|
Tuesday, April 29, 2008 |
|
|
Yottabytes
Back to the MRI question, where are the MRI's -- and CAT scans, X-Rays, and mammogram's -- stored? They used to be on film and the patient would carry them around from specialist to specialist and the hospital would keep football field size storage rooms loaded with them. Progressive hospitals today use a PACS (Picture Archiving and Communication System). The performance and reliability of PACS are critical to a hospital's ability to provide patient care. The PACSs have gotten better and better but physicians are continuously raising the bar. Understandably, CIO's and CFO's are concerned about the fast growth of storage needed as the imaging technology supports higher resolutions, more images per study, and escalating federal and state government storage requirements. Physicians want online access 24x7 from the office, hospital or their home to not only the MRI you had today but the one you had a year ago and maybe ten years ago. Hospitals have tried to cope with the increased demand by offering online storage for very current images and "nearline" storage for those that have been archived. Nearline often means that the image is stored on tape and can be brought online if a special request is made. Increasingly physicians and patients do not feel there is anything "special" about it -- they expect all data to be online all the time just like Amazon. The online retailer has every order they have ever received since the company started in 1995 online and available 24x7. Easy for them some might say. An order for a book is trivial compared to a digital MRI image. How big is a digital MRI image? A recent cervical spine MRI contained 160 images and was approximately 60 megabytes in size. About the same as 200 iPhone pictures or 20 iTunes songs. Let's suppose a community hospital has 25,000 patient visits per year and that on average a patient has two image studies performed. That would be 50,000 times 60 megabytes which equals 3 terabytes. Now let's consider what size storage is available and how much it costs. In the mid 1970's an IBM "disk pack" for a mainframe computer had a capacity of 200 megabytes -- about three MRI's. The entire storage system could contain eight "drives" for a total of 1.6 gigabytes. It seemed like a lot at the time. The cost of the disk drive that the disk pack fit on was nearly $200,000. During the last thirty years the cost has continuously plummeted while the capacity has skyrocketed. The Apple Time Capsule has a capacity of one terabyte and costs $499. IBM has a new storage system that offers up to 1,176 terabytes in a single system. Soon we will be talking about petabytes (1,000 terabytes) and then exabytes, zettabytes, and yottabytes. When I had written a story about yottabytes back in 2005 a reader said the term should be "alottabytes". A yottabyte is 1,000,000,000,000,000,000,000,000 bytes. The bottom line is that there will be plenty of storage to put all our images online. The key challenge is the management of the data -- keeping it secure, backed up, resilient to disaster, and easy to access and manipulate. Many providers will decide to put all the data in the "cloud" and let someone else manage it. Amazon's Simple Storage Service (S3) is the tip of the iceberg. They charge $0.15 per gigabyte per month of storage used. IBM offers a wide range of storage services and also partners with many healthcare information technology companies. The normal reaction would be that having all the images online is too expensive. I think many of us will instead think of it like electricity. Healthcare providers use a lot of electricity and some are beginning to cogenerate their own to save money. One thing they don't do however is consider having some of their electricity "offline" or "nearline". It is online 24x7. That is the way we will soon think of medical images.
Healthcare , Internet Technology , Personal Computing April 29, 2008 01:18 PM |
|
|
Monday, April 28, 2008 |
|
|
Chocolate and Gum
|
|
|
Tuesday, April 22, 2008 |
|
|
Our Medical Records
Patient safety, ballooning costs, and government budget pressures are accelerating the move toward electronic medical records. There are many variations on the theme. EMR's in the hospital, Personal Health Records on a smartcard, and Personally Controllable Medical Records on the web. EMR's are emerging from insurance companies, pharmacies, community doctors, hospitals, regional health information organizations, employers and software companies. It is not yet clear which EMR or combination will prevail. Personally, I will be glad when all my medical information is encrypted and stored on the Internet where I will know that at last it will be safe and under my control. One thing I know for sure is that it is time to make major strides. There will be many participants in making it happen. Government and non-profit organizations such as HL7 must play a key role in establishing standards so that the various kinds of EMR's can be compatible. Most experts agree that Personal Health Records sponsored by healthcare providers have the best chance of success in the short term. Longer term "cloud-based" PHR's such as proposed by Google, Microsoft, and others have great potential but need to overcome trust and privacy concerns of consumers. One provider pilot program that I think has potential is the "SmartCard" at Danbury Hospital in Connecticut. Five-hundred patients in the trial program can insert their smartcard in a kiosk (very similar to an ATM), confirm the identity displayed, and then receive a "ticket" -- just like at the deli -- to await being called for their test or procedure. The patients like it because they don't have to use the "clipboard" and the staff like it because they don't have to ask patients to use the clipboard. The potential goes well beyond just automating the check-in process. The smartcard can be the "carrier" for our electronic personal health record. At some point our mobile phones may take over the task but in the short to medium term the smartcard may prove to be a very effective aid to empowering consumers to manage their own our healthcare. The potential is huge. Upon entering the office of our primary care physician the office system could recognize our smartcard and an exchange could begin which updates the hospital with the latest information on the card, updates the card with any updated test results the hospital may have, and updates the primary care physician with the latest test results plus notes from any specialist consultations. If the primary care physician "writes" a script on his or her system it could be automatically transferred to the smartcard and to the pharmacy system and when the patient gets to the pharmacy the card could be recognized and the prescription would be filled. All the information on the smartcard would be encrypted and accessible only after authentication by the smartcard holder. This could be done using a password or a biometric such as an iris scan or fingerprints. This may sound futuristic to some but similar things are already being done. Denmark began a drive toward paperless hospitals more than a decade ago and is achieving much success. Verified Identity Pass, Inc. has a vision of using smartcards to enable us to breeze through airport security lines. The Fly Clear smartcard contains digitized versions of both your iris scans and fingerprints. There are numerous technical and financial challenges inhibiting the rollout of a smartcard system in a pervasive way. The biggest challenge is that the benefits are "shared" -- neither patient, provider, or payer can justify the cost but collectively everyone wins. It reminds me of the UPC challenge of the 1970's. In spite of large benefits from knowing what got sold and when, the grocery stores were hesitant to invest in UPC scanners because there were no products that had UPC symbols on them (the first product to have a code was a packet of Wrigley's Gum in 1974). The stores found it difficult to justify the cost even though there would be labor savings from scanning versus "ringing up". The package goods manufacturers were also skeptical, despite the benefit of knowing exactly how their products were doing at retail on a timely basis. I remember visiting the M&M Mars candy factory in Hackettstown, NJ in the early 1970's and discussing UPC scanning with the director of product packaging. She said there were not enough benefits to offset giving up the "real estate" on the candy bag to place a symbol for which there were hardly any scanners to scan them. (The first UPC scanner was installed at a Marsh's supermarket in Troy, Ohio in June, 1974). It took strong leadership, competition among retailers, and perseverance to get us to the ubiquitous scanning which we enjoy today. Likewise with personal healthcare records. The benefits are huge -- increased accuracy of information leading to better outcomes and reduction in duplicative procedures, and ultimately personalized healthcare. Physicians will spend less time ordering procedures and medications, liability costs should go down due to fewer errors, increased collaboration will improve caregiving, patients will be able to relocate and take their healthcare data with them, and patients will be able to take a more proactive role in their own health and selection of providers. All it takes is strong leadership, competition among healthcare providers, and perseverance. The glass is half-full, not half-empty. It is likely that in the next five years we will see more progress toward electronic medical records than we have seen in the last twenty. |
|
|
Wednesday, April 16, 2008 |
|
|
|
Tuesday, April 15, 2008 |
|
|
SOA Las Vegas
What attracted the 6,300 people to fly to Las Vegas and fill every hall, ballroom, salon, patio, and restaurant at the MGM Grand? IBM calls it "Smart SOA". I call it The Application Web. Only the most brilliant technical people could come up with SOA as a name for something. Let's see, is it safe operating area, School of the Americas, Skies of Arcadia (a Nintendo game), Society of Actuaries, state of the art, or the Sarbanes-Oxley Act? Nope. The SOA that brought all these people together stands for "service oriented architecture". It is really important. The wikipedia has a comprehensive definition of SOA but basically it represents a new way for companies -- and hospitals, schools, and governments -- to enable their customers -- and suppliers, business partners, and employees -- to get things done on the web. Actually it is isn't new -- the idea has been around for decades -- but now it is really happening. It is so much a part of the vernacular at IBM that they just matter of factly talk about "so a".
Over the last fifty years there has been an explosion of computer applications, but many of them were built in silos and were highly inflexible. In some cases companies thought decentralization was the answer so they allowed divisions and departments to do their own thing. The result was that many have a hodgepodge of incompatible systems that nobody is happy with. The web took things a big leap forward. At last there was a common way (the browser) for accessing and displaying information, even though the applications that run on the server -- that do the pricing, inventory lookups, shipping estimates, invoicing, etc. -- are still proprietary and usually tied to one particular IT vendor or system. The applications have also been very monolithic; i.e. in order to fulfill the expectations of customers on the web the application has to do the whole job. Soup to nuts; present the right price, confirm if the item is in stock, calculate shipping, and confirm the status of the order. Increasingly, customers want to get access directly into the supply chain and see exactly where their order stands. In short, applications have gotten larger and more complicated -- harder, not easier. The SOA services do not all have to be developed or acquired internally. Thanks to the Internet, services can be "rented" from others. For example, suppose that a company called American Specialties Inc. (ASI) specializes in selling American goods for delivery mostly outside of America. They want to create an application to sell their products on the web. The trickiest part of the application is determining the best way to ship the product to ensure it gets there when the customer wants it and at the lowest cost. ASI doesn't’t have the skills to write this particular part of the application and they haven’t bee able to find a vendor with a software package that can do it and which is compatible with the rest of ASI’s software. It turns out that there is another company called Rates and Costs Inc. (RCI), which specializes in the calculation of optimum routes and the associated costs for shipment to places anywhere in the world. RCI offers the calculation as a service on the web and it is the exact function ASI needs to incorporate into their web application. Since RCI follows the SOA standards, ASI is able to see the specifications for RCI’s service – what inputs are required and what output does it produce. RCI could have created their calculation service using any IT platform they choose -- the standards assure that things can work together. The programmer at ASI likes RCI’s program because it performs exactly the right function that ASI needs and the software has already been written and tested! ASI follows the SOA standards to incorporate RCI’s service into their web application. Whenever a user goes to ASI’s web page and needs shipment route and cost information, a link is made behind the scenes to RCI’s web server to get the information. ASI’s customers don’t know, nor will they care, that part of the job is being done by RCI’s server; not ASI’s server. ASI makes an arrangement to pay RCI each time one of ASI’s customers uses the RCI web service. Creating programs by linking to other programs without regard to what programming language was used to create the others’ programs represents a whole new paradigm. It is one of the information technology industry’s holy grails. Standards organizations, such as Oasis, have been attempting for years to create a “neutral” programming environment. The UNIX vendors – HP, DEC, Sun, IBM, Data General, and others – formed various organizations, councils and consortia over the years attempting to bring things together. Progress was made but none of these initiatives achieved real openness and true compatibility across the information technology industry -- until SOA. It is not really new but it is time. Open Internet standards and SOA tools are making it happen. SOA will make it possible for the web to evolve from a web of content to a web of content and applications. SOA will enable server-to-server interaction in addition to browser to server interactions. Servers will negotiate with other servers and even complete transactions by themselves with no direct human intervention. These interactions will replace the paper forms and faxes that flow back and forth from company to company today. E-business evolved to on demand and on demand has evolved to business and IT "alignment". At this stage many enterprises have bought in to the concept but are struggling with how to get there. This is why many web sites don't fully meet our needs -- they are dependent on many independent applications that the enterprise has had for decades and so far have been unable to integrate them. SOA is the new model -- it offers the first comprehensive, standards based way to get the job done. Adoption of SOA will enable the interoperability within the many functions and departments of enterprises and between enterprises that has been a decades long dream. History has shown that adoption of standards leads to an explosion of usage and that will surely be the case with SOA. The SOA standards will enable entire industries to be brought together. Virtual corporations comprised of a federation of smaller ones will enable “hyper competition” on a global scale. How does "Web 2.0" fit into all this? Like a ball and glove. Quite the hot topic in tech circles and among venture capitalists, Web 2.0 is basically a style, a model, an approach, and a philosophy wrapped together. It includes a "lightweight" programming model that is more like web page development than traditional programming. A key element of 2.0 is the blog feed -- a way to allow people to look at a web page but also subscribe to it. Another element is AJAX, a technique built on a collection of Internet standards that produces a rich user experience -- kayak.com is a good example -- with pages that don't "reload", they just change while you are looking at them. Another characteristic of Web 2.0 is that it is a perpetual beta -- users are treated as co-developers. At the conference, IBM announced WebSphere sMash which may turn out to be a really key tool for the evolution toward Web 3.0. Jerry Cuomo, IBM Fellow and CTO for IBM WebSphere, described a broad vision for how "smashups" will extend the web in a major way. The idea is to make it simple to combine content from multiple web sites. For example a travel agency may want to combine the best deals from airlines and hotels along with comments and discussion from tourists all in one "seamless" site. The smashup tool is based on a community project called "Project Zero" that has been underway for a number of months and is now ready to go mainstream. All things considered, IBM really has it's act together with regard to SOA. Every software and services executive at the company is well versed on it and has it baked into their business and development plans. The promise is great and with tens of thousands of software engineers and top management support I think it is fair to expect IBM to continue to deliver on their vision. They have already made dozens of acquisitions to fill in the white spaces and customers are signing up and getting results. There were hundreds of customers and business partners there in Las Vegas to tell their success stories. Nothing is more creditable than having someone else tell your story for you.
Conferences , IBM , Internet Technology , On Demand , Travels , e-Business April 15, 2008 08:25 PM |
|
|
Sunday, April 6, 2008 |
|
|
IBM Happenings: March 2008
|
|
|
Saturday, April 5, 2008 |
|
|
patrickWeb Blogroll patrickWeb Blogroll
Now that we all live in a world of social networking, there are many ways to share not only links to your favorite blogs, but links to your favorite anything. (There are also numerous specialized ways to "tag" stories, pictures, songs and videos). At sites such as del.icio.us you can put all your bookmarks in one place and share them with anyone and everyone. But, it is still ok to have a good old fashioned blogroll. One of the entries in the patrickWeb blogroll is The Guidewire. Guidewire Group is a market intelligence firm that is focused on technology entrepreneurship, early-stage companies, and emerging technology markets. The insight they have developed over the years is quite valuable to their clients and the community that has built up around them. The Guidewire Group analysts meet with hundreds of innovative companies each year and a subset of the companies ends up launching their product or service at a DEMO conference. There are a number of stories about the DEMO conferences here in this blog but over at the The Guidewire blog there is quite a buzz going on. There are always debates about the future of emerging technology and whether we are living in a post-bubble or pre-bubble period . Now there is a debate about the future of the emerging technology conferences. The latest story is called Let's Get Real: Business is Not Personal.
|
|
|
Friday, April 4, 2008 |
|
|
Geocaching Update
There are many GPS receivers on the market now with some breaking below the $100 mark. The hot area is GPS for cars. Many new cars offer built-in units as an option but the "after" market is much larger. One of the newest entries is Dash. The new Dash Express claims to be the first two-way, Internet-connected GPS navigation system. The device delivers traffic and destination information in a new way. You can look up somewhere that you want to go using the Internet and then have that "waypoint" delivered via cellular or WiFi signal directly to your Dash. You then select the new waypoint and the GPS will guide you to your destination. The Dash can also show you the location of all other cars nearby that have a Dash. That makes it a good proxy for traffic but what would be much better would be if all the GPS manufacturers got to together and agreed on a standard for information sharing so that each GPS could actually show the "total" traffic in the area, not just traffic of those cars that have a Dash. I have been using GPS devices for quite a few years and have or have had most of the manufacturers. On the trike, I have the TomTom Rider. On other bikes I have Garmins. For the last few years I have been using a Magellan for geocaching. They make a really nice device but I don't like their software. GPS is becoming ubiquitous but the formats for the data storage and data interchange with PCs is a Tower of Babel. Magellan is not alone -- the entire GPS industry thrives on proprietary formats that they think help them maintain market share but actually constrict the market and confuse customers. Thankfully, there is a great piece of software called GSAK (geocaching Swiss Army Knife) that is indispensable for anyone who wants to exchange GPS data with their PC. I highly recommend it. GSAK allows you to download thousands of caches from geocaching.com to an easy to use desktop application. You can then sort them, search them, organize them in various ways, see all the logs of those of have found (or not) the caches, and a Google Map to show exactly where the caches are. Once you are ready to pack up and head for the trails, GSAK allows you to easily transfer the selected cache information to your GPS. As soon as the backorder gets filled I will be ready to go geocaching with the new Garmin Colorado 400t. Looks like the Colorado will be a rugged and advanced handheld and it will be pre-loaded with detailed topographic maps with a 3-D map view, a high-sensitivity receiver, barometric altimeter, electronic compass, an SD card slot, picture viewer and a bright color display. I will be reporting on whether it is as good as it sounds or not. Meanwhile two of my Magellans went on eBay this week and hopefully they will make a new geocacher somewhere happy.
|
A reader of the 
In the 
It is 11PM. Do you know where your medical records are? Most likely they are scattered across multiple doctors' offices in 
My apologies for the serious typo in my latest post. I can not explain it. Thank you to the many who were kind enough to send timely email about the error which enabled me to fix it quickly. There are some things the spell checker doesn't catch.
The 
In a nutshell, 
The 
As of this week there are 57,370 people who have accounts at