18 August 2022

Scales of failure: CovidSafe App vs My Health Record

Apps Government MHR

Two years, 17 close contact identifications only, a mass of people’s very personal data parked somewhere and $21 million later, the CovidSafe app is being shut down. Anything to see here? 

Government needs to be able to build or oversee the build of tech, and be competent at it. Technology is at the centre of (potentially) far better experiences for citizens’ engagement with key government services departments such Centrelink, the NDIS, Medicare, the RMS and the ATO. But these are obvious core infrastructure requirements for any modern government.  

When we start reaching beyond the provision of such core services, things often go horribly wrong as far as government building and maintaining tech infrastructure is concerned.  

Is there a line that maybe government shouldn’t be crossing in trying to help us manage our everyday lives? If there is, how do we know we’ve crossed such a line, and what can do we do when we realise we’ve crossed it? 

Consider the relative scale of two high-profile health tech fails: the CovidSafe App, which cost $21m and has only now been cancelled by the new Labor government, and My Health Record (MHR), with which we continue to persist, with much the same energy and cost we have persisted with for the last 10 years.  

In the case of the MHR we’ve just decided to fork out at least another $300m or so ($100m for Accenture, the rest likely for the Australian Digital Health Agency to keep managing it) over the next three years, after spending more than $2 billion on it already over 10 years. This despite there being no formal measurement ever provided of ROI to the community of the project, to the point where last year the agency running it was directed by the Liberal government to develop a “meaningful use” index of the project.  

Presumably, the intent of such an index is to demonstrate some ROI to the public. But there’s an obvious problem. Would the agency whose existence revolves largely around a project that is more than 10 years old and costs more than $2 billion really develop an effective measure of ROI?  

The government had a chance to mothball the MHR about six years ago when it reviewed the whole project and installed a new head of the Australian Digital Health Agency, but instead it decided the perceived issue until then of low usage could be solved by forcing the public to opt-in to having an MHR. That made the numbers look much better. But it didn’t move the dial much at all on meaningful use by the people that counted: doctors and patients.  

In the meantime, investment in the project over time has nearly doubled. 

The more the government sinks into the project, the harder it gets to admit there is something fundamentally wrong. After all, how could you spend more than $2 billion and still not have a lot to show for it?  

Who wants to pull that trigger? 

The pressure on government to keep doubling down is enormous and without any other clear digital health strategy. Even the new Labor government appears to be folding to that pressure by agreeing to the “business as usual” Accenture contract extension for three years. 

You might consider that the CovidSafe App fail at $21m is not in the league of the MHR at close to $2.5billion, but in terms of percentage return on investment, the CovidSafe App is probably a much bigger failure than the MHR.   

At least with the MHR we have a lot useful learnings, and some infrastructure and data sets that would be ongoing and useful in any new digital health strategy that we probably wouldn’t if not for the project. For example, Individual Healthcare Identifiers are going to be vital to any ongoing digital health strategy, and having a lot of the population opted in to access and provide data isn’t a bad thing.    

If the government were to eventually recognise that a lot of the money being poured into the MHR is (a lot of) good money after bad now, and that this money could be far better spent on developing distributed data sharing capabilities within the system via much more modern and appropriate web and cloud data sharing technologies, then it’s likely that a lot of what has been built around the MHR would continue to be applied to any new strategy and focus.  

The MHR just wouldn’t be the centre of digital health strategy anymore, sucking away money and obvious opportunity for more effective means of data sharing between health providers and patients. 

Compared to MHR, $2.5b or not, there are no redeeming features of the CovidSafe App project at all.   

That the project has only last week been officially shelved poses some interesting questions about when and how government recognises its health tech building failures, and how they stop them within the context of the politics surrounding such projects.  

The stats on the CovidSafe App are very telling of something obviously wrong from an early point in a project that was never stopped for some reason. It might be useful for government to stop and ask itself if it is a microcosm of the same dynamics that have been in play, but at a far greater scale over a far longer time period, for the MHR? 

Here are some of the basics on the CovidSafe App that tell a very strange story of persistence in the face of obvious dysfunction and failure: 

  • Total cost at closure: $21million 
  • Launched: April 2019 
  • Immediately identified tech issues of phone Bluetooth capability for both major phone platforms – locked phones couldn’t talk to each other and most phones were locked. 
  • Other issues identified with functionality soon after – it didn’t seem to work at big functions, for instance. 
  • Not updated until late 2021 to track the Delta variant – largely too late to have any effect on Delta spread. 
  • In its total operational life it only identified 17 close contacts that had not already been picked up by manual contact tracers. 
  • At launch it was stated it would require at least 40% or 10 million registrations to work. It only ever achieved 7.9 million registrations. Six million registrations were achieved in the first few weeks of the app, after which it never looked like getting to 10 million. 
  • It also relied on users consenting to their positive test results to be used but only 800 users ever consented to this. Again, early on. 
  • It failed to uncover any close contacts during massive covid spikes in NSW and Victoria in September 2021. 
  • The cost of the app until Sept 2021 was $10m. Despite all of the above, the  government then literally doubled down to take the total cost to $21m.  
  • Of the additional money spent, $7m was spent on advertising, again, despite very clear evidence the app had already failed in every respect and had numbers that did not make it workable ever. 
  • New government elected in May 2022. 
  • App not formally decommissioned until August 16, 2022. 

Although the MHR has been a project run over 15 years at a cost of $2.5billion or so, and the Covid Safe App has cost $21m over only two years, there were some eerie parallels in how the government persisted with each project. 

The eeriest might be that at the midpoint of each project, despite obvious signs of failure, the government’s reaction was to double down in an attempt to force success. 

In the case of the CovidSafe App, announcing a failure of $21m in the end, as embarrassing as the above list is, wasn’t that big a deal. 

Obviously, applying the same review and logic to the MHR is proving a lot harder for the new government. 

But there appear to be some very clear patterns of behaviour that the government would do well to look at in the failure of the Covid Safe App, and apply any learnings it might get to its much larger health tech plays such as the MHR, which seems to have displayed very similar characteristics.