Every time you read about health care IT investments, the 2 big things that standout are the budgets, measured in hundreds of millions, and the time, measured in years for roll-out. The big issue that is rarely raised is, what effect has past investments had on health care?

An immediate standout is the data generated. Modern data systems are very powerful and capable of serious number crunching. Given the popularity of hospitals, especially for the sick, there is plenty of data generated by people afflicted by complex and not so complex needs.

On the face of it, data systems should represent a godsend for busy clinicians, but the complex world of health care, with its numerous issues does present a double edged sword. The trouble with data, is that it has to be understood and acted upon. The more data, means the more resources required in order to do something. The conflict arises with clinicians gaining more complex patients (drugs, injuries, chronic disease etc) who now generate more data (pathology, medication, care plans, tests) for the clinician to hunt down across multiple systems and to make sense of.

With all the issues facing health care, this is not to discredit data systems, who knows what the world would be like without them. But! Having an ultra powerful tool does not represent gaining a feasible solution. While it looks good to media outlets, that spending on health care IT is into the hundreds of millions, I would like to see more questions raised on the delivery of clinically relevant solutions.

As often can be the case, one generic term can give birth to a range of different definitions, and the world of Decision Support in health care is no different. The term decision support is typically used widely in health care to describe any assistance with decision making such as a paper document, graphs, or the advice from an advanced computerized inference engine. So as with all things, there are some things that are more useful than others. To make it more simple, I think the starting point is to define decision support as tow categories, ones that are passive and those that are active.

Passive decision support systems apply to systems which wait for users to request advice. Alternatively, a system may be active if it monitors for emerging problems and pro-actively alerts clinical staff to minimize risk before harm occurs. 

There is however a third category and one which I consider places the whole world of decision support systems into a completely different league, and that is: computer systems that can infer new information from data and actively support decisions, in stark contrast to approaches that simply present or summarize data.

Without clinical experience, it is difficult for many people to understand what a quantum leap this third category represents. Historically, work in clinical decision support has been in the areas of passive data presentation (e.g. guideline pdf documents), medication interaction checking and care plans. The time you spend fumbling with extensive (and rather complex) charts representing treatment pathways, is often extensive. 

In the context of delivering vast improvements to health care (computer systems that can infer new information), only current medication interaction checking can be classed as a decision support system. Items such as care plans act only as a reference tool and cannot be considered as a decision support system per se, seeming they do little to resolve critical issues facing health care, something that a fully-fledged Decision Support System should provide.