A new landmark in computer vision

Science fiction books and movies have long imagined that computers will someday be able to see and interpret the world. At Google, we think computer vision has tremendous potential benefits for consumers, which is why we're dedicated to research in this area. And today, a Google team is presenting a paper on landmark recognition (think: Statue of Liberty, Eiffel Tower) at the <a href="http://www.cvpr2009.org/">Computer Vision and Pattern Recognition (CVPR) conference</a> in Miami, Florida. In the paper, we present a new technology that enables computers to quickly and efficiently identify images of more than 50,000 landmarks from all over the world with 80% accuracy.<br /><br />To be clear up front, this is a research paper, not a new Google product, but we still think it's cool. For our demonstration, we begin with an unnamed, untagged picture of a landmark, enter its web address into the recognition engine, and <span class="Apple-style-span" style="font-style: italic;">poof</span> — the computer identifies and names it: "Recognized Landmark: Acropolis, Athens, Greece." Thanks computer.<br /><br />How did we do it? It wasn't easy. For starters, where do you find a good list of thousands of landmarks? Even if you have that list, where do you get the pictures to develop visual representations of the locations? And how do you pull that source material together in a coherent model that actually works, is fast, and can process an enormous corpus of data? Think about all the different photographs of the <a href="http://images.google.com/images?q=golden%20gate%20bridge">Golden Gate Bridge</a> you've seen — the different perspectives, lighting conditions and image qualities. Recognizing a landmark can be difficult for a human, let alone a computer.<br /><br />Our research builds on the vast number of images on the web, the ability to search those images, and advances in object recognition and clustering techniques. First, we generated a list of landmarks relying on two sources: 40 million GPS-tagged photos (from <a href="http://picasaweb.google.com/">Picasa</a> and <a href="http://www.panoramio.com/">Panoramio</a>) and online tour guide webpages. Next, we found candidate images for each landmark using these sources and <a href="http://images.google.com/">Google Image Search</a>, which we then "pruned" using efficient image matching and unsupervised clustering techniques. Finally, we developed a highly efficient indexing system for fast image recognition. The following image provides a visual representation of the resulting clustered recognition model:<br /><br /><a onblur="try {parent.deselectBloggerImageGracefully();} catch(e) {}" href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgfCuAeLMg9C8JTj9ko8qyuRzlzYsNWPiqPxHXI2h86xZVdWo5_71MvMkEwLSeSj2-pE6dn_Ie7qHZvkhyphenhyphenF7U0K6dXMIBdo5fweTVm8m0RhkZdbTOJr2nroDPk_7ixn-9JVdQZp08KN7QLR/s1600-h/Acropolis+Cluster.png"><img alt="" border="0" id="BLOGGER_PHOTO_ID_5350151754492326706" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgfCuAeLMg9C8JTj9ko8qyuRzlzYsNWPiqPxHXI2h86xZVdWo5_71MvMkEwLSeSj2-pE6dn_Ie7qHZvkhyphenhyphenF7U0K6dXMIBdo5fweTVm8m0RhkZdbTOJr2nroDPk_7ixn-9JVdQZp08KN7QLR/s400/Acropolis+Cluster.png" style="display:block; margin:0px auto 10px; text-align:center;cursor:pointer; cursor:hand;width: 370px; height: 400px;" /></a><br /><div style="text-align: center;"><span class="Apple-style-span" style=" font-style: italic; font-size:10px;">In the above image, related views of the Acropolis are "clustered" together, allowing for a more efficient image matching system.</span><br /></div><br />While we've gone a long way towards unlocking the information stored in text on the web, there's still much work to be done unlocking the information stored in pixels. This research demonstrates the feasibility of efficient computer vision techniques based on large, noisy datasets. We expect the insights we've gained will lay a useful foundation for future research in computer vision.<br /><br />If you're interested to learn more about this research, check out the <a href="http://www.google.com/googleblogs/pdfs/google_landmark_recognition.pdf">paper</a>.<br /><br /><span class="byline-author">Posted by Jay Yagnik, Head of Computer Vision Research</span>

A new landmark in computer vision

Science fiction books and movies have long imagined that computers will someday be able to see and interpret the world. At Google, we think ...

Đọc thêm »

Bungled Brachytherapy, Computer Interfaces and Other Mysteries At The Philadelphia Veterans Administration Hospital

The Philadelphia VA Hospital has had some problems with <a href="http://en.wikipedia.org/wiki/Brachytherapy">brachytherapy</a> (implantation of small radioactive pellets for prostate cancer) recently. Multiple procedures were bungled, including wrong placement and incorrect dosages.<br /><div><br /></div><div>In this story, there apparently is a computer involved. From the New York Times article "<a href="http://www.nytimes.com/2009/06/21/health/21radiation.html?pagewanted=1&_r=1">At VA Hospital, a Rogue Cancer Unit</a>", June 20, 2009:</div><div><br /></div><div><div color="#333333" style="font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;"></div><blockquote><div color="#333333" style="font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;">... The hospital suspended the brachytherapy program on June 11 last year. By then, 45 substandard implants had been found.</div><div style="color: #333333; font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;">Two days later, the Joint Commission, which helps set standards in the hospital industry, surveyed the Philadelphia V.A. and on the next day accredited the hospital. “This organization is in full compliance with applicable standards,” the Joint Commission said.</div><div style="color: #333333; font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;">The commission said that it had no indications of the problems in the brachytherapy program when it arrived at the hospital and that its surveys are not detailed enough to have uncovered the flawed implants.<b><i><span class="Apple-style-span" style="color: red;"> [Or, apparently, flawed computers - ed.]</span></i></b></div><div color="#333333" style="font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;">Soon after, the N.R.C. sent its own inspectors to Philadelphia. And the more the inspectors looked, the more they found. All told, 57 of the implants delivered too little radiation to the prostate, either because the seeds missed the prostate or were not distributed properly inside the prostate. Thirty-five other cases involved overdoses to other parts of the body. An unspecified number of patients were both underdosed in the prostate and overdosed elsewhere.</div><div style="color: #333333; font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;">From December 2006 to November 2007, the nuclear commission found, 16 patients received seed implants in Philadelphia <b>even though computer interface problems prevented medical personnel from determining whether those treatments had been successful.</b> The V.A.’s radiation officials knew of the problem but took no action, the nuclear commission charges.</div><div style="color: #333333; font: 15.0px Georgia; line-height: 22.0px; margin: 0.0px 0.0px 15.0px 0.0px;"><br /></div></blockquote></div>I am uncertain how "computer interface problems" (in the <a href="http://www.philly.com/philly/news/homepage/20090621_Feds_see_wider_woes_in_VA_s_cancer_errors.html">Philadelphia Inquirer</a>, they were referred to as "glitches") prevented medical personnel from determining treatment success over several years. I would be most interested in hearing more about these "interface problems."<br /><div><br /></div><div>Here are some questions:</div><div><br /></div><div><ul><li>Did these problems involve the VistA system?</li><li>What was interfaced to what, exactly?</li><li>Who did the interfacing?</li><li>What regulatory authorities validated the interfaces and systems? </li><li>If no regulatory agency was involved, why not?</li></ul></div><div><div><br /></div><div>I have some familiarity with odd events at the Philadelphia VA Hospital. (Not including the fact that I spent a few months there as a medical intern in the early 1980's). In the mid 1990's I took my father there for evaluation for increased service-related disability. He had been treated for skin lesions in the Army in WW2 and after by the VA with <a href="http://en.wikipedia.org/wiki/Fowler%27s_solution">Fowler's solution</a> (an arsenical) and as a consequence of this (even then-outdated and dangerous) treatment, had developed widespread basal cell carcinomatosis over a major portion of his body, with chronic bleeding and discomfort. I accompanied my father to the exam but did not identify myself as an MD, only as his son.</div><div><br /></div><div>My father was seen in an evaluation by several physicians and students (arsenic-caused basal cell carcinomatosis is quite rare now) in my presence, and he was then handed his (paper) chart to take with him back to the main desk. I told my dad I wanted to look at the note. The note by a physician who'd seen him stated (paraphrasing):</div><div><br /></div><div><blockquote>"Mr. Silverstein said he'd taken more than the prescribed dose of arsenic for years, and even shared it with his wife."</blockquote></div><div><br /></div><div>My father and I were shocked and dumbfounded. He'd said no such thing, and being a retired pharmacist of 40+ years, thought anyone making such a statement would have had to have been insane. (My mother had even harsher words when she heard about this.)</div><div><br /></div><div>Needless to say, I was upset. I confronted the physician who wrote the note, but that physician would not change it, bizarrely claiming that they remembered my father telling that story in a visit<i> several years prior. </i>Needless to say, such a claim violated all the precepts of medical information integrity of which I was familiar.</div><div><br /></div><div>In an initial attempt to counteract this disability exam sabotage, I actually crossed out the statement in the chart, writing "this is untrue" or words to that effect and signed my <i>own</i> name. </div><div><i><br /></i></div><div>The head of the Philadelphia VA Hospital would not return my calls on this matter.</div><div><br /></div><div>That is, until <a href="http://rockefeller.senate.gov/press/record.cfm?id=291882">Jesse Brown, then Secretary of the Department of Veterans' Affairs</a>, inquired directly a few days later.<br /><br />Unknown to the VA examiners, my father had been sent for the disability exam <b>after reporting problems to Sen. Jay Rockefeller's office about prolonged delays in having his case heard</b>.  Sen. Rockefeller's staff had set up the exam!<br /><br />Sen. Rockefeller's staff was rather upset at the story I reported upon my return to New Haven about the bizarre chart fabrication at my father's VA disability exam, and apparently relayed the story up the chain of command, as it were. </div><div><br /></div><div>After that, the head of the Philadelphia VA Hospital <b>really, really</b> wanted to speak to me and called me at Yale several times. He wanted to set up a phone conference between me, himself, and the doctors who'd seen my father. I told his administrative assistant that there was "nothing to talk about", and that the false statement in my father's chart would be removed. <b>Period.</b></div><div><br /></div><div>I think it was. My father's increased disability was granted in the end, but what was going on here with disability exams was never fully investigated to my knowledge (having done disability exams myself years prior for the regional transit authority, police, fire etc. in Philadelphia, I suspected an "incentive" program to deny vets a disability determination). If I had not examined my father's chart, we might have never known a reason for his being turned down. </div><div><br /></div><div>A culture of honesty and accountability seemed lacking then, and seems lacking now.</div><div><br /></div><div>Let's hope the investigation of these brachytherapy failures and "computer interface problems" is a bit more thorough than in my father's situation.</div><div><br /></div><div>-- SS</div><div><br /></div></div>

Bungled Brachytherapy, Computer Interfaces and Other Mysteries At The Philadelphia Veterans Administration Hospital

The Philadelphia VA Hospital has had some problems with brachytherapy (implantation of small radioactive pellets for prostate cancer) recen...

Đọc thêm »

Improving Patient Safety In The EU: HIT Should Be Classified As Medical Devices. And, Can We Drop the "Massive Cost Savings" Fable?

Here is some sense on healthcare IT from the European Union, echoing a number of themes I've written about at Healthcare Renewal and my <a href="http://www.tinyurl.com/hit-misadventure">academic site</a> on HIT difficulties and failures.<br /><br />These themes revolve around the consideration of healthcare information systems (increasingly used to mediate many aspects of medical care delivery) as medical devices requiring regulation by transparent and impartial authorities - i.e., without conflicts of interest; the importance of regulation due to the adverse consequences caused by ill designed and/or poorly implemented HIT; the need for validation and postmarketing surveillance of the performance of these systems; and vendor accountability.<br /><br />From the <a href="http://www.lakemedelsverket.se/english/">Swedish Medical Products Agency</a>:<br /><span style="font-weight: bold;"><br /></span><blockquote><a href="http://www.lakemedelsverket.se/english/All-news/NYHETER---2009/Improving-patient-safety-in-the-EU-Many-Medical-Information-Systems-should-be-classified-as-Medical-Devices/">Improving patient safety in the EU: Many Medical Information Systems should be classified as Medical Devices </a><br /><br /><span class="datetext">Thursday, June 18, 2009</span><br /><br />Correctly functioning Information Technology (IT) systems for health care is a prerequisite for ensuring patient safety. However, applicable regulation is rarely applied, mainly due to the lack of useful guidelines on how to classify the IT-systems <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[guidelines whose formulation has been resisted for many years by the HIT industry, I might add - ed.]</span> A national working group in Sweden has prepared a guidance report that will help increase compliance with the Medical Device regulation in the EU, thus improving patient safety.<br /><br />An increasing number of serious incidents, with IT-systems involved, have been reported from the health care sector in recent years <span style="font-style: italic; font-weight: bold; color: rgb(255, 0, 0);">[but largely covered up in part due to contractual gag clauses on HIT sales and resultant violations of healthcare leadership fiduciary responsibilities - </span><span style="font-style: italic; font-weight: bold;">see <a href="http://www.hcrenewal.blogspot.com/2009/03/health-it-hold-harmless-and-defects-gag.html">this link</a></span> - <span style="font-style: italic; font-weight: bold; color: rgb(255, 0, 0);">ed.]</span> In most of these cases, existing regulatory regimes have not been applied, either by manufacturers or by health care providers. As a result, the assessment of safety issues has become unclear because the manufacturer’s responsibility is vague. <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[It's not vague; it's contractually absent - ed.]</span><br /><br />For this reason, in 2008 the Swedish Medical Products Agency (MPA) initiated a national working group with the aim of providing guidelines for the classification of IT-systems intended to be used in health care. The working group has now prepared a report, intended to serve as guidance for both manufacturers and end-users. At present, there are no guidelines of this kind in Sweden, or the rest of Europe.<br /><br />“Many parties, including the European Commission, have expressed interest in our work. We believe that it will improve the quality and safety of medical information systems. <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[Finally -ed.] </span>A common approach to product classification is key” says Lennart Philipson, Scientific Director.<br /><br />One important conclusion from the working group is that the Medical Device Directives is the most suitable regulation to be applied.<br /><br />“Our market surveillance of medical information systems will be based on the conclusions in the report from now on. If a system falls within the definition of a medical device, the requirements for CE-marking <span style="font-weight: bold; font-style: italic;">[<span style="color: rgb(255, 0, 0);">Conformité Européene, a</span> <a href="http://www.biz-services-savvy.com/ce-mark.html">branding mark</a> <span style="color: rgb(255, 0, 0);">- ed.]</span></span><span style="color: rgb(255, 0, 0);"> </span>shall be applied” says Mats Ohlson, Chairman of the working group.<br /><br />Manufacturers need to adapt to applicable standards when it comes to quality systems, risk management and usability. They will also need to establish post-market surveillance routines to regularly assess products on the market.<br /><br />The working group consisted of members from the Swedish Medical Products Agency, the National Board of Health and Welfare, the Swedish Association of Local Authorities and Regions, Swedish Medtech, the Swedish Standards Institute, Swedish Electrical Standardization and notified bodies through Intertek Semko.</blockquote><br />A report entitled "<span style="font-weight: bold;">The Medical Products Agency’s Working Group </span><span style="font-weight: bold;">on Medical Information Systems: Project summary</span>" is available in English translation at <a href="http://www.lakemedelsverket.se/upload/foretag/medicinteknik/en/Medical-Information-Systems-Report_2009-06-18.pdf">this link</a> (PDF).<br /><br />Download and read the entire report.<br /><br />Highlights:<br /><br /><blockquote><p>A general opinion of the health care providers represented in this Working group is that from a patient safety point of view, i<span style="font-weight: bold;">t is desirable that stand alone software and systems intended to, directly or indirectly, affect diagnosis, health care and treatment of an individual patient <span style="font-style: italic; color: rgb(255, 0, 0);">[what I have referred to a "mediating aspects of care" - ed.]</span> </span>shall be regulated under a <span style="font-weight: bold;">Product Safety Regulation</span>.<span style="font-weight: bold;"> </span><br /></p><p>The Working group has not been able to define any other appropriate regulation than the Medical Device Directive when it comes to the definition of such systems. Adherence to the Medical Device Directive is important so that manufacturers and clients can work with the same intentions. <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[I would add that the intentions are not exactly the same; the manufacturers' intentions are, by necessity, to maximize profits - ed.]</span> This project report serves as one step towards the clarification regarding which conditions that shall apply.<br /><br />The Working group believes that <span style="font-weight: bold;">software intended for a medical purpose must be regarded as a "device" <span style="color: rgb(255, 0, 0); font-style: italic;">[indeed - ed.] </span></span>and expressions such as "project", "service" and similar must be avoided describing a Medical Information System. This approach has advantages since a product safety regulation, such as the Medical Device Directive, can be applied and the product will have:<br /><br />• a defined intended use<br />• defined and documented specifications<br />• a manufacturer with a clear responsibility until the delivery is accepted<br />• a controlled ”Post Market” surveillance<br /><br />Furthermore, the Working group believes:<br /><br />• that the Medical Device Directive shall be applicable for software and systems that fulfills the definition of a medical device<br />• that the manufacturer of the intended systems shall follow the appropriate<span style="font-weight: bold;"> </span><span style="font-weight: bold;">validation method</span> <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[e.g., as in the pharma industry's IT - ed.]</span> in the Medical Device Directive<br />• that the level of risk shall be assessed, motivated and control the classification and verification method<br />• that applicable standards shall be followed for design, verification and validation regarding software and information systems that are medical devices<br />• that there is no legal requirement that manufacturers of devices in the lowest device classes must have certification for their operational processes to fulfill the Medical Device Directive. The Medical Products Agency has in these cases no mandate to formally demand certification, such as for instance ISO 13485. However, in reality it is still necessary to have some form of a functional quality management system to manage a proper design process for software systems <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[due to the fact that patients, ethical issues, and human rights issues are involved - ed.]</span> as well as the necessity to fulfill the Post Market Requirements including Vigilance reporting.<br />• that the prerequisite for a successful and fair application of the regulation assumes that health care providers demands that such systems shall be CE‐marked and be regarded as medical devices<br />• that a controlled and standardised verification method when performing installation work in a user’s network, supported by the manufacturer, is an essential supplement to the manufacturer’s CE‐marking and a prerequisite for an acceptable safety level when using the device.</p><p><br /></p> </blockquote>This section gets to the heart of the matter regarding electronic medical records:<p></p><blockquote><p><br /></p><p><span style="font-weight: bold;">8.2.3 Electronic patient records </span><br /></p><p>Electronic patient record systems are <span style="font-weight: bold;">more than an archive for just documents</span>, they also have features for compiling and transferring information to be used for decision making. <span style="font-weight: bold;">Patient records are the basis for documentation and planning of diagnosis, therapy and follow‐up, especially when it comes to drug prescriptions and recommendations. <span style="color: rgb(255, 0, 0); font-style: italic;">[These caveats include CPOE systems as well - ed.]</span></span><br /></p><p>It is becoming more common that electronic patient record systems and other systems are interconnected, for instance imaging systems or laboratory systems. It is obvious that such systems <span style="font-weight: bold;">s</span><span style="font-weight: bold;">hould not be regarded as “purely administrative”</span>; instead they have the characteristic features that are typical for medical devices. They sort, compile and present information on patients’ treatments and should therefore be regarded as medical devices in accordance to the definition.</p><p>Since the electronic patient record system often replaces/constitutes the user interface of “traditional” medical device systems, the call for 100% accuracy of the presented information is increased. Patient record systems have crucial impact on patient safety, <span style="font-weight: bold;">and this has been proven to be the case after a series of incidents that has been reported to the Swedish National Board of Health and Welfare. </span><span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">[It would be interesting to know more about this "series of incidents" - ed.]</span></p></blockquote><p> </p><br />Finally, in the June 20, 2009 Wall Street Journal article "<a href="http://online.wsj.com/article/SB10001424052970204005504574235751720822322.html">The Myth of Prevention</a>", Abraham Verghese, Professor and Senior Associate Chair for the Theory and Practice of Medicine at Stanford, echoed <a href="http://hcrenewal.blogspot.com/2009/06/wharton-on-healthcare-it-can-i-go-home.html">several Wharton professor's doubts</a> about the cost savings and ultimate value of electronic medical records, touted as the cybernetic savior of healthcare:<br /><br /><blockquote>... I have similar problems with the way President Obama hopes to pay for the huge and costly health reform package he has in mind that will cover all Americans; he is counting on the “savings” that will come as a result of investing in preventive care and investing in the electronic medical record among other things. It’s a dangerous and probably an incorrect projection. </blockquote><br /><br />HIT can be a facilitator (not a replacement or revolutionary tool) to physicians, if done well, and not overdone to the point of costing far more than <a href="http://hcrenewal.blogspot.com/2006/10/70-million-for-electronic-medical.html">entire new hospitals or hospital wings</a>. There is no replacement for a physician with a strong background in the preclinical and clinical sciences, ethics, critical thinking, and the scientific method. Computers are not a miracle. They can help provide facts, but cannot fundamentally turn a poor physician into a Marcus Welby. Verghese on that issue:<br /><br /><blockquote>To come back to my favorite painting: a computer cannot take the place of the doctor in Fildes’s painting <span style="font-weight: bold; font-style: italic; color: rgb(255, 0, 0);">["The Doctor", 1891, see below - ed.];</span> an electronic medical record (EMR) may or may not save money (it won’t be anywhere as much as is projected) but what it will do is ensure that we doctors, nurses, therapists, particularly in hospitals will be spending more and more time focused on the computer, communicating with each other, ordering and getting tests, buffing and caring for our virtual patient—the iPatient is my term for this phenomenon—while the patient in the bed wonders where everybody is. Having worked exclusively for the last seven years or so in hospitals that have electronic medical records (EMR), I have felt for some time that the patient in the bed has become an icon for the real focus of our attention, the iPatient. Yes, electronic medical records help prevent medication errors <b><i><span class="Apple-style-span" style="color:#FF0000;">[only when done well, see<span class="Apple-style-span" style="color: rgb(0, 0, 0); "> <a href="http://jama.ama-assn.org/cgi/content/abstract/293/10/1197">Koppel</a></span><span class="Apple-style-span" style="color: rgb(0, 0, 0); font-style: normal; font-weight: normal; "> <b><i><span class="Apple-style-span" style="color:#FF0000;">- ed.]</span></i></b> and are a blessing in so many ways, but they won’t hold the patient’s hand for you, they won’t explain to the family what is going on.</span></span></i></b></blockquote><blockquote><br /></blockquote><br /><div style="text-align: center;"><a onblur="try {parent.deselectBloggerImageGracefully();} catch(e) {}" href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjAE7RytgExhui-vW_XuPeuePNenYs0LVWL85fksBIphZhtsoPXzESdZdoVldOUoYuqqKqVmW2uGDV82HsTUkchsibY6aJzv3lvyeIhU5gKvh6fyafpDfR-TSGIoJXE0TzeNEDmvD4qZiuk/s1600-h/the_doctor.jpg"><img alt="" border="0" id="BLOGGER_PHOTO_ID_5349443605127209442" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjAE7RytgExhui-vW_XuPeuePNenYs0LVWL85fksBIphZhtsoPXzESdZdoVldOUoYuqqKqVmW2uGDV82HsTUkchsibY6aJzv3lvyeIhU5gKvh6fyafpDfR-TSGIoJXE0TzeNEDmvD4qZiuk/s320/the_doctor.jpg" style="margin: 0px auto 10px; display: block; text-align: center; cursor: pointer; width: 320px; height: 214px;" /></a><span style="font-size:85%;">"The Doctor", 1891 (Sir Luke Fildes, 1843-1927)</span><span style="font-size:85%;"><span style="font-family:arial;"></span></span><br /></div><br /><br />I will go a step further. It is now clear that the promises of massive savings from electronic medical records, usually originating from vendors based on hearsay or cherry-picked, favorable but flawed studies that dismiss a <a href="http://hcrenewal.blogspot.com/2009/03/reader-on-why-we-have-busywork.html">growing body</a> of <span style="text-decoration: underline;"></span><a href="http://www.ischool.drexel.edu/faculty/ssilverstein/failurecases/?loc=other">other literature</a> (such as in Cerner CEO Neil Patterson's hyper-ebullient piece <a href="http://nealpatterson.cernerblog.com/2009/02/26/the-abcs-of-systemic-healthcare-reform.aspx">here</a>), have been <a href="http://hcrenewal.blogspot.com/2009/05/machinery-behind-healthcare-reform-how.html">deliberately manufactured</a> by a lobby of industry pundits.<br /><br />It is only now that I more fully comprehend the origins of and motivations behind the "we will revolutionize medicine" fervor exhibited by HIT company CEO's, such as I experienced and wrote about starting back in the 1990's. From my 1999 essay "<a href="http://www.ischool.drexel.edu/faculty/ssilverstein/informaticsmd/infordef1.htm">What is Medical Informatics, and Why is it an Important Specialty</a>":<br /><br /><p></p><blockquote><p>... A letter I had written <span style="font-style: italic; font-weight: bold; color: rgb(255, 0, 0);">[in Feb. 1999 - ed.]</span>, "<a href="http://web.archive.org/web/19991005222815/http://www.healthcare-informatics.com/issues/1999/02_99/postmark.htm">Broken Chord</a>" published in the journal <a href="http://www.healthcare-informatics.com/">Healthcare Informatics</a>, further amplifies the skills and experience issue. The letter addressed issues of skills, insights, and roles of MIS personnel in healthcare settings. I described a Microsoft Healthcare Users Group conference attended predominantly by healthcare MIS staff and vendors, where I observed a panel discussion moderated by the CEO of HBOC<b><i><span class="Apple-style-span" style="color:#FF0000;"> [before he was found to be cooking the books - ed.]</span></i></b> and composed of several other industry CEO's. The panel was discussing how they would revolutionize healthcare through their leadership in information technology. </p><p> During the Q&A period I asked the audience how many really felt they would revolutionize medicine through their leadership in IT. Several hundred--almost all in the audience--raised their hands. I then asked how many had ever taken care of patients or examined any textbook of medicine, such as Harrison's Textbook of Internal Medicine or the Merck Manual. A minority of hands went up. This suggested, in my view, a striking deficiency of knowledge, experience and insight on the part of the de facto clinical information technology leadership, complicated by rather cavalier attitudes regarding the essential role of clinical expertise. </p></blockquote><br />I believed what I was hearing to be <span style="font-weight: bold; font-style: italic;">irrational</span><span style="font-style: italic;"> </span>exuberance, not <span style="font-style: italic; font-weight: bold;">manufactured</span> exuberance.<br /><br />Quality improvements are possible under the most controlled of environments, such as <a href="http://www.ischool.drexel.edu/faculty/ssilverstein/failurecases/?loc=cases&sloc=Cardiology%20story">here</a> (such projects are not easily portable, unfortunately). Massive cost savings? When organizations are spending upwards of $100 million just for system implementation, not including ongoing lifecycle costs, massive savings projections from HIT are risible.<br /><br />One would think that after Madoff, <a href="http://www.washingtonpost.com/wp-dyn/content/article/2009/06/19/AR2009061900078.html?hpid=moreheadlines">Stanford</a> and other recent financial confidence scandals, that bright people would be more skeptical of "massive returns, always and forever" promises, but that does not appear to be the case. With regard to mass savings, I make the prediction that in 2014, as <a href="http://www.ischool.drexel.edu/faculty/ssilverstein/failurecases/?loc=cases&sloc=UK">in the UK today</a>, our government, healthcare management, and medical professionals will realize, Madoff style, that they've been had.<br /><br />So, Health IT Industry executives, pundits, and consultants, can we drop the grandiose "we will revolutionize medicine!" and "massive savings" myths, and focus on facilitating clinicians and on quality improvement, please?<br /><br />-- SS<div><br /></div><div>6/22/09 <b>Addendum</b></div><div><b><br /></b></div><div>The U.S. Food and Drug Administration (FDA) and Sweden's Medical Products Agency are <a href="http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108646.htm">no strangers</a>.</div>

Improving Patient Safety In The EU: HIT Should Be Classified As Medical Devices. And, Can We Drop the "Massive Cost Savings" Fable?

Here is some sense on healthcare IT from the European Union, echoing a number of themes I've written about at Healthcare Renewal and my ...

Đọc thêm »

Scrushy Owes $2.88 Billion for Damages

As <a href="http://online.wsj.com/article/SB124533875598827811.html">reported by the Wall Street Journal</a>, here is another reminder about just how bad the leadership of health care can be,<br /><blockquote><br />Richard Scrushy was hit with a staggering $2.88 billion civil judgment in a suit brought by HealthSouth Corp. shareholders, one of the largest findings ever from the era of massive corporate scandals.<br /><br />The plaintiffs said that the former chairman and chief executive helped artificially inflate HealthSouth's earnings for at least six years through an accounting scam uncovered in 2003.<br /><br />Lawyers said <em>Thursday's judgment appears to be the largest financial penalty ever levied against a single executive</em>.<br /><br />Mr. Scrushy was once <em>the poster child of highflying CEOs</em>. He was paid <em>salary and perks of as much as $40 million a year</em>, performed at company events in his country band and <em>traveled on one of a fleet of corporate jets</em>.<br /><br />Mr. Scrushy <em>hosted parties on the Chez Soiree, his 92-foot yacht</em>, and arranged to pay for breast augmentation for a member of a company-sponsored singing group.<br /><br />Judge Horn said <em>Mr. Scrushy either knew or should have known about the fraud</em>, giving no credence to Mr. Scrushy's repeated insistence that subordinates perpetrated the scheme without his knowledge.<br /></blockquote><br /><br />Note that Scrushy had is already a convicted felon, "as he is two years into a seven-year sentence in federal prison on the bribery conviction." He has gone from being a "poster-child" for the high-flying health care corporate CEO to being a poster-child for liars and criminals fraudulently leading health care organizations.<br /><br />So I get to say it again: instead of focusing on the reform of health care finance, we should be focused on the reform of health care leadership. We have set up a system that rewards imperial health care CEOs with the the lavish life-style and power that used to be reserved for the aristocracy. We have attracted health care leaders who value their own wealth and power ahead of patients' welfare and safety. To improve health care, we need responsible, ethical, honest stewardship of health care organizations. If we do not improve health care's leadership, it will not matter whether we finance it via private insurance, public options, or single-payers.

Scrushy Owes $2.88 Billion for Damages

As reported by the Wall Street Journal , here is another reminder about just how bad the leadership of health care can be, Richard Scrushy w...

Đọc thêm »

Synthes Indicted

Last month, we <a href="http://hcrenewal.blogspot.com/2009/05/synthes-settles-agrees-to-stop-giving.html">posted</a> about a legal settlement in which device manufacturer Synthes agreed to stop creating conflicts of interests by paying physicians who performed its trials with company stock. Synthes is back in the news, and not in favorable terms. As <a href="http://www.philly.com/inquirer/health_science/daily/20090617_Indictment_says_W__Chester_firm_tested_cement_on_patients.html">reported by the Philadelphia Inquirer</a>,<br /><blockquote><br />A Swiss company with major operations in West Chester illegally tested its bone cement on about 200 people, three of whom died, according to a 52-count indictment issued yesterday by the U.S. attorney in Philadelphia.<br /><br />Synthes Inc., a producer of orthopedic products that employs about 1,400 in Chester County, <em>did not tell any of the patients that they were participating in experimental surgeries</em>, the indictment said.<br /><br />Federal prosecutors also <em>accused Synthes executives of lying to the U.S. Food and Drug Administration </em>about whether they had tested the product for uses that the agency had not approved.<br /></blockquote><br /><br />Particularly striking were allegations that the company's CEO was directly involved in the decisions.<br /><blockquote><br />The FDA had approved Norian XR to fill bony voids or defects in some parts of the body, but not in the spine. The agency insisted that Synthes conduct clinical trials, which are lengthy and expensive and require FDA oversight, if it wanted to expand the uses of Norian XR to treat spinal fractures.<br /><br />But, <em>a company executive identified in the indictment as 'Person No. 7' decided not to conduct clinical trials. Instead, the person said the company should 'get a few sites to perform 60-80 procedures and help them publish their clinical results,'</em> according to the indictment.<br /><br />The report never explicitly identifies 'Person No. 7,' but says he was <em>the chief executive officer and a large shareholder</em>. <em>A company spokesman identified Hansjörg Wyss as the CEO while the tests occurred.<br /></em><br />Forbes Magazine lists Wyss, a Swiss citizen, as a West Chester resident and <em>the Philadelphia area's wealthiest person, with a net worth of $5.7 billion</em>. Last year, <em>Wyss gave the largest amount from a single source in Harvard University's history when he donated $125 million to create a bio-engineering institute</em>.<br /><br />Wyss owns about 40 percent of outstanding Synthes stock. The company, which had about $805 million in first-quarter sales this year, specializes in orthopedic devices, including plates and screws.<br /></blockquote><br /><br />The Inquirer article included more detail about how the company allegedly disregarded risks to patients as it pushed the surgeons to try Norian XR on them, without disclosing that its use was experimental,<br /><blockquote><br />Instead of conducting trials with FDA oversight, <em>Synthes executives convinced about 50 surgeons to test Norian XR in people with spinal fractures. Synthes sales representatives often were in the operating room</em> during the surgeries, including those where patients died, prosecutors said.<br /><br />When company executives began these surgical tests in 2002,<em> they knew that Norian XR caused blood clots in test tubes and that clots became lodged in the lungs in tests with pigs</em>, the indictment said.<br /><br /><em>Synthes continued the experimental surgeries even after two patients died on the operating table</em>, prosecutors said. The company stopped the tests after <em>a third patient died</em> in January 2004.<br /></blockquote><br /><br />Again, it seems that the current leaders of health care all too often put more priority on personal profits than on protecting patients' health and safety. So instead of all the talk about reforming how health care is financed, maybe the conversation should be about reforming how health care is lead, and finding ethical, honest leaders who put patients' welfare ahead of their own wealth.

Synthes Indicted

Last month, we posted about a legal settlement in which device manufacturer Synthes agreed to stop creating conflicts of interests by payin...

Đọc thêm »

Managed Care Executives Will Not Limit Rescissions

We have posted before about how certain health care insurance companies/ managed care organizations in California were found to have cancelled individual health insurance policies after the people holding them made substantial claims, supposedly rationalized by minor errors or omissions in the information the people supplied to the companies on their individual applications found after the claims were made. Several companies were subsequently disciplined by the state government for these "rescissions." See posts on rescissions by WellPoint <a href="http://hcrenewal.blogspot.com/2007/03/wellpoint-fined-1-million-for-canceling.html">here</a>, <a href="http://hcrenewal.blogspot.com/2007/07/more-than-half-of-wellpoints-policy.html">here</a>, and <a href="http://hcrenewal.blogspot.com/2008/07/wellpoint-settles-but-has-it-become-too.html">here</a>, and by Health Net <a href="http://hcrenewal.blogspot.com/2007/11/health-net-statements-inoperative-in.html">here</a>.<br /><br />Executives of several such companies testified before a US congressional committee recently, with remarkable results, as reported by Lisa Girion <a href="http://www.latimes.com/business/la-fi-rescind17-2009jun17,0,5870586.story">writing in the Los Angeles Times</a>. First, by way of background, the article suggested that rescissions are widespread:<br /><blockquote>An investigation by the House Subcommittee on Oversight and Investigations showed that health insurers WellPoint Inc., UnitedHealth Group and Assurant Inc. <em>canceled the coverage of more than 20,000 people, allowing the companies to avoid paying more than $300 million in medical claims over a five-year period</em>. </blockquote><br />Furthermore,<br /><blockquote><br />The committee's investigation found that <em>WellPoint's Blue Cross targeted individuals with more than 1,400 conditions, including breast cancer, lymphoma, pregnancy and high blood pressure</em>. And the committee obtained documents that showed <em>Blue Cross supervisors praised employees in performance reviews for rescinding policies</em>.<br /><br />One employee, for instance, received a perfect 5 for 'exceptional performance' on an evaluation that noted the employee's role in dropping thousands of policyholders and avoiding nearly $10 million worth of medical care.<br /></blockquote><br /><br />Nonetheless, the executives were not prepared to abandon rescissions:<br /><br /><br /><blockquote><br />Executives of three of the nation's largest health insurers told federal lawmakers in Washington on Tuesday that <em>they would continue canceling medical coverage for some sick policyholders</em>, despite withering criticism from Republican and Democratic members of Congress who decried the practice as unfair and abusive.<br /><br />The executives -- Richard A. Collins, chief executive of UnitedHealth's Golden Rule Insurance Co.; Don Hamm, chief executive of Assurant Health and Brian Sassi, president of consumer business for WellPoint Inc., parent of Blue Cross of California -- were courteous and matter-of-fact in their testimony.<br /><br />But <em>they would not commit to limiting rescissions to only policyholders who intentionally lie or commit fraud to obtain coverage</em>, a refusal that met with dismay from legislators on both sides of the political aisle.<br /><br />Sassi said rescissions are necessary to prevent people who lie about preexisting conditions from obtaining coverage and driving up costs for others.<br /><br />'I want to emphasize that rescission is about stopping fraud and material misrepresentations that contribute to spiraling healthcare costs,' Sassi told the committee.<br /><br />But <em>rescission victims testified that their policies were canceled for inadvertent omissions or honest mistakes about medical history on their applications</em>. Rescission, they said, was about improving corporate profits rather than rooting out fraud.<br /></blockquote><br /><br />Also,<br /><br /><blockquote><br />Late in the hearing, Stupak, the committee chairman, put the executives on the spot. <em>Stupak asked each of them whether he would at least commit his company to immediately stop rescissions except where they could show 'intentional fraud.'<br /><br />The answer from all three executives:<br /><br /><strong>'No.'</strong></em><br /></blockquote><br /><br />This is just amazing. Here are executives of three of the country's largest for-profit health care insurance companies/ managed care organizations asserting they will continue cancelling peoples' policies after they make claims, simply because of minor errors or omissions in the application documents that have no bearing on the particular claims, or on discrepancies between information provided by patients and doctors, given that patients may not have full access to or understanding of what is in their charts. What good is an insurance policy that is liable to be cancelled as soon as one makes a claim on it? <br /><br />Whatever these companies are selling, it is not insurance. Basing our health care system on their products is foolish, but that is what we are doing, and what many political leaders would continue to do. <br /><br />In my humble opinion, if we are going to have a system based on privately provided insurance, that insurance has to be honest, unlike what some of our largest insurance companies seem to be peddling.

Managed Care Executives Will Not Limit Rescissions

We have posted before about how certain health care insurance companies/ managed care organizations in California were found to have cancell...

Đọc thêm »

Google translates Persian

Today, we added Persian (Farsi) to <a href="http://translate.google.com/?sl=fa&tl=en#">Google Translate</a>. This means you can now translate any text from Persian into English and from English into Persian — whether it's a news story, a website, a blog, an email, a tweet or a Facebook message. The service is available free at http://translate.google.com.<br /><br />We feel that launching Persian is particularly important now, given ongoing events in Iran. Like <a href="http://www.youtube.com/blog?entry=2soHkPYkGJw">YouTube</a> and other services, Google Translate is one more tool that Persian speakers can use to communicate directly to the world, and vice versa — increasing everyone's access to information.<br /><br />As with all machine translation, it's not perfect yet. And we're launching this service quickly, so it may perform slowly at times. We'll keep a close watch and if it breaks, we'll restore service as quickly as we can.<br /><br />We've optimized this service for translation between Persian and English. But we're working hard to improve Persian translation for the additional 40 languages available via Google Translate. If you see something you think is incorrectly translated, we invite you to click on the "contribute a better translation" link and we'll learn from your correction.<br /><br />The web provides many new channels of communication that enable us to see events unfold in real-time around the world. We hope that Google Translate helps make all that information accessible to you — no matter what language you speak. So please visit <a href="http://translate.google.com/?sl=en&tl=fa#">Google Translate</a> and try it out.<br /><br /><span class="byline-author">Posted by Franz Och, Principal Scientist</span>

Google translates Persian

Today, we added Persian (Farsi) to Google Translate . This means you can now translate any text from Persian into English and from English i...

Đọc thêm »