Table 2.

Outcomes, research questions, sources of data and specific metrics

	Outcome category	Research questions	Sources of data	Metrics
1	Deaths from pneumonia; overall pediatric deaths in health facility	Is there a difference in pneumonia case fatality rates and overall pediatric CFR from pre to post implementation of improved oxygen therapy and solar power?	Admission record books	Pneumonia case fatality rate: Pneumonia deaths / pneumonia admissions (%). Paediatric case fatality rate: Paediatric deaths / all pediatric admissions (%).
2	Referral / transfer	Is there any difference in referral rates from pre to post intervention?	Admission record books	Paediatric transfers / all pediatric admissions (%).
3	Patient characteristics and response to oxygen therapy – effectiveness of oxygen therapy using the method we have designed to treat hypoxaemia	What are the conditions associated with hypoxaemia in remote rural health facilities? What is the response to oxygen therapy when oxygen is given using solar–powered oxygen concentrators? What is the duration of hypoxaemia in children managed in remote rural health facilities?	Standardised admission record data	Diagnoses associated with hypoxaemia (proportions) disaggregated for neonates and children >1 mo. Response to oxygen therapy (median change in SpO2 in the first 30 min; and proportion responded / not responded, ie, proportion with persisting SpO2<90%, or severe signs of respiratory distress 30 min after commencing oxygen). Days of oxygen therapy. Duration of hypoxaemia for neonates and children >1 month.
4	Maintenance of oxygen equipment	Are concentrators maintained well, are problems identified and is appropriate action taken? What proportion of concentrators undergo weekly maintenance and performance checking? What problems are identified? What proportion of concentrators are functioning well after 1, 2 and 3 years since installation?	Oxygen concentrator performance log–books. On–site checks on support and monitoring visits	What proportion of oxygen concentrators undergo weekly maintenance and performance checking? List of problems identified and action taken. Number / proportion of concentrators providing >85% oxygen and reliable flow rates as checked by oxygen analyzer at 1, 2 and 3 years.
5	Health workers knowledge and skill of oxygen therapy	What is the oxygen knowledge and skill of the health workers in remote health facilities? Does this improve with training and CQI	Oxygen competency tests – repeated measures	Repeating oxygen knowledge and skill tests at 12 monthly intervals.
6	Reliability, efficiency and adequacy of solar power	Solar power output (quantitative kW hours per day), adequacy of this power for running concentrators and other equipment needed by the health facility, and any problems identified.	Tristar 60 Amp controller	Solar power output: average kW hours per day produced by each solar system in the previous 6 mo. Occurrence of breakdown of solar output, such that concentrators are unable to function.
7	Training outcomes	What training was done and what are the perceived training needs?	Project records	Record of training courses conducted: formal; in–house as part of CQI
8	Sustainable processes	Is it sustainable in PNG to have concentrators run off solar power as the source of oxygen in remote areas?	Mixture of the above data sources, and documentation of relevant events, meetings and occurrences	How would we measure sustainability? Equipment sustainability (ie, proportion of concentrators and oximeters still functioning well at 1, 2, 3 years). Energy sustainable (proportion of solar power systems still functioning well at 1, 2, 3 years). Health worker sustainability (proportion with good oxygen knowledge and skills at 1, 2, 3 years). Financial sustainability (examples where national, provincial or local governments have invested in other facilities in a similar way). Policy sustainability (occurrence of a national oxygen policy that incorporates these principles). Documentation of other examples of engagement by policy makers.
9	Wider benefits	Does CQI in rural health facilities improve wider outcomes? Care seeking by parents? Health worker morale?	Above data and qualitative assessments	Care seeking: Number of children brought to health facilities (admission activity number per year for 3 years, compared with pre–intervention baseline data). Women having confidence in the health services: Number of babies born at health facilities (births per year for 3 years, compared with pre–intervention baseline data). Qualitative assessment of what health workers think of the oxygen project, whether and how it helps them in their work, what problems it may have resulted in, what could be done to help them better serve their communities, how they feel about their work, what their hopes are for their careers.

CQI – continuous quality improvement, CFR – case fatality rate, SpO₂ – arterial oxygen saturation