Intervening on Preventable Infections with IT

One of the most common preventable healthcare-acquired infections is the urinary tract infection (UTI), most often acquired because of issues around urinary catheterization. Hundreds of thousands of preventable UTIs occur in patients every year, but there is a great deal of clinical and process complexity around trying to improve the rate of UTI infections in hospitals.

One thing is clear: programs aimed at aggressively reducing the incidence of urinary tract infections can be successful, given the right types of care delivery process analysis and change, as well as the appropriate and timely use of clinical information systems by front-line clinicians. That point was brought home recently when a group of University of Michigan physician researchers examined the problem and published a meta-study of the issue, entitled “Systematic Review and Meta-Analysis: Reminder Systems to Reduce Catheter-Associated Urinary Tract Infections and Urinary Catheter Use in Hospitalized Patients,” along with co-authors Mary A.M. Rogers, M.D., Michelle Macy, M.D., and Sanjay Saint, M.D., in the September issue of Clinical Infectious Diseases.

Jennifer Meddings, M.D., of the University of Michigan Health System, led the research. A practicing physician with a background in both internal medicine and pediatrics, Meddings practices within the UM Health System and also does regular clinical research. She spoke recently with HCI Editor-in-Chief Mark Hagland regarding what was learned in the study, and the implications of the research for the healthcare clinical IT audience.

Healthcare Informatics: One of the core process problems, it seems, is that clinicians-both physicians and nurses-tend to leave urinary catheters in place too long, and at risk to patients, for a variety of reasons. Can you discuss some of the clinical and process reasons for this problem?

WE FIND THAT WE HAVE TO EXAMINE WHY IT'S SO TEMPTING FOR NURSES AND PHYSICIANS TO PLACE URINARY CATHETERS EVEN WHEN IT'S NOT CLINICALLY CALLED FOR.

Jennifer Meddings, M.D.: Yes, the current default is for the urinary catheter to stay in place, unless several steps happen. The first step is that the doctor has to realize the catheter is still in place; step two is that the doctor has to decide to take it out; then, in step three, the doctor has to write an order for the catheter to be taken out; and step four is that the nurse has to work it into her or his daily work schedule in order for it to come out. Our team of researchers was interested in what studies had been done around interventions such as reminders and stop orders in this area. There are two flavors of reminders: one is a simple reminder; and flavor number two is an actual stop order, in which the catheter will come out unless the doctor wishes for it to stay in; and that's analogous to restraint orders. So our research involved a literature review of what researchers had found. And then we were able to pool the data, and the results were a meta-analysis.

HCI: And, importantly, you and your colleagues found that when reminder and/or stop-order systems are put in place, 52 percent of the time, an infection was averted, right?

Meddings: Right.

HCI: That's quite significant. But you also found that the culture of clinical practice has to be examined in an organization, in order for change to take place, correct?

Meddings: Exactly. In June, I spoke to a national conference of wound, ostomy and continence nurses, called the WOC Conference, in Phoenix. I spoke to them at great length, and whenever I speak with people about these types of reminders, I tell them that the reminders themselves will not likely be successful unless organizations address issues such as the convenience of putting in urinary catheters as a substitute for continence care.

In fact, nurses and physicians are often very well-meaning in placing urinary catheters. And there are some cases where urinary catheterization is very appropriate, such as when a patient has an open wound on their backside that has to be guarded against moisture. But often, that's not the case. So we find that we have to examine why it's so tempting for nurses and physicians to place urinary catheters even when it's not clinically called for, when such catheterization takes place, for example, because these patients will require more linen changes and more baths. So you have to give nurses more power.

And when hospitals are trying to implement catheter-removal protocols, they have to specifically address some of these issues, including the people-power issues, but also with patients who are morbidly obese and hard to turn in bed, and with patients with dementia. In some cases, these issues can be addressed using a method called prompted voiding, or through the use of continence garments or bedside commodes. We recommend that clinicians get together and make sure that the resources they need to provide the care in place of the catheter, are available; and in doing so, they have to address a lot of the workflow issues that make the catheter right now very tempting and convenient.

HCI: So when it comes to establishing reminder systems, reminders around catheterization need to be embedded into the EMR in some way in order to be successful, right?

Meddings: Yes, for hospitals that have an EMR. One of the nice things we found in this study is that several hospitals worked verbal or written reminders into their ordering processes. So we don't think this can only be successful in an automated environment. But creating an order with a reminder or stop order is a little trickier than appears at first glance. Many times in EMRs, if you put in a reminder that will be generated with a catheter order or antibiotic order, etc., you have to remember that that reminder pops up on the screens of the primary care team members. We found when we initiated this that the reminder would pop up on the screen of whoever was viewing the record at the moment, so it could pop up on the screen of a physical therapist, for example, and then never reappear.

So the first thing we learned is that the reminder has to be made to pop up on the screen of the correct person. The next thing we found is that, in some institutions and units, it's common practice to place catheters without orders, especially in fast-moving environments such as emergency departments. The intention is good, because they want to care for patients. Now, for units like the ED in which orders are not created in every instance, you can get around that by using a sticker that you put on the back of the catheter bag, so that the person putting it in has to sign the sticker and indicate the reason, which encourages the catheters to go in for an appropriate reason. That's a low-tech solution, but works in environments that are fast-paced.

But in most units, there is time to put in a full set of orders. And in those units, we like to specify that the patient actually needs a catheter. It would seem apparent that there should be a logic for it, but I can tell you that I've worked in units where that is not the case. But by requiring a physician order, that will assign responsibility to a physician. And that's good, because when you put catheters in without orders, it's hard to change practice.

Here at the University of Michigan Health System, we've actually created a multidisciplinary catheter-associated UTI committee, and we look at hospital-wide rates of catheter-associated UTIs, and design interventions to improve our practice overall in this area. And one of the first challenges this committee looked at was that we had a simple electronic reminder in place, but it didn't seem to be impacting infection rates. And what we learned is that sometimes, you'll design a computer order entry in your head as a clinician, and then it gets sent to the programmers, and then the actual order that comes out may not be what you expected. And it wasn't the fault of the programmers; we didn't understand how it would actually function. We studied this, and found that it was very easy to order an indwelling or Foley catheter, but it was very hard for us, even on this committee, to find alternatives, such as an intermittent straight catheter, a short catheter that goes in and out and doesn't stay in the patient; or what's called an external, or condom, catheter, which male patients can actually use, since it doesn't go inside the bladder.

WE'RE VERY INTERESTED IN NURSE-EMPOWERED STOP ORDERS. THAT TYPE OF ORDER WILL POP UP ON THE NURSE'S SCREEN, AND THEN THE NURSE WILL BE GUIDED BY THE STOP ORDER.

And these are two nice options that have a lower risk of infection, but if it's hard to find those orders, they won't be written. So you have to make sure you don't inadvertently make it easier to order an ingoing catheter rather than alternatives. And once you start using an EMR, you can inadvertently increase your rates of UTI by having a catheter order be embedded in standard inpatient orders. Only a minority of patients needs a catheter; and it's not appropriate for a standard default order set to have a catheter in it.

And you can create so-called ‘smart orders,’ which really provide some of the greatest value in an EMR. Once you've written an electronic catheter order, you can do some very intelligent things, such as guiding physicians and nurses to choose a specific timeframe (such as every 24 hours, or 48 hours, or post-op day one or two, for example), and then embed a reminder in the system so that the EMR asks the clinician whether they want to keep the catheter in, when that period of time has elapsed. And you can direct it at either the physician or nurse, with the option of touching the screen to keep it in. You can also create a stop order, which, when it's generated, orders that the catheter be taken out, unless the patient meets a criterion or set of criteria at the time the reminder is generated and the clinician sees that reminder. That type of order is very analogous to restraint orders or antibiotic restriction orders; those are orders that do have stops in place, and physicians are familiar with those.

What's more, once you choose to embed a stop order, you can choose to have the stop order directed either at doctors or at nurses. We're very interested in nurse-empowered stop orders. That type of order will pop up on the nurse's screen, and then the nurse will be guided by the stop order. It will say, your patient still has a urinary catheter in. Please remove the catheter, if your patient no longer meets the following criteria. And the nurse will get this in the morning, and the nurse does not have to contact the doctor to get the doctor to write an order, because in most cases, there isn't a clinical reason for it to be required for the physician to do a stop order. And all of this can be designed to take into account the exceptions, such as when a patient has had a certain kind of surgery or has been on a very specific medication.

HCI: What are your recommendations to IT people?

Meddings: My recommendations are to be careful in the design of the catheter orders and catheter stop orders. I recommend a stop order if they have the capability in the system, rather than reminders, because reminders can be ignored, and the stop orders cannot. But we suggest that the IT people work with the clinicians in designing the order and really look at it carefully before they implement it, because often what they design turns out to be not the final product. And it's not as simple as it appears, particularly when it comes to taking previously handwritten orders and automating them within the EMR.

The last thing we advise people on, and which we're still learning about ourselves, is that just because you've written a nurse-empowered stop order doesn't mean that nurses will be comfortable removing catheters without physician contact. So we recommend that nurse leaders work with floor nurses on issues around their personal discomfort. This often involves not just a learning experience for them, but a new skill or practice, and they need to be comfortable with that new protocol in order for it to succeed.

Healthcare Informatics 2010 November;27(11):47-50