When public health officials track the outbreak of a virus, like H1N1, it takes time to get the story right. They have to collect and assemble data from institutions scattered across the country, a process that can be, well, slow.

For instance, at the CDC’s FluView website, you can see statistics for influenza trends across the country. But today’s “weekly influenza report” was assembled with data from the week ending 7 May 2011. Or put another way, the latest information is already 11 days old.

It seems crazy that sometimes the information we desperately need is the most difficult to get, but it’s all too often true. You can up-to-the-minute details on the location of your neighborhood’s taco truck, but if you want flu data, you’ll have to wait about 2 weeks.

Read more…

My inbox flooded with links to the report released by NIH (and evangelized by TIME) stating that lifestyle interventions (diet, physical activity, mental exercises, etc.) may not be that effective in preventing Alzheimer’s Disease.

Before I mount my full counterattack, I need to carefully read through the studies the meta-analysis cites.  Still, a quick glance at the exclusion criteria of the meta-analysis reveals the authors limited their review to studies using patients over the age of fifty.  So really, these results imply that lifestyle modifications may not prevent, delay, or treat Alzheimer’s Disease if you start these changes later in life.

My second point is that all lifestyle modifications are not created equal.  Scientific evidence in animal studies suggests that of all interventions, aerobic exercise is our best chance of staving off cognitive decline.  In fact, this meta-analysis also found some correlation between exercise and preserving or improving cognitive ability.

There’s a good article in The Economist that discusses the failures of the drug industry to find a solution to treating Alzheimer’s Disease.  One particular quote resonates with my feelings on the NIH report:

Another fundamental problem is that, whatever is causing the damage, treatment is starting too late. By the time someone presents behavioural symptoms, such as forgetfulness, his brain is already in a significant state of disrepair. Even a “cure” is unlikely to restore lost function.

Obesity (determined by BMI) and blood glucose levels are by far the best predictors of whether a person will develop diabetes. Yet doctors are always on high alert for new biomarkers that may be more sensitive indicators of which patients will develop diabetes in the near future.

The idea of using biomarkers to predict diabetes is not entirely new. Glycated hemoglobin (HbA1C) values are now routinely being monitored to screen for at-risk patients. However, a new study in PLoS One shows that several new biomarkers in the blood may further our understanding of exactly who’s at risk for diabetes, and increase our knowledge of the etiology of the disease.

Veikko Salomaa and colleagues from the Department of Chronic Disease Prevention at the National Institute for Health and Welfare in Helsinki, Finland, tested nearly 13,000 people and found almost 600 cases of diabetes during routine follow-up exams.

According to the study, low levels of adiponectin, and high levels of apoB, C-reactive protein (CRP), and insulin, increase the chance that a woman will develop diabetes. When these factors were measured, proper diabetes prediction increased by 14% compared to when doctors only use classic risk factors, such as BMI and blood glucose levels, to predict disease.

The biomarkers that best predicted diabetes in men were low adiponectin, and high levels of CRP, interleukin-1 receptor antagonist, and ferritin. Accounting for these biomarkers led to a 25% increase in correct diabetes detection in the cohort.

read the study here.

In Chapter 6 of The Decision Tree, “Screening for Everything”, Thomas talks about the human papilloma virus (HPV), the virus that causes cervical cancer. Traditionally, doctors detected HPV by looking for irregular cells in the pap smear. But now, a cheap ($5) test can detect and analyze the DNA of the virus, determining if it is the high- or low-risk type, which can determine the likelihood of a patient developing cervical cancer.

One problem remains: you still have to get women into the clinic to be tested. However, a new study in the British Medical Journal shows that home testing is not only a reality, but it may actually boost compliance rates. Roughly 28% of women using the home testing kit, which consisted of a simple cervicovaginal lavage, effectively screened themselves, while only about 17% of women required to go into the doctor’s office for screening showed up.

The HPV DNA test is primarily looking for the high-risk virus serotype, and the authors of this study claim that home screening kits have the same sensitivity as the doctor’s protocol when specifically looking for the aggressive virus.

Special thanks to Lindsay Crouse for bringing this to my attention. In her email to me, she brilliantly summed up the significance of home HPV testing:

While screening has been tremendously successful in Western countries at reducing cervical cancer cases and deaths, the obstacle of reaching all women through screening remains. Currently, if a woman is to be screened for cervical cancer, she must visit a health care provider for a gynecological exam. If she is unable or reluctant to do that, whether due to transportation, cost, or comfort issues, she is less likely to get screened at all, and is consequently at increased risk for developing cervical cancer. More than half of such cancers are typically diagnosed in women who do not get screened regularly.

[this is a post I did for TheHealthcareBlog, crossposted here]

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 Ottawa 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.

Calculator image via Flickr by Ian Ruotsala

News of 23andMe launching an online community for expecting mothers, and a few recent journal articles talking about the dangers of environmental polychlorinated biphenyls (PCBs) on both prenatal and neonatal brain development, have got me thinking about the future of in utero preventive medicine.  Lesson: I guess it’s never too early to begin implementing a Decision Tree…

A quick note that my latest story for Wired, on the emerging science of early detection of cancer, is now on stands (and online).

The story focuses on the Canary Foundation, a Silicon Valley-based nonprofit that’s funding an innovative approach to cancer research: strictly focusing on developing two-step tests that will spot various cancers in their earliest stages, when the odds of successful treatment are highest.

My effort here was to explore how early detection – which sounds obvious on its face; of course we should find cancer early – in practice creates a series of riddles and/or paradoxes. For instance, when you’re looking for something floating in the bloodstream (a molecular signal of early cancer), how can you be sure it’s present in high enough volumes early enough to be worthwhile as a test? Or: What if a test is great at spotting cancers that, paradoxically, may not actually be lethal, and thus may not merit immediate treatment? What I find admirable about the Canary Foundation approach is that they don’t look at finding a protein or a DNA signal as the be-all/end-all of a valid test – it’s just the beginning the a statistical parsing that may or may not result in something clinically useful.

If it’s not obvious, the connection to the decision tree thesis is this: Finding disease early, when treatment choices are various and have more promise of success, is a far better position to be in than waiting for symptoms and late-stage treatments. My hunch is we’re going to be moving towards more and more screening tests for more and more conditions. The challenge will be striking a balance between good tests that and the expense of too much screening and too many false signals.

Oh, and a shout-out to Wired’s design department, helmed by Scott Dadich, which always does an ace job turning some rather sober writing on my part into something alluring and cover-worthy.