Oxygen CoLab: Building the case for oxygen concentrators and the service-based models under which they operate

Our evidence framework, September 2023

​

Hello 👋

2

In this deck, we share the emerging evidence framework for the Oxygen CoLab.

​

We want to understand and make the case for oxygen concentrator and service-based models as a delivery solution alongside other ways of providing medical oxygen to patients in LMICs. We believe that if we update and innovate concentrators to work better in challenging environments, and find the right services to support facilities to use them at the right price, they can fill a much needed gap in medical oxygen access at lower-levels of care.

​

We are innovating and ‘learning by doing’: testing updated products and innovative services in 62 facilities in Kenya, Uganda, Nigeria, Tanzania and India. Our evidence comes from real world implementation, rather than dedicated research.

​

We also want to partner with others to use new evidence to strengthen the health system: removing or reducing barriers which make it harder for facilities to get and use oxygen equipment to save lives. Getting people to use our evidence is at the heart of our strategy.

​

We have valuable learning about oxygen delivery using concentrators from our past work. This is the first time that we have explicitly stated what we want to learn going forward, and invested in data collection and analysis. This deck show the evidence we think is needed to achieve our goals.

The problem we are solving

> 1 million deaths occur every year due to hypoxemia in low- and middle-income countries

Almost all of these deaths are preventable.

5

During COVID, there was significant investment in increasing access to oxygen within higher-level facilities that could deliver critical care - but this progress left behind lower-level facilities that vulnerable patients rely on as the point of first referral for care.

​

Oxygen concentrators are proven to save lives in low-resource settings when implemented properly, yet their potential has yet to be realised at scale.

• Current concentrator models, designed for the Global North, are poorly adapted to resource-limited contexts that may have unreliable power supplies, harsh environments, and shortages of specialised maintenance staff. All of this makes scaling oxygen concentrators challenging.
• Additionally, donations and conventional equipment purchasing do not include maintenance services, creating a lack of ongoing vendor accountability. This means suppliers lack incentives to provide ongoing training and maintenance–or to offer a mix of technologies tailored to the needs of each health facilities unique setting.

Lower-level facilities are the point of first referral for care, but they have been left behind in terms of investment. This is where the concentrator plays a role.

🔮 The better future we want to see in 2030

6

In 2030, oxygen concentrators and service-based-models are a sustainable option for providing accessible and affordable oxygen at the right place, at the right time, for the right people across LMICs.

​

🚀 To get to this better future, we have five big goals by 2025

1. Selected health care providers are able to use oxygen concentrators consistently at the right time, for the right patients in LMICs
2. Fit-for-purpose oxygen concentrators exist and are ready for market, with a strong pipeline of next generation devices
3. Selected national governments and donors invest in and increase procurement of fit-for-purpose oxygen concentrators under services-based models (e.g. OaaS) and integrate concentrators and wrap-around services in health policies and plans
4. Inform emerging global consensus on the optimal niches and use cases for fit-for-purpose concentrators within overall low-resource / LMIC oxygen ecosystems
5. Support new capital investments to scale oxygen delivery via fit-for-purpose concentrators under OaaS models in low-resource settings / LMICs.

The work that we do, and the evidence we gather and share with the world, will help us achieve these goals in partnership with governments, NGOs, donors, and companies.

We are bringing people together to deliver on the promise of oxygen concentrators to expand access to these low-resource settings. A promise that has gone unfulfilled for over 15 years.

​

​

To realise the full potential of oxygen concentrators, we are investing in innovating the products for low-resource settings, developing service models which keep concentrators working for health facilities, and helping governments and companies adopt these models to improve access to oxygen for vulnerable patients.

​

Product Innovation

We are investing in key innovators and manufacturers to research, develop and lab test fit-for-purpose oxygen concentrators in accordance with UNICEF’s industry-validated Target Product Profile (TPP).

​

Oxygen-as-a-Service

We are testing Oxygen-as-a-Service models which allow health facilities to pay vendors for oxygen availability, rather than purchasing equipment outright, through a portfolio of on-the-ground pilots.

Evidence into action

We are generating evidence about the facilities concentrators are best able to serve while ensuring evidence changes policy and increases investments in oxygen concentrators delivered through services-based models (nationally and globally).

To get there we need to learn more

8

Tomorrow, we want to know:

• What are the facility archetypes where concentrators have a role to play?
• How the concentrator fits in alongside other delivery solutions: what works, where, how and why
• Whether updated, fit for purpose concentrators add value above current solutions, and if there is potential demand for them
• Whether service based models (Oxygen as a Service, or O2aaS) add value, and what works, where, how, and why
• What the potential demand for service based models is, and whether current models are scalable.

(We have broken this down into more detailed research questions, which you can see later on).

Today, we know that:

• Concentrators are 25-75% cheaper than cylinders
• In primary health centres, concentrators can provide independence from unreliable supply chains, and flexibility due to their size and easy deployment
• Concentrators play an effective role in other parts of the health system as a backup or part of a mixed source systems
• Concentrators need reliable power and ongoing maintenance, which increases costs and hinder use
• Concentrators are inefficient for treating large volumes of patients
• Patients benefit from using concentrators for low flow applications (e.g. chronic conditions)
• In rural settings, concentrators are also used for high flow applications (e.g. acute conditions like pneumonia).

The changes we would like to see by 2025

9

Our evidence will help us understand and make the case for concentrators and service based models in national and global spaces. We want our evidence to lead to action on ⚖ Policy, 📊Financing and 📐Product development.

We want to see:

What we want to learn about oxygen concentrators

Our hypothesis is that…

​

If we know the right mix of:

​

​

​

​

​

​

​

​

across different contextual constraints

​

​

​

​

​

​

​

AND we accelerate innovation to the concentrator products themselves

​

AND we collectively use this evidence to advocate for changes to policy and investment in oxygen concentrators

​

THEN reliable oxygen will be accessible and affordable to patients at the place and time of need

📍WHERE

Community Isolation

Clinical Services

Size of the facility

Existing O2 Supply

Electricity Access

We accelerate innovation to the concentrator products themselves

We collectively use this evidence to advocate for changes to policy and investment in oxygen concentrators

Reliable oxygen will be accessible and affordable to patients at lower levels of care

If we know the right mix of…

across different contextual constraints

Financial Autonomy

Environment

Staffing

We identified eight categories of contextual constraints that we’re learning about

Size

• Small: <50 beds
• Medium: 50 – 99 beds
• Large: 100 beds or more

Staffing

• Limited: Mostly junior and/or unspecialized clinical staff; few doctors if any; few biomed techs if any
• Moderate: Some senior / specialized clinical staff, at least one doctor on each ward, some biomed techs – may have one or two biomed engineers
• Extensive: Many senior / specialized clinical staff, several doctors per ward, several biomed engineers & techs

Clinical Services

• No inpatient services: at most a handful of short-term “stabilization” beds for severe patients awaiting referral
• Basic inpatient services: some or all of: BEmONC, Respiratory, Simple Surgeries, limited A&E
• Comprehensive inpatient services: all basic plus some or all of: Complex surgeries, ICU, NICU, full A&E

Community Isolation

• Extreme: Other health facilities are far away via poor roads / transport infrastructure
• Partial: Geographically remote but other facilities accessible via acceptable transport infrastructure, albeit with potentially long transit times
• Connected: Other health facilities easily accessible within reasonably short transit times

Oxygen Market Access

• Poor: sellers of medical oxygen (in any format) are too far for the facility to do business with them and/or require minimum volumes or payment terms that the facility cannot achieve
• Moderate: The facility can reach and do business with sellers of medical oxygen (in at least one format) but must manage logistics and/or routinely experiences delays in deliveries; the facility may incur price or other penalties in exchange for flexibility on minimum volumes and/or payment terms
• Excellent: The facility is close to sellers of medical oxygen and able to do business with them cost-effectively because of their ability to meet minimum volumes and conform to required payment terms and contractual structures

Electricity

• Unreliable: The facility experiences frequent (>2 daily) and/or long (>2 hours) outages in grid power in addition to frequent, severe voltage excursions
• Intermittent: The facility routinely (daily or less) experiences outages in grid power which are predictable and/or short in duration. Voltage excursions are infrequent and moderate in severity or better.
• Stable: The facility rarely experiences short outages in grid power; voltage excursions are infrequent and mild.

Financial Autonomy

• Low: Facility depends almost entirely on budget subventions / disbursements from external bodies; funds availability and procurement decisions are frequently made by external bodies with limited input from facility management
• Medium: Facility depends in part on budget subventions / disbursements from external bodies; disbursements and procurement decisions may require formal approvals from external bodies
• High: Facility generates and controls the majority of its revenue; annual and/or multiyear budgets and procurement plans may require approval from external bodies but individual disbursements and procurement decisions do not

Environment

• Harsh: Temperature generally exceeds 24 celsius and frequently exceeds 30 celsius for prolonged periods; relative humidity generally exceeds 50% and/or frequently exceeds 80% for prolonged periods; 24-hour mean PM2.5 and PM10 concentrations frequently exceed 35ug/m3 and 150ug/m3, respectively.
• Moderate: Temperature is generally maintained around 20 – 24 celsius but may exceed 30 celsius for prolonged periods; relative humidity may exceed 80% for prolonged periods; Filtration systems maintain 24-hour mean PM2.5 and PM10 concentrations of less than 35ug/m3 and 150ug/m3, respectively.
• Mild: Reliable climate control maintains temperature around 20 – 24 celsius and relative humidity around 30% to 50%; Filtration systems maintain annual mean PM2.5 and PM10 concentrations of less than 10ug/m3 and 20ug/m3, respectively.

Each of our grantees is working under a different set of conditions: we want to understand the suitability of their products and services, including their business models, against these constraints.

We believe concentrators delivered through Oxygen-as-a-Service are appropriate in several facility archetypes* facing specific contextual constraints

Small healthcare facilities serving isolated communities with basic clinical services.

Mid-sized facilities that serve as the first port of call for referrals—responsible for stabilizing and treating critical patients.

Secondary care facilities providing comprehensive and specialty services but struggling with limited staff & fragmented infrastructure.

Small but relatively well-equipped facilities providing comprehensive outpatient and basic inpatient services.

*Even within the four categories there will be a significant degree of diversity. The four categories are broader than where our portfolio currently stands.

Our portfolio of O2aaS: testing service models across the different facility archetypes

Small healthcare facilities serving isolated communities with basic clinical services.

Mid-sized facilities that serve as the first port of call for referrals—responsible for stabilizing and treating critical patients.

Secondary care facilities providing comprehensive and specialty services but struggling with limited staff and fragmented infrastructure.

Small but relatively well-equipped facilities providing comprehensive outpatient and basic inpatient services.

How our evidence can strengthen systems

✅ Aspirations: what we aim to do

​

• We’re concentrating on understanding the specific value add of concentrator products and services. We focus more on how they work in LMICs, than on oxygen’s impact overall.
• We want to learn by doing: gathering practical, actionable insights from real world implementation. We put much of our resources into grants for piloting, and our grantees will collect a lot of data. We don’t just want evidence: we want their models to succeed, and we spend resources supporting that.
• We will use the evidence to tell compelling stories that influence stakeholder uptake of concentrator products and services. We built the evidence framework around our perception of their evidence needs.
• We need some data that tells us which aspects of products or services work and in what conditions. We have quite granular indicators that enable cross-grantee comparison over time.
• We need some data that tells us how concentrators work alongside PSA, LOX, and cylinders. We will have ‘analysis’ indicators, measured once, that will help us understand our data better in context.

​

• We learn primarily from real life innovation and our resource investment reflects that- we invest more in supporting our grantees to learn than on independent research studies.
• Whilst we will have comparators that help us learn about concentrators in different conditions, we aren’t doing a randomised control trial or large scale multi-variate testing (yet).
• We value ‘good enough’ evidence for decision making, over data consistency. Grantees will need to measure some indicators slightly differently or at different times, bearing in mind their internal capacity, resources, and the context they’re working in.
• At the moment, integrating our data and making it interoperable within national health information management systems is beyond our current funding scope.

🚫 Limitations: what we won’t or can’t do

Ensuring the

uptake of evidence

18

Nationally

​

At a national level, we will support grantees with their advocacy efforts across three levels (more detail here)

​

• Country testbeds: we will bring stakeholders governments, grantees, UNICEF and other partners together to review, analyse and act on the evidence about oxygen concentrators and O2aaS
• Advocate for decentralising oxygen access as part of WHO Oxygen Roadmaps: in partnership with UNICEF and others, we will aim to input our evidence on concentrators and O2aaS into oxygen roadmaps
• Grantee support: we will support grantees with the skills and tools they need to participate in national dialogues.

Globally

We identified the following influencing levers in the global oxygen space to help us make the case for the appropriate use of oxygen concentrators.

​

We will aim to:

• Ensure the concentrator has a voice within the new Global Oxygen Alliance (GO2AL)
• Share our evidence on concentrators and service-based models with the The Lancet Commission recommendations for medical oxygen security (due for publication in 2024)
• Raise awareness of concentrators and service based models as a potential solution for countries working on their national oxygen roadmaps

What our evidence will tell you if…

💵 You are planning to purchase oxygen

• You might be a government procurement body, a health policy maker, a facility administrator, deciding what to purchase.
• Our aim is to help you decide whether to buy concentrators, PSA plants and if concentrators, what services you might need to purchase alongside it.

​

📈 You are thinking of investing in, or funding, O2aaS

• You might be an investor in the health technology space, or an international funder, whether bilateral, multilateral or philanthropic foundation. You may be seeking to create impact through your funding.
• Our aim is to help you understand the different ways to invest in oxygen as a service and which one might be best for you.

​

✍️ You are making or implementing Oxygen policy

• You might be an investor in the health technology space, an international funder, whether bilateral, multilateral or philanthropic foundation. You seeking to create impact through your funding.
• Our aim to help you work out where you can most effectively implement oxygen policy and what it should include.

​

​

💵 How we can help you with purchasing oxygen

📈 How we can help you invest in, or fund, O2aaS

​

✍️ How we can help you make or implement Oxygen policy

​

What data will we use to learn about oxygen concentrators and service based models?

We need to collect the right data to answer our questions, without duplicating efforts or overstretching our resources

24

Our evidence framework collects data about:

• patient access to oxygen
• the added value and availability of products, and services
• financing and demand for products and services
• the policy enabling environment

We’ll need help from others to complete the picture.

Patient access to oxygen through concentrators

Do these products and services contribute to improved patient access to oxygen?

• Proportion of likely hypoxemic patients receiving oxygen.
• Changes in referral patterns for children with severe pneumonia because of lack of Oxygen as a result of increased oxygen access.

Rationale: We specifically want to concentrate on the place and added value of concentrators.

We know that PATH, CHAI, the Lancet Commission and others who are well established in the space are already making the case for oxygen as an essential medicine.

So, we will invest more in collecting data about product and services, and less in understanding the clinical impact of oxygen delivery.

However, we still need some data points which tell us whether products and services are changing oxygen access levels.

Added value of fit-for-purpose products

Do the products perform and add value against different contextual constraints?

• Power efficiency of the product.
• Resilience to heat, humidity, and temperature cycling, in accordance with detailed lab testing protocol (this includes performance and performance degradation over time).
• Increased resilience to simulated power outages, in accordance with detailed lab testing protocol.
• Feedback from health worker and facility administrators on product usability, safety, and performance, compared to current products.
• Number of products still in working condition after 6 months of usage in real life conditions.
• Maintenance and repair costs of new products.
• Total comparative cost of ownership of the product (in relation to PSA, LOX, cylinders and other concentrators, in various locations) (Analysis).
• Comparative costs of powering the product (in relation to PSA, LOX, in locations with high electricity costs) (Analysis).

Rationale: These help us understand the cost, energy consumption, usability, repairability and resilience of fit-for-purpose concentrators in different conditions.

Can updated concentrators better cope with dust, humidity and power outages? And are they affordable to health facilities, and can their use be sustained?

We are still building out the detail on these metrics to ensure we express these as bounded, comparable indicators. For example, finding a good way to compare power costs, and maintenance and repair costs. (We have these more precise comparator metrics for services.)

Added value of Oxygen-as-a-Service (O2aaS)

Do Oxygen as a Service (O2aaS) services perform and add value, against different contextual constraints?

• Mean equipment downtime (in hours) per equipment-month
• Number of reported O2 supply failures per equipment-month
• Number of patients that have had their oxygen saturation monitored with pulse oximetry at their first point of contact at facility per 24 hours out of the total number of patients evaluated at first point of contact per facility.
• Feedback from health workers on operational fit within existing workflows and ease of use
• Impact of current and updated clinical and maintenance practices on equipment uptime (selected facilities)
• Avg cost per unit of oxygen provided / used
• Comparative performance of PSA, Cylinder and LOX
• Comparative cost of O2aaS and total cost of ownership of products in grantee target facilities

Rationale: We want to demonstrate the comparative advantage of service based models against established models.

First, we want to understand whether O2aaS maintenance and repair efforts mean machines are more likely to work when they’re needed, and be repaired fast. Equipment downtime, and supply failures are proxies for that.

Second, we want to know if O2aaS training efforts means health workers are better able to use the concentrators, more likely to engage in good practice like measuring saturation at first point of contact. We’ll check that training leads to better practice, and better outcomes.

Third, we want to know if maintenance is better. And fourth: what does all this cost, and how does that compare in the long term to other delivery models? (We believe that service based models account for more real costs than product purchase does.)

Availability of O2aaS now and in the long term

Are services available and ready to be scaled?

• Number of facilities reached by O2aaS grantees
• Percent change in concentrator and services contract sales over time
• Proportion of target procurement bodies issuing tenders for O2aaS contracts
• Repeat customer sales as a proportion of total sales
• Total value of financial instruments committed to O2aaS grantees (as a proportion of grantee targets)
• Value of purchases made across global stakeholders (including in-country actors) for O2aaS contracts and fit-for-purpose concentrators

Are business models for services and new products viable and sustainable?

• Average return on sales (ROS) of O2aaS businesses.
• Grantee customer retention / contract renewal rate.
• Percent change in concentrator, services contract sales over time.
• Repeat Customer Sales as a proportion of total sales.
• Repeat customer sales growth rate.

Rationale: We want to see increased purchasing by health facilities and national procurement bodies.

However, we know that’s risky for them. We hope to test a donor-funded financial instrument that de-risks scaling of service based models for governments. This requires measuring how many facilities/governments purchase services, and whether our instrument makes a difference to that.

In order for governments to purchase O2aaS in particular, they need to see that these services can be provided reliably in the medium term: that the businesses are sustainable. So, we measure things about business finance and operations that purchasers might reasonably expect to see.

We also want governments to see signs that health facilities are satisfied with services, to encourage purchasing: that’s why we measure repeat customer sales over time. It’s part of the sales pitch.

​

Availability of fit for purpose concentrators

Are more and better products available to the market?

• Proportion of targeted manufacturers bringing TPP-compliant concentrator models to market
• Proportion of manufacturers who make updates to product designs in response to lab testing results
• Proportion of target agencies, organisations, and procurers adopting TPP-aligned concentrator specifications
• Estimated value of TPP-compliant (and/or "next generation") concentrator and O2aaS supply model market growth opportunities (absolute and relative to total concentrator market) (Analysis)

Rationale: We want to accelerate the market for fit for purpose concentrators: investing in R&D for the products, and incentivising their production through new specifications, and showing how the market has grown. We want to know: are products available to the market? And will they be in future?

Manufacturers also need to understand whether and how product specifications are changing. They can also tell us whether they think the market is changing, and if that affects their future R&D investments.

​

Funding and purchasing

Are funds available or committed towards concentrator products and services?

• Proportion of target organizations issuing formal /documented positions advocating for funding / purchasing O2aaS
• Proportion of target organizations (including oxygen funders) who make funds available for procurement of fit-for-purpose concentrators / O2aaS

Rationale: These indicators help us understand the enabling environment for oxygen concentrators: what is the donor and government funding environment? Are there more funds available? Are more bodies including concentrators in their plans?

They also help us understand demand- which in turn influences manufacturers and investors to keep going. We want to know who is demanding TPP-alignment; and who is purchasing what.

​

Funding and purchasing

Are governments and health facilities purchasing products and services?

• Proportion of target agencies, organizations, and procurers adopting TPP-aligned concentrator specifications
• Average value of financial instruments secured by services-based supply model companies (e.g. CoLab grantees) (absolute and as a proportion of grantee targets)
• Proportion of target procurement bodies issuing tenders for O2aaS supply contracts

Rationale: These indicators help us understand the enabling environment for oxygen concentrators: what is the donor and government funding environment? Are there more funds available? Are more bodies including concentrators in their plans?

They also help us understand demand- which in turn influences manufacturers and investors to keep going. We want to know who is demanding TPP-alignment; and who is purchasing what.

​

Policy environment

Do guidelines, policy and procurement enable use of concentrator products and services, locally, globally and nationally?

• Number of specific policy reform recommendations about service-based models and fit-for-purpose concentrators developed and vetted with global and public-sector stakeholders
• Proportion of targeted national and subnational policy bodies (both governmental and non-governmental) adopting policy changes in agreement with CoLab recommendations (including via national oxygen strategies/policies)
• Health workers reporting challenges or gaps in current guidelines
• Proportion of relevant clinical guidelines that include up-to-date hypoxemia management content
• Proportion of maintenance guidelines which include adequate content on equipment necessary to enable high-quality hypoxemia management
• Feedback from national governments on service model fit into existing procurement and decision making modalities and workflows
• Value of potential health system cost savings realised via expanded oxygen access (analysis of existing data)

Rationale: We want to know how well we’re doing when it comes to influencing the global dialogue. We also think it’s important to show manufacturers and investors that concentrators are credible and valued, and that there is an enabling regulatory environment for their use. We may find that it’s less relevant to collect data on changes to clinical and maintenance guidelines, because they are not influential instruments in many countries. If that happens, we will drop those indicators.

The types of data we will use to influence the different target actions

Partnerships for evidence: how you can help us

34

Our evidence comes from real world implementation, rather than dedicated research. Because we can only collect data in the facilities we’re working with, there will be gaps in our evidence. To fully understand and make the case for concentrators and service-based models, we need to work with others, to fill those gaps using their data and their research. We also think the case will be stronger, and seen as more credible, if we gather and analyse our collective evidence together.

Thank You