Sentences to ponder (there is a great stagnation)

Over the span of our data, health IT inputs increased by more than 210% and contributed about 6% to the increase in value-added.  Virtually all the increase in value-added is attributable to the increased use of inputs…

Here is much more.  Can anyone find an ungated copy?


The government has mandated the replacement of cheap inputs with expensive but basically equivalent inputs. Additionally, the extra expense crowds out entrepreneurial investment in potentially high value-added IT.

The important factor here is that IT itself is a very low percentage of healthcare cost so no wonder it accounts for a small percent of total value increase.

Let's say, you replaced a $1 bolt on your car engine by a fancy $3 bolt. Net increase in "Engine bolt spending" is 200% but what is the increase in the total value increase of your car?

I can upload an ungated copy, but I guess that would be piracy.

Can't we conveniently assume it falls under the fair use academic sharing exception?

Cannot find ungated copy, sorry. The money quote is here
"Health IT appears to be very productive at the margin. Hospitals appear to under-invest in health IT despite relatively high private returns. Nevertheless, given the current state of the technology and our estimates of diminishing returns, broad expansions in health IT would have a small impact on hospital productivity. This implies that while government funding for increased EMR adoption may be welfare enhancing, they will not transform health care delivery."

Using EHRs, physicians spend 20% more time seeing the same number of patients, and quality of care improvements may exists but are hard to quantify. Because physicians are reimbursed per visit, and do not have "billable hours", marginal productivity, however measured, may thus seem to go up, but it actually goes down substantially

No surprise them that in the MGMA survery for example (, ungated), 30.6% of physician practices reported that their produtctivity went down, not up.

EHRs make doctors see less, not more, patients!


Hmm. I'm surprised about this for two reasons.

First, my mom had three ER visits with admissions during March and April, at two Inova facilities in northern Virginia. I was present only for the last ER visit but the docs made immediate and extensive use of her prior hospitalization records online records to eliminate a battery of tests. In particular since it was already online they were able to rely on previous X-Ray, CAT, and MRI imaging to eliminate what might otherwise have been more intrusive, definitely more expensive, and possibly more risky tests and procedures. Later the same information was available to offsite but in-system physicians who had previously seen her (and given more optimistic prognoses) to determine appropriate care in consultation with her admitting physician. It's hard to imagine that this kind of access was only 6% more efficient than couriered handwritten charts and imaging? But maybe so.

Second, in conversations nearly 20 years ago I sat in on a heated debate between in-laws, one a small-town GP, another a west-coast specialist, and the third an administrator of a midwest regional hospital. The administrator was adamant that docs had to quit stalling and get online, both because he constantly had to deal with what he considered catastrophic information mismatches (e.g. from handwritten prescriptions) and with rising costs overall. The docs were, um, resistant. None of them, incidentally, were (or to the best of my knowledge are) able to type. All are now retired.

My guess is that the biggest cost of health IT is dragging older caregivers (doctors and surgeons, obviously, but also staff) into the 20th Century records-keeping-wise. My GP in-law was so infuriated at having to enter a backlog of patient records that he didn't realize he'd suffered a series of mini-strokes! He hadn't noticed that the difficulty he had typing and tabbing around pages had transitioned to temporary stroke-induced aphasia! (He was only in his 50s, by the way, and otherwise healthy. He quickly recovered from the strokes but retired early because he just plain hates computers.)

One thing I'll guarantee you, though: very few individual hospital systems had the clout to force their doctors to start keeping records online. So if it wasn't for government or (the only institution more consistently and reliably less efficient or caring than government) insurance companies forcing the issue it might still not have happened at all.

For what it's worth I remember similar arguments in industry about the transition from typing pools using manual typewriters to typing pools using (pre-networked) word processors. The cost savings of switching out IBM Selectrics for 1st- and 2nd-generation IBM PCs with Word or WordPerfect can't have been more than 6%. (This was at a time when "tough minded" executives were still bragging they'd fire anyone they saw using a computer with a mouse. Because, you know, memorizing keystroke commands was so much more efficient for... the administrative staff who did all their computer use for them.) But then the real savings didn't come from increased typing-pool efficiencies, it came from doing away with them altogether in favor of direct input by executives and other employees. I expect the same will be true for medical IT -- online record keeping won't become transformational till there's sufficient mass to stop thinking of them just as substitutes for paper-based records.

Anyway, I'm not ruling out that hospital IT isn't currently underwater inputs-to-value wise. I'm very skeptical that no "critical mass" point will be found where added value begins to rise substantially.


Thank you for these anecdotes and your hopes.

Allow me to add another anecdote. I am a physician who thinks that more patients need to be seen at lower cost and at least equal quality of care. I am also a physician who had worked for 10 years in IT industry (process control) before ever seeing my first patient.
It is my conviction that physicians and healthcare in general urgently needs productivity increases. I have trulyoped that EHRs would help, and have been a strong promoter of them in my workplaces.

But the data shows that EHRs do no such thing. And my personal experience has been that EHRs require me to spend more time at the keyboard and less with the patient, making my workday 20% longer seeing the same number of patients, but with less face time with each of those patients.

One reason for our disconnect may be that you see the EHR in isolation. In my view, by being embedded so tightly into the hospital structure, the EHR has become the primary point of control for administrators, both inside and outside the hospital. Many, many regulations from such institutions as JCAHO that could not be enforced previously are now enforced throught the EHR. Where previously, ancillary personnel could document test results and I could sign off on it after review, I now have to fill out those datafields in person.

, wants more patients seen at lower cost, I have been a big beleuever

EHRs require me to spend more time at the keyboard and less with the patient

At least for me, most of my interaction with my PCP and specialists involves them asking questions, ordering tests and modifying prescriptions. As I understand it, they can't bill unless they see me in person. It would seem that interacting with more of your patients electronically would really boost productivity.

1000x this. I am frankly amazed that anyone working in healthcare would not appreciate this point. I don't even work in the field and it is apparent that the vast majority of the patient-doctor interaction is based around information exchange. The whole point of keeping decent records (which paper records have emphatically not been) is to increase the ratio of time spent figuring out what to do with the information you have on the patient to time spent gathering that data by reducing the substantial amount of redundancy currently present in the system.

I don't appreciate the specific dynamics involved in the current system, but I would imagine that it will look like a lot of IT reform projects with substantial amounts of foot dragging by people who don't want to participate in a new system with a leap of productive followed once the whole enterprise gets off the ground and runs smoothly with informed participants. Given that the relevant group that needs to modernize is the entire national healthcare industry and I can see how this may take some time. But there are countries that have already succeeded at this.

Can someone please explain what the sentence means?

At a guess, I'd say that it means there are no efficiency gains from increased use of IT.

From what I know on the technical side (i.e. limited exposure to HL7 and the massive amount of carping about it), I wonder if part of the problem isn't the extremely craptacular solutions being pushed.

Do the 100,000 lines of code inside a MRI or CT-scan machine count as IT?

This paper is a little bit early, and does not address current or anticipated changes.

For example, HMOs are changing their contracts so that they can integrate into the hospital's system and pay based on parameters that are easily monitored through the IT system. Now, I don't know if that shows up on HOSPITALS efficiency model, or on the HMOs. But, its there.

Similarly, I don't know if the savings Medicare will have to monitor to prevent against fraud and abuse will show up on the hospital's efficiencies or the payors.

Nor do I know if you are measuring the efficiency and quality control that comes with being able to quantify outcomes across hospital systems based on the adoption of different protocols. Some of the larger healthcare systems own foundations which monitor outcomes, and their ability to monitor, and identify, best practices can only come about with strong IT systems.

I visited my General Practitioner in the (English) NHS at 5:30 pm on Friday, having booked the appointment at 4:30. She assessed the problem quickly and phoned the local Teaching Hospital to seek the opinion of a relevant specialist: they talked, each consulting a copy of my notes on screen. Apart from her moue at the rather cheesy Mozart that the phone played while she had waited to be connected, all seemed to go well.

I don't advise Americans on how to organise health care, apart from the generalisation "Don't copy the NHS". But the NHS does some things well, sometimes. Of course, I can't yet tell you the outcome.


On the page you link to for "much more" is a box for your email address; enter your .edu email address and they'll send you a link. At least that's how I read it and how it has worked for me using my academic email address. /Steve

why is the nber gating its papers? It must be some violation of nsf policies, does the nsf have anything to say about that?

I would be surprised if they make more than a negligible amount from paid downloads, is their mission to limit the spreading of knowledge?

Since I can't see the entire research I am left to speculate. I hope that the point of the research is that the "regulations" are causing poor implementation of "IT" and not that there is no need for "IT". Because there is no conceivable way that a paper based system, for the whole of the healthcare industry, is better than an electronic one. There can be poor implementation of a system, but this idea that the entire concept is inferior is absolutely ridiculous. I will leave open the possibility that this is somehow that once in a millennium event, like the discovery of the periodic table, that so flies in the face of everything that everyone knows as to be an absolute game changer. The idea that instead of a tri-corder on the starship enterprise that the doctor take out the patients paper file, the last shred of paper left in the universe. And this method will stand the test of time, secure in the unshakable truth that there is no way for doctors to function using anything but paper. And that Moore's Law, continuing to enhance the rest of technology and science by analyzing massive amounts of data in shorter and shorter time frames, has nothing to offer to the world of medicine. But usually when something like that comes along, like a few months back when researchers thought they found something that traveled faster than light but realized it was a malfunction, we probably need to test this idea a little more.

And those IT companies have made billions, according to Forbes magazine. See article on CERNER et al. Government mandates provide economic opportunities.

I did find an ungated copy and there's a lot more interesting titbits in there that Tyler's summary didn't capture. e.g.

.....Hospitals' IT investments are highly productive at the margin. The median long-run net marginal products of IT inputs are $1.04 for IT capital and $0.73 for IT labor......

......Health IT's high marginal product suggests that widespread adoption may generate large productivity gains......

......Our fi ndings suggest that IT inputs are, at the margin, under-utilized..................

........Health IT utilization di ffers substantially by hospital ownership. For-pro fit hospitals utilize 83% less IT capital than not-for-pro fit hospitals.......

Overall their conclusions are far more nuanced than "Medical IT has stagnated". Of course, I quote selectively, so caveat emptor.

I think this is a different but related article; same lead author; the abstracts are somewhat different, but the titles are quite similar:

As an IT guy, not surprised at all. They are generally the wrong inputs.

Again, I have to hammer this point: there is virtually no competition to drive effective use of IT. Implementing good IT solutions is hard, it doesn't happen without a lot of pain, and people generally try to avoid pain.

Longtime reader. First time responder. I am a practicing physician whose group switched to EHR 2 years ago. In the very near future the three major hospital systems will be switching to EHR so I and my partners have to each get passwords and learn new systems for each place.

I have these observations. IT is very useful for looking up test results: xrays, blood work, pathology, dictations of previous histories or surgeries. It is ponderous and useless for daily notes. It takes a long time to record information. The government and NGO regulation authorities mandate a lot of useless documentation. It becomes impossible to find useful information. Pages of template based check boxes make it difficult to find useful information quickly and effectively. Worst of all is that none of the major systems communicate with each other. Why should they? There is no financial incentive to do so. Most major hospital chains implemented EHR for the nurses over the last 5 to 10 years. Their notes have become gigantic and impossible to read coherently. Handwriting problems have been replaced with pages of check box answers that really do not communicate effectively. Rather than improve with each upgrade they have become more ponderous because the administrators who decide on the system's features do not have to pay with daily frustration and headaches the consequences of their choices.

The EHR in my office has been in existence for more than 15 years yet they have not done basic things to allow me to view several parts of a patient chart at one time. I have to open and close pages, keeping the data in my memory to record in the new note. Back and forth; back and forth. It is awful.

The difference between EHR and other forms of IT is that it is government mandated. My vendor does not care that the product is bad. Once I am trapped there is no choice. I cannot buy another product and switch my records. The programs do not communicate with one another. I would lose all my patient data.

I have found the situation painful, frustrating, and hideously expensive. We were promised a significant savings in overhead. Despite having one less full time employee our expenses have increased significantly.

Finally, our employees have been diverted from the productive work of helping people to solve their problems to plowing through data entry tasks.

Comments for this post are closed