« Vacations Are Good For You | Main | Late Night Chutzpah Watch »

July 17, 2007

Do Electronic Medical Records Help?

I often plug the possibilities of electronic medical records to vastly increase care quality while reducing costs, as they have in the Veteran's Administration System. But I'd be remiss not to point to this recent study showing that, in private practices, using electronic health records resulted in merely minor improvements in only two of the 17 tested metrics. And on one metric, the prescription of statins, they actually worsened physician performance. I don't have full access to the study and so can't tell you what hypotheses the researchers attached to their results, but the data is interesting on its own.

July 17, 2007 in Health and Medicine | Permalink


Don't you have to incorporate the EMRs with other technology, like barcodes on prescriptions, etc., to get benefit?

Posted by: Nicholas Beaudrot | Jul 17, 2007 1:30:52 PM

"data" is a plural noun.

Posted by: nolo | Jul 17, 2007 1:47:51 PM

There are two other aspects to consider when evaluating electronic medical records:

* Ease of use when they have to be used really quickly, as in a triage situation.

* Privacy concerns.

From anecdotal experience, the answer so far seems to be that such records systems are at best a work in progress.

These systems are developed by software engineers and not the people who have to use them, and can be breached relatively easily.

Posted by: shaun | Jul 17, 2007 1:54:10 PM


The key piece missing from most EHRs is decision support systems (DSS). In a simple sense, they are/will be less annoying versions of the Microsoft help paper-clip guy: they pop-up with appropriate suggestions on best practice/standard-of-care at the point/time-of-care. Only the crudest forms have really been implemented into some health IT systems, but there is a long way to go. A key obstacle is physician resistance to so-called "cookbook medicine"...

Posted by: wisewon | Jul 17, 2007 1:56:49 PM


The key piece missing from most EHRs is decision support systems (DSS). In a simple sense, they are/will be less annoying versions of the Microsoft help paper-clip guy: they pop-up with appropriate suggestions on best practice/standard-of-care at the point/time-of-care. Only the crudest forms have really been implemented into some health IT systems, but there is a long way to go. A key obstacle is physician resistance to so-called "cookbook medicine"...

Posted by: wisewon | Jul 17, 2007 1:56:57 PM

Aren't medical records just one element of a system of control aimed at getting doctors to diagose and treat "better" in the sense of paying more attention to feedback and prescribing treatment in accord with rules derived from a broader and more objective body of knowledge than the individual practitioner has access to?

Perhaps, in the absence of supervision, better information doesn't help.

Posted by: Bruce Wilder | Jul 17, 2007 2:11:35 PM

Supervision? Of a doctor? Bite your tongue. ;)

I don't think this is surprising, both for what Wisewon points out and because I think doctors have shown themselves surprisingly resistant to bringing information technology into their practices. I wonder if the real change here needs to happen during their med school training, improving their comfort levels with adapting computer technology into the rest of practice (I don't know, that may be happening already, though the med students I see here in Boston don't seem especially IT oriented even now). I agree that one thing to keep in mind here is that the development of software for practice seems very behind the curve of software used in other areas. But again, I think making the software more practice useful is getting more doctors more involved in using it in practices and helping to design what they need. And just as an aside, it always stunned me that pharma companies didn't see this as a natural place to park some of their promo money - docs who spend more time online seem ripe for pharma messaging in a variety of forms. Yes, that's problematic on another level, but you'd think pharma would like the idea. Or at least I would.

Posted by: weboy | Jul 17, 2007 2:20:25 PM

And just as an aside, it always stunned me that pharma companies didn't see this as a natural place to park some of their promo money

Pfizer, Microsoft and IBM had a collaboration a few years back...

Posted by: wisewon | Jul 17, 2007 2:24:06 PM

Thanks wisewon, I didn't know that; a friend of mine who's a Medical Director says that there's very little development - the HMOs would love to get practices more electronic, but the docs resist, is what he says. And pharma, too, is surprisingly old school on this - my old marketing company works with Pfizer on their online work, and they're only just really getting into web-based marketing in the past few years, and the focus is really on Direct to Consumer ways. Everyone seems comfortable letting docs have their way on not being more IT savvy. Then again, who can get a doc to do something the doc didn't think of him or herself? :)

Posted by: weboy | Jul 17, 2007 2:31:19 PM

Does this study refute the cost savings argument?

With regard to patient care, switching to electronic medical records will have major payoffs down the line. Computer-aided diagnosis is only going to improve, particularly once more training data becomes available. Electronic medical records would get tons more people working on this. More immediately, a large (and secure and anonymous) database of medical records would be a huge boost to public health research.

Finally, consider this, when someone is having surgery, they are hooked up to monitoring devices. What happens to all that data? It gets lost, but with electronic medical records doctors (or computer programs) can look it over after the fact. Similarly 3D body scans collect a huge amount of useful data, but without electronic medical records, how can we take advantage of all that information? Furthermore, the information is around forever, if new diagnosis techniques develop, they can be applied to data taking years earlier.

Posted by: EERac | Jul 17, 2007 2:49:38 PM

This is essentially a financial issue for most physicians in practice. Reimbursement decreases every year and that coupled with rising staff salaries and malpractice premiums, puts obtaining and implementing an EMR on the back burner. They are very expensive to implement and productivity goes down initially when it is implimented. When it is not exactly a proven benefit for the practice's bottomline it gets tossed to the wayside.

Posted by: Dingo | Jul 17, 2007 2:49:50 PM

Weboy, who are you insulting, BU, Tufts, or Harvard?

Med students and residents at major academic medical centers are by and large fine with EMR. Many of them also have access to VA systems during their training as well. In fact, old farts at these institutions are also fairly tech-savvy. But even at these centers, EMR slows stuff down. VA clinics can slow to a crawl because of documentation requirements compared to private clinics where "records" consist of scribbling or unedited dictations farmed out to Bangalore.

Those of us trained in EMR would love to never go back, however.

It's not that physicians don't like oversight any more than anybody else, it's that physicians don't like oversight from people who don't understand the unique experience of their job.

Posted by: Pup, MD | Jul 17, 2007 2:52:27 PM

"data" is a plural noun.

Off topic, but this always bugs me...

Data is only a plural noun in certain registers (scientific journals, etc.).

In a casual/conversational medium (such as a blog) data is mass noun with a singular verb.

Posted by: Dave White | Jul 17, 2007 3:13:39 PM

To weboy and Pup, MD:

You’re probably both right in a sense. The older, more established Dr’s would prefer to keep their manual processes because they’re comfortable with them. I can see how younger Doctors would be more inclined to new technology because they’ve grown up with it. We ran into similar problems implementing CRM solutions in the late ‘90’s. Aside from the fact that the value of CRM was largely overstated, the new sales reps where fine using it because they grew up on computers while the older reps were hell bent on keeping their old, black book of business.

In the arena of healthcare, the biggest difference I see it that the Doctor’s a very powerful end user that doesn’t have the type of management oversight that can force these things through. So adoption will only come organically, as younger doctors fill out the field.

Posted by: DM | Jul 17, 2007 3:31:23 PM

I thought part of the VA's big success was there integrated digital records system. Heck, it's even open source:


I'm sure for a single clinic to digitize doesn't help at all.

At my local university hospital I recently saw a nurse (no GPs), a nurse ENT (no MD around), an allergist (another nurse), and had a couple of CT scans. All them made me fill out patient histories. I asked a nurse why I had to fill out the same damn eight page form every time, and she said they're trying to switch to digital records, but it sucks, so each of the clinics in the hospital is reluctant to let their records out. I checked, and they're using a Microsoft product. Yeesh.

Posted by: Chris | Jul 17, 2007 3:48:32 PM

A well-implemented nationwide digital records system would also be an astounding tool for evaluating the effectiveness of treatments as well.

Posted by: Chris | Jul 17, 2007 3:50:28 PM

One problem is that ERM (well-designed ERM, that is) really shines in managing multiple/complex conditions--something the VA does a fair bit of; the VA system is good at making sure that the endocrinologist and the cardiologist are not working at cross purposes.

That benefit doesn't exist if both the endocrinologist and the cardiologist have ERM, but aren't using the same record.

Posted by: SamChevre | Jul 17, 2007 3:50:54 PM

Harvard, actually, Pup. :) I'm near Mass General.

And well, "unique experience" is one way to put it. :)

I'm just joshin' though, really. I wasn't sure what kind of technical training doctors are getting these days; though I think, as mentioned above, the cost issues of implementing EMR systems into many practices is the big hurdle, which is coupled with a reluctance of older docs to change.

Posted by: weboy | Jul 17, 2007 4:05:47 PM

As was pointed out before, these changes have to be made system-wide to be effective. There's no benefit if they're just in one office.

Posted by: Wells. | Jul 17, 2007 4:16:45 PM

No other profession requires anyone to work so hard for so long so inefficiently, only to be turned loose with 200k debt in your early 30s.

And the training, just watch one of those Nature shows where the big baboon beats the shit out of the little baboon until little baboon grows up and beats the shit out of the next little baboon. That's medical training in a nutshell.

Posted by: Pup, MD | Jul 17, 2007 4:33:17 PM

I would be curious to know whether the use of EMRs increased office efficiency such that either doctors were able to treat more cases in the same amount of time or were able to function with less administrative staff.

I also wonder if increased efficiency for the patient (i.e. less time filling out the same forms over and over) resulted in more willingness of patients to actually seek treatment when they needed it due to reductions in patient concern about wasted time.

A study like this also needs to account for the number of years of data available for each patient. The VA might be doing better because they've had EMRs for longer and retain their patients for longer, in which case more detailed case histories more easily accessable might have value that was overlooked in this study.

And finally, access to medical records stored in databases might make large-scale data analysis of things like comorbidity (at the general medical research level) and hospital treatment outcomes (at the local level) easier and more effective.

Posted by: Galen | Jul 17, 2007 4:38:14 PM

I think the reason the VA has done well with this is the employment model they have. They are able to make decisions like that. The majority of physicians are still private practice. The majority of EMRs that have been implimented are largely multispecialty groups or hospital backed practices. EMRs will continue to be more and more common primarily because most physicians coming out are favoring a hospital backed employed model. There will not be a landslide to use them unless they are entirely paid for by Medicare.

Posted by: Dingo | Jul 17, 2007 4:57:01 PM

Just a note here -- EMRs have potential benefits beyond their direct effects on patient care. They also have contributed enormously to the VA's ability to conduct health services research, as investigators have much easier access to much better and more detailed patient records, and can use this info to evaluate treatments, procedures, care models, etc. This is one of the main contributing factors to the VA's health services research operation becoming one of the best in the world. EMRs would also facilitate post-market drug efficacy studies and that sort of thing. In short, they would be a tremendous help with efforts to make our health system more efficient and more evidence-based.

Posted by: Andrew | Jul 17, 2007 5:26:00 PM

If you gave away a free system to docotrs that interfaces with every area hospital and every other group practice so that the entire community has one seamless integrated electronic record....

no docotr I know of would say no to that.

The key word in the above paragraph is FREE. Why would doctors want to spend 200k on a system in which they have zero guarantees that it will become the standard model for the other practices and hospitals in the community? It makes zero sense for them to outlay that kind of money on such a risky venture.

Posted by: joe blow | Jul 17, 2007 6:51:12 PM

One of the key problems with implementing EMR's at the individual or small group practice level is the mismatch between costs and benefits. Doctors absorb the costs of implementation – both capital costs and the inefficiencies associated with startup while most of the benefits accrue to payers – fewer unnecessary tests and less inappropriate care to reimburse. Even if the docs can ultimately get by with one fewer administrative FTE per doctor, the overall value proposition is far from compelling.

There is more value to be captured within hospitals, I think. In that setting, very sick patients are often seen by five or six or more doctors who often don't know what tests each is ordering or what drugs each is prescribing. EMR's could largely eliminate both duplicate testing and adverse drug interactions. Moreover, there is more likely to be adequate support staff that can do much of the data entry and data retrieval that is required to make the system work. Hospitals are also more likely to be able to afford the capital cost to buy the equipment and software in the first place as well as train the staff.

Posted by: BC | Jul 17, 2007 7:14:10 PM

The comments to this entry are closed.