≡ Menu

Mr Obama, smarten these systems!


Like millions of my fellow Americans I listened to our new President today. As I did I was struck by the opportunities for decision management to deliver the smarter systems that will be critical with some of the priorities President Obama laid out in his speech. There were four commitments he made that struck me particularly – about the power grid, about healthcare, about government and about regulation.

“We will build the roads and bridges, the electric grids and digital lines that feed our commerce and bind us together”

This is a classic lead in to an infrastructure-based stimulus – spend government money on things that will help the economy and the nation. Given decision management’s focus on “software” rather than “hardware”, it might not be obvious what the role of decision management would be in this element of the plan but there are several. First look back to the start of the speech where he said “… the ways we use energy strengthen our adversaries and threaten our planet.” It seems to me that it is not enough for us to invest in our electric grid and digital lines, we must do so in a way that changes our energy usage patterns. One of the most straightforward ways to do this is to invest in what is known as smart metering. Smart meters measure power use like traditional meters but track exactly when the power was used. This allows power suppliers to introduce variable pricing – charging more for power when demand is high and less when demand is low. Not only does this allow consumers to manage their power costs more effectively it has been estimated that it would mean we could avoid building any new power generation capacity for a significant period and could retire some of our worst polluting generating capacity.

Smart Meters, of course, aren’t that complicated but they change the kinds of systems the utilities need. Instead of billing being a fairly mindless calculation, now the rules for billing need to be applied. Given how people and businesses are these rules will change constantly – utility companies will tweak the rules to manage their workload and maximize their profit, cities and states will require specific rates for certain people at certain times or will require that certain times should always get the best prices, large electricity buyers will negotiate their own contracts and so on. Unless there is a system that can make a smart enough decision on the back end, this constantly evolving set of rules will cause problems.

The longer term consequences are just as interesting. Suddenly there is all this data about who uses what power, when. Data mining to find trends, segments with specific behaviors and to predict how people might react to specific price changes will become the norm. Integrating the results of this data mining back into the pricing models and the pricing decisions will start the process of moving towards optimal pricing, balancing the needs and behaviors of consumers with the intent of regulators and the profit of utilities. A smarter grid driven by smarter billing.

“…wield technology’s wonders to raise health care’s quality and lower its cost.”(my emphasis)

Improving healthcare quality is something with which technology, especially information technology, is often associated. Yet too much of the discussion focuses on electronic medical records (EMR) and physician order entry (POE) systems and not enough on care. While its useful to make sure the information needed is available to physicians and that their instructions can be rapidly and accurately passed on to pharmacists and other medical professionals, none of this is focused on improving care. To improve care we must improve the decisions about care that physicians make. This means both checking those decisions for errors like incorrect dosage or drugs that will interfere with each other and making sure good decisions are made in the first place. EMRs and POEs can only begin to address these quality issues if they focus on point of care decision-making. This kind of clinical decision management involves systems that are working on behalf of the medical professionals to make decisions when there is no staff available (in automatic monitoring of patients, for instance) and when medical professionals don’t realize that a decision needs to be made (or when they appear to have forgotten). Clinical decison management systems can use all the information in a patient’s record (not just the amount that can be read in 2 minutes between patients) and compare it with all the tests being carried and out and all the symptoms being reported to help improve care. Clinical decision management also addresses both the increasing need for patients monitor their own health and do so from home and the costs of having people come to hospitals unnecessarily. Clinical decision management can also:

  • Check for drug interactions (rules-based analysis of all the prescriptions known at the point/time of checking)
  • Check doses and instructions against patient data (weight, age, sex etc), applying the rules specific to the drug
  • Identify potentially fraudulent patient identities using rules and analytics like neural nets to find unusual patterns
  • Recommend follow-up activities using rules about drug procedures and predictive models for things like likelihood of following a protocol.

Clinical decision management can reduce costs and improve results while also improving the patient’s experience and there are not many technologies that can say that.

Of course care is only part of the healthcare system – there is also insurance. As someone who grew up in the UK with “socialized” medicine (otherwise known as “decent care for everyone”), the whole US health insurance process can seem a little bizarre. What is clear, even to me, is that smarter systems could make a real difference to the cost of healthcare. First, the decision to pay a claim can and should be automated more effectively. Not only can costs for claims processing be reduced by eliminating unecessary manual review, more fraud can be detected at lower costs. The credit industry has shown that fraud decision management systems can reduce fraud by huge percentages – 80% or so – and the limited adoption of these kinds of systems (combining explicit rules and policies with sophisticated neural networks) in healthcare has shown similarly effective results. With all the money flowing into Medicare, Medicaid, Workers Comp and just plain old health insurance it would be astonishing if there were not fraud and pretty organized fraud at that. Detecting this fraud before it is paid is orders of magnitude more effective than the established “pay and chase” mentality. Every dollar of fraud found and driven out of the system is a dollar saved to pay for healthcare.

Fixing our healthcare system will require many steps, but using smarter systems to drive fraud out of the system, improve processing efficiency and improve care should be part of it.

“…not whether our government is too big or too small, but whether it works” (my emphasis)

It is clear from President Obama’s previous comments and the immediate use of whitehouse.gov on inauguration day that this is an administration that plans on using technology to improve government. Much of the interest has been in his very public commitments to transparency. But there is another side to using technology, and smarter systems, to improve the government works. Systems smart enough to automically process tax returns while being easy to change to reflect new regulations can make the tax system work better (the IRS is working on a tax decision management application as we speak). These systems can also help detect fraud, just like in healthcare, and determine who should be targeted for audit bringing in the tax revenue that is owed by scofflaws and lowering the burden on those of us who are honest. Systems that can handle the approval of permits or eligibility for programs and benefit without manual intervention, without delay and without getting out of date as the regulations change make government work more efficiently and less expensively. Not only that, they also allow citizens to get the anwers they need to build their businesses and run their lives when they want to not when a government office happens to be open. Not just e-government but smarter government.

“Its power to generate wealth and expand freedom is unmatched, but this crisis has reminded us that without a watchful eye, the market can spin out of control”(my emphasis)

We have all watched the market meltdown over the past few months. Smarter systems will be essential if government agencies are to monitor the market more closely. Capitalism thrives when it is dynamic so the systems government uses to regulate it must also be dynamic and easy to change. Regulators must be able to change how they check transactions, what early-warning signs they want to look for and much more. Only a smarter, more agile set of systems will do. If they have to handle the real-time transaction load of modern capitalism manually or with hard-to-change systems they will inevitably fall behind. And we know where that goes…

A technology-aware administration can also push companies to adopt systems that will help monitor keep a watchful eye on the market. Too often companies avoid taking responsibility for monitoring their business in real time by arguing it is too hard to do. Regulators are forced to tolerate monthly reporting after the fact. Yet the technology for smarter systems that monitor and manage every decision not only exists, it works well and has proven again and again. Not only should regulators make sure their own systems are smarter, they should make sure the companies they regulate also have smart enough systems.

So if our new President wants a better grid, cheaper and better health care, government that works and a market that is watched but not hamstrung then smarter systems and decision management should be on his list.

Thanks to hurriyet.com.tr for the text of the speech.


Comments on this entry are closed.

  • Mark Eastwood January 21, 2009, 7:51 am

    One other thing to add to your comments about medical care; smarter systems to monitor test results. Using rules the university of Virginia has a system that gets clinical results (I forget the name but its now part of GE Medical). The system looks for test results that are significantly different than prior tests on the same patient or changes of x% in y% time period or dangerous levels of some test result… say PT (clotting factor) drops or rises to some dangerous threshold. The idea is that the system pages the dr/pharmist/nurse to alert. The system alerts when the same (or similar class/type) of drug is administered by different people using different methods (pill/shot/IV).