Trying to link ward-level hand hygiene compliance and healthcare-associated infections

I've spent 15 years studying hand hygiene, so I obviously think it is critical to safe healthcare delivery. As I mentioned in my prior post on hand hygiene and HAI, sometimes these posts are difficult to write. The difficulty stems from the inertia required to confront dogma, while simultaneously bracing for the inevitable criticism. And of course, I could be completely wrong. Often times, dogma is correct.

But as I've gazed out of my office in the old Singapore CDC (soon to be replaced by a shiny new NCID), I've had moments to consider various causal diagrams linking hand hygiene to various outcomes, like CLABSI. (see below or source) If you carefully examine (click to expand) this or other causal models you see that hand hygiene is there, but it is only one of many possible causes of CLABSI. So, strictly speaking hand hygiene is in the causal pathway to CLABSI development. That's the dogma and it's true, to a point.

But let's move on to my contention: ward-level or ICU-level hand hygiene compliance changes can't be linked to reductions in HAI. For example, no amount of raising hand hygiene from 0% to 100% can be associated with reductions in HAI, such as CLABSI. It's just not mathematically possible. Sure, some study might show such an association, but I wouldn't believe it.

So to borrow a strategy used often by fellow blogger Mike, I'll use math(s).

Most facilities monitor hand hygiene compliance with direct observation. On the ward level, we reported that less than 30 opportunities/ward/month are collected. That was in 2012, so let's say things are much better and we observe 100 opportunities. This actually doesn't matter - you could observe 10,000 per month with an automated system, but let's stick with 100 opportunities.

Now, let's estimate how many opportunities are related to HH moment #2 (before aseptic procedure). Most estimates that I've seen are close to 10%. And how many of HH moment #2 involve directly manipulating a central venous catheter - let's go with 2%. You can estimate a lower or higher rate depending on ward acuity, but I'm going to stick with 2% since the vast majority involve peripheral lines. So, 2% of 10% is 0.2% or 0.002.

So, how many months of 100 observations/month are required before we witness one opportunity where the HCW touches a CVC? Answer: 5 months.

Now over those 5 months, let's assume we have observed a hand hygiene compliance of 50%, so 250/500. Let's also assume the worst and say that HCW were 0% complaint with CVC-related moment #2 in those 5 months. Now, let's assume they were 100% compliant over the next 5 months after we targeted a hand hygiene education program to moment #2. Our compliance would increase to 251/500 or 50.2%. Any other increases in hand hygiene would not be in the causal pathway for CLABSI, so even if compliance shot up to 80%, we would only care about the 0.2% increase.  In fact, this highlights why it's difficult to link ward-level hand hygiene compliance to reduced CLABSI, since most of the increase does not involve CVC-related moment #2. It's almost all noise.

And if you still want to install the automated monitoring system, you can multiply the numerator and denominator by 100, and still have 25000/50000 (50%) with an increase to 25100/50000 (50.2%). And if your hand hygiene education was super successful and compliance increased to 80% (40000/50000), it would still be true that only 100 of the 15000 additional compliant opportunities would be CVC related. 100/15000 is 0.67%. Thus, CVC related hand hygiene opportunities are a needle in a haystack.

I encourage you to check my math, choose different rates or numbers and correct me in the comments below or on Twitter. Sadly, it's hard to link ward (or ICU) level hand hygiene compliance to ward-level CLABSI rates. But as I've said before, keep washing your hands and monitoring hand hygiene compliance in your hospital. No one wants a CRE outbreak.

Hand hygiene doesn't prevent healthcare associated infections

Not all transmission leads to infection and not all infections are preceded by transmission. Hand hygiene prevents transmission, not infection.

....some posts are just hard to write.

One of the persistent beliefs in infectious diseases and infection prevention is that hand hygiene compliance prevents healthcare associated infections. Perhaps this harkens back to Semmelweis and the prevention of puerperal fever through hand disinfection. Of course, if puerperal fever was a CDC HAI and clinicians didn't wear gloves, we could still say hand hygiene prevents HAI. However, that's not the current reality.

CDC defines HAI as CLABSI, CAUTI, SSI and VAP. We can even consider hospital-onset BSI and almost any other infection we can track using CMS or EMR data and monthly aggregate hand hygiene compliance is not a significant component in the causal pathway for the development of an HAI.

Sure, hand hygiene/sterile gloves before catheter insertion and hand antisepsis prior to invasive surgical procedures are standard practice. However, when I talk about hand hygiene compliance, I mean monthly hand hygiene on room entry/exit or following the WHO 5 My 5 Moments during care on medical wards and in ICUs. And yes, there are instances where Moment #2 - before clean/aseptic procedure could potentially reduce CLABSI, but the proportion of CLABSI caused by such breaks in moment #2 pale that occur outside of the insertion bundle pale in comparison to those prevented with the highly effective CLABSI bundle. Otherwise, monthly aggregate hand hygiene compliance would have been included in the CLABSI bundle. It wasn't.

Let's discuss SSI prevention. Do we really think that interns and nurses practicing hand hygiene on the wards prevents SSIs to any measurable extent compared to pre-operative CHG bathing or peri-operative antibiotics?  No, I didn't think so.

How about we look at this another way. If you were called by a CT surgeon because of an outbreak of SSI in CABG patients or an outbreak of CLABSI in her ICU, would you first (or second or third) start a hand hygiene campaign? I assume no and thus, you don't think hand hygiene prevents SSI or CLABSI. 

Thus, for all practical purposes, we won't be able to do studies associating improved hand hygiene compliance on the wards or ICUs with reduced infections. Even when such studies are done and do show an association, they have minimal basis in causal reality. Requiring hand hygiene bundles and intervention studies to show reduced HAI is incorrect and counterproductive. Since hand hygiene on wards and ICUs is not in the causal pathway for HAI incidence, we shouldn't expect hand hygiene to prevent them.

But all is not lost. Hand hygiene does prevent MDRO transmission (and indeed transmission of susceptible pathogens) in healthcare settings. Hand hygiene is critical to tackling the MDRO crisis but these benefits aren't currently captured by CMS and most EMR systems. To document the benefits of hand hygiene, we would need to complete surveillance for important pathogens on admission and discharge and document acquisition or transmission. This is expensive and likely not necessary nor feasible.

Keep your heads up and continue to drive hand hygiene compliance. Continue to do hand hygiene surveillance and improvement studies! Hand hygiene is critical to MDRO prevention and likely the future of healthcare. Just stop it with the HAI target.

Addendum: This post was written in response to the question: "Do you care about increases in monthly hand hygiene compliance if you can't document reduced HAI?" I would answer yes. Hand hygiene is an important clinical outcome in itself and requiring HAI reductions is a trap. Don't fall into that trap. I've attempted to explain why here.

Addendum 2: In response to this post, others have mentioned CDI as an HAI that could be targeted with hand hygiene interventions. As Dan mentioned back in 2013, CDI might not be the optimal target since a minority of cases appear to be related to in-hospital transmission. This was shown back in 1994. Stewardship might be a more appropriate intervention for CDI prevention.

Public Reporting - Do we need big brother?

I was fortunate to be an invited speaker at the 2016 ECCMID meeting in Amsterdam last week. My topic was "Monitoring Process of Care: Do We Need Big Brother?" I used the opportunity to take a big picture view of public reporting of HAI and MDRO data in the US and Europe and a closer look at the selection of process versus outcomes measures for reporting. I've posted my slides below and you can also listen to my talk on ECCMID's website. As I believe Dan stated earlier, I hope that all meetings evolve to allow free/open access to presentations like ECCMID has.

Where did all the pediatric HAIs go?

In keeping with the pediatric theme this week, there is a nice study just published in Pediatrics outlining the general trends in Pediatric HAIs between 2007 and 2012. Stephen Patrick and colleagues from Vanderbilt and Boston Children's Hospital used CDC NHSN data from 173 NICUs and 64 PICUs to track CLABSI, VAP and CAUTI rates.

The good news is that CLABSIs declined from 4.9 to 1.5 per 1000 central-line days in the NICU and from 4.7 to 1.0 per 1000 CL-days in the PICU. There were also significant reductions seen in VAP in both NICUs and PICUs. CAUTIs were not adequately reported in NICUs; however, in PICUs the authors found that CAUTI rates did not change significantly over the 6-year study period. I've included the NICU and PICU figures below. Importantly,  CLABSI rates were twice as high and VAP rates were over three times as high in very low birth weight ( < 1500g) infants.

Apart from highlighting recent successes in reducing HAI, I think the main messages (as I've already stated elsewhere) are that HAI are still very common in VLBW infants and that rates remain above zero in all pediatric populations. Thus, despite cultural changes, implementation of insertion bundles and technological innovations, such as antimicrobial-coated catheters, we still need additional evidence-based methods to prevent HAI in children. While we might take this occasion to rest on our laurels and celebrate past successes, we should instead increase funding for development of HAI prevention interventions, particularly in VLBW populations.

And while policymakers might think we can mandate further HAI reductions through such things pay-for-performance, it is clear from these data that those tools have already done their job yet there's much more hard work to do. Notice the almost flat rates in both figures over the past 3-4 years. We need more than quality-improvement goals. Without scientifically proven methods to further reduce HAI, I suspect rates will remain stagnant. That would be unfortunate.

Left to our own devices

The big news today in healthcare-associated infection (HAI) prevention is the publication of the CDC’s Emerging Infections Program (EIP) point prevalence survey of HAIs, which includes burden estimates and an update on the epidemiology of HAIs in the US, circa 2011. Simultaneously, CDC released an update on national and state-level progress in HAI prevention. The CDC’s press release provides the bottom line messaging around these data: (1) we’ve made progress, and (2) we still have a long way to go.

One of the most important messages can be found in the abstract of the EIP point prevalence survey paper: 

“Device-associated infections (i.e., central-catheter–associated bloodstream infection, catheter-associated urinary tract infection, and ventilator-associated pneumonia), which have traditionally been the focus of programs to prevent health care–associated infections, accounted for 25.6% of such infections.”

The device associated infections (DAIs), particularly CLABSI, are also the HAIs for which the most progress has been made in prevention. Why? Because we have prevention approaches that have been tested and implemented in most US hospitals. A common theme around the remainder of infections (now the great majority of HAIs) is that we have far less understanding about how exactly to prevent them (case in point: non-ventilator associated healthcare-associated pneumonia). 

If we expect to see further substantial reductions in HAIs, we’ll need more funding to support prevention studies for HAIs that aren’t device-associated, and for studies of prevention approaches that address HAIs that are beyond the reach of our rudimentary approaches to DAI prevention (e.g. bloodstream infections sourced to gut or skin in high risk patient populations like burn or bone marrow transplant). 

The assumption that we already know how to prevent most HAIs is patently ridiculous, and over the next few years we will see rates plateau as we gain the maximal benefit from improved hand hygiene and DAI prevention bundles. The next phase of infection prevention will require novel approaches.

One way of Getting to Zero

Favorite quote: "I can disagree with gravity but it doesn't mean that it no longer applies to me."

 

Re-designing rooms and re-engineering care to reduce infections

There's an article in the Wall Street Journal on the hospital room of the future. It notes elements of architectural and product design that can reduce the transmission of infection. Take a look at the graphic above. You'll note, for example, that on entering the patient room the sink lights up in red to remind healthcare workers to wash their hands. There are a number of other examples of using design to reduce infection in the graphic. One not specifically designed for infection prevention that may still be useful in preventing infection is the large video monitor on the footwall of the room. Let's say I'm a consultant and I have influenza. I could still interact with the patient without being physically present, not feel guilty about staying home, and not risk transmitting influenza to the patient. Obviously I could not perform all the elements of a physical exam, but not all follow-up visits require an exam, though I have heard rumors that some doctors perform perfunctory exams for billing purposes only. Moreover, does every consultant and medical student need to exam the patient every day? Think of how many contaminated stethoscopes and dirty white coats are touching patients every day with the majority of those interactions adding no value to the care of the patient. Perhaps one physician could examine the patient and the rest skype in.

I've been thinking lately that we really need to carefully exam all the things we do in hospitals and then engineer out the opportunities for transmission of infection. I was reminded of this by a paper in Transactions on Healthcare Systems Engineering, which examines something as simple as how ICU nurses cover patients for each other while on breaks. This paper points out that by providing a structured coverage arrangement the number of unique persons interacting with any given patient are significantly reduced, which potentially reduces transmission of infection. I bet there are many such  examples where re-engineering patient care could reduce the potential for infection transmission.

Graphic:  Wall Street Journal

HICPAC lays down the letter of the law

This week's Annals of Internal Medicine has a paper entitled "Public Reporting of Health Care–Associated Surveillance Data: Recommendations From the Healthcare Infection Control Practices Advisory Committee." It's a somewhat misleading title, as the paper isn't about public reporting, it's about what happens in hospitals before HAI data are publicly reported. The primary message is this: if a case meets criteria for HAI sensu stricto, it's an HAI, dammit! Don't go asking a doctor for her clinical opinion on whether it's an infection and then erasing that case from your line list if she says, "doesn't look like an infection to me."

I have mixed feelings about this. On the one hand, in order to allow for valid inter-facility comparisons of infection rates, everyone has to play by the same rules. I think we all get that. But it's incredibly frustrating to review a case that is clearly not an HAI, yet be forced to label it as such. And it's not an uncommon occurrence. Last week, my IPs brought me a case of a patient admitted with pneumonia and because the patient's condition worsened after admission, we were forced to label the case a possible VAP. This morning, an IP from another state emailed me a case of a patient who was admitted with an infected wound, went to the OR, and subsequently developed a surgical site infection, which had to be categorized as an HAI. The IP notes, "the same patient, without the surgery, would have a community acquired wound infection and would not be counted as an HAI."

Even more commonly, hospitals with large oncology populations see many cases of bloodstream infections with enteric flora in patients who just so happen to have a central line. While I give CDC credit for now allowing us to classify cases as mucosal barrier injury related bloodstream infections, it's of little value, as these infections are still publicly reported as CLABSIs. And to add insult to injury, those poor IPs in Pennsylvania have to send patients a letter telling them they suffered an HAI that wasn't an HAI.

All of these problems with post-ascertainment veto and adjudication are occurring because the stakes are high. Most of the time, this is done in good faith, I believe. There is a big push in hospitals to hold staff accountable for adverse events, and it really stings to have the finger pointed at you for an event that was not preventable, or maybe not even an event. It undermines the credibility of IPs and hospital epidemiologists with clinicians when you call a single positive VRE blood culture in a neutropenic leukemic patient a CLABSI. To mitigate that, I find myself appearing at committee meetings to explain that while this case technically meets the criteria for CLABSI, all evidence tells us this is an infection not related to the central line, and is in fact, not preventable. Some hospitals keep two sets of books--the official publicly reported set, and the internally "correct" set.

So while I agree with HICPAC in spirit, this paper only addresses a part of the problem. It ignores the fact that we need definitions with more specificity, and those definitions are needed now. The mucosal barrier injury infection definition has been validated, so let's use it for public reporting. Task IPs to send in descriptions of cases where the definitions are not working, catalog them, categorize them, and start fixing the definitions in a timely manner. Allow for relatively rapid tweaking of definitions instead of acting as if definitions are carved in stone and represent some absolute truth. While we will never be able to have perfect case definitions in the murky world of medicine, fixing the underlying problem to the degree that it can be fixed would decrease the drive for post-ascertainment veto and adjudication. Or how about embracing adjudication but have it occur at a central level? In a world of electronic communication it wouldn't be that difficult.

And one last thing: I think that any official statement from HICPAC should be in the public domain, not behind a journal paywall. Perhaps this paper is also posted somewhere on CDC's website, but I was not able to locate it, if it is indeed there.

New CDC report on antibiotic resistance

Yesterday, the CDC released a new report on antibiotic resistant infections (full text here). It's a nicely done, mostly nontechnical report that seems primarily designed to raise awareness. There's not much new in the report except some updated estimates on the impact of these infections, which probably explains why the New York Times relegated its coverage to page 13. That piece, which quotes Eli, notes that the impact estimates are lowish, and CDC officials acknowledge that the estimates were intentionally conservative. As we've noted previously in this blog, we need more funding for prevention studies, new drugs, and better stewardship to protect the drugs we currently have.

Annual cost of HAIs in the US = $10 billion

A new paper in JAMA Internal Medicine aims to give us an update on the impact of healthcare associated infections in the US. Given all the focus on HAIs, particularly with the rollout of value based purchasing, this is an important study. The investigators performed a systematic review, which included papers that were published this year, to determine attributable cost estimates, used NHSN data for incidence estimates, then performed Monte Carlo simulation. Total cost of the 5 major infections (CLABSI, CAUTI, VAP, SSI and C. difficile) was estimated to be $9.8 billion per year. The table below shows the breakdown by infection type.

Eli may want to comment on the methodology of the study and the validity of the results, but I suspect these numbers will be cited frequently.

The copper kerfuffle, continued:

Several months ago we invited a guest post from Stephan Harbarth and Matthias Maiwald, a post that questioned the biological plausibility of a recently published clinical trial of copper surfaces that claimed a more than 50% reduction in the rate of healthcare-associated infection (HAI) and/or MRSA-VRE colonization. Now Harbarth, Maiwald and Stephanie Dancer have published a more extended critique of the study in a letter-to-the-editor in the September issue of Infection Control and Hospital Epidemiology. I encourage you to read the letter, and the reply from the authors, and make up your own mind about the validity of the findings and the transparency with which the authors reported their outcomes.

My two cents: the authors’ have been caught out in a case of selective reporting (or at least egregious obfuscation) of their outcomes, and it would probably have been better for them not to issue a reply (mostly because the reply is not persuasive). The reply claims that “any HAI” wasn’t reported because it would have also included some patients with colonization, and the development of infection versus colonization may be biologically different. But it is still OK to report the outcome “HAI and/or colonization”? They also make the case that Harbarth, et al cannot question the biological plausibility of their findings because they themselves have also argued that the environment is a source of HAI pathogens. This of course misses the point entirely. Harbarth, et al aren’t arguing that the environment has no role, they are arguing that it is implausible, given what we know about the pathogenesis of HAIs, that the environment has the major role suggested by the findings of this study.

A well-designed, persuasive, multicenter, randomized controlled trial that demonstrated a greater than 50% reduction in HAI by changing high-touch surface composition should have been published in a very high impact journal (e.g. JAMA, NEJM, Lancet), and the findings should have been front-page news in major media outlets. Alas, that didn’t happen with this study, for some of the reasons outlined in the letter by Harbarth, et al.

The Environment and HAI – Where does Biological Plausibility Come In?

This May's Infection Control and Hospital Epidemiology (ICHE) contained a randomized trial of copper-coated surfaces in ICU settings which reported a 50-70% reduction in several aggregate outcomes that included hospital-acquired infections and colonization with MRSA and VRE.  In this guest blog post, physician-scientists Dr. Matthias Maiwald from the KK Women’s and Children’s Hospital in Singapore and Dr. Stephan Harbarth from University of Geneva Hospitals in Geneva, Switzerland question the plausibility of these findings and put them in the larger context of what actually causes HAIs.

In 1965, Sir Austin Bradford Hill published a landmark paper, entitled “The Environment and Disease: Association or Causation?” in which he outlined what would become known as the “Bradford Hill Criteria.” The “Hill Criteria” help distinguish association from causation in epidemiological research. One of nine criteria was biological plausibility. Quoting: “It will be helpful if the causation we suspect is biologically plausible. But this is a feature I am convinced we cannot demand. What is biologically plausible depends upon the biological knowledge of the day.” As commented elsewhere, the spirit of this criterion is to check whether the proposed causation violates any of the known laws and facts of science of biology, and as Hill outlines, this depends on currently available knowledge. It is said that Hill did not intend the criteria to be applied rigidly in the sense of a checklist approach; instead, he regarded them as “viewpoints” that would merely help in the assessment.

Fast-forward to the May 2013 Special Topic Issue of ICHE concerning the role of the environment in infection prevention. In the issue’s introduction, Weber and Rutala quote figures from a 1991 article by Weinstein concerning the biologically plausible sources of healthcare-acquired infections (HAIs): “patients’ endogenous flora, 40-60%; cross infection via the hands of personnel, 20-40%; antibiotic-driven changes in flora, 20-25%; and other (including contamination of the environment), 20%.”

In the same issue, an article by Salgado and colleagues caught our attention. This clinical trial compares 614 patients randomly placed into standard ICU rooms or into rooms where 6 frequently-touched items (e.g. bed rails, overbed tables, intravenous poles, etc.) had been replaced with copper alloy surfaces. The measured primary outcomes, according to the paper’s methods, were:
   (a) any HAIs and
   (b) colonization with methicillin-resistant Staphylococcus aureus (MRSA) or vancomycin-resistant enterococci (VRE). Besides HAI and colonization, outcomes presented in the results section included the numbers of patients who had
   (c) both HAI and colonization,
   (d) HAI and/or colonization,
   (e) HAI only but no colonization (i.e. number of patients who had HAI minus the ones who had both HAI and colonization), and
   (f) colonization only but no HAI.

Are you confused? Separate data for outcomes in each trial arm were only reported for (d-f) but not (a-c).

For HAI and/or colonization (d), the article reported what amounted to a 49% reduction in the copper rooms vs. non-copper rooms (21 vs. 41 patients; p=.02), for HAI only (e) a 62% reduction in the copper rooms (10 vs. 26; p=.013), and for colonization only (f), a 67% reduction (4 vs. 12; p=.063, NS). What was was not reported were the numbers of patients with (a) HAI and (b) colonization, listed separately for each trial arm, but the article concluded – in the discussion – that copper surfaces in rooms reduced the risk of HAIs by more than half. Conventional wisdom, however, would suggest that (a) any HAIs and (b) any colonization events, would be the most biologically relevant outcomes, and that it may not be so informative to combine these two events (under d) in the same statistical calculation, because they are biologically very different from each other. So, we extracted the missing numbers from the other numbers presented and arrived at (a) HAIs 17 vs. 29, and (b) colonization, 11 vs. 15 events. Putting these into our statistics calculator, they were – non-significant.

Now, let us revisit possible transmission routes in hospitals. We have: (i) endogenous transmission, from within the patient’s own flora, (ii) exogenous transmission via direct transfer, (e.g. as in handborne without surfaces), and (iii) exogenous transmission via surfaces and secondary transmission from surfaces onto the patients. If we look at (a) HAIs and (b) colonization with MRSA or VRE, then all three pathways can lead to HAIs, while only the two exogenous pathways can lead to colonization. If there is a >50% reduction of HAIs through copper surfaces (pathway iii), this would mean that the overall proportion of transmission from pathways (i) and (ii) plus the proportion of transmission from the remaining non-copper surfaces in the copper-treated rooms among pathway (iii) among all HAIs together would contribute less than 50% to overall HAIs.

The obvious question comes to mind: is that consistent with the known proportions of the different pathways leading to HAIs? The preliminary answer, given the Weinstein data (see above), would be, “given the biological knowledge of the day, apparently not.” It is also noted that the overall numbers of HAI and colonization events in the present article are relatively small.

Finally, anyone of us engaging in research can accidentally have outcomes that are surprising or do not quite add up with existing knowledge in the field. That is, in our opinion, where the intended purpose and scope of a discussion section of an article comes in, and where the Hill Criteria provide important food for thought. As one of us has put forward (Teleclass Feb. 7, 2013) on a different occasion and concerning a different topic, we would welcome the broad application of a check for biological plausibility when findings from clinical trials – and even systematic reviews – are reported. But we are not confident that our voices will be heard.

Image of Sir Austin Bradford Hill, source: toxipedia

HAI stories

Via the CDC's Safe Healthcare blog, a healthcare-associated infection story worth reading, told from the perspective of Brenda Helms, Infection Preventionist. 

We can't predict HAI with ICD-9 codes and it's only going to get worse

I'm getting ready for a chat with a reporter concerned with issues surrounding HAI surveillance. During my preparation, I thought again through the issues of code-based algorithms (e.g. ICD-9) and I've come to the conclusion that they are useless for assessing the burden of HAIs and HAI trends and it's only going to get worse.

One area we (and many others) have looked at is the utility of ICD-9 code-based algorithms (ie administrative codes) for detecting HAIs efficiently. A key metric frequently reported by researchers is the sensitivity of a specific code or code algorithm, which is great if the purpose of the algorithm is to improve the efficiency of detection by manual methods. Thus, if the sensitivity is high-enough, you could use the code-based algorithm to reduce the number of charts that require an IP's review. If you are using codes in this way, great!  I have no problems with that.

However, many are now using code-based algorithms to track trends in specific HAIs and measure the burden of disease. My general feeling on these is that they should be completely avoided for several reasons:

1) No matter how sensitive the algorithm is, all we care about here (since we are not validating with manual review) is the positive predictive value (i.e. the proportion of all code-positive patients that actually have the HAI of interest)

2) The PPV is very low for almost all HAI algorithms

3) If we are doing our job and lowering the incidence of HAI per admission in our hospitals the PPV by definition will only get worse (given a fixed sensitivity and specificity)

To show you why I have these concerns I have constructed two 2x2 tables evaluating an excellent hypothetical code-based algorithm for UTI with a sensitivity and specificity set at 95%.  In this first 2x2, I have evaluated the performance of the algorithm when the HAI has a 5% incidence per admission (i.e., 5% of the admissions had a UTI). You can see that such a great algorithm with a high-prevalance of disease, has a poor PPV of 50% - like flipping a coin.

Now, assume we have done an amazing job and cut our HAI rate down to 1%.  Given the same hypothetical algorithm, our PPV is now a horrible 16%. Thus, as we get better at preventing HAIs, we get worse at detecting them using code-based algorithms. Are you comfortable saying UTIs are increasing or decreasing or are associated with a certain level of excess costs, when only 16% of the UTIs in your estimation are actual (true positive) UTIs?  Me neither.