Most air purifying devices have not been tested on humans, and little is known about their potential harm.

A large scoping review of engineering measures against airborne infections, from ventilation and filters to UV irradiation, ionizers, and “plasma” purifiers, was published in the Annals of Internal Medicine. The authors reviewed 672 studies from 1929 to 2024 and found a gap between marketing and science: only 57 studies (about 8-9%) even tested whether such solutions reduce the incidence of disease in people; another 9 - on “guard” animals. Most publications measured only air (particles, “harmless” microbes, surrogate markers), and potential harmful by-products (for example, ozone) were almost not assessed.

Background of the study

In the wake of COVID-19, the question of “how to make indoor air safer from viruses” is no longer purely an engineering one: aerosol transmission accounts for most outbreaks in enclosed spaces, meaning measures like ventilation, filtration, and UV disinfection have become a broad public health policy issue. The CDC explicitly recommends “aiming for ≥5 air changes per hour (ACH) of clean air” and making “clean air” a core part of respiratory virus prevention, along with vaccination, especially in schools, clinics, and offices. This reflects a shift in focus from surfaces to the air we share.

On the professional standards side, a key milestone was the publication of ASHRAE Standard 241 (2023), the first standard to set minimum requirements for infectious aerosol management in new and existing buildings: how to combine outdoor air intake with recirculated air purification, and how to design and maintain systems to reduce the risk of transmission. The standard moves the conversation from the realm of “gadgets” to the realm of building system design and operating procedures.

At the same time, the scientific base for “engineered” interventions turned out to be heterogeneous. A recent scoping review in the Annals of Internal Medicine collected 672 studies (1929-2024) and showed a gap between laboratory metrics and clinical outcomes: the vast majority of studies measure airborne surrogates (particles, viral RNA, “harmless” microbes in chambers), and there are very few trials to reduce real morbidity in people. This does not mean that the technologies “don’t work,” but it emphasizes that schools and hospitals need field RCTs and quasi-experiments that take into account efficacy and safety.

A separate hot topic is ultraviolet. The “far” UV-C zone of 222 nm is actively promoted as a method of disinfection “in the presence of people”, but a couple of recent studies have shown that such lamps generate ozone and secondary oxidation products under certain conditions; therefore, in addition to the benefits, side effects need to be measured in real rooms. For classic UVGI systems (upper-room/ducted solutions), there is also a lack of clinical trials, although the reduction of contamination and inactivation of aerosol pathogens are demonstrated reliably in models and chambers. Bottom line: the potential is high, but implementation standards must be based on honest field data.

How the study works (and why you can trust it)

A team from the University of Colorado, Northwestern, the University of Pennsylvania, and several CDC/NIOSH sites systematically searched MEDLINE, Embase, Cochrane, and other databases for primary studies, with a second reviewer duplicating the data extraction. The resulting basket included 672 papers: about half examined pathogen inactivation (405), with fewer examining removal (filtration; 200) and dilution/air exchange (ventilation; 143). The outputs were dominated by airborne outcomes: viable nonpathogenic organism counts (332 studies), nonbiological particle mass (197), or viable pathogens (149). A key gap was the rare assessment of harm (chemical byproducts, ozone, secondary reactions). The project is registered with OSF and funded by NIOSH.

What are “engineering controls” and where are they subtle?

The authors include in engineering measures everything that physically changes the air and its movement paths: ventilation/dilution, filtration (MERV/HEPA), UV disinfection (including 254 nm and "far" 222 nm), photocatalytic oxidation, ionization/plasma, combined hybrids. According to the summary data of media retellings and author's comments:

• 44 studies were found on photocatalysis, but only one tested for reducing infections in humans;
• on plasma technologies - 35 works, not a single one involving people;
• on nanofilters (capture + “kill”) - 43 works, also without testing on humans;
• A common problem with portable "cleaners" has been the almost complete lack of real clinical outcomes.

The main conclusion

The review does not say that “purifiers don’t work.” It says that most of the science is still about air, not about people. That is, we often know how a device reduces the concentration of particles or harmless microbes in a chamber, but we don’t know whether it reduces real infections in classrooms, hospitals, and offices. And even worse is the safety: ozone and other byproducts that some devices (from individual UV lamps to “plasma”/ionizers) can generate are rarely tested. Independent studies have previously shown that, for example, some GUV systems (222 nm) can cause the formation of ozone and secondary aerosols - this requires a direct assessment of the benefit/harm in real rooms.

Why is this important right now?

The COVID-19 pandemic has shifted the conversation about ventilation and air purification from the realm of engineering to public health. Schools, clinics, and offices are pouring money into technology, not always distinguishing between silver bullets and marketing. A new review sets the bar: we need real-world testing with real-world outcomes—incidence of disease, exposure of people to viable pathogens, and adverse effects—not just surrogates like CO₂ or dust.

What can already be done "in practice"

Focus on the basic principles:

• ensure adequate air exchange and fresh air supply;
• local filtering (High Efficiency Particulate Air filters/HEPA purifiers) where appropriate;
• control sources: reduce crowding, masks during outbreaks, regular cleaning.

Be careful with "miracle boxes":

• prefer devices with independent field testing rather than just chamber testing;
• avoid technologies that can generate ozone, aldehydes and other reaction products unless there is transparent safety data;
• require manufacturers to provide full reports: testing methods, operating conditions, maintenance, noise, energy consumption.

Look at the system, not the gadget: proper ventilation + reasonable density of people + hygiene are often more profitable than single “magic” solutions.

What is missing in science (and what requires a review)

• Randomized and quasi-experimental studies in schools, health care facilities, offices, where the end point is cases of infections or, at a minimum, exposure of people to viable pathogens.
• Standardization of outcomes (common clinical and “airborne” metrics) and fair classification of technologies (inactivation/removal/dilution) for comparability.
• Systematic harm accounting: ozone, secondary VOCs/aerosols, impact on vulnerable groups, economic/energy cost.
• Independence of expertise: transparent financing, blind verification of results, replication.

Who is this news addressed to?

• For school and hospital managers: focus on ventilation and verifiable filters; require independent field data before purchasing.
• HVAC engineers: Help customers differentiate between “dilution,” “removal,” and “inactivation” when selecting solutions for a room scenario.
• For home buyers: If you're buying a portable "virus" purifier, check for real-world testing and no ozone generation; remember that open windows and basic maintenance still work.

Viewing Limitations

The authors excluded non-English language publications and “gray literature,” and the scoping design itself describes the field but does not provide meta-estimates of effect. However, the scale (672 studies), the multidisciplinary team (academic + CDC/NIOSH), and the convergence of findings with independent news analyses make the picture robust: real-world clinical data on “cleansers” are rare, and safety is less well studied than it should be.

Study source: Baduashvili A. et al. Engineering Infection Controls to Reduce Indoor Transmission of Respiratory Infections: A Scoping Review. Annals of Internal Medicine, online August 5, 2025. https://doi.org/10.7326/ANNALS-25-00577