When Technology Isn't the Answer

Technology is supposed to improve outcomes and efficiency especially when it comes to “health-information technology” (HIT). But it’s not always that simple. Zachary Meisel, a doctor/writer, argues that two recent studies of HIT in Pittsburgh and Philadelphia revealed some complications. “In Pittsburgh, medications were given too frequently because the computer used standardized dosing times to order medication (as opposed to using the time of the first dose to calculate time to the next dose),” Meisel writes. “In the Philadelphia study, many of the problems arose from what are known as human-machine interface flaws. For example, doctors would sometimes assume that a display of?standard doses were?suggested doses specific to the patient being treated at that moment (not the same thing!).” Meisel also worries about the loss of communication between doctors and others (radiologists for example), as such communication often leads to better patient care. [%comments]

Drill-Baby-Drill drill Team

95% of all medication is self administered by the patient or family. And patients miss doses, skip doses, forget doses, mix up their medications, travel and forget to bring their medicines, and try to save money on medicines.

Technology does not touch this.

The dosing regimen is only as robust as its weakest link.


There is a technology problem there in so much as the user interface isn't well designed if those are the problems they are having. But really that is a human factors issue.

But there is also a plain human issue -- the people aren't trained.


It's an interesting point and shows the potential growing pains as we try to adapt tech to new applications...but I find the headline a bit misleading. It suggests that tech isn't the solutions, where really Meisel is saying that badly designed tech isn't the answer.

If we look at the examples you mention, you have one where the programming is wrong in the application of dosing times and another where the basic design of the app lead doctors to make a wrong assumption about how it works. These aren't example of tech not being the answer, but humans not being able to apply the tech in the right way.

The argument that tech might take away communication is a good one though. No matter what, you can't replace the value of meeting face-to-face and spending time going over something in person.

Laura L. Nelson, Ph. D.

This is an example of the technology running our lives rather than our using the technology wisely to improve our lives.

I would like to meet the people who said that info tech would make our lives easier. I would like to go postal with them, because just as Reddy Kilowatt lied about virtually free energy in the 1960's, those experts who pushed us into wiring everything from customer service to (un)edited encyclopedias were sadly mistaken.
Paperless office? Give me a break. ROTFL.

This is all yet one more example of human technological ingenuity outrunning humans' rational choices. We can. Should we? The question of should vs. shouldn't is too rarely asked. Many people who are now unemployed are victims of this unthinking adoption of new technologies. Children who spend their time in virtual worlds rather than the real world are victims of this new technology.

Yes, we have the "information technology." It does not mean that we do not have to THINK, people, THINK about what we are doing with it and what it is doing with us and to us. Too often, the software runs us rather than us running the software.


Steve Cades

Anyone who has used a complex piece of software can explain the problem described here: Insufficient attention to users' reactions; too much design by software developers. No disrespect intended toward developers; just a very different mindset than those of users.
For hospital applications: Pay closest attention to the reactions of nurses; secondarily to physicians. It's the people who do the direct patient care, distribute meds, oversee feeding, etc., who will spot the ambiguities and plain problems with the machine-user interface. Every time a new or modified system is introduced, hire a team of anthropologists or psychologists to shadow the nurses and note every grunt and grimace, and then ask "what have we got wrong?"
At a personal level, I wish I had a dollar for every time I've had to interpret a message from a technician, or a modified screen on a revised application, for my highly intelligent, but non-technical wife. I understand the economic drivers of sub-optimal user interfaces, but by now the whole industry should see the advantages of "sweating the details" as Apple, driven by Steve Jobs, does.



There's no such thing as a problem with technology. It's either a problem with the people designing/manufacturing the technology or a problem with the people using the technology.

Eric M. Jones

"It's a poor workman who blames his tools...."

Ian Kemmish

The "loss of communication" point appears to be based on anecdotal evidence with a sample size of one. The good doctor might at least say how many lives he believes were saved by his informal chats with the radiologists, and whether anybody at all can corroborate this opinion.


operator error

Jeremy Wong

I work in health IT, so I thought I'd add my two cents. A couple of points. First, the obvious ones:

1. As always, technology doesn't do anything by itself. If it is poorly-programmed or has poor usability, it will be no better (and often worse) than the preceding technology (Likewise, no one would argue that a "more powerful" phone or computer would improve your life, if you don't know how to use it.).

2. It depends who you ask. The article calls out functionality known as computerized physician order entry (CPOE), which is simply the electronic routing of physician orders (e.g. prescriptions, lab tests, x-rays, etc.). While I'm not familiar with the research cited, I know there is a multitude of research indicating that CPOE dramatically reduces costs and medical errors.

Second, the more detailed response:

1. On the whole, electronic medical record systems (EMRs) are not very well-designed. They must be better "synced up" with the way hospital staff do their work. In the grand scheme of things, these systems are in their infancy, and there are numerous clinical, billing, and registration scenarios that they are ill-suited for. Inpatient EMRs - that is, those used in hospitals - are particularly ill-suited for many tasks because hospitals tend to use the same EMR across all its departments. As a result, the complexity and flexibility demanded from these systems is staggering. On the whole, these systems aren't quite up to the task.

1b. Conceptually, there is a way around this, which is sometimes called a "best-of-breed" approach. This simply means allowing each clinical department to independently select the EMR that is best for them. However, this leaves the hospital's IT department to effectively integrate the systems and ensure that all data is completely transferrable between the numerous systems...which is exactly why hospitals tend to choose one EMR for the entire hospital......but this is a completely different conversation.)

1c. In my opinion, ambulatory EMRs (that is, for your "average" small-practice doctor) generally work much better because the end-users typically have a greater say in selecting and implementing the system, and the practice can choose the EMR that is best-suited for their medical specialty.

2. These particular issues - seemingly involving a poorly-designed CPOE module and poor training - can be solved easily. However, these only illustrate a tiny aspect of the challenges involved with implementing an effective EMR.

The bottom line is this. No, technology is not THE answer to improving healthcare. And, yes, the poor implementation of poorly-designed technology can have a negative impact on healthcare quality. As with any complex problem, there are no easy answers.



Sounds like an issue with Knowledge Management.

Technology is several cases only magnifies efficiencies and inefficiencies already present in the system.


One other consideration - when dealing with health IT, its essential to understand the business motivation of the vendors of such systems, especially those involving medications.

Pharmaceutical companies view health IT as an opportunity to ratchet up the amount of medication prescribed and dispensed, but will apply this lever behind-the-scenes using technology vendors as the proxy.

Therefore, a lot of systems are designed to increase prescribing activity -- and their design can be influenced, subtly or overtly, by incentives offered by the pharmaceutical industry.

You'll never find health IT software that tells the user they're over-prescribing to a given patient, and yet over-use and misuse of medications is a huge problem and contributor to adverse outcomes.

The reason is the vendors are all on the payroll of big pharma.


freakconomics changed my way of thinking and analysing the facts in the way it was never before.
i am going to tell you why my laptop has been stolen:
only one of my two laptops has wi-fi connection since first one is the older one.so the new one i used to keep with me,since it has the real so called portability feature,while other one in the guest room since it has to stick with the short length extension wire modem.each one of my mates uses new one which lead to neglect of the older one and gone unnoticed even after being theft by thief.it came to notice only when one decided to play game which needed more systems on LAN.
so the only cause why my laptop went theft,is the unavailability of the wi-fi connection in the laptop which lead to neglect of it service and eventually went theft.
if i am right its the same way S.D .levitt guided us to think the root major cause of the problem in his master piece FREAKONOMICS.



I find Jeremy Wong's comments interesting, but I am concerned that the research he has read about computerized patient order entry, is not relevant to most system, just to the best systems that were refined over 3 or 4 decades. So, if you are at Beth Israel Hospital in Boston, or any of the hospitals in the Intermountain system, you're in good shape. If you have a less refined system, get used to having a crazy order for ampicillin because a doctor typed in 2695 kg to do weight-based dosing on a baby (Oops, that should have been 2.695 kg), or to have your nurse come to ask if you really wanted insulin on the patient in room 4 and not room 5 (because the user interface uses a tiny and illegible font).

Don't generalize from "best practice" systems to say that most health information technologies improve care. I do not want to go back to the days of walking down to the lab and looking up results in a shoe box. But, given costs of $50,000 to $100,000 for a private practice office, to hundreds of millions for a mid-size healthcare system, the systems should work lots better.


Jeremy Engdahl-Johnson

Where does $20 billion in waste associated with medical errors fit on list of year's big quality stories? http://www.healthcaretownhall.com/?p=3337


I have been using EMR for over 15 years now. I am still waiting for a system that works. Office (out patient) systems work mostly because they focus on documenting the examination and billing for the visit. Hospital (inpatient) EMRs are hell.

Most order entry systems do not come close to the efficiency of an order sheet, written clearly and read by the nurse or pharmacist meant to carry out that order. You will have more success getting docs to write legibly and communicate their orders better, than you will getting them to use a system that requires you to click through 10 pages of options just to order a regular diet. Or one that cannot understand what I mean by NS (normal saline) and abdominal series (a specific set of xrays). Or one that cannot intuit that if I am writing "ROUTINE" labs I mean them to be done as has been routinely done for 50 years in US hospitals, during the morning phlebotomist rounds.

The entrepreneurial nature of the US economy has dictated a disorganized healthcare IT industry. What we really need a politburo to tell us we will all use a Microsoft Windows-equivalent to do EMR. To have various companies selling their brand of EMR leads to problems with sharing information and managing upgrades. It leads to some companies using NS in their order entry programs, others normal saline and yet others saline. I get to deal with each one of these systems when I go to 3 different hospitals I work out of. Rather than remember the quirks of each, it is much easier for me to use my pen and let the clerk who knows the particular CPOE program of their employer translate for me. So what CPOE has accomplished at most hospitals is that we now have a high school educated clerk translating for the nurse and the pharmacist what I write, Whereas before computer order entry, the pharmacist and nurse got to read the order themselves.

One argument against EMR that I have rarely seen mentioned, but one that I think plays a significant role in the current systems being made too complex, bulky and cumbersome, is the information overload that we expect of /require from medical records. The systems must record that order from three days ago, written by mistake and promptly deleted, for an xray that was never done. They must record the normal temperatures, blood pressures and heart rates of thousands of patients, which become obsolete right after the next set of normal vital signs is documented. They must record each doctors repetition of the same data, over and over again. And I am not sure if it is the fault of EMR or overbearing regulatory focus on documentation, but it is not unusual these days to see nurses tethered to COWs (computers on wheels) rather than at patient's bedside.

We build the current crop of EMR to meet the requirements of tort attorneys, rather than good, efficient patient care.



"It's a poor workman who blames his tools...."

This phrase is a prime example of how a cliche demonstrates the utter ignorance of the utterer.

If tools didn't matter, professionals wouldn't spend so much time trying to find better ones.

Let's all use rocks and not complain that they don't work well. After all, we've been using rocks for tens of thousands of years.

Regina Woolley

Communication has broken down in modern medicine and I think the practitioners are very comfortable with it that way.

Consider this real scenario. An older patient has multiple doctors who do at best a mediocre job of communicating with each other. The patient goes in the hospital for an acute episode. The hospitalist takes over and may or may not communicate with the existing doctors, but probably won't. The patient is treated and discharged to a nursing home. The nursing home is completely cut off from all the previous doctors, consisting of a bunch of nurses and probably just one part-time doctor who meets the legal requirements of signing off on medicines, but who is otherwise unfamiliar with the patient. If the patient ever makes it back home, which in many cases they won't, then it is up to the patient and their family to update the original doctors on what has happened.

Ian Callum

At some point machines will provide medical care to most of the population. Hospitals and doctors are too expensive.

Drill-Baby-Drill drill Team

Simple Solution to a Complex Problem:

No missed doses, no under doses, no trips to the pharmacy, no out-of-pocket expenses, no under-serving the poor.

Patient: I have chronic headaches, low energy, unexplained sadness, high cholesterol, smoke, and have strong family history of heart attacks and cancer. And I fear I will die before my 55 birthday.

Doctor: Drink three cups of water everyday. PERIOD.