What happens when you put 15 venous access experts in a room? Magic. That is, the Michigan Appropriateness Guide for Intravenous Catheters.
Hospitalist Vineet Chopra, MD, FACP, recently gathered clinicians from various countries and specialties, along with 1 patient, to debate and decide on the appropriate indications for venous access devices, including peripheral intravenous catheters and peripherally inserted central catheters (PICCs). Their recommendations were published by Annals of Internal Medicine on Sept. 15.
Dr. Chopra, who is an assistant professor of internal medicine and research scientist at the University of Michigan in Ann Arbor, recently spoke with ACP Hospitalist about the development of the recommendations and their implications for hospitalist practice.
Q: What motivated this project?
A: Vascular access is literally the most common procedure performed in hospitalized patients. Yet, it's never thought about that way and almost always taken for granted. It is an invasive procedure, and yet the research behind it has not been well organized. When you look closer, you begin to see that people act in very different ways with respect to who gets an IV, who gets a central line, who gets a PICC. This variation brought our attention to the topic.
We started trying to quantify this variability in hospital medicine a few years ago in a paper that we published in the Journal of Hospital Medicine. We conducted surveys of hospitalists in terms of their opinions, experiences, knowledge, and beliefs regarding PICCs, and they were all over the map. For example, some providers stated they would use that device preferentially for getting venous access in patients who had difficult-to-cannulate veins, and there were others who used a much more restrained approach. The risks of harm with various vascular devices are actually quite different. We don't really have good guidance as to what should be used when. So a document that defines best practices really seemed necessary.
Q: How do you think current use compares to what was deemed appropriate by the panel?
A: It's very hard to define what current use is. We now have 47 hospitals across Michigan working collaboratively to collect data regarding use of PICCs and outcomes. There are large gaps between what is deemed appropriate and what is currently ongoing. A classic example—1 of the recommendations deemed appropriate by the panel was [to] not place PICCs for durations of less than 5 days, unless you need to give an irritant or a vesicant that you can't give through a peripheral IV. Yet in data we have of over 5,000 PICCs, about 25% are removed within 5 days, which is astonishing. It's over 1,000 devices when you do the math.
Another challenge, it turns out, is that we physicians aren't very good at documenting the indications for these devices. For instance, a lot of them read generically, like “venous access” or “physician request.” Yes, of course, it's venous access and requested by a physician, but why a PICC specifically? If you don't know why the device was placed in the first place, it's hard to make a decision whether it was an appropriate placement or not.
It also seems as if there are 2 main instruments in the toolbox for most hospitalists when it comes to IV access: it's either an IV or a PICC. Our focus in the document was to really shine light on other devices that could also be used to meet your clinical needs but may not cause as much as harm. A classic example is midlines, which are used often in nursing home settings. They can be used to draw blood. They can be used for infusions for most antibiotics. Available data suggest that they don't cause DVT [deep venous thrombosis] nearly as often as PICCs, and they're not associated as often with infection, and yet they're highly underused. We think there are thus lots of areas for improvement and opportunities. The appropriateness recommendations help codify that in a way that's not been done before.
Q: Were you surprised by any of the conclusions the panel came to?
A: The surprise wasn't so much the conclusions but the process. It was incredible how extremely collaborative it was. I was also surprised by the intense debate that transpired in specific areas—especially on when to use a device and which device to use. The evidence in this area is quite fragmented and much of it has been generated by nursing. As you can imagine, with the cultural shifts between nursing and physicians, there's lots of disagreement about what device, when, and why.
One important issue I also didn't see coming was the difference between Europe and America. We had a number of international panelists who were from countries where nurses are privileged to put in devices other than an IV or a PICC. For instance, they can put in non-tunneled central lines and some even place ports and tunneled catheters. Some of our European panelists said, “You are asking us to make a decision among certain devices, where none of them are appropriate in our opinion.” Some of the lack of agreement turned out to be based on these nuances.
Q: What were the biggest areas of disagreement among the participants?
A: One of the disagreements we had early on was the use of PICCs for short-term periods in patients who had no other veins or needed frequent blood draws. The evidence here isn't clear. But … hospitalists often care for patients that need blood tests to aid in decision making or venous access so that treatments can be given in an expeditious fashion. Ultimately the panel agreed devices other than PICCs may be safer and could achieve these goals. But the process of discussion and deliberation was intense.
There were other areas where there were lots of nuances. For instance, in ICU patients, the evidence suggests that the risk of PICC-associated DVT is 2-and-a-half times higher [than with] traditional central venous catheters. So, recommendations for PICC use in ICU settings are different than those in non-ICU patients. However, panelists deliberated about what happens if the ICU doc is not comfortable putting in a central line. Do we really want somebody who is not comfortable being dinged for doing what they thought was safer given their skill set? Some panelists said, “If the ICU doc doesn't feel comfortable putting a line in, they have no business being in the ICU.” Perhaps, but there are ICU docs out there who often are the only critical care physician for several facilities and simply can't place lines in everyone they see. So, the panel recommended avoiding PICCs in ICU patients unless the duration of use is 15 or more days or the operator is not comfortable placing the central line.
Q: How do you get these recommendations into practice?
A: We want to start testing this in our real-world laboratory, the Hospital Medicine Safety Consortium, which now consists of 47 hospitals across Michigan. We want to figure out what works, what doesn't, how, and why. What we ideally would like to do is create a toolkit, “Here's the top 5 recommendations from this panel of ways to improve your practice around venous access and PICCs,” and actually make it simple. Hospitalists can choose 1, 2, 3, maybe all 5 of these recommendations and then track the impact of these changes.
Because vascular access is so multidisciplinary and increasingly it's nursing and radiology that are putting in these devices, we need to reach out and partner with our colleagues. A lot of our panelists who were from those specialties have actually volunteered to write op-eds and viewpoints that crystalize recommendations from the master document in Annals and create a concise bullet-point list for their specialty. I think this type of outreach and buy-in is key to start to translate these recommendations into practice.
On a larger scale, this must involve professional societies as well … including MAGIC in future guidelines. We need payers and hospitals to be able to implement this and look at cost savings and overall safety outcomes.
Without this last piece, we're likely not to get traction, despite all of the outreach we're doing. Time will tell. Some of the recommendations are dense and difficult to operationalize—when should you call the interventional radiologist, for instance, versus when should you simply place a PICC at the bedside. There are 5 or 6 different criteria and some may apply to some hospitals, some may not. But simple things like “Don't put a PICC in somebody who is on dialysis without speaking to nephrology first” are easily operationalizable and easily quantifiable.
Q: Do you see appropriateness of PICCs eventually being a quality measure?
A: Yes. What we would like to do, especially for our collaborative, is give hospitals a dashboard. We have started doing this for benchmarking purposes … feeding back things like the number of PICCs that were placed for less than 5 days, how many PICCs developed CLABSI [central line-associated bloodstream infection], how many developed VTE [venous thromboembolism]. With these data, I can begin to see how hospital use of PICCs might be appropriate versus inappropriate becoming a quality measure, much the same way we do with urinary catheter device use and complications. What would be really compelling would be to tie that to outcomes … rates of CLABSI and VTE, and that will be a principal focus of our collaborative.
Q: How did the inclusion of a patient on the panel affect the end result?
A: So much of the use of venous access is driven by patient needs and desires. So we brought in the patient's voice to help understand this aspect. We were fortunate to find a patient here at Michigan who had undergone a number of vascular access procedures and was very articulate, eloquent, and agreeable to coming on the panel. Their role was to moderate discussions and they really influenced conversations when the evidence was not clear but there were clearly patient preferences to consider. For instance, if someone needs blood draws 4 times a day and you can't find a vein, do you really think the patient would be happy or comfortable with your putting in a central line in the neck, versus a PICC or an ultrasound-guided IV in the arm? When these recommendations were being discussed, several of the expert panelists turned to the patient and said, “What do you think?” It allowed us to say, in areas where there was controversy or lack of clarity, that a discussion with the patient about that individual context should take place. Many of the recommendations thus include this perspective, which I really think helped balance out the document in a neat way.
Q: What lessons should hospitalists particularly take from MAGIC?
A: In internal medicine and family medicine training, vascular access really isn't taught. Most residents learn central line placement by seeing 1, doing 1, then teaching 1. Many of us fail to recognize that there are skills, knowledge, and a host of other devices available that can be useful. For hospitalists, this document is particularly relevant because it serves as an important learning opportunity. MAGIC shows that there is an art and science to vascular access and there is an evidence base, often not used, to support decision making. I certainly hope that this impresses upon our specialty the need to be more reflective about choices when it comes to venous access and to think a lot about both short-term and long-term risks and benefits.
Q: Do you think these recommendations will change much based on future research?
A: I sincerely hope so. This is an important first step, but it's just that, a first step. There are going to be newer devices that come out that we haven't even thought about or included. There will be new evidence regarding harm and benefit, and if and when that comes to light showing 1 device may be better than another, then I absolutely think we should look back on MAGIC and modify recommendations. MAGIC is a great way of organizing what we know, but I sincerely hope it's a living document. Only then will it serve to best take care of our patients.