes: Narrative descriptions of handover, ward rounds, and nursing rounds, with specific emphasis on how pulse oximetry is used, and decision making for oxygen therapy.

Sub-study 5: Indepth interviews and focus group discussions Research questions: 5a) How do patients/caregivers perceive pulse oximetry and oxygen therapy within their broader care experience? 5b) How do healthcare workers, managers and technicians perceive oxygen therapy and the provision of oxygen-related care within the broader care provision experience?

Population:

• Patients and caregivers enrolled in sub-study 3 (patient journey mapping).
• Healthcare workers (bedside), managers (including clinical and non-clinical managers) and technicians in each health facility. We will select staff with direct responsibility for wards caring for children <15 years.

Analysis/outcomes: Reflexive thematic analysis of interviews and focus group discussions.