by Jessica Richmond
While some have been successful in signing up for health care on healthcare.gov, many others have been frustrated, and citizens and the media continue to raise complaints and frustrations about this poorly functioning website.
With the President of the United States making national news statements about The Affordable Care Act being “more than just a website” and promising to resolve the issue, the question remains: How can we heal HealthCare.gov? And importantly, how long will that take?
It’s important to note that technology projects of this scale are complex to say the least.
So, what would the technology doctor order for a recovery plan for HealthCare.gov? Here are some steps:
1. Evaluate the Patient: In a previous post I mentioned the critical importance of testing and testing often. We know HealthCare.gov is sick, but without additional testing and data, it’s impossible to know if there is one root cause or multiple. Before that is determined, any efforts to remediate issues are premature.
2. Assess the Symptoms vrs. Source Problems: We have all heard about the issues with load and the long wait times and errors, but those are all symptoms of the problem, not the diagnosis. The site needs to be thoroughly tested to determine at what layers and points the breakdown is occurring. When testing, they should pay special attention to integration points as these are the biggest risk items.
3. Get Multiple Opinions: A single diagnosis isn’t the best approach to complex technical issues at this scale. When my team was brought in to solve a large problem recently, I created two separate engineering teams to evaluate and diagnose the problem. The teams worked independently and then we compared results at the end. This helps reduce risks associated “groupthink” and explores multiple options - giving us the best chance of finding the true root cause.
4. Don’t hide the Diagnosis: There is no therapeutic privilege in witholding information about what ails healthcare.gov - with the exception of confidential data and personally identifiable information (PII). Open sourcing the problem or parts of the issue and getting feedback from the community is a very effective way to grow public sector code bases. For example, The White House is open with the We The People codebase on GitHub.
5. Acknowledge that Trial and Error Still Has a Role: With complex code bases that are experiencing performance issues, the best way to resolve the issue is to take a strategic approach and then iterate over time using rapid prototyping approaches and frequent retesting. This allows us to know what is working, what is causing other issues, and how to iteratively improve.
6. Staff Correctly & Iterate Frequently: Don’t try to solve a big problem all at once. Make incremental changes and control the environment. Losing control of the environment and process will result in more problems. So while throwing resources at a problem may be tempting, don’t allocate resources until you have your testing and quality assurance approaches established. Then when you make even a small improvement - release and release quickly. When the world is watching, even the smallest improvements count.