Value Analysis
|
QUICK TAKE>>> |
|
From randomized controlled studies to meta-analysis research covering data from a wide range of sources, there is a vast amount of information on anti-infective central venous catheters (AICVC). These studies have examined everything from when these devices should be used to the various types of coatings used in AICVCs. So how can value analysis coordinators, infection control professionals and nurses best make sense of all the research when making purchasing decisions? Here's a no-nonsense approach to understanding the evidence. |
Last October, the Centers for Medicare & Medicaid Services (CMS) stopped reimbursing hospitals for eight "never events," or incidences that, according to The Leapfrog Group, Washington, D.C., "should have never happened and could have been prevented." While saving millions of dollars for Medicare (and the state Medicaid programs and private insurers that follow Medicare's lead), each denial of reimbursement will cut into hospitals' bottom line. The medical technology marketplace has responded with a flood of new technologies to prevent never events, ranging from computerized provider order entry systems to external fixation pins.
But many of these devices have supply chain and value analysis professionals scratching their heads, trying to figure out what's effective and what's marketing speak. Budgets are tight, and no one wants to waste money on technologies that don't work. Sales reps are glad to give out copies of studies that support their product, of course. Direct access to the medical literature can provide a more balanced view, but with thousands of journal articles being added each year, who has time to find all the relevant articles, let alone make sense of the conflicting claims?
Evidence-based health technology assessment is an efficient and reliable way to find out if the products that are being marketed for never events actually work. Evidence-based assessments are based on objective data from clinical trials, rather than on customer surveys, anecdotal experience, or expert opinion. The data are collected through a systematic search of the scientific literature and then critically analyzed to ensure that the study design and outcomes are accurate and trustworthy. Assessments also consider ancillary costs, so that supply chain and value analysis professionals can identify the most cost-efficient solutions.
Anti-infective options
Medicare's list of never events includes central line-associated bloodstream infections (CLABSI), one of the most serious health care-associated infections. Acutely ill patients, particularly those in intensive care units (ICUs), are dependent on intravenous access for administration of blood products, medication, total parenteral nutrition and monitoring, often through use of central venous catheters (CVC). Unfortunately, CVCs can also become a route for infection directly into the bloodstream. Although the exact cost in terms of lives and dollars is debated, the October 2008 Infection Control and Hospital Epidemiology estimated the cost at $3,700 to $29,000 per incident. Now that CMS has added these infections to the never events list, this is a cost hospitals will have to absorb.
Several strategies have been developed to reduce CLABSIs. Clinical studies such as O'Grady et al. (2002), Ranji et al. (2007), and Marschall et al. (2008) demonstrate that some of these interventions for which the evidence is strongest, such as adherence to standardized protocols for CVC insertion and maintenance, are widely accepted as the basic first-line requirements for preventing CLABSIs.
In addition to these core measures, a number of special products—catheters, caps and bandages—coated or impregnated with antimicrobials have also been developed for prevention of CLABSI. Anti-infective central venous catheters (AICVC) are among the best studied of these devices. There have been numerous randomized controlled trials consistently showing that many AICVCs prevent bacterial colonization of catheter tips, and several meta-analyses of those trials have found that AICVCs as a group also reduce the incidence of CLABSI.
Since it seems clear that AICVCs reduce infections, that's all you need to know, right? As is so often the case, the truth is more complex. Let's look at how a careful examination of the evidence can put the decision to implement AICVCs in context.
Do the parts equal the whole?
Meta-analysis, when well done, is generally considered to be one of the highest levels of evidence. A meta-analysis takes the data from multiple studies, puts it together, and then recalculates the result as if it was a single trial. The advantage of this approach is that it gives you more data to work with and makes it easier to demonstrate statistical significance. This is very useful in the case of AICVCs because bloodstream infections are rare and it would be difficult (and very expensive) to do a single trial large enough to prove a statistically significant difference between AICVCs and standard CVCs.
Although the companies that produce different AICVCs have done a number of trials involving hundreds of patients in which patients who had AICVCs consistently had fewer infections than patients who received uncoated CVCs, the difference between the AICVC and the regular CVC was statistically significant in only a few of those trials.
However, the weakness of meta-analysis is that it assumes that all the trials are essentially studying the same thing. In fact, the trials included in these meta-analyses covered a wide range of different antimicrobial coatings, including external chlorhexidine-silver sulfadiazine, double-sided chlorhexidine-silver sulfadiazine, minocycline-rifampin, silver-platinum-carbon, miconazole-rifampin, silver, heparin and benzalkonium chloride. Since these coatings are made of different materials, it is unlikely that they all have equal efficacy. Once they have all been combined into one meta-analysis, it's difficult to determine if one type is more effective than the others, and the average may not apply to all of the individual types.
Pooling data proves revealing
For some AICVC models, enough studies of the individual types have been done that it is possible to pool only similar studies and still derive statistically significant results. So although it's not clear whether the first generation of externally coated chlorhexidine-silver sulfadiazine CVCs was effective in preventing infection, a study published in the December 2008 Health Technology Assessment shows that we can say with some confidence that the second generation devices (coated both internally and externally) are effective. In addition, minocycline-rifampin CVCs have also been studied in enough patients that, when the studies are pooled, a statistically significant decrease in infection rate is demonstrated.
This was shown by studies in The Journal of Antimicrobial Chemotherapy in March 2007, Intensive Care Medicine in December 2007, The Journal of Hospital Infections in February 2008, and Health Technology Assessment in December 2008.
The other types of AICVCs included in the meta-analyses have not been studied in enough patients to reach firm conclusions. That does not necessarily mean that they are not effective, or that they are less effective than the better tested models, but it does mean that claims that might be made about their impact on infection rates have not yet been proven.
Timing makes a difference
Another question is when AICVCs should be used. AICVCs cost more than untreated CVCs, so you want to make sure that the extra money you spend will result in a greater improvement in outcomes than if you spent it some other way, say installing extra hand-washing stations or hiring more nurses. To get at this answer, we need to think about background rates and the number needed to treat (NNT).
NNT is the number of patients you would need to treat in order to prevent a single case of whatever it is you are trying to prevent. This, in turn, is linked to the background rate of infection—how many infections you would normally expect to have if you didn't do anything. If you have a very high rate of infection, say 10 percent, and using AICVCs prevents all infections, then AICVC use will prevent 1 infection in every 10 patients, at a median cost of $18,000, or whatever a CLABSI costs at your hospital according to Health Technology Assessment in 2008.
However, if you have already reduced your CLABSI rate to 0.1 percent, you would need to use AICVCs on 1,000 patients to prevent 1 infection. Depending on the extra cost of the AICVC and what else you might do to bring the rate down further, this may or may not be a good investment. So when the sales rep promises a 75 percent reduction in infections, you need to ask yourself "75 percent of what?" and "75 percent compared with what?"
Although most of the published trials have fairly high CLABSI rates in the control group, the incidence of bloodstream infections has been falling across the U.S. since 2001. According to a study published in Intensive Care Medicine in 2008, the average infection rate was 5 per 1,000 catheter days from 1992 to 2004, but dropped to 2.9 infections per 1,000 catheter days by 2006. In the February 2009 issue of Journal of the American Medical Association, a similar drop was noted in CLABSI caused by Staphylococcus aureus types, including methicillin-resistant Staphylococcus aureus (MRSA).
It is not clear what is driving this improvement, but it is probably a combination of better awareness and monitoring, resulting in greater vigilance and a wide range of interventions. According to a study published in the October 2008 Infection Control and Hospital Epidemiology, many of these interventions—education, hand hygiene, avoidance of the femoral site, maximum sterile barrier precautions, chlorhexedine skin antisepsis, catheter care and prompt removal—have achieved widespread acceptance and are often combined into a "bundle."
For these core measures to work, they need to be part of a continuous quality improvement initiative that includes a systemwide commitment, with checklists and enforcement. However, if done right, they can have a major impact. A January 2007 Agency for Healthcare Research and Quality (AHRQ) review of such studies found that reductions range from 50 percent to 80 percent.
In a well-publicized study in the October 2004 issue of Critical Care Medicine, Berenholtz and colleagues reduced the CLABSI rate in an ICU from 11.3 infections per 1,000 catheter days to zero per 1,000 catheter days, demonstrating that a zero infection rate is a feasible target through process improvement, without resorting to specialized devices. Despite these findings, the June 2007 Mayo Clinic Proceedings noted that as of 2005, only 32 percent of non-VA hospitals surveyed had implemented these basic measures, compared with 38 percent that were using AICVCs.
A recent study illustrates why this matters. Schuerer et al. (2007) describes how a surgical ICU reduced their CLABSI rate by 70 percent to 3.3 infections per 1,000 catheter days through implementing the basic steps described above. They then introduced second-generation chlorhexidine-silver sulfadiazine CVCs to see if they could bring the infection rate down further. The introduction of AICVCs did reduce the CLABSI rate by another 30 percent to 2.1 infections per 1,000 catheter days, but the difference was not statistically significant.
They concluded that AICVCs do have a modest impact on CLABSI, but the impact may be difficult to demonstrate once the infection rate has already been lowered through other interventions. Had they started with the AICVCs before introducing the other measures, the 30 percent reduction likely would have been significant. Therefore, you may not want to introduce AICVCs as part of a comprehensive package along with other first-line interventions because the NNT and the cost-benefit analysis for these catheters may be less favorable after you implement basic infection control measures.
Which patients?
A related question is whether AICVCs should be used for everyone or reserved for specific populations. Most of the published trials of AICVCs deliberately selected patient populations at high risk for infection, such as cancer patients, in order to ensure that they would have enough events (i.e., infections) for the results to be statistically significant. However, just as the actual impact of AICVCs may be less after instituting other infection control measures, the impact will also vary according to how likely that particular population is to become infected.
Reserving the more expensive AICVCs for high-risk groups such as cancer patients may make sense, or the efficiency of having a single contract may mean that switching the entire hospital to AICVCs is more cost-effective. The final decision will vary from hospital to hospital.
The example of AICVCs illustrates that health technology assessments based on a systematic and critical analysis of all clinical trials provide a stronger basis for decisions about new technologies than does use of single-trial results, opinion pieces, or vendor-provided literature. Hospital supply chain and value analysis professionals should be prepared to use an evidence-based approach when sifting through the flood of technologies being offered to prevent never events. In the current economic climate, it's more important than ever to thoroughly evaluate clinical benefit and cost-effectiveness when adopting new devices.
This article first appeared in the August 2009 issue of Materials Management in Health Care.
To respond to this article, please click here.
