My friend Mark Smith, who runs the California HealthCare Foundation, once wryly observed, “Have you ever noticed that the doctors who talk about how much fun primary care is only practice it one afternoon a week?” I may have become the hospitalist version of Mark’s Ivory Tower internists, but I’ll take my chances.
I just finished a two-week stint on the wards. I loved it, but I only do clinical ward attending a couple of months per year – I realize that those of you who see patients all the time may not be quite as jazzed. Nevertheless, one of the advantages of doing clinical medicine sporadically is that it leaves me well positioned to notice changes in clinical care, for the same reason that people who see your kids every six months are struck by their growth spurts.
Today, some thoughts on information technology and its impact on day-to-day hospital care.
Let me begin by describing the IT environment of my hospital, UCSF Medical Center. We don’t yet have Computerized Provider Order Entry (it’s been coming “next year” for the past 5 years), but we do have an increasingly functional Electronic Medical Record (our vendor is GE, which bought the foundering IDX company a few years ago). Our EMR contains all physician and nursing notes, as well as vital signs, labs, and x-rays.
A few minutes ago, I finished “writing” 10 progress notes. The notes were created in a homegrown application called “Note Writer” – an internet-based template system that prompts attendings and residents to fill in named fields. It asks if I gave pneumovax and whether the patient has a Foley catheter (I’ve blogged previously about how docs can easily subvert such decision support, but it’s better than nothing.) More impressively, yesterday I absently wrote, “patient received 4 U PRBCs” (packed red blood cells). The computer, remembering that “U” is a Joint Commission “high risk abbreviation,” balked, helpfully suggesting that I convert “U” to “units.” I simply clicked “OK” and – voila – the substitution was done.
After I expressed my awe, a resident sitting next to me asked a great question. “I can understand why a handwritten “U” would be a risky abbreviation,” he said. (He is right; there have been dozens of deaths when an order “Give 10U Insulin” was mistaken for “Give 100 Insulin.”) “But why would it be dangerous on the computer?” My weasely answer: “For now, we’re in this half-way world between paper and computers, and there is no way that either we or the regulators will be able to manage separate rules for each. So, even though it seems silly to flag “U” and “MSO4” on the computer, I can’t see a way around this.” He nodded the nod of someone thinking “another moronic rule” and went back to his notes.
I was also impressed that Note Writer made it easy to create each field (history, physical, x-rays), even allowing one to retrieve today’s labs and vital signs and paste them into special boxes. Before the program launched last year, we wrote our notes in MS Word and copied them into the EMR’s Daily Notes field. Virtually everybody duplicated their previous day’s Word document and modified it slightly, leading to page after page of regurgitated history, like barnacles accreting on a ship’s hull. This was particularly egregious and annoying in long-stay ICU patients, whose notes routinely contained daily reports of echos done months ago, a CT scan showing a distended bowel in a patient whose colectomy was last year, “We plan to have a family meeting soon” when the family has long since returned to China. (Read this for a hilarious JAMA satire on this Copy and Paste phenomenon.) Anyway, I was impressed that Note Writer has mostly eliminated copying and pasting by making it easier to create fresh, updated fields. A pretty apt description of what good IT is supposed to do: make it easy to do the right thing.
When people think of healthcare IT, they generally focus on computers, but it’s broader than that. Take the simple act of paging. Everybody in our hospital now wears alphanumeric pagers, linked to an internet-based paging system (“PagerBox”). With this combo, most of today’s pages contain information rather than just a call-back number. I just reviewed my last 15 pages: 12 of them were text messages; most didn’t require a call back. This is a transformative technology: a huge time saver for everybody.
So it appears that computers and other forms of clinical IT have improved our lives and the lives of our patients. Is there a downside?
Down the hall from my office on the 9th floor of the UCSF Medical Center is the Chatterjee Residents’ room, named after our revered cardiologist/sensei Kanu Chatterjee. It has a breathtaking view of the Golden Gate Bridge, eight computer workstations, a half-dozen chairs and a tattered couch. There are usually some fermenting chips and burritos on a low table in the middle. The room is filled, morning to night, with residents discussing clinical issues, their attendings’ foibles, movies, sports, childcare, and more. It is our residency’s Corner Bar, minus the Coronas (I hope).
Several years ago, I nearly sold my firstborn to get our medical service’s patients on the hospital’s 14th floor, the only one big enough to handle more than 50 patients (several other services wanted to consolidate their patients on this floor too). My argument at the time: our service averages about 90 patients (twice the size of the second biggest adult service), so having our patients on a single floor would markedly improve interdisciplinary (particularly doctor-nurse) communication. Why? After the docs saw their patients, they’d linger to write their notes. A nurse could tap an intern on the shoulder, and voila: instant collaboration. “Joe, can you take a look at Mr. Smith’s belly?” “Meg, Mrs. Garcia’s daughter is here – can you chat with her for a minute?”
Well, guess what. Since the docs no longer have to be on the 14th floor to “write” their notes, do you think that they hang out in a poorly ventilated, cramped charting room? Or that they wander down five floors to chillax with their buddies, stretch out on the couch, and write their notes in what, for them, is a far more collegial and pleasant setting. I can’t blame them, but it means that the minute they’re done seeing their patients they disappear from the floor, losing that magical interdisciplinary opportunity.
And, since the system can be accessed remotely, after a long day I might not only leave the floor; I might leave the hospital, writing my notes from my favorite easy chair, half an eye on Jon Stewart quipping on my Tivo.
Overall, I have no doubt that our IT systems have improved patient care. Notes are far more legible and far more likely to be read. Interns can collect their patients’ overnight vital signs and labs by pressing “Print Rounds Report” rather than spending 30 minutes on a frustrating Easter Egg hunt for flow charts. I don’t have to stop what I’m doing to answer a page intended only to give me a nugget of information.
But it will be vital to understand some of the unexpected consequences – particularly this issue of “dis-location” from the site of care (which I first described two years ago in the New England Journal) – and thoughtfully develop new methods and venues to promote interdisciplinary collaboration. It is too important to lose without a struggle.
Wednesday: the impact of “dis-location” on radiology, or “does anybody under 40 remember radiology rounds?”
***
No, I haven’t forgotten about the Office for Human Research Protections and the Michigan checklist study. Several of the inpatient and critical care societies are about to roll out a huge letter writing and advocacy campaign, and a steady drumbeat of articles in newspapers and blogs (all critical of the OHRP ruling) continues. Hopefully we’ll hear soon that the Department of Health and Human Services is asking OHRP to rethink the crazy way it regulates quality improvement. Keep up the pressure!
do you think the algorithms designed into the emr will fall under the same scrutiny as checklists in the icu? one could argue that they are no different? i am sure they are collecting data to demonstrate the reduction in miscommunication and errors since initiation of these safety mechanisms.
Can we hear more about the letter writing campaigns from the inpatient and critical care societies? Any particular details?
Bob,
Your insight is brilliant and I am impressed that such a busy physician has the time to provide such a comprehensive blog. Medical IT is important. However, having worked with CPOE and EMR, it is no surprise that complex care altering CPOE devices have been coming to UCSF “‘next year’ for the past five years”. While you have yet to experience the disruption to medical care by CPOE, what you have described as your foray (“Note Writer”) into EMR that, as you described, limits the interdisciplinary component, your patients’ medical care has been altered. CPOE devices have been thought to be a panacea for improving mediical care as you hear the presidential candidates debate and heed our government’s commands. However, not one of the commercially available CPOE devices has been scientifically evaluated for safety and efficacy and impact on the overall care of patients. You have astutely pointed out unintended consequences of health IT. In an earlier blog, your headline asked “Can Checklists Kill?” The science by Pronovost says no. To pose your same question (does CPOE kill?) about CPOE devices, the answer is not known, but what data is out there showed its implementation to increase mortality (by altering care) by threefold of critically ill children whose median age was nine months (“Pediatrics”, December 2005). A study by Koppel (JAMA, March 2005) reported that CPOE facilitated 22 new types of medication errors. You and your readers might also seek to peruse the Drexel website by Scot Silverstein listing 43 cases of “Common Examples of Healthcare IT Difficulties”.
This is merely another perspective, albeit a minority and probably an unpopular one, considering the blinding and preconceived enthusiasm for CPOE and EMR. I often wonder how many hospitals and hospital administrators have equity interests in the company whose CPOE device they purchased. Show me the science. In my humble opinion, it is a fallacy to assume the hypothesis that (unlike Pronovost’s checklists) costly CPOE and other EMR care altering devices will save lives is true, until robust science proves it to be true. Thus, when CPOE devices are implemented, the patients whose care is altered, indeed become unwitting subjects in this research, and should be protected, by OHRP or the FDA or both.
Best regards and continue to enlighten us,
Menoalittle
We need a good set of standards on the design of a clinical IT systems. Since most systems are highly configurable these standards need to focus on how the hospital will build and use them as well as what capabilities the software offers. Some systems out there are focused on charting to capture the highest billing level rather than meaningful clinical information (TSystem is one that comes to mind). When you follow another medical professional you should be able to get a picture of what is going on rather than a bunch of check boxes that generate canned words that sound the same with every patient and inflate what was actually examined or done. While these types of designs make documentation easy its not that helpful when you are reviewing someone elses notes (unless your the biller) and constitutes a slippery slope to documentation fraud.
We need a good set of standards on the design of a clinical IT systems. Since most systems are highly configurable these standards need to focus on how the hospital will build and use them as well as what capabilities the software offers.
The problem of developing standards is as much or more about getting the hospitals to agree to and apply them as it is about the design of clinical IT systems. The high degree of configurability is a red flag; it signals that site-specific variation is not only allowed, but demanded by much of the market, perhaps especially the larger institutions.
The prevalence and effectiveness of IT in business, particularly in areas of early adoption like accounting, has gone hand in hand with the development and adoption of well-articulated standards for the relevant business processes. There is a particular discomfort with the imposition of such standards in clinical data-gathering, decision-making, and documentation.
See this article on recent comments by Don Simborg in a piece written for JAMIA.
I’ve got just a tiny lament. We’ve got EMR, pretty much at the state where everybody else is, which is to say everything but the “soon-to-come” order entry features. But as a person who’s collected interesting and fine fountain pens for thirty-five years, I kind of miss using them to write my notes. It was a satisfying ritual in an often chaotic day.