Friday round-up
Federal investigators have released a new report stating that only about 1 in every 7 hospital errors are reported. Adverse events ranging from infections to excessive bleeding to even death are supposed to be reported through systems present at almost every hospital in the US. The systems often allow for anonymous reporting, in order to encourage hospital staff to cooperate. However, as it states in the article: "organizations that inspect and accredit hospitals generally "do not scrutinize" how hospitals keep track of medical errors and other adverse events". And if the accrediting bodies do not scrutinize the process (and results thereof), there isn't much incentive for staff to report. It also states in the article that no new federal regulations regarding this are expected.
The National Science Foundation has released the 14 winners of the "Digging into Data" challenge. The 14 winning projects all involve innovative ways to use data analysis and natural language processing (NLP) to enhance research in the humanities and social sciences. Those interested in large-scale data mining and investigation should read through the winning projects, as they all sound extremely interesting. I think my favorite might be the analysis of newspaper reporting on the 1918 flu pandemic (and not just because most of the PIs are from my alma mater!), in order to see how such reports affected public opinion and the idea of "authority" during the outbreak. I will be eagerly anticipating their results. The 14 winning research projects are sharing nearly $5 million in funds.
Dr. Matheson Harris has written a brief and clear tutorial for patients (with some help from the Chicago Tribune) on how to spot a good doctor (and a bad one!) and how to be a good, educated patient. I really like a lot of what is said here, and agree with the vast majority of it; though it might be a little harsh to tell patients not to go see a doctor who can't see them within a few days. I think it greatly depends on the type of doctor you're seeing, and what the appointment for. I make my dermatology appointments a year in advance because it is so hard to get an appointment at the practice (widely considered one of the best in the nation). And many women in the state of Pennsylvania can tell you about the difficulties in getting an OB-Gyn appointment due to the shortage of those specialists. But the gist of the advice and guidance here is very strong, and all patients should read and take it to heart.
Updates on Dr. Watson
Dr. Watson, I presume?
It looks like Watson-esque computers could begin to be used in nations like India (with a rapidly-growing population and a dearth of doctors and healthcare providers to serve it) within 1-2 years. Seems like the perfect place to start: both in affluent areas of the country where people happily take on new technologies (at times even quicker than in the West), or in poorer areas where medical assistance is desperately needed.
I particularly like this explanation, by Dr. Manish Gupta, of IBM:
“A doctor is essentially doing something analogous to what a person does on a quiz show: questions that are asked on the quiz show are parallel to a patient coming in and describing his problem. There may be medical test reports, descriptions of the problem and the current state. In confusing or ambiguous situations, doctors rely on literature they had read in school or current medical developments and their own experience before coming out with one or more hypotheses (of what the disease could be).”
Similar to the way physicians have adopted lab testing equipment, smart phone apps, and Up to Date, for example; these "Watson, MDs" will soon be utilized as tools of the healthcare trade.
And at the University of Maryland School of Medicine in Baltimore, steps are being taken to this end for the US market as well. Speech recognition programs are being integrated with the question-answering and data-mining capabilities of Watson to develop an "examining room" or "bedside" version. Early adopters may be able to start to utilize the tool in 2013.
Acronym Soup: NLP and CLU in EHRs
This article from KevinMD.com has taught me a new acronym: CLU, or Clinical Language Understanding. CLU is a way to use Natural Language Processing (NLP) in the medical world. In other words, it's like a "medical encyclopedia" inside an NLP system:
"CLU works off of a complete, highly granular medical ontology, which has been tuned to relate and identify all kinds of medical facts so that the underlying NLP engine can “understand” what the caregiver is saying. For example, CLU knows that “cancer” is a “disease” and would auto-populate the EHR with that information. CLU knows that “amoxicillin” is an “antibiotic;” this knowledge is a direct impact of the ontology. .. Clinical Language Understanding allows doctors to be efficient with documentation, helps to ensure patients’ medical records are comprehensive and are not reduced purely structured content created by point-and-click templates..."
I'm in the process of creating a special interest group related to NLP here at Resilient Ambassadors HQ, and I think this might be the focus of one of our first conversations.
Slides
A few notes to the viewer:
- Please feel free to use and share these slides. If you do use them, give us some "props" via citation or attribution. It doesn't have to be fancy. We appreciate it!
- These slides and the content within them only represent the thoughts of the listed author, and do not reflect the beliefs of any overarching organization or entity. Or, if they do, it we didn't do it on purpose. Boy, this is getting complicated.
- Just one more thing: While each of the ambassadors certainly strives to fact-check our research and ensure the accuracy of the information in our slides, none of us is claiming to be an expert. These slides are as much a documentation of our learning as they are (hopefully) informative for others.
Have at it!
