I watched a Midwife trying to enter my partners test results into what I believe was a CERNER system at my local hospital. It was painful. The UI is text based, looks pre windows 3.1 and breaks all the basic rules of UI design. Evidently, they also had more serious technical issues under the hood:
http://bit.ly/4vLPtE http://bit.ly/9NQKi8
It took them months to sort this out and wasted a huge amount of the medical staff's time. I personally experienced the problems with appointments disappearing. Also, the nurses were having to print out test requests and give them to patients to take to appointments with instructions about what they information they needed to bring back because they couldn't make the requests through the system or access the results of the tests they had requested.
In short, I don't think the success of this company is based on the quality of its software or the timeliness of its delivery. Maybe this is always a problem where the end users are not the customer. The company that builds the system is chosen based on tenders from a few suppliers. As a consumer of a system like this it's hard to identify what you are actually buying. The purchasing of such a system will be managed by a senior executive who will have little experience of the pitfalls of the system in use before they are contractually committed to paying out significant sums of money. There is no market connection between the quality of the software and the end users. I suspect that companies make money in these markets by gaming the bidding process and milking every loophole in the contract to add extra time and cost, not by actually producing a better quality product than their competitors. It's not just a public sector purchasing problem either. Just look at any large construction project, these are also typically be late and over budget. Is there some kind of 'complexity threshold' at which point the market is not driving the quality or cost of the product because you are not sure what you are buying and for how much?