healthcare technology

The big hairy audacious goal of most every data scientist I know in healthcare is what you might call the Integrated Medical Record, or IMR, a dataset that combines detailed genetic data and rich phenotypic information, including both clinical and “real-world” (or, perhaps, “dynamic”) phenotypic data (the sort you might get from wearables).

The gold standard for clinical phenotyping are academic clinical studies (like ALLHAT and the Dallas Heart Study).  These studies are typically focused on a disease category (e.g. cardiovascular), and the clinical phenotyping on these subjects – at least around the areas of scientific interest — is generally superb.  The studies themselves can be enormous, are often multi-institutional, and typically create a database that’s independent of the hospital’s medical record.

Duke University Medicine is using geographical information to turn electronic health records (EHRs) into population health predictors. By integrating its EHR data with its geographic information system, Duke can enable clinicians to predict patients' diagnoses.

According to Health Data Management, Sohayla Pruitt was hired by Duke to run this project; “I thought, wow, if we could automate some of this, pre select some of the data, preprocess a lot and then sort of wait for an event to happen, we could pass it through our models, let them plow through thousands of geospatial variables and [let the system] tell us the actual statistical significance,” Pruitt says. “Then, once you know how geography is influencing events and what they have in common, you can project that to other places where you should be paying attention because they have similar probability.”

iHealth Beat explains that the system works by using an automated geocoding system to verify addresses with a U.S. Postal Service database. These addresses are then passed through a commercial mapping database to geocode them. Finally, the system imports all U.S. Census Bureau data with a block group ID. This results in an assessment of socioeconomic indicators for each group of patients.

“When we visually map a population and a health issue, we want to give an understanding about why something is happening in a neighborhood,” says Pruitt. “Are there certain socioeconomic factors that are contributing? Do they not have access to certain things? Do they have too much access to certain things like fast food restaurants?”

Duke is working to develop a proof of concept and algorithms that would map locations and patients. They are also working on a system to track food-borne illnesses.

After decades as a technological laggard, medicine has entered its data age. Mobile technologies, sensors, genome sequencing, and advances in analytic software now make it possible to capture vast amounts of information about our individual makeup and the environment around us. The sum of this information could transform medicine, turning a field aimed at treating the average patient into one that’s customized to each person while shifting more control and responsibility from doctors to patients.

The question is: can big data make health care better?

“There is a lot of data being gathered. That’s not enough,” says Ed Martin, interim director of the Information Services Unit at the University of California San Francisco School of Medicine. “It’s really about coming up with applications that make data actionable.”

The business opportunity in making sense of that data—potentially $300 billion to $450 billion a year, according to consultants McKinsey & Company—is driving well-established companies like Apple, Qualcomm, and IBM to invest in technologies from data-capturing smartphone apps to billion-dollar analytical systems. It’s feeding the rising enthusiasm for startups as well.

Venture capital firms like Greylock Partners and Kleiner Perkins Caufield & Byers, as well as the corporate venture funds of Google, Samsung, Merck, and others, have invested more than $3 billion in health-care information technology since the beginning of 2013—a rapid acceleration from previous years, according to data from Mercom Capital Group. 

Background: To date, health research literature has focused on social network sites (SNS) either as tools to deliver health care, to study the effect of these networks on behavior, or to analyze Web health content. Less is known about the effectiveness of these sites as a method for collecting data for health research and the means to use such powerful tools in health research.

Objective: The objective of this study was to systematically review the available literature and explore the use of SNS as a mode of collecting data for health research. The review aims to answer four questions: Does health research employ SNS as method for collecting data? Is data quality affected by the mode of data collection? What types of participants were reached by SNS? What are the strengths and limitations of SNS?

Results: The inclusion criteria were met by 10 studies and results were analyzed descriptively to answer the review questions. There were four main results.

(1) SNS have been used as a data collection tool by health researchers; all but 1 of the included studies were cross-sectional and quantitative.

(2) Data quality indicators that were reported include response rate, cost, timeliness, missing data/completion rate, and validity. However, comparison was carried out only for response rate and cost as it was unclear how other reported indicators were measured.

(3)The most targeted population were females and younger people.

(4) All studies stated that SNS is an effective recruitment method but that it may introduce a sampling bias.

Conclusions: SNS has a role in health research, but we need to ascertain how to use it effectively without affecting the quality of research. The field of SNS is growing rapidly, and it is necessary to take advantage of the strengths of this tool and to avoid its limitations by effective research design. This review provides an important insight for scholars who plan to conduct research using SNS.

more at http://www.jmir.org/2014/7/e171/

Information that may offer medical insights has been locked away in the filing cabinets of doctors' offices.

Researchers at IBM, Berg Pharma, Memorial Sloan Kettering, UC Berkeley and other institutions are exploring how artificial intelligence and big data can be used to develop better treatments for diseases 

But one of the biggest challenges for making full use of these computational tools in medicine is that vast amounts of data have been locked away — or never digitized in the first place.

The results of earlier research efforts or the experiences of individual patients are often trapped in the archives of pharmaceutical companies or the paper filing cabinets of doctors’ offices.

Patient privacy issues, competitive interests and the sheer lack of electronic records have prevented information sharing that could potentially reveal broader patterns in what appeared to any single doctor like an isolated incident.

When you can analyze clinical trials, genomic data and electronic medical records for 100,000 patients, “you see patterns that you don’t notice in a couple,” said Michael Keiser, an instructor at the UC San Francisco School of Medicine.

Given that promise, a number of organizations are beginning to pull together medical data sources.

more at http://recode.net/2014/06/07/medicines-big-problem-with-big-data-information-hoarding/

What does 2014 hold? According to Eric Schmidt, Google's executive chairman, it means smartphones everywhere - and also the possibility of genetics data being used to develop new cures for cancer.

Schmidt says there's a big change - a disruption - coming for business through the arrival of "big data": "The biggest disruptor that we're sure about is the arrival of big data and machine intelligence everywhere - so the ability [for businesses] to find people, to talk specifically to them, to judge them, to rank what they're doing, to decide what to do with your products, changes every business globally."

But he also sees potential in the field of genomics - the parsing of all the data being collected from DNA and gene sequencing. That might not be surprising, given that Google is an investor in 23andme, a gene sequencing company which aims to collect the genomes of a million people so that it can do data-matching analysis on their DNA.

(Unfortunately, that plan has hit a snag: 23andme has been told to cease operating by the US Food and Drug Administration because it has failed to respond to inquiries about its testing methods and publication of results.)

Here's what Schmidt has to say on genomics: "The biggest disruption that we don't really know what's going to happen is probably in the genetics area. The ability to have personal genetics records and the ability to start gathering all of the gene sequencing into places will yield discoveries in cancer treatment and diagnostics over the next year that that are unfathomably important."

It may be worth mentioning that "we'll find cures through genomics" has been the promise held up by scientists every year since the human genome was first sequenced.

So far, it hasn't happened - as much as anything because human gene variation is remarkably big, and there's still a lot that isn't known about the interaction of what appears to be non-functional parts of our DNA (which doesn't seem to code to produce proteins) and the parts that do code for proteins.

Researchers at Vanderbilt University Medical Center have used natural language processing technology in an electronic medical records system to identify patients with multiple sclerosis and collect data on traits of their disease course.

The work is significant, researchers say, because much remains unknown about the course of the disease, which varies widely among patients. “Most research studies have focused on the origin of the disease, partly because of the difficulty in ascertaining sufficient longitudinal clinical data to study the disease course,” according to the study published in the Journal of the American Medical Informatics Association.

“Electronic medical records may provide such a tool. We have previously shown that genomic signals of MS risk may be replicated using EMR-derived cohorts. In this paper, we evaluated algorithms to extract detailed clinical information for the disease course of MS.”

The study used algorithms based on ICD-9 codes, text keywords and medications to identify 5,789 patients with MS, and collected detailed data on the clinical course of the patients’ disease to measure progression of disability. “For all clinical traits extracted, precision was at least 87 percent and specificity was greater than 80 percent.”

Researchers uncover new ties between genetics and skin cancer by mining patients’ medical records.

Usually, studying the relationship between DNA and disease involves comparing the genomes of thousands of people with a disorder to the genomes of thousands of people who don’t. These studies can be expensive and may take years, requiring researchers to identify patients, enroll them in the study, and collect the genomic data.

The face of medical care is rapidly changing thanks to major advancements in the capture, proliferation, and analysis of medical data. Technologies like the electronic health records (EHRs) and personal health records (PHRs) are drastically improving the way data is aggregated and shared.

Now the hope is that big data analytics will help to make sense of seemingly endless streams of medical information.

As many doctors are painfully aware, outcome-oriented care is no longer a buzzword but a reality. The Center for Medicare and Medicaid Services has started to implement a program where payments are based on the ability of providers to meet key National Quality Strategy Domains (e.g. care criteria). Public payers are testing this new methodology, and private payers are expected to soon follow.

These big data analytics applications can also be relevant for the FDA, which may want to see how drugs perform in a non-test environment to ensure the appropriate patient populations are receiving the drug. I also expect pharmaceutical companies to actively scour this data to track drug efficacy post-release or identify markets that could “benefit” from increased penetration.

I am eager to see how the data evolution improves outcomes for doctors and patients.

When an imaging run generates 1 terabyte of data, analysis becomes the problem

Today's neuroscientists have some magnificent tools at their disposal. They can, for example, examine the entire brain of a live zebrafish larva and record the activation patterns of nearly all of its 100,000 neurons in a process that takes only 1.5 seconds.

The only problem: One such imaging run yields about 1 terabyte of data, making analysis the real bottleneck as researchers seek to understand the brain.

To address this issue, scientists at Janelia Farm Research Campus have come up with a set of analytical tools designed for neuroscience and built on a distributed computing platform called Apache Spark. In their paper in Nature Methods, they demonstrate their system's capabilities by making sense of several enormous data sets. (The image above shows the whole-brain neural activity of a zebrafish larva when it was exposed to a moving visual stimulus; the different colors indicate which neurons activated in response to a movement to the left or right.)

The researchers argue that the Apache Spark platform offers an improvement over a more popular distributed computing model known as Hadoop MapReduce, which was originally based on Google's search engine technology. 

more at http://spectrum.ieee.org/tech-talk/biomedical/imaging/can-computing-keep-up-with-the-neuroscience-data-deluge/

It's been tough to identify the problems that only turn up after medicines are on the market. An experimental project is now combing through data to get earlier, more accurate warnings.

No one likes it when a new drug in people's medicine cabinets turns out to have problems — just remember the Vioxx debacle a decade ago, when the painkiller was removed from the market over concerns that it increased the risk of heart attack and stroke.

To do a better job of spotting unforeseen risks and side effects, the Food and Drug Administration is trying something new — and there's a decent chance that it involves your medical records.

It's called Mini-Sentinel, and it's a $116 million government project to actively go out and look for adverse events linked to marketed drugs. This pilot program is able to mine huge databases of medical records for signs that drugs may be linked to problems.

The usual system for monitoring the safety of marketed drugs has real shortcomings. It largely relies on voluntary reports from doctors, pharmacists, and just plain folks who took a drug and got a bad outcome.

"We get about a million reports a year that way," says Janet Woodcock, the director of the FDA's Center for Drug Evaluation and Research. "But those are random. They are whatever people choose to send us."

Big data offers breakthrough possibilities for new research and discoveries, better patient care, and greater efficiency in health and health care, as detailed in the July issue of Health Affairs. As with any new tool or technique, there is a learning curve.

Big data analytics technology has been able to find patterns and pinpoint disease states more accurately than even the most highly-trained physicians.

The human brain may be nature’s finest computer, but artificial intelligences fed on big data are making a convincing challenge for the crown. In the realm of healthcare, natural language processing, associative intelligence, and machine learning are revolutionizing the way physicians make decisions and diagnose complex patients, significantly improving accuracy and catching deadly issues before symptoms even present themselves.

In this case study examining the impact of big data analytics on clinical decision making, Dr. Partho Sengupta, Director of Cardiac Ultrasound Research and Associate Professor of Medicine in Cardiology at the Mount Sinai Hospital, has used an associative memory engine from Saffron Technology to crunch enormous datasets for more accurate diagnoses.

Using 10,000 attributes collected from 90 metrics in six different locations of the heart, all produced by a single, one-second heartbeat, the analytics technology has been able to find patterns and pinpoint disease states more quickly and accurately than even the most highly-trained physicians.

more at http://healthitanalytics.com/2014/07/07/case-study-big-data-improves-cardiology-diagnoses-by-17/

Imagine going to the doctor with an infection and being sent home with a course of drugs. Unknown to your doctor you actually have two infections. If you take the drugs will the other infection go away by itself? What if you take the drugs and the other infection gets worse? This quandary faces those treating patients with multiple infections.

A new study led by former University of Sheffield PhD student Dr Emily Griffiths, in collaboration with the universities of Edinburgh, Liverpool and Zürich, has taken a novel approach to understanding this problem, shedding light on how multiple parasites interact within humans.

The study compiled a list of many of the parasites that infect humans, another list of the parts of the body consumed by each parasite, and also information about how the immune system responds to each parasite. This information was used to construct a large network of multiple infections in humans - a bit like a food web of infections inside the human body.

Building this network revealed some previously unknown patterns, something that could pave the way for new treatment strategies which help tackle multiple infections. For example, groups of parasites often share similar parts of their host, and these groups are prime candidates for coordinated treatment.

Dr Griffiths, who carried out the research during her PhD in the Department of Animal and Plant Sciences at the University of Sheffield, said: "After studying the fascinating range of hundreds of different infections that can occur in the same person at the same time, we've shown that we could better treat patients if we know what parasites are eating inside our bodies.

"Our web has revealed the ways hundreds of parasites could live together, which means that we can develop new coordinated treatments that help fight more than one infection.

Nathan E. Wineinger, Ph.D., a research scientist at Scripps Genomic Medicine discusses how mhealth data is moving towards big data at a rapid pace.

Mobile health (mHealth) technologies allow for the generation of intensive care unit medical information, literally, in the palm of your hand. A smart phone can be transformed into a mobile heart monitor to diagnose atrial fibrillation, and continuous glucose monitoring has revolutionized the way diabetics manage their blood sugar levels. The digitization of human health through noninvasive devices and sensors can provide meaningful measures of individual wellness outside of a clinical environment.

With the increasing burden of chronic diseases, analyzing and understanding trajectories of care is essential for efficient planning and fair allocation of resources: authors Nicolas Jay, Gilles Nuemi, Maryse Gadreau and Catherine Quantin propose an approach based on mining claim data to support the exploration of trajectories of care.