0 Recommendations

In-Room Recirculating UVGI Air Disinfection: Measuring and testing efficacy of UVGI systems

By John Van Adrichem | 02 Oct, 2008 Last edited by Julia Fischer-Mackey on 12 May 2010

The acceptance of in-room recirculating UVGI systems with a Merv 8-11 filter is improving. The URV rating or the CADR rating can be demonstrated mathematically but, Hospital administrators want to see, feel, notice and measure the improvement of their interior spaces. My question is how can this be most effectively demonstrated? What type of test protocol and over how long should the test program run? Long period test programs must account for variations in outside environmental changes due to seasons and the coming and going of patients would alter the in-room microbial concentrations. Could a short test program be set up to allow the interior microbial environment to reach steady state based upon the CADR? What should be measured to show improvement? Should the testing be restricted to the inlet and outlet of the appliance?


This discussion is about measuring and testing efficacy of UVGI systems. If you wish to contribute to this topic, please reply here.

A new discussion on protection rate with surgical masks and N95 respirators for healthcare workers has spun off in the replies below. For ease of reference and participation, we have re-posted this as a new discussion with relevant replies here: https://www.ghdonline.org/ic/discussion/what-protection-with-surgical-masks-a.... Please use this link and click on ‘Reply” to contribute to this second discussion. Thank you.

Attached resource:
  • UVGI in Ventilation Ducts & Recirculating Room Air Disinfection Units: Part 1 (download, 1.6 MB)

    Summary: This presentation covers:
    * UVGI applications
    * Hospital-acquired infections
    * University & dormitory applications
    * HVAC coil cleaning
    * Forced air UV systems
    * Relative humidity effect

    This document supported the 2008 "Engineering Methods for the Control of Airborne Infections: An International Perspective" course that took place at the Center of Continuing Professional Education at Harvard School of Public Health July 14-25, 2008.

    Source: Immune Building Systems, Inc.

    Publication Date: July 14, 2008

    Language: English

    Keywords: 2008 Harvard Engineering Methods Course, Engineering Controls, Harvard Airborne IC Course, Training, UVGI, UVGI Rating Value



Edward Nardell, MD Moderator Replied at 9:13 AM, 2 Oct 2008

Just a few comments on this topic. Acceptance of in-room units by hospital administrators has never been the problem - they tend to love them because they appear to be a simple, no construction, plug-in solution.

In fact, overly enthusiastic acceptance in many countries in Eastern Europe is a problem. As we discussed in the course, they often are not likely to be effective simply because the CADR is relatively low. Still, it the CADR is sufficient and the air mixing good, they can contribute to equivalent air changes in an isolation or procedure room - especially in low-volume rooms. I applaud hospital administrators who want proof of efficacy before going for these. However, testing with viable organisms is obviously not an option in a hospital; even avirulent surrogates like serratia cannot be aerosolized around vulnerable patients. Testing by reducing ambient organisms has never worked well in the reports that I have seen. I don't know enough about Merv 8-11 filtration characteristics to know whether testing the room clearance of inert airborne particles with a particle counter is possible. However, I would argue that the ultimate test for efficiency must be room clearance of an appropriate airborne surrogate, not just inlet/outlet testing. Consider an in-room unit with a highly efficient filter that moves little air. Testing across the unit may yield spectacular results, but the effect on room concentrations of a contaminant may be minimal. This is in fact the situation we find with many units in Eastern Europe.

Paul A. Jensen, PhD, PE, CIH Moderator Replied at 4:12 PM, 2 Oct 2008

Ed's point "the ultimate test for efficiency must be room clearance of an appropriate airborne surrogate, not just inlet/outlet testing" is the key response to this issue. Whether we are dealing with an in-room air cleaner, mechanical ventilation, or upper-room UVGI, it needs to work in synergy with the environment. In addition, it needs to be designed properly. For instance, an HVAC system with poorly designed supply diffusers which are located near an exhaust grill, would lead to short circuiting and great inefficiencies. Another example would be UVGI. If you have upper-room UVGI with a poorly designed fixture (i.e., little output, in a corner of a room with little air mixing), there is little chance of it inactivating a significant number of airborne microorganisms. Now, let's return to the room air cleaner discussion. The Association of Home Appliance Manufacturers (AHAM) is an organization that verifies the testing results of home appliances such as room air cleaners (www.cadr.org). AHAM "certifies" the CADR. Generally, we think of air changes per hour (ACH) or airflow rate per patient/occupant. For instance, the CDC TB IC GLs recommend 6-12 ACH for a TB isolation room (depending on whether it pre-dates 1994 or is new/reconstructed). The Russian norms require 80 m3/hour/patient (47 cfm). Let's assume we have a room that is 3 m x 5 m x 2.4 m (9.8' x 16.4' x 8'). 6-12 ACH would require an exhaust airflow rate of 216-432 m3/hr (129-258 cfm). If we assume two patients are in this room, the required airflow rate per Russian norm would be 160 m3/hr (96 cfm). Let's now look at the CADR for a typical in-room air cleaner . . . To achieve 12 ACH, we would need a CADR of 258. CADR of 258 means that the airflow rate of a unit in terms of particle-free air is 432 m3/hr (258 cfm). Also, the room needs to be well mixed and the air cleaner needs to work in synergy with the room. In general, there are few air cleaners that would provide this level of CADR. Finally, I am not saying that an appropriate design with UVGI and MERV-rated filters will not work; I am merely dealing with the technology within a room air cleaner. Another discussion topic may be the use of UVGI and MERV-rated filters in a recirculating ventilation system!

S. Mehtar Replied at 5:13 AM, 4 Oct 2008

I have two major concerns relating to Africa.
1. High level of dust in the atmosphere- how does this effect UV operating systems.
2. Maintenance of such pieces of equipment. There is currently insufficient capacity generally in healthcare facilities to look after these items of equipment.
Any ideas?

S. Mehtar Replied at 5:21 AM, 4 Oct 2008

Can anyone point me in the right direction for accurate information on the following: When a healthcare worker wears a surgical mask, how much reduction occurs in the infectious particles inhaled by the person wearing the mask when faced with a case which is not wearing one?

This obviously differs by distance from the source but is there a graph one can refer to? How long does this protection last? I know that for ordinary bacteria it is approximately 10 mins (until the barrier is broken by respiratory moisture?

This become relevant when one advises on patients being examined or routine procedures in undiagnosed cases or uncomplicated TB where the patient is not wearing a mask.

What similar protection is known about N95 respirators which are not properly fit tested? (probably similar to a surgical mask)

Practical problems at the moment.

Edward Nardell, MD Moderator Replied at 1:04 PM, 4 Oct 2008

1. Re. the effect of dust on recirculating in-room UVGI or filtration

Yes, dust would definitely affect performance. Again, I am not, and I know Paul Jensen does not, generally advocate the use of these devices for the reasons I stated below. Your point about dust in the dry season in Africa is a good one. Most UVGI units have filters to keep the dust from accumulating on the lamps, which should help, but under extremely dusty conditions such filters become "full" and airflow diminishes. The answer is to regularly (and safely - with the fixture off) wipe off the lamps with alcohol, and to change clogged filters. However, on a recent visit to Botswana with Paul, small UVGI units placed over the head of the bed were designed to make it impossible (or extremely difficult) to change the filter or clean the lamp. So, your concerns are legitimate.

A more general issue is that ANYTHING that is done technically to reduce TB transmission indoors requires some attention or maintenance. Even advocates of simple open windows acknowledge that one needs to be sure that the windows are not closed at night for security or because of cold air.

Therefore, IC programs should be designed with maintenance in mind, with extra lamps for UVGI, with extra filters, and with appropriate resources to hire and train personnel whose job it is to maintain the equipment. At Partners In Health we have local engineers trained in Haiti and Rwanda to maintain mechanical equipment. Funding from the GATM and other sources should include these provisions. We must build the capacity to maintain critical technical equipment in Africa
and elsewhere - even if we start small. There is a proverb that says, "it is better to light one candle than to curse the darkness".


Edward Nardell, MD Moderator Replied at 1:32 PM, 4 Oct 2008

I am hoping Paul will comment on Dr. Menthar's question too, which is about masks and respirators, not UVGI. This question comes up all the time and merits discussion here.

Some general comments about surgical masks as protective devices for the wearer. By convention in TB circles, we refer to surgical masks as face and nose covering to protect the operating field - or general environment when worn by a patient. Although they may resemble N95 or equivalent respirators, these devices are not designed to produce a tight face seal, but rather to prevent large respiratory particles expelled during breathing, coughing, or speaking, from falling onto operating fields or evaporating into airborne droplets. Air can usually easily get between the filtering facepiece and the face, often around the bridge of the nose or at the sides of the mouth. Even respirators that are designed to reduce face seal leak still have as much as 20% face seal leak in actual use, even after successful fit testing, surgical masks could have 30-50% face seal leak and are not fit tested.
This is not a matter of moisture so much as a matter of design and materials - no clip around the nose - only one elastic band, or sides that bulge out in some designs.

As noted above, a good N95 or equivalent respirator can do much better than a surgical mask if properly fit tested, properly worn each time, and not used when they finally loose their integrity or the elastic bands become too weak. Still, even when new, expect at least 10% face seal leak. To get a better seal requires more permanent respirators and for maximum protection, positive air purifying respirators (PAPRs) are used for such hazardous procedures as asbestos removal. These are expensive and require filter changes.

There is not a lot of evidence that fit testing improves respirator performance even though it makes some sense that it should. The selection of a respirator design from several different styles to fit
variations in face shapes and sizes is probably more important. If nothing else, however, fit testing demonstrates to the user that there are ways to wear the respirator that are more likely reduce leakage and that should help. I would say that a non-fit tested N95 is still more likely to provide more protection than a surgical mask because of its design features. There are some respirators that perform extremely well on a variety of face shapes without fit testing, but they have other downs sides including high cost.

I know Paul and other respirator experts will have comments.


S. Mehtar Replied at 4:33 AM, 5 Oct 2008

Dear Ed

Thank you so much. I love a good discussion. The information is invaluable and we are aware of the shortcomings both of surgical masks and ill-fitting respirators. So, the next question that comes to mind is-- how much reduction is adequate? 10%? 50%?. Evidence suggests that one needs approximately 60 infectious particles to cause infection (obviously not disease).Therefore, if one knows the particular load, can one estimate the reduction in particles for the wearer? In other words, what level of leak is acceptable in clinical practice given the distance from the patient?

One of the problems we have with respirators fitted to people with flat noses is that there must, by definition, be more leak than with a "roman" nose! An interesting thought. So, should those with flat noses have an extra bridge support with something like a foam cushion, for example?

Next question. What is the effect of covering the nose and mouth with a hanki or similar?
Any thoughts?

Edward Nardell, MD Moderator Replied at 9:38 AM, 5 Oct 2008

Dr. Mehtar now raises two difficult questions:

1) how much protection from respirators is adequate?

2) what is the infectious dose of TB?

Here is a detailed response to the first of these challenging questions. Others are invited to join in. The other will follow.

How much protection can be generalized beyond respirators - how much protection is adequate from natural ventilation, mechanical ventilation, UVGI, or from a combination of all these when they are available? Does the answer vary if the risk is drug susceptible or MDR/XDR TB? What if
those exposed are confined against their will and have no choice about exposure, like prisoners or refugees? What risk can health care workers be expected to assume as part of their work, and what is the hospital or health care systems responsibility versus personal responsibility?

There are no easy answers to these questions as we quickly enter the discipline of risk assessment and avoidance, money spent per life or DALY gained, etc.

However, a paper I published in 1991 on the theoretical limits of protection of building ventilation and another published later with Kevin Fennelly on the additive benefit of respirators and building
ventilation may be useful.

The 1991 paper (PubMed abstract: http://www.ncbi.nlm.nih.gov/pubmed/1907115?ordinalpos=50&itool=EntrezSystem2....) was based on an actual outbreak investigated by the Massachusetts Health Department when I was TB Control Officer in the mid 1980s.

A worker in a welfare office returned from a month-long holiday sick and got progressively sicker while at work for the following month before being diagnosed with cavitary, smear positive TB. Two rounds of contact TST testing showed that 27 of 67 initially negative co-workers had converted with strongly positive TSTs (only 1 had active disease at the time), for an infection attack rate of 40% among that cohort.

The building had been the subject of air quality complaints resulting in measurements of the ventilation, and I was asked how much the poor air quality had contributed to the outbreak. To address this question I turned to a well-established mathematical model which estimated that to produce that rate of infection, given measured building ventilation rates. The paper concludes that had the ventilation been twice as good (possible but not easy to achieve), approximately 13 instead of 27 workers would have been infected. Had the ventilation been 4 times what it really had been (not feasible), still 7 workers would likely have been infected. If it were possible to invest the resources (expensive re-construction and energy costs) to have 2X or 4X the existing ventilation, would that amount of protection be enough? Is it an acceptable risk if 7 of 67 workers exposed for 30 days to an moderately infectious co-worker become infected? Again, we are not talking active disease, but exposure to drug susceptible TB, so they were given INH prevention and only the one secondary case occurred. Had this been Kwazulu Natal, South Africa, however, with XDR TB and many the workers
HIV-infected, suddenly the risk of infection becomes a serious death threat and primary prevention all the more critical.

The reason for these findings is that almost all interventions work by removing some proportion of what organisms remain in the air. In the case of building ventilation, each increase in ventilation removes fewer and fewer of the remaining organisms as they become more dilute.

Unfortunately, there is no lower limit of risk, just a lower probability of infection. This is the subject of the next of Dr. Mehtar's questions. What about respirators?

In another paper (The relative efficacy of respirators and room ventilation in preventing occupational tuberculosis, 1998, PubMed abstract here: http://www.ncbi.nlm.nih.gov/pubmed/9801283?ordinalpos=36&itool=EntrezSystem2....), Fennelly and I used the same approach to show that respirators/masks and ventilation (indeed any interventions) work together to reduce risk. If we reduce risk by half by ventilation by 2X increase, and had poorly fitting respirators with a protection factor of only 50%, the net effect is protection equivalent to 4 times the equivalent ventilation. This may or may not be adequate, depending on the risk (duration, susceptibility, drug resistance, etc).
While it may be difficult to double ventilation night and day in very hot or cold climates by natural or mechanical means, having respirators with protection factors closer to 90% should be possible if they are well designed, properly fit tested, and properly put on each time.

So, my long winded answer to Dr. Mehtar's great question is to strive for at least 90% protection from respirators.

I am out of words for the moment and you are likely out of time so I will come back to Dr. Mehtar's second great question on the infectious dose of TB, above, another time.

Other opinions?

Edward A. Nardell, MD

Edward Nardell, MD Moderator Replied at 9:48 AM, 5 Oct 2008

Dr. Mehthar also asks about the effect of covering the nose or mouth with a hand or tissue and I can respond to that one quickly.

There are no quantitative studies yet, but at the AIR facility in South Africa we are planning to study the effects of a surgical mask on patients on reducing transmission for sentinel guinea pigs who serve as surrogates for highly susceptible health care workers. In principal the hand, tissue and surgical mask all work the same - to block large respiratory particles before they are expelled into the air and
evaporate into airborne infectious droplets. Dr. Richard Riley who developed the concept of large respiratory droplets evaporating into the dried residua (infectious droplet nuclei) believed very strongly in the effectiveness of simple cough hygiene measures, but there is no
objective proof - yet.

The first paper on the use of surgical masks to prevent airborne transmission was for flu during the 1918 pandemic, and some evidence of efficacy was offered, although it would not meet current standards of evidence.

In conclusion, cough hygiene makes sense theoretically, clearly cannot offer 100% protection, but like administrative controls, ventilation, UVGI, and use of respirators, is part of an overall infection control


Paul A. Jensen, PhD, PE, CIH Moderator Replied at 8:30 AM, 6 Oct 2008

Thanks so much to Professor Mehtar and Dr. Nardell for the great exchange here. In addition to the second paper to which Ed referred (Fennelly KP, Nardell EA. The relative efficacy of respirators and room ventilation in preventing occupational tuberculosis. ICHE 19:754-759, 1998.), I refer folks to on by Mark Nicas (Nicas M, Neuhaus J. Variability in respiratory protection and the assigned protection
factor. JOEH 1: 99-109, 2004.)

Ed and Kevin (1998) concluded "These modeling data suggest that the risk of occupational tuberculosis probably can be lowered considerably by using relatively simple respirators combined with modest room ventilation rates for the infectious aerosols likely to be present in isolation rooms of newly diagnosed patients. . . . "

Mark and John (2004) examined seven published half-mask respirator studies and two published PAPR studies and concluded "Given these results and related considerations, we recommend that the current half-mask APF be reduced from 10 to 5." Thus, the standard N95 or FFP2 filtering facepiece (disposable?) respirator provides 80% or better protection when properly fitted on a properly trained person.

So, my answer to Dr. Mehtar's great question is to strive for at least 90% protection from respirators and expect 80% or so with a good respiratory protection program.


Paul J

This Community is Archived.

This community is no longer active as of December 2018. Thanks to those who posted here and made this information available to others visiting the site.