Open source lights the way to better hospital and clinical care in the COVID era

By Mary Anne Patton

Like most people with diabetes, I’ve spent a bit of time over the past few months wondering how things would go if I was admitted to hospital with COVID-19.

Thanks to open source solutions OpenAPS, AndroidAPS and Loop, I’m in a better position now than I was two and a half years ago, with my glucose levels, to withstand the onslaught of the virus.

There are plenty of people in the same boat feeling grateful too, using either DIY or off the shelf closed loop systems. And as usual with diabetes, everyone’s doing the best they can with the resources they have, in the lives they find themselves in, to keep things as healthy as possible.

But what if we end up on a sliding scale IV infusion in hospital? What if our devices are removed from us and we’re in no state to control them or advocate for ourselves? Would nursing staff be able to manage our diabetes? Would they need to avoid contact with us as much as possible to conserve their personal protective equipment?

I feel sure my endocrinologist would support me to keep using my closed loop and monitoring devices if possible.

But would it be possible?

I’ve been reassured by the thought that my partner could monitor me remotely and liaise with my endocrinologist or ward staff if he gets an urgent predictive alert indicating that action needs to be taken to prevent a dangerous blood glucose situation.

But could he?

Galindo et al, in the Journal of Diabetes Science and Techology, state that Nightscout would be helpful for monitoring in-patients in the COVID era and could help inform treatment. “However the logistics surrounding software implementation, data security, and lack of FDA approval create barriers to its use in the hospital setting at this time”.

It got me wondering how open source solutions like Nightscout, created by the diabetes patient community, who’ve always understood the need for real-time remote monitoring of all aspects of diabetes treatment, not just CGM, might be further developed. What would need to change for them to become approved solutions for mainstream deployment, or, at least, used as models to show what is possible and needed from industry, right now?

Nightscout

With a Nightscout dashboard at a nurse’s station, in handheld devices, or at remote locations staffed by diabetes specialists, clinicians could be alerted in real-time to:

• Blood glucose too high
• Blood glucose too low
• Urgent high or low predictions
• Loop not working for patients with closed loop solutions

A glance at the Nightscout display would show:

• Amount of insulin on board
• Carbohydrates on board
• Current basal rate of insulin pump
• How the closed loop is making its predictions
• How fast and by how much glucose levels are changing
• when insulin pump site & reservoir need changing
• when CGM sensor needs to be replaced
• … and much more…

COVID-induced adoption of CGM

The FDA approval of Dexcom and Libre for hospital use in April was a needed step forward, giving clinicians formal approval to monitor glucose remotely without finger sticks. Stories of doctors taping Dexcom receivers to the walls outside COVID wards came trickling in. And industry partnerships forged ahead with in-patient continuous glucose monitoring solutions.

But why not fast track solutions that enable viewing of all diabetes-specific information in an integrated, vendor-agnostic platform? It’s all important for glycaemic regulation, and glycaemic regulation has been shown to influence hospital outcomes. Many biometric markers are monitored in real-time. Why isn’t it the same for diabetes factors?

Silos or integration?

Hospitals, diabetes clinics and telehealth providers can utilise vendor-specific platforms for each insulin pump and CGM, but many of these still require special downloading cables, or viewing either glucose or insulin data separately rather than together, or are retrospective not real-time.

Having separate platforms for each system is cumbersome and inefficient for staff who are already run off their feet. We need an integrated system.

Multi-device platforms

Tidepool’s solution, developed by people with a close association to diabetes or diabetes themselves, has some very nice features for tracking diabetes data. It has reports combining CGM, insulin and other data, for an extensive range of diabetes devices. Glooko is another multi-device platform used in many diabetes clinics. But these are not, currently at least, real-time solutions. Could real-time functionality be added to them?

T1Pal

One development I’m keeping my eye on at the moment is T1Pal, a start-up from early Nightscout developer and OpenAPS contributor, Ben West.

Ben has created a managed Nightscout service for a monthly subscription fee, with added security and privacy controls, which he hopes will allow it to gain Part 11 FDA clearance for use as a “secondary display device”. The aim is to make Nightscout simpler to set up for people who don’t want, or are unable to, pursue a DIY solution.

I’ve been trying out T1Pal’s Nightscout service for the past few months. Although I don’t have quite the same degree of customisability as I do with my regular Nightscout system, and for people who use the Dexcom app it is only available in the US, one feature I really love is the way it allows me to create multiple share URLs that are easily revokable and keep the “real” Nightscout URL hidden. I’ve been using these for telemedicine consults.

I like to imagine a solution using Nightscout, or a similar real-time monitoring system, that combines data from the devices we choose to use because they work for us, might open up options for more people to access between-visit, just in time, clinical care at some point, especially in this new era of telemedicine.

I also wonder if they could help meet the needs of hospital clinicians involved in diabetes care?

FHIR / interoperability

Hospital systems are moving towards using the open standard, HL7 FHIR, for interoperability. Neither open source or commercial diabetes solutions have adopted this standard so far, but wouldn’t it be ideal if future diabetes wearables, monitoring software and hospital systems could talk to each other?

Sensotrend

In Finland, patient-led start-up, Sensotrend, has already taken a step in this direction. Its CE marked software uses FHIR to enable Nightscout data to be automatically uploaded into the Finnish personally controlled electronic health record system, Kanta PHR. Michael Rinnetmäki created the system in conjunction with long-term Nightscout developer, Sulka Haro, and others. In a 2019 presentation, Michael explains the constraints, enablers and potential of Sensotrend, and stresses that high levels of citizen trust in the Finnish government alleviates privacy concerns.

It’s compelling to imagine a future where FHIR-based Sensotrend data, could allow epidemiologists to identify disease patterns and treatment factors that influence outcomes in situations like COVID-19. I like to imagine a future where this information is fed back into creating optimal treatment protocols for people with diabetes.

Where to from here?

Type 1 diabetes is an extremely arduous condition to manage at the best of times. People with diabetes need access to the pumps, CGMs, algorithms, monitoring software, and automated insulin delivery systems, that streamline the process as much as possible. We need the ability to control who has access to our data and when. Clinicians we trust need access to this data as painlessly as possible to collaborate with us to fine-tune our settings. Automated decision support tools and just in time care are just around the corner and promise to boost physical outcomes and psychological wellbeing.

New legislation from the US Department of Health and Human Services mandates that from January 2021, systems must give patients full access to their own data via Patient Access APIs in order “to prevent information blocking and other anti-competitive behaviours” and enhance interoperability. “Standardized APIs will have a particular impact on third-party health app developers.”

So, it’s time.

Let’s embrace the solutions that tech savvy people with diabetes themselves, and their families, who have a deep understanding of the nuances of the condition, have gone out of their way to create. Let’s learn from them and create solutions that maximise interoperability and boost outcomes for everyone concerned.

Next time we are side swiped by a pandemic, let’s be ready for it.

References

Patton, M. (2020) COVID-19: Today’s diabetes therapy is tomorrow’s outcome data. myartificialpancreas.net

Galindo, R. J., Aleppo, G., Klonoff, D. C., Spanakis, E. K., Agarwal, S., Vellanki, P., … Pasquel, F. J. (2020). Implementation of Continuous Glucose Monitoring in the Hospital: Emergent Considerations for Remote Glucose Monitoring During the COVID-19 Pandemic. Journal of Diabetes Science and Technology, 14(4), 822–832. 

Braune, K. et al. (In press) Open-Source Technology in the NICU: A Case Study on Real-Time Continuous Glucose Monitoring in a Neonate with Transient Congenital Hyperinsulinism. Journal of Internet Medical Research.

Experiences from the Finnish Kanta PHR Ecosystem

Demanding Better Diabetes Care in Hospitals (2020) DiabetesMine

KORE and Dexcom Partner to Bring Innovative Solution in Response to Global Pandemic (2020) IoT.Business.News

Tidepool for Telemedicine (2020) www.tidepool.org/clinicians/telemedicine

Glooko (2020) www.glooko.com

T1Pal (2020) www.t1pal.com

Patton, M. (2020) Nightscout and T1Pal for telemedicine, via Youtu

DeBronkart, D. A tale of two patients: the difference #FHIR hopes to make with free-flowing data. (2019) www.epatientdave.com

HHS’ final interoperability rules standardize APIs for patient health data access through apps (2020) mobihealthnews.com

Clinical experiences keeping infusion pumps outside room for COVID-19 patients. (2020) Institute for Safe Medical Practices.