Why Calculators Are the Future of Healthcare

Thomas goetz

Want to know the future of medicine and healthcare in one sentence?

For my money, it goes like this: The real opportunity in healthcare is to combine our personal data with the huge amount of general biomedical and public health research, in order to create customized information that’s specific to our person and our circumstance. We need relevance, and the right information at the right time will help us make better choices for prevention, helping us stay healthier longer, it’ll help us navigate diagnosis, letting us select screening tests that are useful and not unnecessarily fearful, and it’ll let us make better decisions on care and treatment – when we’re trying to choose among various treatments to find our way back to health.

It’s in the last category – care and treatment – that I wrote a recent post at the Huffington Post about one man’s story with prostate cancer. Tom Neville got a diagnosis and then had to struggle to find information to help him make sense of what to do. Ultimately, he chose surgery, but the difficulty of the choice led him to create Soar Biodynamics, a company that offers decision-making support for men assessing their prostate health.

You can read his story here and learn more about his tool here, but for the purposes of this post I wanted to consider the kind of decision-making tool he created. It’s called a nomogram, and it’s one of my favorite discoveries in researching The Decision Tree.

A nomogram is basically a calculator – a way to assess our risk or outcome for a particular condition. A nomogram starts with an interface where a few telling datapoints can be entered, and then turns to an algorithm that crunch those numbers together with broader data about the condition. The result is a statistical prediction – the prediction can concern the outcome of the disease, or it can be a recommendation for particular treatment (a medical nomogram is not to be confused with mathematical nomograms, which are tools for calculating geometrical something or others).

The Framingham Risk Calculator, which calculates your risk of heart disease, is a kind of nomogram.

Memorial Sloan-Kettering Cancer Center, the research institute and hospital in New York City, has developed almost a dozen nomograms for a range of cancer conditions. There are tools for predicting the spread of breast cancer, a tool for assessing lung cancer risk among smokers, a tool for predicting the prognosis after colon cancer surgery, and more. Dr. Pierre Karakiewicz at the University of Montreal Hôpital Saint-Luc has developed nomogram.org, which offers prediction calculators on four different types of cancer.

Nomograms are one of the best examples of Decision Tree thinking, the sorts of tools that are easy for patients and doctors alike to use and understand—particularly when they’re available online and free of charge. They’re brilliant and auspicious because the turn research around so that it faces the patient: An individual can interrogate medical science for how it applies to his specific circumstances, rather than having to navigate through stacks of research papers and findings for some wisp of relevance.

Nomograms are especially powerful when they’re combined with a screening test, because they help people understand what to make of the test and point to what to do with the result. They immediately customize the clinical data, be they nanograms-per-milliliter figures or spots on mammograms. Nomograms let patients ignore the inscrutable repository of jargon that is medical research in favor of something personal, something real, and something to go on. They allow us to make sense of a screening test’s result, and allow us to take some measure of meaning from it.

The University of Texas at San Antonio, for instance, has developed a prostate risk calculator that lets a man enter his PSA level along with his age, race, family history, and a couple of other metrics and churns out his risk of developing prostate cancer. Importantly, the calculator also calculates the risk of a high-grade cancer, accounting for the fact that not all prostate cancers are lethal. The value of such a tool, says Ian M. Thompson, professor and chairman of the department of urology at the University of Texas Health Science Center at San Antonio, who developed the calculator, is that it turns the PSA figure from one isolated data point into one of many inputs. “We need to build in characteristics about the person, their age, their race, their family history,” says Dr. Thompson. “It’s not just what one test tells us.”

Nomograms, of course, are no substitute for a doctor’s definitive assessment and treatment (or better yet, more than one doctor). And they are only as good as the data that goes into them; if they’re not kept up to date on the latest information and research, they can lead people astray. But especially for conditions where we have some agency – where we can take actions today that can enhance our tomorrow – they are a terrific tool.

The catch with nomograms is that they must be developed one disease at a time, which means they don’t scale up so well. Each one takes a great deal of work and expertise. But if I had millions of dollars for philanthropy, I’d spread it around to smart researchers across a lot of fields where nomograms could help people assess their risk for disease and potentially take actions today. It would be money well spent.

Thomas Goetz is the author of The Decision Tree: Taking Control of Your Health in the New Age of Personalized Medicine. The executive editor at Wired Magazine, you can follow him on Twitter twitter.com/tgoetz.

Note: This article has been edited to reflect a correction. Dr. Pierre Karakiewicz is with the University of Montreal, not the University of Ottawa, as previously stated.

12 replies »

  1. The initiative taken for the concern is very serious and need an attention of every one. This is the concern which exists in the society and needs to be eliminated from the society as soon as possible. The above statement is seen to be contradictory. The situation is very critical and need an experience complainer to resolve it. This action proof to be a win, win situation. This is a true art work, which will be a success story.
    Best Savings Account

  2. Also like to add the ROI calculator which helps in evaluating the prospects of installing an EHR system at your practice.
    I think ROI is very important factor that should be duly considered when look achieve a ‘meaning use’ out of a EHR solution. Though one may get vendors providing ‘meaning use’ at a lower cost, their ROI / savings through the use of their EHR might be pretty low when compared to costlier initial investment.

  3. What will help propel a movement towards patients/ consumers using predictive analytic tools is when the majority of providers, clinics, and hospitals have an electronic medical record system in place. While this technology is often expensive, many providers are leveraging stimulus funding from ARRA and partnering with larger hospital systems in their network to obtain licenses at a discount. Payors really dominate this space. They have the incentive to keep costs down. Wellpoint’s Resolution Health has a patient decision support tool. http://www.resolutionhealth.com. At this time, anyone using a predictive analytical tool is using claims data (after you visit the doctor information). When this changes to encounter data and combines real-time bio-monitoring data from apps like Wii Fit, bluetooth enabled glucose, heart rate, and blood pressure devices that one can pick up from a local pharmacy, then a patient decision support tool will have ROI to the patient, and likely to the payor insuring them. I hope this type of technology will support the relationship between patient and provider.

  4. Tom, I’ll gladly use any calculator you devise that Dr. Nortin Hadler might endorse.
    So I don’t expect to be calculating any time soon.

  5. So patients should be using these calculators from home and determine what care they want to get. Then they can see if they have any coverage for their needs and wants and then decide to go ahead or not.

  6. right, Steve – so we need *two* tool sets: one to navigate the decision, and another to execute the behavior. The good news is these are popping up here/there, but by making these explicit targets hopefully more people/organizations will start filling in the gaps.

  7. Thomas — I really like the point about relevance as key to health decisionmaking. Health decisions then need to lead to actions so I’d also argue that the for every nomogram, it would be great to have a companion tool that, once you made a decision, could help you act on that decision. So for example, if you decide to improve your diet, it would be nice to have a tool that factors in your specific health conditions and predispositions, and then makes recommendations that also factor in your preferences (e.g. type of cuisine, interest in cooking) and the practical realities of your situation. These realities might include, for example, the kinds of foods you can find in your neighborhood, how eating fits in to your daily routines, or how much money you have to spend on food.

  8. Well, yes, the datasets are limited to basic demographics and claim data, but I think they continuously run these things and aggregate results to estimate expenditures for their pools.
    What would be nice is to license their basic algorithms, update for more personal data points, like diet and exercise and other things insurers don’t have, and create a website. People could enter their various conditions (not just one disease) and circumstances and it would give them a complete assessment of what’s likely to happen, with various paths possible (like what if I stop smoking next year or what if I stop taking those statins). It could also tie into a medical knowledge database and provide personal information and education materials customized just for you (Don Kemper’s HL7 Info-Button). The killer app would be to tie it all together with a PHR or a patient portal and maybe genetic profiles and so much more. Talk about patient empowerment….
    Can we get some seed capital and get started ? 🙂

  9. Great point about insurance companies! Life insurance pretty much invented this game. But what their algorithms lack are new details about the individuals; the personal datasets are both limited in scope and by time (far as I know, they don’t come back for more info once they issue a policy).

  10. Don’t insurance companies have tons of these algorithms already implemented for their risk assessment?
    Maybe we should ask them to release the algorithms to the public domain and have someone put a cool Web 2.0 UI in front.
    Or maybe this could be a new revenue source for insurers, now that most people can’t afford their traditional products :-).