Clarification Appended: January 14, 2011
Recently, while I was working an overnight shift in the emergency department, two paramedics wheeled an elderly woman into the busy ER. She was clearly very ill: her eyes were sunken and her mouth was parched, she was slumped over and unable to do much more than moan. The paramedics told me that her family members, who had stayed home (not uncommon!), wanted to make sure we knew that she wasn't usually "like this" and that she had recently been hospitalized at a different facility where many tests and "other stuff" had been done. Unfortunately, all her records were locked up at the other hospital's medical-record room, which was closed in the middle of the night.
We had to start from scratch. We ordered a CAT scan of her brain to look for stroke, put a catheter in her bladder and gave her a chest X-ray to look for infection, and applied a rectal exam to look for bleeding. We may have ended up doing all of this anyway, but having more information about her recent hospitalization would clearly have allowed us to be more efficient and directed in her care. My colleague, another doctor, turned to me and said, "I cannot wait until HIT [industry shorthand for 'health-information technology'] makes this problem goes away."
The "problem," of course, is that medical errors and excess costs increase when health information isn't portable or easily accessible. The conventional wisdom is that electronic medical records, electronic prescriptions and electronic order-entry systems save costs and lives. Since 2009, the federal government has invested over $20 billion into improving HIT. And this month, the federal government will start doling out dollars to doctors' offices and hospitals to encourage them to adopt electronic health records. On its face, this makes absolute sense who, after all, would argue that more information isn't essential to improved care at lower cost? During the last presidential campaign, both Senators John McCain and Barack Obama called for HIT enhancements as key to fixing health care.
But an overzealous push toward HIT can also lead to unintended consequences. In Pittsburgh, a study showed that the rollout of an electronic system for doctors' orders was associated with increased odds of infant deaths in an intensive-care unit. And a study conducted in Philadelphia demonstrated that computerized physician-order-entry systems facilitated medication-prescribing errors. What was going on? 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). 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!). Or doctors sometimes picked the wrong medication or dose because they didn't know that all of their options did not fit on a single computer screen and that they could click through to a second screen for more choices.
Remember the giant whiteboard used to keep track of patients and their room numbers on the TV show ER? The whiteboard is a great example of a low-tech device that can easily be customized to maximize patient care. A colleague of mine who lectures widely on patient safety shows a picture of an ER whiteboard where nurses put teddy-bear magnets next to the names of children in order to distinguish them from adult patients. Similarly, social scientists have observed how doctors and nurses gather around the whiteboard to discuss patients and their care in an effective manner. But with efforts to maximize patient privacy and go high tech, the whiteboard has largely been replaced by computerized electronic tracking systems which often reside on small monitors and frequently have no way to display information as efficiently and elegantly as a teddy-bear magnet.
Of course, some of these problems will diminish as HIT systems themselves become better designed and more flexible (where I work, the computerized tracking system is customizable and has a big screen display so that doctors and nurses can easily gather around it). And implementation problems (when people are asked to do things differently than the way they have always done it, they mess it up at first) will get better as everyone gets used to using HIT. But there are some insidious factors that come with electronic health information that can be difficult to identify and measure, especially when the human element is removed from health care by technology. Here is an example: not that long ago, if I ordered an X-ray on a patient, I had to walk into the radiology-reading area to look at it. Often there would be a radiologist sitting there in the dark room; we would talk about the patient in a way that would often lead to closer examination of one part of the X-ray. Sometimes this conversation would lead to a cooperative reconsideration of the findings. But now, unless I have a specific question or concern, I don't have to go back to the radiology reading room HIT allows me to look at both the X-ray and the radiologist report almost instantly on my desktop computer in the ER. Sure, it is efficient and helps me see more patients and spend more time on other emergency tasks. But the result, I worry, is that I may be more likely to miss something important. Experts in the field of medical informatics attribute this error to the loss of feedback. In other words, communication should be more than just transferring information it needs to be about getting people to act in the right way. In my X-ray case, HIT leads to standardized and reliable information exchange but unintentionally takes away the conversation, and the opportunity that conversation presents to improve patient care.
So should we scale back our efforts to expand and improve HIT? Absolutely not. My elderly patient who arrived by ambulance (who, it turned out, had a colon and blood-stream infection) and the many patients like her, need to have their medical information made available in a comprehensive, reliable and efficient way. But we also need to be aware that electronic health information is not a panacea. The federal government has issued strict criteria for electronic health records called "meaningful use" which will seek to expand access to information and minimize disparities in care. But explicit plans to minimize unintended consequences of health information technology appear to be limited. HIT can change the way we deliver health care in ways that may be both hard to monitor and sometimes worse for our patients. And it is essential to keep an eye out for these influences.
Dr. Meisel is a Robert Wood Johnson Foundation Clinical Scholar and an emergency physician at the University of Pennsylvania. The Medical Insider, his column for TIME.com appears every Wednesday.
Clarification: The original version of this story has been updated to make clear that the federal government does not yet have many explicit plans to minimize unintended adverse consequences of health information technology.