CDC’s Research Program on Hospital-Acquired Infections (HAIs) Minimizes Airborne Transmission

By Lambert Strether of Corrente.

This post originated in a valiant but doomed effort to determine whether modes of transmission (airborne, fomite, etc.) could be shoehorned into ICD-10-CM (International Classification of Diseases, Clinical Modification)[1], coded as Y95 (“Nosocomial condition,” where nosocomial is what we said before “Hospital Acquired Infection” (HAI) became a thing). My tentative conclusion is that it can, even if the Association for Professionals in Infection Control and Epidemiology (APIC) vehemently asserts that it should not[2]. Getting airbone transmission coded would, of course, make it visible, the first step to creating incentives to eliminate it, thereby saving lives in this airborne pandemic, and the next one.

That said, what I might call the coding establishment, the institutional matrix in which ICD-10-CM is embedded, maintained, and modified, and which credentializes medical coding, is so complex that I coudn’t work out how to operationalize my brilliant idea, or even if it was brilliant. (There are a number of establishments that must be jackhammered before the reality of airborne transmission becomes an institutional norm, the coding establishment being but one). I couldn’t even find a worked example of Y95 data, though doubtless it’s out there somewhere in training materials. 

All this makes me sad, because as readers know, I’m a big fan of classification systems, and ICD-10-CM, together with ICD-10-PCS (“Procedure Coding System”) together provide, at least in the United States, an practical (though not formal) ontology for diseases, health conditions, and clinical “encounters” (“encounters” seeming to be the word for what happens when a procedure is performed[3]).

In any case, I began to get a crawling feeling in the pit of my stomach the maintainers of ICD-10, the Centers for Disease Control (CDC) and the  Center for Medicaid Services (CMS), didn’t consider airborne transmission of infectious agents in a hospital setting to be an HAI at all. So I went to CDC’s HAI page, and found it did not. I then went to CDC’s “research gaps” material to find out if they were even looking into the question. They were not. Then I found CDC’s toolkits for investigating HAIs. They did not. So in this post, I will quickly establish that yes, airborne infection is an HAI, at least as a layperson (that is, a dull normal patient seeking to escape a hospital setting alive). Then I will look that CDC’s HAI pages, as just described. In conclusion, I will speculate on why airborne tranmission is having such a hard time getting traction in, well, the HAI establishment[4].

Airborne Tranmission is an HAI

From The Royal Society’s Interface, “Airborne transmission of disease in hospitals” (2009):

Tuberculosis (TB; Mycobacterium tuberculosis) is clearly transmitted in the air and can be a source of outbreak in hospitals. Healthcare workers infected with TB can spread the infection widely and extensive screening of patients and other staff may be necessary. Similarly Norovirus is transmitted by aerosol and is difficult to contain in a hospital ward without sufficient single rooms with en suite toilets. Historically, natural ventilation was seen to be beneficial in hospital wards and was part of hospital design. With the advent of sealed high-rise buildings and forced ventilation, expensive negative pressure rooms have been introduced to house patients with infections thought likely to be transmitted by aerosol. The spread of tuberculosis among HIV patients was a recent dramatic example of the problems with enclosed rooms and prisons…. MRSA disseminates widely throughout the ward and is commonly found in dusty, inaccessible high surfaces. Clostridium difficile spores are thought to spread in the air and can be found near a patient carrying the organism.

From École polytechnique fédérale de Lausanne (EPFL), “Airborne Infectious Disease Transmission in Hospitals: Significance and Control by Ventilation” (2023):

[N]osocomial infections have significantly contributed to the overall transmission of various notorious respiratory infectious diseases these years. It was estimated that up to 20% of infections of SARS-CoV-2 (the responsible virus of COVID-19) in inpatients and 73% in healthcare workers (HCWs) may be due to nosocomial transmission in the UK [4]. A huge fraction of Middle East respiratory syndrome (MERS) cases was linked to the healthcare setting, ranging from 43.5% for the nosocomial outbreak in Jeddah, Saudi Arabia, in 2014 to 100% for both the outbreak in AL-Hasa, Saudi Arabia, in 2013 and that in South Korea in 2015 [2]. A well-known nosocomial outbreak of severe acute respiratory syndrome (SARS) that originated from ward 8A at the Prince of Wales Hospital in Hong Kong eventually infected 69 HCWs, 16 medical students, 53 patients, and a number of visitors [5]. Other cases include chickenpox, seasonal flu, Methicillin-resistant Staphylococcus Aureus (MRSA), and many more.

From Infection Control Today, “Rethinking Airborne Pathogens: WHO Proposes New Terminology for Disease Spread” (2024): 

Health care settings have been the primary source of the spread of respiratory illnesses, and the response to stopping this spread can be suboptimal. This was exemplified by the 2003 SARS-CoV-1 outbreak in Toronto. A report of this outbreak, published by the National Academies of Sciences, discusses transmission by droplets and wearing protective equipment during droplet-generating procedures, but not for exposure to all patients with respiratory illnesses…

A recent commentary in the Lancet by Trisha Greenhalgh and colleagues stated that handwashing has been the mainstay measure to prevent disease transmission. Strategies to prevent the spread of pathogens through the air, such as wearing a fitted N95 mask and improving indoor ventilation, are often “”ignored or downplayed.””

Many feel we need to wipe the slate clean and effectively start over.

And specifically for Covid, in the United Kingdom, or at least Wales:

Shockingly, at least 77% of patients in hospital with Covid in Wales 🏴󠁧󠁢󠁷󠁬󠁳󠁿 caught it WHILE THEY WERE THERE.

It’s probably a similar story in England, but they don’t publish the numbers anymore. 🙄

What’s it going to take for hospitals to reinstate infection controls, like masks? https://t.co/WfyXv3gDgE pic.twitter.com/LTeujucsAo

— Cat in the Hat 🐈‍⬛ 🎩 🇬🇧 (@_CatintheHat) April 10, 2024

(The Cat in the Hat account publishes those useful summaries of the UK’s Covid Inquiry.) In the United States, from JAMA, “Is Nosocomial SARS-CoV-2 Still Worth Preventing?” (2023)[5]:

Tens of thousands of hospital-acquired SARS-CoV-2 infections have been documented, and these are just the tip of the iceberg.2 Many infections go undiagnosed because they are asymptomatic, paucisymptomatic, occur early in hospitalization and are misattributed as community exposures, occur after hospital discharge, or are otherwise untested. During surges, up to 10% of hospitalized patients with SARS-CoV-2 and other respiratory viruses may have acquired their infection in the hospital.2,3 Likewise, we have good data that nosocomial respiratory viral infections can be prevented. Masking health care workers is associated with a 50% to 60% decrease in hospital-onset respiratory viral infections

Finally, nosocomial Covid seems to increase other nosocomial diseases. From Medscape, “Hospital-Acquired Infection Rates Remain High Post Pandemic” (2024):

Hospital-acquired infections associated with antimicrobial resistance (AMR) increased during the COVID-19 pandemic by 32% and remain 13% higher post pandemic compared with prepandemic levels, showed a study of US hospital data that compared AMR levels before, during, and post pandemic.

The largest increase in infections was seen with those caused by gram-negative, carbapenem-resistant organisms including Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacterales, where levels during the pandemic rose by over 50% compared with the prepandemic period and remained 35% higher than prepandemic levels in 2022.

Intriguing. Perhaps CDC should look into this.

CDC’s HAI Materials Systematically Avoid or Minimize Airborne Transmission of Infectious Agents

Here I will look at three CDC documents: (1) its “progress report” on HAI; (2) its document on “research gaps”; and (3) its forms to fill out during an “Outbreak Investigation.”

(1) Progress Report

This page, “Current HAI Progress Report” (updated November 25, 2024) summarizes CDC’s current HAI efforts:

The 2023 National and State HAI Progress Report provides data on central line-associated bloodstream infections (CLABSIs), catheter-associated urinary tract infections (CAUTIs), ventilator-associated events (VAEs), surgical site infections (SSIs), methicillin-resistant Staphylococcus aureus (MRSA) bloodstream events, and Clostridioides difficile (C. difficile) events.

This is, in 2023 (!), there’s no data on Covid as an HAI, or on airborne transmission generally. (While it is fair to say that both MRSA and c. difficile can spread “through the air,” we see that what is being measured is “infection types,” and not the mode of transmission.)

(2) Research Gaps

Now let’s look at the page for “Research Gaps in Patient and Healthcare Personnel Safety“:

At CDC, the Division of Healthcare Quality Promotion (DHQP) conducts and funds research and innovation designed to address HAIs and AR [Antibiotic Resistance”] that are potentially preventable but lie beyond the reach of current prevention strategies.

DHQP developed the so-called “evidence review” used in CDC’s first attempt to gut hospital masking protection for patients; see NC here[6]. Apparently, however, for DHQP, airborne transmission of infectious diesase is either not “potentially preventable” or “lie[s] beyond the reach of current prevention strategies.” Odd! Be that as it may, let’s look at the PDF that can be downloaded from this page First, the words “airborne” and “aerosol” do not appear anywhere in the document. Second, the word “ventilation” does not appear exactly where one would expect it to. Third, a detailed look provides some encouragement. From page 10:

Built environment

The buildings in which modern health care is delivered constitute the patient care environment. This includes all fixtures and furniture, the water supplied and drained away, even the air circulated down corridors and hallways and into patient and procedure rooms. These environments can either promote or interrupt effective care, and can become contaminated, either from pathogens that usually colonize and infect patients, or more environmental-adapted organisms that can fill the role of opportunistic pathogens.

Questions in this domain include:

• What are design and organizational factors that make it easier for healthcare personnel to take correct actions?

• Are there ways to engineer the built environment to reduce risks of pathogen contamination and transmission?

• Are there ways to change processes of care to reduce the risks of environmental contamination and pathogen transmission?

• What are approaches that will best enhance our understanding of pathogen persistence and survival in healthcare environments (including surfaces, fomites, and water/wastewater systems) to inform transmission risks?

We see a glimmer of hope with “even the air,” but the glimmer fades with “surfaces, fomites, and water/wastewater.” Third, the emphasis, amazingly, is still on handwashing. From page 12:

Pathogen transmission

Hand hygiene by healthcare personnel, environmental cleaning, correct reprocessing of shared devices, and selective use of isolation including cohorting and barrier precautions (i.e., gloves, gowns) remain the primary proven tools for preventing transmission, and yet are difficult to implement and can have low levels of personnel and patient acceptance.

And:

•  How long do specific pathogens retain infectivity when suspended in air of varying humidities and temperatures?

•  What is the optimal air quality in various parts of healthcare facilities to reduce transmission of respiratory infections?

•  What is the most ideal implementation and prioritization of current tools (e.g., hand hygiene, barrier precautions, masking, environmental cleaning) to prevent transmission of pathogens within and between healthcare facilities, considering clinical effectiveness, cost, acceptability of the intervention to different patient and clinician communities, and ease of implementation? For example, are there circumstances where gowns do not add significant value to gloves and/ or hand hygiene in preventing transmission? Or are there better ways to implement barrier precautions that would reduce transmission?

Needless to say, that five years into a pandemic whose infectious agent is tranmitted through the air, we are still posing the question “What is the optimal air quality in various parts of healthcare facilities to reduce transmission of respiratory infections?” is an utterly damning failure on the part of CDC and DHQP in particular.

Of course, if we were gathering data on airborne transmission when there is an HAI outbreak, that might help build a case. But we’re not.

(3) HAI Outbreak Investigations

CDC provides “toolkits” for investigators of HAI outbreaks. Needless to say, the investigators have a form of fill out. First, we have (PDF) “Healthcare-Associated Infection (HAI) Outbreak Investigation User’s Guide for Completing the Abstraction Form.” I have read it carefully. There is nothing there that would assist in the epidemiology of airborne transmission. (When the word “ventilation” is used, it means the patient is on a ventilator). For example:

For example, if investigating an outbreak of respiratory infections, it may be helpful to have a record of respiratory therapists and nurses involved in the patient’s care during the outbreak period.

It would be even more helpful know if the Health Care Workers (HCWs) were wearing surgical masks, respirators, or raw-dogging it. It would also be nice to know about open doors and windows, HEPA filters, and HVAC generally.

We also have the “abstraction form” itself, “Healthcare-Associated Infection (HAI) Outbreak Investigation Abstraction Form.” I have read it carefully. The form gathers data on Blood Products, Devices, Point of care testing/injections/infusions, Invasive Procedures, and Consult Services. There is nothing relevant to airborne tranmission whatever.

Conclusion

But why? Central line-associated bloodstream infections, catheter-associated urinary tract infections, ventilator-associated events, surgical site infections, methicillin-resistant Staphylococcus aureus, bloodstream events, and C. difficile events are all, as it were, first-class HAIs, whereas airborne infections aren’t classed at all. Yet, as we know from the ongoing Covid events, of all these HAIs, only airborne infections create the possibility of pandemics, and so only they present Taleb’s “risk of ruin” (as 2020’s HCWs know all too well). So why? Here let me quote a great slab from World Health Network’s complaint to the HHS on HICPAC:

B. Competition for Funding with Rival Siloes. WHN writes:

Furthermore, members of HICPAC, recognized for their expertise in areas such as bloodstream infections, sepsis, sharps injuries, hand hygiene, fomite transmission, sterilization and disinfection, antimicrobial resistance, and Ebola, are funded specifically for their work in these fields and would not be funded for airborne transmission prevention. This creates a potential conflict of interest which may interfere with a decision to shift the focus of infection prevention to airborne diseases, which is required to deal effectively with the hospital-based transmission of COVID-19. Such a shift could threaten the funding that supports their salaries, research, staff, and programs, as well as their positions of authority in infection prevention and control [IPC], and that of their colleagues. This inherent tension is compounded by similar conflicts of interests among CDC officials responsible for nominating HICPAC members and setting the committee’s agenda, including the current and former HICPAC Federal Officers and the director of NCEZID.

I don’t know anybody who has an issue with threatening IPC. Do you? (And if these two sections make HICPAC and CDC seem like a snakepit of self-dealing, well, it looks like that’s what it is. It would also be interesting to know if the CDC Foundation is hooked into this “inherent tension” at all.)

C. Perverse Incentives in Fee-for-Service Systems from Hospital-Acquired Infections. WHN writes:

HICPAC’s Charter mandates providing guidance on “prevention, and control of healthcare-associated infections” Therefore, committee members that are compensated for encouraging spread of infection (or compensated for being knowingly or willfully ignorant of the science of infection control in a healthcare setting), are in conflict of interest with HICPAC’s objective.

More specifically, it is well established that direct payment systems can lead to perverse incentives against the prevention of hospital-acquired infections (HAIs). In fee-for-service payment models, hospitals are reimbursed for services provided, including the treatment of HAIs. In such a system, hospitals can generate more revenue by providing additional care to treat these infections, rather than by preventing them in the first place. Many members of HICPAC are from hospital management, and as such, have direct financial interests that conflict with the prevention of HAIs. This conflict of interest, long recognized in the context of other HAIs, must now also be addressed for COVID-19 and other airborne diseases.

In short, at the CDC/DHQP level (not just HICPAC) “first class” HAIs like catheter-associated urinary tract infections are funded, and “no class” HAIs like airborne infection agents are not. Further, the incentives work against funding airborne HAIs at a “first class” level. Too many lives might be saved. Rule #2. 

NOTES

[1] American Academy of Professional Coders (APIC): “ICD-10-CM is used for medical claim reporting in all healthcare settings and is a standardized classification system of diagnosis codes that represent conditions and diseases, related health problems, abnormal findings, signs and symptoms, injuries, external causes of injuries and diseases, and social circumstances.”

[2] APIC (PDF), “APIC Position Paper: The Use of Administrative (Coding/Billing) Data for Identification of Healthcare-Associated Infections (HAIs) in US Hospitals,” 2010.

[3] Anticipating a future post, it’s not clear how infectious agents that “move like smoke” can be fit into an encounter model. That may be one reason why there is so much institutional resistance to the reality of airborne transmission; the paradigms of science are easy to shift, compared to the paradigms of administrators.

[4] “Establisment” being a word that merits careful definition, but not today.

[5] The answer to the question posed in the headline: “However, we have reached a point at which we can be selective about both the measures we choose and when we implement them. Masking, admission testing, and visitor screening are likely the highest yield interventions. The close correlation between community respiratory viral incidence and nosocomial infection rates suggests that selectively deploying these measures during periods of heightened community transmission could abort the plurality of nosocomial transmissions. We believe strategic use of selective prevention measures during periods of increased transmission remains warranted.” No mention of ventilation, naturally; nor consideration given to the lag between infection increases and “when we implement” the measures (two weeks, IIRC, for CDC’s infamous “green map,” absolutely unacceptable in a pandemic situation of exponential growth.

[6] I’m starting to feel like this is all connected.