Technology

Not long ago, Stanford computer scientist Debashis Sahoo, PhD, told investigators at the Stanford Institute for Stem Cell Biology and Regenerative Medicine that in a few seconds he could find many of the important stem cell genes that the researchers were used to finding only after spending millions of dollars and years in the lab. “We laughed and said, ‘That’s impossible,’” recalls Irving Weissman, MD, director of the institute, in a recent video. But Weissman went ahead and gave Sahoo information about two key genes - and within a few seconds, Sahoo had used his desktop computer to scour the world’s public gene databases, analyzed that information with the computer algorithm he had designed, and come up with over a dozen genes new genes that were involved in the development of certain kinds of cells. That search, Weissman estimates, saved his team a decade of work and about $2.5 million.

More details are shared in the video above. And as a reminder, big data - and the ways in which people like Sahoo are mining through vast amounts of publicly available information to further research and advance health care - is the focus of a Stanford/Oxford conference being held here later this month.

Previously: Atul Butte discusses why big data is a big deal in biomedicine and Mathematical technique used to identify bladder cancer marker

Daily text messages may be an effective option to help children with asthma manage their symptoms and reduce doctor visits, according to recent research from the Georgia Institute of Technology.

In the study (.pdf), pediatric patients with asthma were randomly assigned to three programs: one group received text messages on alternate days, another received text messages daily and a third served as the control and did not receive any text messages. Participants ranged in age from 10 to 17 years old, owned a mobile phone and could read at the fifth grade level. The text messages asked patients questions about their symptoms and provided health information about asthma. Futurity reports:

Over four months, the intervention groups received and responded to SMS messages 87 percent of the time, and the average response time was within 22 minutes. After the study, the research team analyzed patients who had follow-up visits with their physician and found that sending at least one text message a day, whether it was a question about symptoms or about asthma in general, improved clinical outcomes.

“The results indicate that both awareness and knowledge are crucial to individuals engaging in proactive behavior to improve their condition,” [said Rosa Arriaga, PhD, who led the study].

The findings are noteworthy in light of past data showing texting is teenagers’ preferred method of communication, they get an average of 3,339 texts a month, and previous research showing they are amenable to receiving health information via text message.

Previously: CDC explores potential of using smartphones to collect public health data, Promoting healthy decisions among teens via text and Craving a cigarette but trying to quit? A supportive text message might help
Photo by Summer Skyes 11

Society is increasingly becoming more data-driven. Noting the power of vast reservoirs of public information, the federal government launched the Big Data Research and Development Initiative — a $200 million commitment to “greatly improve the tools and techniques needed to access, organize and glean discoveries from huge volumes of digital data.” And the National Institutes of Health expanded its stake in the federal initiative in hopes of speeding up the translation of biomedical discoveries into bedside applications.

In an effort to bring together innovative thinkers from information-technology corporations, startups, venture-capital firms and academia to capitalize on the wealth of opportunities using data-mining in biomedicine, Stanford Medicine and Oxford University are sponsoring a three-day conference from May 22-24. Curious to know more about the event and promise of big data, I reached out to Atul Butte, MD, PhD, Stanford systems-medicine chief and the conference’s scientific program committee chair. Below he shares why he’s passionate about how data-mining can transform scientific research and health care and discusses the conference program.

A recent Stanford Medicine article called data-mining the “fastest, least costly, most effective path to improving people’s health” that you know. Can you explain why you believe this to be the case?

Data-driven science, or data-mining, works faster and effectively because we are already sitting on billions of measurements made across the health system! Every time a physician orders a medication, every time a nurse or pharmacist dispenses a drug, every time a blood test is performed, every x-ray or CT scan that’s performed… all of this information ends up in a database today. So the part of science or innovation that involves collecting the measurements is actually the easiest part now, because the measurements are already there, just waiting for the right question to be asked.

In the same article, you said “hiding within [existing] mounds of data is knowledge that could change the life of a patient, or change the world” - and that if you didn’t analyze those data or show others how to, you feared no one will. How did you grow so passionate about this area?

I think we in the biomedical field make these measurements, but we often don’t realize how these measurements can interrelate or be used together. Our example from one of our recent articles was on our use of two big sets of public data. One set covered the molecular changes seen in tissues affected by diseases, and another set covered the molecular changes seen in cells treated by drugs. We realized that we could partner just these two public data sets together, to get new ideas of what other diseases might be treatable by these drugs. And, we could do this in a purely computational approach – an approach that is nearly infinitely scalable to more diseases, more investigators and more ideas. When I see hard working investigators working tirelessly to make highly accurate and significant measurements, but so few people taking advantage of that data, I can’t help but be passionate!

Earlier this year, you published a study, which involved combing through large amounts of data, to find that beta carotene may protect people with a common genetic risk factor for type-2 diabetes. Can you describe other recent findings that have stemmed from researchers’ use of this “big data” approach?

Stanford professor Russ Altman, MD, PhD, and his team recently showed how search engine logs can be mined to discover side effect of release drugs that might not have shown up during the initial clinical trials on those drugs. Similarly, Nigam Shah, MBBS, PhD, assistant professor of medicine, showed how similar side effects for drugs are sitting in physician clinical notes. Both text-based clinical notes and search engine logs are massive sources of big data that to date have barely been tapped for medical research.

What was the catalyst for launching the Big Data in Biomedicine conference?

The Li Ka Shing Foundation has played the leading role in bringing us together with Oxford University in planning events on big data. Our first, smaller conference was held in Oxford last November. Based on the success of that event, we realized we could host a larger conference at Stanford and open it up to the public. We couldn’t have done this without the support of the Li Ka Shing Foundation.

Continue Reading »

The number of gene-expression data sets available in public databases has climbed rapidly over the past decade, allowing researchers to spot disease trends without doing time-intensive experiments in the laboratory. The “big data” deluge promises to accelerate the process of understanding disease while driving down the costs of developing new therapies.

To underscore the wealth of opportunities for scientists who can mine these continuously growing databases in innovative ways, Stanford Medicine and Oxford University are sponsoring a three-day conference next month on big data’s role in biomedicine. The event will be held May 22-24 at the School of Medicine’s Li Ka Shing Center for Learning & Knowledge and will feature keynote speeches from Anne Wojcicki, CEO and co-founder of the consumer-genomics company 23andMe, and David Ewing Duncan, author of Experimental Man. In addition, attendees will hear from more than two dozen speakers representing large information-technology corporations, startups, venture-capital firms and academia.

In a release, Stanford systems-medicine chief Atul Butte, MD, PhD, who is the conference’s scientific program committee chair, commented on the motivation for hosting the conference and what participants will learn at the event:

We’re bringing together people from academia, industry, government and foundations who want to learn more about how big data can drive innovation for a healthier world… We expect that attendees will walk away from this with a strong understanding of the latest tools and technologies available for studying and using big data in biomedicine, of where the unmet medical needs are and how they can be addressed with these approaches, and of what the tractable next steps are that they can take to become innovators.

Additional program information and registration details are available on the conference website.

Previously: Mining data from patients’ charts to identify harmful drug reactions, Strength in numbers: Harnessing public gene data to answer a diverse range of research questions, Mining medical discoveries from a mountain of ones and zeroes, The data deluge: A report from Stanford Medicine magazine, Stanford’s Atul Butte discusses outsourcing research online at TEDMED and Health-care experts discuss opportunities and challenges of mining ‘big data’ in health care
Photo by Dwight Eschliman

Recognizing the value of mobile devices in conducting public health research, the Centers for Disease Control and Prevention has launched a project to examine the feasibility of collecting data using smartphone-based surveys and text messages. A recent post on Mobihealthnews offers more details:

The groups aim to first send the surveys to US residents aged 18 to 65 nationwide with questions related to smoking habits and alcohol consumption. Following the survey, the smartphone users will be asked to participate via text message in the feasibility study, which includes a survey immediately following the first and then another one a week later. The texts will include links to the survey on a mobile-friendly site.

Some of those who participate in the initial outreach that are non-smartphone users will be asked to participate in another study, the text message pilot, which will conduct the surveys one question at a time via text.

The study aims to evaluate, among other things, the response bias of data collected from the smartphone users on the mobile site to those responses collected via text.

Previously: Survey shows more than a quarter of American adults are mobile health users and CDC binge-drinking study demonstrates cell phones’ value in research
Photo by Jhaymesisviphotography

My daughters spent their first few days of life in the neonatal intensive care unit, and I won’t soon forget padding down the long hospital hallways, decked out in my flimsy gown and fluffy blue slippers, every two hours to go visit and feed them. As emotional as this time was for me and my husband, I recognize it would have been even more so if I wasn’t able to see my baby - which is why I think a new program at a Los Angeles hospital is so cool. Called BabyTime, the Cedars-Sinai program uses iPads to connect parents with their premature or ill newborns.

readwrite’s Brian S. Hall reported yesterday:

Mothers who are confined to recovery rooms following delivery, typically because of a cesarean section or other complications, often can’t see their newborns in the intensive care unit for 2-3 days. “With BabyTime, the new mother can now see their baby in about 2-3 hours,” Yvonne Kidder, a clinical nurse in the hospital’s Neonatal Intensive Care Unit (NICU), told me:

“BabyTime’s been wonderful. For mothers, to see their baby, this absolutely lessens their anxiety. For the fathers, who can become overwhelmed with all the information they are receiving, BabyTime bridges the gap and allows for a direct line between mother and caregivers.”

Previously: The emotional struggles of parents of preemies

A past report from the Pew Internet & American Life Project shows that older adults have enthusiastically embraced social media tools. Now comes new research indicating that social media and online communities can provide valuable support for elderly patients in managing their health. Consumer Affairs reports:

“For me, it was interesting to learn that there is evidence for a large potential of social media in clinical practices,” said [Dr. Anja Leist of the University of Luxembourg]. “Older adults can use social media to access health-related information and engage in patient-to-patient-doctor conversations. There are many online forums where people in difficult life situations, such as informal caregivers of a spouse with dementia or individuals with depression, can exchange thoughts as well as receive and provide social support.

“Other positive consequences are that lonely adults can overcome loneliness through contact to family and friends and other users with similar interest,” Leist said.

However, researchers cautioned that several challenges need to be addressed before social media can be used in a clinical setting to help manage patients treatment, such as protecting personal health information and assisting seniors in identifying accurate online sources for medical information.

Previously: Study shows Internet can help raise awareness about cancer prevention, A look at social-media use among psoriasis patients and Patient online peer group offers community, drives research

Frustrated by inconclusive tests, strange symptoms and a lack of answers from their health-care providers, some patients have turned to the online community for answers about perplexing illnesses. And a new web-based tool, called CrowdMed, aims to make it even easier to diagnose medical mysteries. The New Scientist reports:

Anyone can join CrowdMed and analyze cases, regardless of their background or training. Participants are given points that they can then use to bet on the correct diagnosis from lists of suggestions. This creates a prediction market, with diagnoses falling and rising in value based on their popularity, like stocks in a stock market. Algorithms then calculate the probability that each diagnosis will be correct.

In 20 initial test cases, around 700 participants identified each of the mystery diseases as one of their top three suggestions.

…

The goal is to help people who come down with any of around 7000 “rare diseases” as defined by health agencies in Europe and the US. In Europe alone, 30 million people have a rare disease, 40 per cent of whom either go undiagnosed or are misdiagnosed at some point.

As the popularity of using the Internet to answer health questions grows, it will be interesting to see how services such as CrowdMed, and search engines like FindZebra, even further redefine the doctor-patient relationship.

Previously: The importance of curation and communities when crowdsourcing clinical questions, New search engine designed to help physicians and the public in diagnosing rare diseases, Report shows 35 percent of U.S. adults turn to the Internet to diagnose a medical condition, Dr. Google: Threat or menace? and Patient self-diagnosis: From the browser to the exam room
Photo by Ryan Brooks

Health-care providers know there’s a wealth of valuable information trapped in the hand-written notes on patients’ charts. But the challenge of collecting and interpreting the data on a large scale remains to be solved. Now researchers at Stanford have taken a step forward in mining patient-based information by using existing language-analysis methods to identify drug side effects in advance of the Food and Drug Administration issuing official alerts.

My colleague writes in a release:

Although their application is new, their information-gathering methods are based on well-established text processing techniques. It’s also simpler and faster than current strategies used in the same arena, said [engineering research associate Paea LePendu, PhD, the lead author of the paper]. Content is first grouped via “ontologies,” which are information graphs organized by associative relationships instead of a rigid linear structure. For example, melanoma is a kind of skin cancer, and so is Kaposi’s sarcoma; by knowing “skin cancer” encompasses both kinds of cancer, the search process picks up this medical knowledge. The system also de-identifies patient information in the process, so sensitive data, such as names and addresses, doesn’t get revealed. With these methods, LePendu said, the technique allows them to process 11 million clinical notes in about seven hours on hardware no different from a laptop computer — a pace that other programs can’t match.

The information is also current: It’s generated from what is observed and recorded in the hospital or doctor’s office. That’s an advantage over the FDA’s AERS reports, which rely on patients and health providers to make the additional effort to report adverse events.

The researchers developed the computerized method to sift through the contents of clinical notes in electronic medical records and used it to examine how often specific drugs and diseases were mentioned in roughly 10 million notes for about 1.8 million patients over 15 years. The goal was to organize these notes into a data-mining substrate they refer to as a patient-feature matrix. “Everyone is excited about the prospect of ‘big data’ mining on electronic health record data,” Shah said. “We demonstrate it in practice.”

Previously: Researchers mine Internet search data to identify unreported side effects of drugs
Photo by The National Guard

Earlier today I wrote about a breakthrough method called CLARITY, pioneered by Stanford psychiatrist/bioengineer Karl Deisseroth, MD, PhD, for rendering intact tissue samples transparent. Above is a video clip showing off the new method’s capabilities. First you’ll witness a “fly-through” of a complete mouse brain using fluorescent imaging. The immediately following clip - it’s spectacular! - provides a three-dimensional view of a mouse hippocampus (the brain’s brain’s memory hub), with projecting neurons depicted in green, connecting interneurons in red, and layers of support cells, or glia, in blue.

Note that in both cases, there was no need to slice the tissue into ultra-thin sections, analyze them chemically and/or optically and then laboriously “sew” them back together via computer algorithms in order to reconstruct a 3-D virtual image of the biological sample. All that was required, after performing the necessary hocus-pocus, was to ”send in the stain” (i.e., use histochemical means to paint different cell types different colors) and move the sample or camera lens or shift the latter’s focal length. Nice trick. With big implications for biomedical research.

Previously: Lightning strikes twice: Optogenetics pioneer Karl Deisseroth’s newest technique renders tissues transparent, yet structurally intact, Visualizing the brain as a Universe of synapses and A federal push to further brain research