I started the week with a blog about the Limitations of Administrative Data, so it's fitting to end the week with lessons learned and next steps.
e-Patient Dave, his doctor Danny Sands, Roni Zeiger from Google, and I spent many hours in online and phone conversation about the data elements in healthcare that are of greatest use to e-patients. Since the American Recovery and Reinvestment Act requires patients be given access to their electronic data, I have wanted to share all data with patients, both clinical and administrative. It's clear from our discussions that sharing billing data with patients is unreliable for clinical history, and it was a mistake to do that.
Administrative data is a coded summary of the clinical care that lacks perfect specificity and time references i.e. just because you had a diagnosis of low potassium 5 years ago does not imply it is a problem today.
Thus, we must be careful about what data we send to PHRs and how that data is presented to patients. Here's the action plan that Dave, Danny, Roni, and I developed to optimize the PHR experience for e-patients:
Problem List
This is useful clinical information as long as clinicians keep it current. Danny has done that with Dave's data, so it's Dave's best current source of relevant diagnoses and ongoing treatment.
Plan
1. Remove our ICD9 administrative data feed from Google so that the clinician's problem list is the only data which populates the Conditions area
2. Continue to improve our problem list functionality in webOMR so that it maps to SNOMED-CT, enabling Google and other PHR vendors to provide medical information and decision support based on a controlled vocabulary instead of just free text
3. Change the BIDMC Google Health Upload screen from "Diagnoses" to "Problem List"
Medication List
Name (with NDC coding), Dosage/Frequency, Prescription, Date provides good "data liquidity" of active medications. We will continue to investigate the utility of sending inactive medications.
Allergy List
Name, reaction, and level of certainty of the reaction has worked well. However, Google Health does not display the detailed reaction information. We will either insert this information into the Google Allergy notes or work with Google to add a new field.
Procedures
We do not currently send procedures to Google Health, nor do they appear in Patientsite. However, Dave feels they may be useful to e-patients. We will add Procedure name and date as a pilot
A great week of discussion with many lessons learned. We look forward to our ongoing work with e-patients, doctors, and Google.
Thanks for the dialog, John. It was clearly constructive and productive. Here's to life!
ReplyDeleteI just want to add one thing so readers are clear about it. You say "Administrative data ... lacks perfect specificity." I'm not sure what that means but I think it's doc-speak for what I would phrase as "can come across as *completely wrong* information," which is the reality, as we've discussed. :–) Such data is just not fit for the purpose.
ReplyDeleteOur shared goal is to put patients in charge of their medical reality, and in that regard I have something of a Harry Truman "plain speaking" streak. :–)
In any case thanks again for the dialog and what I think is a good decision.
One last note before heading to work: I just posted an update on e-patients.net, with kudos. Thanks again.
ReplyDeleteThanks to Dave, Roni, and Danny Sands for a great dialog and a great outcome.
ReplyDeleteOur first changes - the removal of administrative data from our transmissions to Google - will go live next week.
ICD-9, -10, -11 are not only used for billing purposes, so simply not using this coding systems doesn't solve the problem where the CCR and the CCD permit the use of these codes. The solution is to differentiate the use of ICD-9 codes when they are used to convey Problems in CCR and CCD. Ditto for procedures. Provider and payer systems are generating CCRs and CCDs using ICD-9 codes derived from EHRS and payment/claims systems.
ReplyDelete"Don't",
ReplyDeleteInteresting thought. I may be wrong but I think John was only a decision to stop transmitting unfiltered/unedited ICD codes as if they were clinical history. I'd be interested in learning more about the other uses you're referring to.
My interest is in helping patients create a good electronic record of the data that they'll need for various reasons. In this discussion we haven't gotten into those specific use cases, but I'm all ears.
It's also clear to me (from my day job as well as this topic) that a data vocabulary by itself is not a complete answer. Well defined workflows, policies and practices are required for entry and editing of data, including audits and ability to track changes.
As of noon today, we turned off all use of ICD-9 codes from administrative systems in our CCR and CCD data exchanges (I tested this with my own medical record to verify it works). We're using an internal controlled vocabulary called BI-98 (Beth Israel Problem Set vocabulary 1998) that is part of the National Library of Medicine's Metathesarus. I spoke with NLM folks this morning and they will be sending us a mapping of BI-98 to SNOMED-CT, so we'll soon be sending SNOMED-CT in our CCRs and CCDs.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThis has been a great blog posting that is very timely as state Medicaid programs look at claims-based EHR/PHR options. I look forward to how it moves forward. Perhaps it may evolve into some sort of "best practices" guide that could be used by states, large payers, and others who may face similar data hickies.
ReplyDeleteIt took just 17 days from Dave's first post to John's announcement this afternoon. This should get the first prize for responsiveness.
ReplyDeleteOn April 1st, I commented on thehealthcareblog:
"To me the moral of e-Patient Dave story is that efficient healthcare reform can only occur if and when we personally own and verify our health related data.
Secondly, Dave story shows why European countries are using SNOMED mapping instead of ICD-9-CM coding. SNOMED mapping contains more than 95,000 terms for the disorders and findings hierarchies, while the ICD-9-CM disease classification has fewer than 10,000 codes."
I am just an informed lay person. But as we can see, lay people can understand more than what health professionals have traditionally accepted. There are certainly many more informed and engaged lay people than we realize. Involving them in a crowdsourcing experiment to improve the BIDMC EHR interoperability could be interesting, don't you think?
Well said, Gilles.
ReplyDeleteJust further shows how complicated all of our medical data really is... (Often walks that fine grey line between black and white). I enjoy the discussion. :)
ReplyDeleteIt was Sir Francis Bacon who warned us about the "idols of the marketplace." These are those ideas or beliefs that are so attractive that they are clung to even when they have been demonstrated to be invalid. The laudable goal of putting the patient in charge of their data has be shown to be important since the patient knew immediately that there were errors and had them corrected. However this took the interaction with professionals and emphasized that data sharing is a more intimate process than was recognized. Data collection's purpose must be clear and for personal health data it should be an interchange between that patient and the provider who is giving the care. If other practitioners are allowed to give out data not in that context, havoc will proceed, since the fundamental question arises, namely, "Who is responsible for the information and its timely accuracy?" The goal should not be open ended sharing from various sources but only in the venue when the provider-patient relationship is fully engaged. As my experienced predecessor in practice once told me as he encourage me to personally see patients in the ER asap, "Go see them, Don. Sometimes they make mistakes."
ReplyDeleteIt was Sir Francis Bacon who warned us about the "idols of the marketplace." These are those ideas or beliefs that are so attractive that they are clung to even when they have been demonstrated to be invalid. The laudable goal of putting the patient in charge of their data has been shown to be important since the patient knew immediately that there were errors and had them corrected. However this took the interaction with professionals and emphasized that data sharing is a more intimate process than was recognized. Data collection's purpose must be clear, and for personal health data it should be an interchange between that patient and the provider who is giving the care. If other practitioners are allowed to give out data not in that context, havoc will proceed, since the fundamental question arises, namely, "Who is responsible for the information and its timely accuracy?" The goal should not be open ended sharing from various sources but only in the venue when the provider-patient relationship is fully engaged. As my experienced predecessor in practice once told me as he encouraged me to personally see patients in the ER asap, "Go see them, Don. Sometimes they make mistakes."
ReplyDeleteJohn,
ReplyDeleteBe careful with utilizing NDC codes as a way of uniquely identify medications. NDC codes can be recycled. Therefore, albeit unlikely, this could create another unforeseen issue. RxNorm should be investigated.
Procedures are incredibly important and almost always overlooked in EHRs and EMRs. When they are implemented they tend to be done in a suboptimal fashion without regard to the just how complex the problem can be. However, with appropriate insight into the details of this complex issue it can be well controlled and turned into a simple problem.
Kosherfrog,
In response to your statement
"To me the moral of e-Patient Dave story is that efficient healthcare reform can only occur if and when we personally own and verify our health related data."
This is a very sticky wicked. First, there is a significant population of patients that may not have enough education to participate at this level to 'verify' their data. Second, it could be a slippery slope that puts an unfair burden on the patient for problems that they may experience. because THE PATIENT failed to check his/her information. This could translate into a complicated web of legal issues.
To continue:
"I am just an informed lay person. But as we can see, lay people can understand more than what health professionals have traditionally accepted."
I do agree with your statement here. But please understand that we do have a very large society of people who do not have your level of education or intellectual capacity. We should not disenfranchise this group of typically lower socioeconomic class of individuals by creating systems out of their reach.. We need to plan systems that help them.
JFS
JFS and Gilles (Kosherfrog),
ReplyDeleteYes, we who participate in blogs (never mind we who WRITE them) are way outside the norm.
Yet the issue still remains that we need a far more functional system, and I agree with Gilles that people can learn to comprehend a lot more than we're accustomed to thinking, especially if things are expressed in plain language, and often even when they're not.
Especially, it's amazing how much people can learn to understand when a life is on the line. I urge, urge, urge people to read Chapters 3-5 and 7 of the e-patient white paper (wiki of Ch3 is here) to understand this.
I don't want to trivialize it but to a certain extent it's like the logic of 50 years ago that said of course women can't handle a business job; they have no experience in business, and no training in professions.
My wife, now 62, was one of the first women in the Michigan State veterinary school and was chastised for "taking up a seat" because all she was gonna do was go get pregnant. 20 years later she was head of the state VMA, evaluating those guys. :)
Still, looking at my records this week, I couldn't comprehend the line items nor put them in context, without my doc's help. That's why the participatory medicine movement talks about physicians as coaches. A coach speaks to the person who's on the court in whatever language works for that player.
Just to chime in, it isn't so much that patients should HAVE to be a participant but that they should have the CHOICE to participate.
ReplyDeleteIn a slightly different context, I was talking to my daughter is a data geek (maybe even more than her mom) and does this for a living at a large teaching hospital system (med school). Her worry is that a patient will see "data" without being fully informed by a knowledgeable medical person (i.e. docter) and misread said "data". Then, she said, what if that leads a patient to commit suicide, for example, when that could have been avoided if a conversation had been ongoing with patient's doctor with better understanding when the patient learned of the whatever-it-might-be?
I can see both sides of the debate with her, but again I told her it should be my choice. And I think she agreed I should have a choice, but only after a doctor/qualified medical person had reviewed it first.
These are issues that are going to arise. Not only will the matter of choice be an issue, but timing will also be (and already is) an issue.
Thank you Robin for your thoughtful comment!
ReplyDeleteThis discussion now parallels what was discussed 14 years ago when I started ACOR. One of the first services we offered was a direct access to the Health Proffesional version of the PDQ statements (the documents prepared by the National Cancer Institute about different types of cancer and different aspects of the treatments).
In those days many doctors came screaming at me saying that we were providing very dangerous access to information that should only be seen by professionals. Robin's daughter argument was used quite a few times. Interestingly if you go to the NCI website today you will find both the patient and health professional version sitting side-by-side.
I fully expect that we will witness similar reticence about patients gaining full access to all their health data. I expect that in a few years this will become the norm, and that just like on the NCI site it will be your CHOICE to read a simple or complex version of your health data. Maybe some provision for this choice should be added to the CCR/CDA schemas.
I am playing devils advocate here: While I have a lot of respect for all individuals involved in this case, and while I fully understand why the billing codes were removed in this particular situation, I can't help thinking that the Medicine 2.0 way of doing things (open, participatory, collaborative) would actually be to leave all the information up (with the right disclaimers/warnings, of course) and let people digg/rate/annotate their information, feeding that information back to the hospital. Given that the discharge codes are also used for research etc., wouldn't it be nice to involve patients in getting these kind of data "corrected"? Isn't it what "e-patients" want? Yes, not all patients would be able and willing to do, but the underlying problem is not solved by removing it from people's eyes. On the contrary, what excites me most about PHRs is that people often only realize the mess health care is in when having access to this kind of data. So I am taking a slightly contrarian view here, arguing that this is perhaps an overly paternalistic move, where we rely on gatekeeping (intermediaries) instead of apomediaries?
ReplyDeleteWe'll, I am sure we can discuss this further at the Medicine 2.0 conference, where e-Patient Dave happens to be a keynote speaker.
Gunther, you hit a nail on the head. I don't mean to suggest that I don't WANT to see what's in my billing records - I just don't want it interpreted as clinical history.
ReplyDeleteTo the contrary, you bet your BOOTIES I want to see, because what's in those records can be used against me, at least in the US. Insurance companies feed that stuff into the MIB, which used to stand for Medical Information Bank, and now for some reason doesn't stand for anything.
The intention of the MIB is supposedly to prevent fraud (prevent a person from withholding a condition). But it's one-sided: they don't make it easy to get a copy of your records, and they consider themselves not at all responsible (and certainly not legally liable) for any errors.
Here's what I wrote about it in my blog's anniversary retrospective last November:
"The final quarter of my first year started with an unpleasant inquiry into something I think is one dirty little secret in our insurance-dominated healthcare system: the "MIB" is a databank run by the insurance industry, to protect it against fraud by you, but with no protection at all for you. This racket needs reform like the credit rating agencies needed reform, to level the playing field.
"After my first post about it, to my amazement, an MIB employee contacted me, and in an extended email exchange, simply argued with me and wouldn't answer whether anyone is responsible for the accuracy of the information they have about you. You should request your free copy of your record.
"9/5: What's in *your* MIB?"9/27: Part 2 (after the email exchange)"
I've heard that the MIB is subject to the Fair Credit Reporting Act, so I'm sure we have a right to see the information that can be used against us. See the Consumer Reports video blog post that's linked from those articles.
The Digg-style item is interesting. Heck, if I own the data, why shouldn't I be able to "opt in" for that feature?
Have e-patients the right to access inaccurate information?
ReplyDeleteAs a clarification:
ReplyDeleteI applaud the forward progress of giving patients unfettered access to their medical information. I wanted to put out a couple of points that I think gets missed by this group of sophisticated people.
That is, we need to be tuned in to the millions of people that have less capabilities than readers of this blog. You are may likely not be aware of the great educational disparity between you and many Americans because we tend to travel in circles that keep us within our own socioeconomic bubble. Let us not forget those less fortunate.
Second, whenever something new is introduced it comes with consequences. Or.. as Col. Lilly said "Give me a drug without any side-effects, and I'll give you a drug without any effect at all." We should be aware of the potential pitfalls, and do our best to reduce their impact.
JFS
"That is, we need to be tuned in to the millions of people that have less capabilities than readers of this blog. You are may likely not be aware of the great educational disparity between you and many Americans because we tend to travel in circles that keep us within our own socioeconomic bubble. Let us not forget those less fortunate."
ReplyDeleteSomeone helps these folks with everything else. Medical records are still just records. Whoever helps them with their banking, their finances (limited or not) will help them with this. Those who do not have that help may not participate. That shouldn't prevent those of us who want to participate from doing so. And socioeconomic circumstances don't limit a person's mind. I have first-hand evidence of that. They may need a free source of computer access, although I think that will be just like phones/cell phones in the near future. Everyone has one. And speaking of those phones, the technology is rapidly becoming such that these EHR/PHRs will be easily accessible with them. Don't limit your thinking about what is. Plan for what is going to be (SOON!).
As a sidenote, smart cards can hold tons of information. Why not develop a way to update them with every office visit (swipe and go?) and then develop readers for home and/or other pertinent places. Swipe and read. Quick and instant. I predict we'll carry all our information on these smart cards (and probably just one card) in the near future. They may even be part of what we used to call a phone. Think bluetooth compatibility.
Joseph,
ReplyDeleteyou are of course right to keep on mentioning all those who are/will be on the wrong(?) side of the digital side. This is a topic that is always close to my heart, since I learned years ago that ACOR users were over 95% college grads, 97% caucasians and clearly doing better than the average financially.
But I have to quote, verbatim, a comment made by my friend Susannah Fox a couple of days in response to a similar question on thehealthcareblog. Susannah works at the Pew Internet Project which studies the social impact of the internet. She leads the Project's health research and oversees the Project's digital strategy, making her the foremost expert on the specifics about health:
--------------------------
"You are right: until everyone not only has access (to technology, to information, to health care), but knows how to use it, we are playing at the margins.
But I'll quote once again from my favorite remarks by Dr. Mark Smith, president and chief executive officer of the California HealthCare Foundation: "transform care within the box, within the confines of where we are and where people living with chronic conditions are" and worry about full participation later.
Recent immigrants, people with less education, and our "oldest old" are certainly at risk of being left out of the Health 2.0 revolution, but I've also seen examples of great work being done to serve safety net populations especially if they have "second-degree" internet access through a relative or a friend.
Yes, let's acknowledge that this is cutting-edge stuff (and my survey data will confirm that in spades) but so was the Web nearly 10 years ago when 46% of U.S. adults had access and only 5% had broadband at home.
---------------------
OK. I failed again at delivering optimal communication.
ReplyDeleteRobin, I am not opposed to PHRs and getting patients more involved in their care. Its great. I love it. I support it. I practice it every day.
I was attempting to make a point that the revolution is only part of the solution to the U.S. problems with health care delivery. I did not mean to say that just because your are in a low socioeconomic class that you can't your mind is limited. However, the fact that education is important to understanding does in fact play a role in helping people help themselves. That being said. My father, a PhD in Chemical Engineering, has a very very hard time understanding much of anything he is told by his doctors. I translate everything for him.
In summary. PHRs-good. I support them.
In reference to your 'smart cards'. First.. there's nothing smart about them. Second, it relies upon patients actually carrying them. Third, it would need a standardized language and format for them to be widely applicable across disparate health systems (this does not yet exist). Fourth, its been in the works for over 4 decades without any success. ... etc. Smart cards are a failure. I don't see them succeeding 'in the near future'. A webcentric system may work. This may in fact be the PHR. But there are many road blocks and pitfalls here as well.
JFS
Joseph,
ReplyDeletesmart cards are a failure in the US ONLY! They are a great success in many other parts of the world where people don't care that the technology is not American and use them for various purposes!
The entire French health care system uses a smart card (now in its second generation) called Carte Vitale. With the help of these cards over 1 billion transactions where recorded, billed and payed in 2007.
I'm sorry that I haven't found a page written in English but I'm sure you can get this one translated:
http://fr.wikipedia.org/wiki/Carte_Vitale
There are MANY things to learn about efficient health care from outside of the USA.
So sorry to be so late to this discussion... hopefully not too late to add my buck two eighty (that's 2 cents, plus inflation)...
ReplyDeleteI see a couple of hiccups in a couple of arguments here:
1. We can't make all or nothing assumptions for patients OR payers OR providers. SOME patients will get it, SOME will realize the importance, SOME will want access to varying degrees of their information. SOME, hopefully MOST, will eventually get it. I've suggested for years that begin teaching kids to advocate for themselves in 7th grade health class. Still think it makes sense. So decisions about access and contents need to be made for those who will embrace, not for the "some" who won't ever want to.
2. Discounting anyone's ability to participate in his/her own healthcare and records is wrong-minded. If anyone had told me 5 years ago today I would be deciphering greek letters in lab test reports, I would have laughed out loud. Five months later I had deciphered them to the point where I proved I didn't have cancer. Not bad for a girl who barely passed high school biology.
3. I have this feeling that some of the fear on the part of providers (in general, no one in particular) is that this kind of participation from patients will a. take up more of their (unreimbursable)time and b. uncover some questionable (?) billing practices. Kinda like an income tax audit, on a much broader, grander scale. Uh-oh! That's what we get when we begin to uncover ICD and CPT codes and look 'em up! I wonder who got paid for what?
4. I have been a victim of the MIB. About two years after my cancer misdiagnosis, I applied for life insurance and got turned down. That's when I learned about the MIB, got my records, and sure enough - my records were a mess. Just like the credit bureaus can make one's life a living hell, so, too can the MIB. So now I teach people how to access and correct all their records.
5. EMRs, and patient access and wish to correct are like a train with no brakes running downhill. You might as well jump on board rather than stand at the foot of the hill.
OK - more than my buck two eighty, but I feel better now.
Trisha Torrey
Every Patient's Advocate
Hi, Dr. Sucher...
ReplyDeleteI guess I am talking more about something like this rather than today's generation of "smart cards". I tend to think in a "future sense", not the limitations of today. Actually, this video is accomplishable in the here-and-now. We just need a vision.
I truly appreciate your dialog. Thank you.
This has been the most intellectual fun I've ever had on a blog. Thank you Robin, kosherfrog, e-Patient Dave and.. oh yes.. Dr. Halamka.
ReplyDeletekosherfrog... I don't discount that smart cards are useful somewhere (germany included). As well, there is much to learn from the French health care system and their NHI. But there is much to fix in our U.S. system for a smart card to work. Therefore, I submit that smart cards are a tool that doesn't fit in our particular tool box. I also submit we need a new tool box. Finally, I think that smart cards do have a lot of pluses, but I also think that there has been revolutionary changes already and to come that make them less of an impact. which leads me to Robin.
Robin. Love the future stuff. Love the video link. I am submitting a grant to the NIH this week on intelligent video displays for Intensive Care Units. There is so darn much that we can do with the technology, but we are so hampered by the incredible waste of current resources in a broken system that we will have significant difficulty ever bringing these concepts into reality. There's so much work to do to fix this morass. Roll up your sleeves!
JFS
Joseph (Dr. Sucher),
ReplyDeleteAs you say this is highly intellectually stimulating. But I think these discusssions are also very helpful for preparing a better future.
I completely agree with your assessment that the system here is so dysfunctional that smart cards won't work. Unless we finally get close to a single payer system, which may happen within a few years if all else fails.
I am sure your patients are very happy to have you as a physician. Just like all of us on this blog, you belong to a minority :-)
And a big thank you to John Halamka who lets us express opinions who are not always, I am sure, to his liking. If we can keep up this conversation for the next couple of years or more, and find ways to bring in the voices of many more informed and engaged patients we (as we in crowdsourcing) should be able to constructively advise HITSP and the PHR woking group of CCHIt which so far doesn't have any real patient representation.
e-Patient Dave showed us that his story is just the tip of huge iceberg. While everybody talks non-stop about interoperability it certainly looks like present systems are not able to deliver the promised data liquidity.
As Rahm Emanuel said "You never want a serious crisis to go to waste". At the very least let's build an effective system for the millions of Americans who want to take charge of their health and health care. And build in enough flexibility to provide easier information for those who do not plan to become experts in their own care.
Gilles just steered me to a terrific April 3 post by Microsoft's Peter Neupert. It very clearly lays out the rationale for handing data to patients, which John Halamka has been supporting. It also expresses that there's not much reason to delay, and makes clear that the way to do it is not just to port the data but to add metadata.
ReplyDeleteIt may be that Microsoft, due perhaps to its massive resources and enterprise-level health products, has had the luxury of working through these thoughts more deeply. I don't know. (I did not just endorse Microsoft!) I'm just sayin', that post nicely covers the issues we've been discussing here.
Everyone has heard the story about the guy whose life was saved because he was NOT wearing his seatbelt. Did everyone stop wearing their seatbelt, because this happened to one guy - NO? This is a knee-jerk reaction and it is irresponsible to make this decision based on one person’s experience and opinion. How was this guy so surprised? The same information showed up on his EOB, and in health plan PHR. Is there something more to this story!!! Why all the publicity?
ReplyDeleteJohn – how many people are happy with the information that shows-up? Can you report on that for us too?
Jay, I don't get it – I'm "this guy" you mention; how are the data errors I found analogous to wearing a seatbelt?
ReplyDeleteYou ask "how was [I] surprised?" The story is here.
Great dialogue and learning experience. I wanted to add in that while ICD codes are not ideal, they still have some clinical utility - so don't need to throw out the baby with the bathwater. In othte words, I hope Dr. H and his team are not just turning off ICD and finishing, but are also looking to see if/how they can have value. This would be particularly important since most other institutions don't have much problem list vocabulary (eg ICD may be where all the data is).
ReplyDeleteSo what might you do to better use claims data? Years ago, during my residency - I did a project involving "smart display" of claims data. So instead of just vomiting up the data, we built a system to "prioritize" the data - the codes used the most for a patient went at the top and how recent it was used. So if you had an ICD code for Hypertension 24 times in your record, including one in the past year - it was at the top of your problem list. If you had hypokalemia once frmo 5 years ago - it would be at the bottom of the list. We had an alternate display of just the top 10 diagnoses. We presented these to Cardiologists as they saw their patients and found about 80% accuracy - not perfect, but pretty good. Similar studies have shown similar results, and of course the insurance and disease mgt companies also do similar things. Maybe you could create a google app to help do this with ICD data moving forward?
A few years ago I discussed the notion of "change" with David Gustafson, professor of Industrial Engineering and Preventive Medicine at the University of Wisconsin-Madison. At the time I learned that the students enrolled at his program needed to demonstrate the implementation of one change in an institution or organization during their time at the University. Just a single one.
ReplyDeleteWhen I talked to e-Patient Dave the morning his story was featured on Boston.com I suspected "change" was on the way but never anticipated it would be so rapid. This speed of change is inspiring and very instructive, regardless if the decision to change Google Health was right, or if the process of change was participatory enough.
Kuddos to all of YOU involved changing healthcare, one change at a time.
Kuddos to all of US!
There is another good reason to get the cleanup of data on the fast track. State Medicaid programs, and others, are moving towards the use of predictive modeling tools to better identify high-risk patients and enable us to target interventions. These tools use claims data as a primary source. If there is faulty data, it leads to faulty modeling and targeting.
ReplyDeleteThanks, all. Terrific conversation. I'm heartened to see such great exchange among "working docs," smart patients (including some with day-job data skills), and policy dweebs. :–)
ReplyDeleteBut Carlos, you mention "the decision to change Google Health." As far as I know, Google didn't (yet) change anything; as I said in my original post, my concern (if you want to call it that) was with BIDMC sending billing data as a proxy for clinical history.
Dave,
ReplyDeleteThanks for your message
I may have misread John Halamka's comment when he said "Our first changes - the removal of administrative data from our transmissions to Google - will go live next week."
Certainly not a "change" Google Health has made yet but a "change" nonetheless. Looking forward to seeing the "change" implemented next week (today is April 20th)
Carlos
The assumption here seems to be that everyone is served by medical personnel that utilize an EHR from which electronic records can be extracted. Nothing could be further from the truth. 80+% of the citizens of this country have no such records. For those people, "administrative data" is the only way that a view of their care across multiple providers can be ascertained. If the CPT & ICD9 codes are incorrect, the probable culprit is the provider that submitted them...not the insurance company that paid for the care. Finally, the physicians I talk to tell me that the most important part of the PHR record is medications. Interestingly, most Payer Based PHRs get this information from their Pharmacy Benefits manager which is the same place that RXHub gets the data and RXHub is what most major electronic prescription vendors and health information exchanges use as well. And, this could be classified as "administrative data" as well since it is based on claims for medications that are actually filled as opposed to what was prescribed by a physician (about 10% of prescriptions are never filled).
ReplyDeleteThe point is, there is no question that hard clinical data is superior. But when it is not available, which is the case more often than not, "administrative data" derived from claims is a valid and useful substitute.
Folks - we're talking with Google about changing the way allergies are displayed in Google Health, but the change to Conditions is just a change on the BIDMC side - we're sending clinical problem list data and not billing data at this point.
ReplyDeleteI will be at Health 2.0 on Wednesday afternoon and I'll be sure to discuss all the issues you've raised in these comments.
Hi JerryB - good to meet you.
ReplyDelete> The assumption here seems to be
> that everyone is served by medical
> personnel that utilize an EHR
What are you referring to? Maybe I missed something I didn't see anyone assuming that. From what (modest amount) I've seen on other blogs, everyone in the healthcare game seems to be well aware that EHR adoption is in the single digits (<10%), except for family doctors, where it's over 40%.
> For those people, "administrative data" is
< the only way that a view of their care across
> multiple providers can be ascertained.
Well, sure, but if the data's wrong, the composite view will be wrong. :-)
> If the CPT & ICD9 codes are incorrect, the
> probable culprit is ...not the insurance company
Perhaps I missed something again, but I don't think I saw anyone suggesting an insurance company was a culprit in any way. Are you referring to something specific?
> when [clinical data] it is not available,
> "administrative data" derived from claims
> is a valid and useful substitute.
Again, with respect, I must say that's only true when the data is accurate. :) Saying who's the "culprit" does nothing to change that, eh?
I hope you had a chance to read my posts on e-patients.net that started all this juicy discussion. It sure opened *my* eyes.
I ran a little program and discovered that this post wins, by far, the Oscar of the most commented post of the entire blog.
ReplyDeleteFurther analysis shows there is no real surprise since 3 of the 5 most commented posts have to do with PHRs.
This little factoid just shows that there is clear interest about PHRs among many of us.
I thought it would be a good idea to summarize and organize a lot of what has been said, from the patient side about this whole story and how it relates to forthcoming laws and system reform. See
"Applying Participatory Principles to EHRs/PHRs" http://e-patients.net/archives/2009/04/applying-participatory-principles-to-ehrsphrs.html
This has been a fascinating blog item with some real experts on the topic contributing. Bravo!
ReplyDeleteOn the issue of smartcards and the comparisons of adoption in the US and European countries such as France and Germany, lets not forget the radical differences between the way healthcare is delivered in the US and those countries. The meagre rate of adoption of smartcards in the US is not due to any inherent weakness in the capability of the smartcard technology. Instead it has to do with the fractured healthcare delivery model present in the US. Here in the US we have an open healthcare system with a myriad of payors and providers which have seriously impacted the standardization of information that occurs during medical transactions. One benefit of a nationalized healthcare system is that it is possible to gain standardization by dictum. The VA system in the US vaguely approximates what is being done in France in that it is possible to implement standards that reach the broadest base of patients in the US.
ReplyDeleteMuch of the discussion and effort regarding PHR/EHR HIE and RHIO interoperability and standaridization stem from our need to refederate the disparate information silos that stem from our private sector healthcare model.
This comment has been removed by the author.
ReplyDelete