Fifty pages of audit, 62 million dollars and zero patients. That was the bill in 2017 for a system that was meant to help doctors at the MD Anderson cancer centre in Houston make decisions about cancer treatment.1
It did not treat patients badly. It did not treat them at all. Not a single person.1

This ending is well known. The road to it is more interesting, because it can be walked backwards, and at every point you can ask who saw what at the time.
A year earlier, when more than sixty million had already been spent, the system was sitting in a test environment and could not reach live data. The pilot ran on the old hospital information system, and in the meantime the hospital had moved to a new one. Without live data it could not be used clinically.12
One step back. Earlier still, the scope of the project changed several times, from one type of leukaemia to another and then to lung cancer. The pilots in partner hospitals, on which the whole business case of the project rested, never got off the ground.2
Continue reading
Read the full article for free
Enter your email and the full article opens straight away. No payment, just a subscription to TORVAKO Insights.
- Full articles with all the sources they draw on
- Once a month, an overview of new articles and one story from practice
- You will be the first to receive TORVAKO study results
Already a subscriber? Enter your email and the article will unlock on this device too.