[Reblog] Will Getting More Granular Help Doctors Make Better Decisions?

From the 8 February 2015 post at The Health Care Blog

Will Getting More Granular Help Doctors Make Better Decisions?

Excerpt (longish post)

But, there are many things that data will never do well.  For certain things, physician heuristics may lead to better decisions than any predictive model.

Heuristics are shortcuts, based on experience and training that allow doctors to solve problems quickly.  They are pattern maps that physicians are trained to recognize. But, heuristics have a reputation for leading to imperfect answers: Wikipedia notes that heuristics lead to solutions that “(are) not guaranteed to be optimal, but good enough for a given set of goals…. (they) ease the cognitive load of making a decision.”  Humans use them because we simply can’t process information in sequential binary fashion the way computers do.

It would be a mistake to call heuristics a sad substitute for big data.  Some cognitive scientists have made the argument, and I think they’re right, that heuristics aren’t simply a shortcut for coming to good-enough answers. For the right kinds of problems, heuristically generated answers are often better than the those generated by computers.

How can this be?

I often think of the following cartoon in Randall Munroe’s superb recent book, What If? Serious Scientific Answers to Absurd Hypothetical Questions.  In trying to compare human and computer thinking, he rightly notes that each excels at different things.  In this cartoon, for example, humans can quickly determine what they thought happened.  Most people can tell you that the kid knocked over the vase and the cat is checking it out, without going through millions of alternate scenarios.  Monroe notes that most computers would struggle to quickly come to the same conclusion.

So, from the perspective of an emergency doctor, here are the three leading problems with the applied use of complex analytics in the clinical setting:

• 1. The garbage in, garbage out problem.  In short, humans regularly obfuscate their medical stories and misattribute causality. You need humans to guide the patient narrative and ignore red herrings.

• 2. If we want to be able to diagnose, screen and manage an ER full of runny-nosed kids with fevers, we simply can’t afford the time it takes for computers to sequentially process millions of data points. The challenge is at one simple and nuanced: allowing 99% of uncomplicated colds to go home while catching the one case of meningitis. It’s not something that a computer does well: it’s a question of balancing sensitivity (finding all true cases of meningitis among a sea of colds) and specificity (excluding meningitis correctly) and doctors seem to do better than computers when hundreds of cases need to be seen a day.

• 3. There is a problem with excess information, where too much data actually opacifies the answer you’re looking for. Statisticians call this “overfitting” the data. What they mean is that as you add more and more data points to an equation or regression model, the variability of random error around each point gets factored in as well, creating “noise”. The more variables, the more noise.

The paradox is that ignoring information often leads to simpler and ultimately better decisions.

…

Related articles

• It takes more than great algorithms to make personalization work
• Sanjay Bakshi: Daniel Kahneman vs Gary Klein

How Text Messages Could Change Global Healthcare

This October 24th Popular Mechanics story includes

• How text messaging is used to coordinate health care by health care professionals in rural areas across long distances
• How text messaging in Haiti was used to locate victims in search and rescue efforts despite language barriers
• Camera phones as diagnostic aids

Excerpts

The notion that SMS could revolutionize healthcare first entered Nesbit’s mind in 2007, when he was still a Stanford undergrad. He’d just met Dickson Mtanga, a community health worker in rural Malawi who was walking 35 miles to deliver handwritten patient charts to the nearest hospital. Nesbit biked out to Mtanga’s village one day, only to discover that his cellphone got a better signal there than it did on Stanford’s campus in Palo Alto, Calif. All those bars of service jumped from the phone’s screen and slapped him across the face: These far-reaching GSM networks, he realized, could connect doctors and patients like never before.

Armed with a $5000 grant, a backpack full of old phones, and a laptop running a GSM modem and the open-source group-texting software called FrontlineSMS, Nesbit started working with the hospital and community health workers to coordinate patient care. The system they put in place allowed Mtanga and others to text in the information on those medical charts rather than making the hours-long trek. Patients could text their symptoms to doctors, cutting down on unnecessary visits for minor ailments and freeing up space for those in need of serious care. Within six months of the system going live, the number of patients being treated for tuberculosis doubled, more than 1200 hours in travel time were eliminated, and emergency services became available in the area for the first time. The operating costs in those six months: $500, Nesbit says

The explosion of cellphone use around the world has inspired a flood of new ideas about how to use that tech to improve healthcare. Besides Nesbit’s Medic Mobile, there are also ideas to turn camera phones into cheap diagnostic tools for vision problems or malaria, for example.

Patty Mechael, executive director of the U.N. Foundation’s mHealth Alliance, keeps tabs on these new techs. They all face major infrastructure hurdles, such as the lack of reliable energy sources to power phone chargers in some developing countries. But another, less tangible challenge is figuring out what mobile health programs are actually working and worth scaling up, and which ones aren’t. “What we have in mHealth are millions of flowers blooming, in many ways. Lots of pilots are being done throughout the world, many of which are reaching populations of a few thousand each,” Mechael says. “We’re at a tipping point where people are starting to say, ‘Okay, we need to be a bit more strategic, collaborative, cohesive.’”

Nesbit is among the voices calling for a more focused approach to mobile health. A wave of angst washes over his face when I ask if there’s too much hype surrounding mobile health, if it’s too saturated of a field. Hype is good, he says. What’s bad is hype that’s disconnected from implementation. All the media coverage and promises made about mobile health in recent years, he says, make it seem as if millions of health workers in developing nations have already integrated their phones into their daily practice. In reality, only about 20,000 have done so. Medic Mobile has SMS systems operating in 14 countries, and that number will jump to 20 in the next six months. Only a few thousand people are using Medic Mobile’s programs today, but the nonprofit just rolled out its first SIM card application, which can be used on virtually every mobile phone in existence. By 2015, Nesbit expects to have 500,000 community health workers using SMS applications to link patients with doctors.

If he hits those numbers, ubiquity really will be the killer app. 

5 Ways Your SmartPhone Can Diagnose You >>>

Read more: How Text Messages Could Be the Future of Healthcare – Popular Mechanics

Related articles

• How Your Discarded Phone Can Improve Global Health (newser.com)
• BellVoz Launches a New Communication Service, Direct SMS, Allows Customers to Send International Text Messages from their Mobile Phone, at a Lower Cost (prweb.com)
• Mobile Medicine (andrewsullivan.thedailybeast.com)
• How Mobile Phones Are Saving Lives in the Developing World (mashable.com)

Point-of-Care Diagnostics for the Developing World

This hour long video talk is from the University of Washington.

 

(The Molecular Medicine Program at the University of Washington also provides additional webcasts from their public lectures, seminars and special courses at http://depts.washington.edu/molmed/webcasts/index.html)

From the Web site

People living in the developing world suffer greatly from many illnesses, many of them caused by infectious agents. These people usually do not have access to stable power or clean water, let alone the best diagnostic tools. What can we do to bring the high-tech diagnostic methods used in the developed world to those with fewer resources? Dr. Paul Yager, Professor and Acting Chair in the Department of Bioengineering, explains how microfluidics, a new technology for manipulating small volumes of fluids, is enabling the development of a small portable and inexpensive system for detecting pathogens far from the centralized laboratory. This system could soon have an impact on global health.

Related articles

• A Quick, Cheap Diagnostic Test for HIV and Other Infections (nextbigfuture.com)
• Diagnostic Development Unit Is A Star Trek-Style Sick Bay Bed (psfk.com)

More Being Prescribed Psychiatric Medications With No Diagnosis

From the 4 August Medical News Today article

59.5% of antidepressant prescriptions were made with no diagnosis in 1996, in 2007 the figure rose to 72.7%, researchers reported in Health Affairs. Antidepressant drugs are today the third most commonly prescribed class of drugs in the USA.

Nearly 8.9% of the American population had at least one antidepressant prescription during any given month during the period 2005-2008.

A good proportion of this growth in antidepressant prescription has been by non-specialist providers whose patients were not diagnosed by a psychiatrist.

Read the entire article

Related articles

• Prescriptions for antidepressants increasing among individuals with no psychiatric diagnosis (medicalxpress.com)
• More Antidepressants Prescribed by Non-Psychiatrists (psychcentral.com)
• Antidepressants Overprescribed in Primary Care (psychcentral.com)

Is Preventive Medicine Really Overtreatment?

Is Preventive Medicine Really Overtreatment?

From the NPR Health News item

In Overdiagnosed: Making People Sick in the Pursuit of Health, Dr. H. Gilbert Welch argues that modern medicine is looking too closely for disease, and that unnecessary screenings, MRIs and CT scans turn healthy people into diseased patients, by revealing often harmless abnormalities….

….

But overtreatment isn’t just a problem for patients diagnosed with cancer. It could sometimes be a problem for healthy people, as my next guest writes in his book “Overdiagnosed: Making People Sick in the Pursuit of Health,” because even healthy people are subject to more and more tests every time they visit the doctor.

Think about it, what do you do? You get the normal tests. You get your cholesterol level, maybe your liver test if you’re doing statins, you have a PSA, you have a body scan, tests that are often they often result in treatment. And because the traditional dogma is, as my next guest writes, more early diagnosis means better medical care, which means more treatment; and more treatment means better health.

But is that traditional view true? Is it accurate? Should we still be thinking about it that way? Are all these tests and treatments actually improving our health or are we looking too hard for disease?….

…

FLATOW: Why is it because doctors can do all these diagnoses, all these tests that they do, do them?

Dr. WELCH: Well, certainly, part of it is what’s possible, and what’s possible is, of course, changed dramatically over the last year. But it’s also part of our ethos, if you will, that it’s always a good thing to look for early forms of disease. And, of course, that message just been sent out to the public through the media and other sources that, of course, the thing you want to do is look for early forms of disease.

But the truth is there are really two sides to the story. I think patients are used to thinking of treatments as having side effects, but so does testing. And the side effect of looking for early forms of disease is that we find, virtually, all of us have some. That’s because we all harbor some abnormalities. And we never know which patients are those that have abnormalities that are going to cause problems in the future. So we tend to treat everybody we find with an abnormality and that means we’re just treating some patients who can’t benefit from our treatment because they were never going to develop the problem at hand if they’re overdiagnosed.

FLATOW: But how do you say to the person, you know, that maybe in the minority, as you say, that you may have saved that person’s life by overdiagnosing them? Is that worth of maybe one in a hundred cases?

Dr. WELCH: Well, I think that’s the question we all need to face. And, you know, sort of, traditionally, doctors have focused on the one out of a thousand we might help by looking for early forms of disease. But we haven’t really asked the question, what happens to the other 999? And this problem was really demonstrated to us in prostate cancer screening, which is really a poster child for the problem of overdiagnosis.

20 years ago, a simple blood test was introduced. And 20 years later, over one million Americans have been treated for a cancer that was never going to bother them. That test was the PSA, or prostate specific antigen. And it turned out an awful lot of men had abnormal PSAs. Many were found to have microscopic cancers far more than whatever suffer from prostate cancer.

Now, you might say, does it matter? Yeah, sure it matters because most of these men were treated with either radical surgery or radiation. And roughly a third suffered side effects of treatment generally related to bowel, bladder or sexual function. Even a few have died from it.

So this is a problem. It’s a matter of finding the balance between the question of just how hard we should be looking for problems in well patients……

Related Articles

• Prostate cancer from a patient perspective (kevinmd.com)
• Health Blog Q&A: H. Gilbert Welch, Author of ‘Overdiagnosed’ (blogs.wsj.com)
• everybody has cancer (breastcancerblog.wordpress.com)
• Cancer breakthrough — or nightmare? (cnn.com)
• Overdiagnosed: Making People Sick in the Pursuit of Health (beaconbroadside.com)
• Prostate Cancer Treatment: Hormone Therapy and Side Effects (everydayhealth.com)
• Watchful Waiting and Prostate Cancer (everydayhealth.com)

Microsponges from seaweed may save lives

Microsponges from seaweed may save lives

Rice University scientists refine process at heart of diagnostic bio-nano-chip

From the February 9 2011 Eureka news alert

Microsponges derived from seaweed may help diagnose heart disease, cancers, HIV and other diseases quickly and at far lower cost than current clinical methods. The microsponges are an essential component of Rice University’s Programmable Bio-Nano-Chip (PBNC) and the focus of a new paper in the journal Small.

The paper by John McDevitt, the Brown-Wiess Professor in Bioengineering and Chemistry, and his colleagues at Rice’s BioScience Research Collaborative views the inner workings of PBNCs, which McDevitt envisions as a mainstream medical diagnostic tool.

PBNCs to diagnose a variety of diseases are currently the focus of six human clinical trials. McDevitt will discuss their development at the annual meeting of the American Association for the Advancement of Science (AAAS) in Washington, D.C., Feb. 17-21.

PBNCs capture biomarkers — molecules that offer information about a person’s health — found in blood, saliva and other bodily fluids. The biomarkers are sequestered in tiny sponges set into an array of inverted pyramid-shaped funnels in the microprocessor heart of the credit card-sized PBNC.

When a fluid sample is put into the disposable device, microfluidic channels direct it to the sponges, which are infused with antibodies that detect and capture specific biomarkers. Once captured, they can be analyzed within minutes with a sophisticated microscope and computer built into a portable, toaster-sized reader.

The biomarker capture process is the subject of the Small paper. The microsponges are 280-micrometer beads of agarose, a cheap, common, lab-friendly material derived from seaweed and often used as a matrix for growing live cells or capturing proteins.

The beauty of agarose is its ability to capture a wide range of targets from relatively huge protein biomarkers to tiny drug metabolites. In the lab, agarose starts as a powder, like Jell-O. When mixed with hot water, it can be formed into gels or solids of any size. The size of the pores and channels in agarose can be tuned down to the nanoscale.

The challenge, McDevitt said, was defining a new concept to quickly and efficiently capture and detect biomarkers within a microfluidic circuit. The solution developed at Rice is a network of microsponges with tailored pore sizes and nano-nets of agarose fibers. The sponge-like quality allows a lot of fluid to be processed quickly, while the nano-net provides a huge surface area that can be used to generate optical signals 1,000 times greater than conventional refrigerator-sized devices. The mini-sensor ensembles, he said, pack maximum punch.

The team found that agarose beads with a diameter of about 280 micrometers are ideal for real-world applications and can be mass-produced in a cost-effective way. These agarose beads retain their efficiency at capturing biomarkers, are easy to handle and don’t require specialized optics to see.

McDevitt and his colleagues tested beads with pores up to 620 nanometers and down to 45 nanometers wide. (A sheet of paper is about 100,000 nanometers thick.) Pores near 140 nanometers proved best at letting proteins infuse the beads’ internal nano-nets quickly, a characteristic that enables PBNCs to test for disease in less than 15 minutes…….

###

Co-authors of the paper included first author Jesse Jokerst, a National Institutes of Health postdoctoral fellow at Stanford University; postdoctoral students James Camp, Jorge Wong, Alexis Lennart, Amanda Pollard and Yanjie Zhou, all of the departments of Chemistry and Biochemistry at the University of Texas at Austin; Mehnaaz Ali, an assistant professor of chemistry at Xavier University; and from the McDevitt Lab at Rice, Pierre Floriano, director of microfluidics and image and data analysis; Nicolaos Christodoulides, director of assay development; research scientist Glennon Simmons and graduate student Jie Chou.

The National Institutes of Health, through the National Institute of Dental and Craniofacial Research, funded the research.

Read the abstract at http://onlinelibrary.wiley.com/doi/10.1002/smll.201002089/abstract***

*** For suggestions on how to get this article for free or at low cost, click here