Are we sleepwalking into PPE paralysis?

“When the facts change, I change my mind. What do you do, Sir?” – John Maynard Keynes

sleepwalkers-figure-roof-sky-royalty-free-thumbnail

Link to Dental Elf

Bottom line answer

As we approach a return to clinical practice policy makers need to be mindful that the dental profession is already highly proficient in cross infection control, and the benefits of new elaborate PPE protocols regarding aerosol generating procedures may be marginal in the light of low disease prevalence. If we need 300,000 participants in a  study it may be impossible to practically demonstrate  significant benefits to patient safety from the perfect PPE model compared to the harms created by expense and access.

 

Background

The most important element of Keynes’s famous quote relating to the current coronavirus pandemic is the word ‘fact’. Facts at this moment in time are constantly changing as this disease spreads, and what was true two weeks ago is now a distant memory. One major problem we face as far as healthcare policy is concerned is a lack of accurate base-rate data on the prevalence of the disease in the population, this was initially addresses by Professor Ioannidis (Ioannidis, 2020). His article in STAT caused quite a lot of online controversy (Bastian, 2020, Taleb, 2020) but what it did is highlight the difficulties of being objective and human. At this moment in time it is extremely difficult not to be influenced by the problems of base-rate neglect, loss aversion, and availability bias via the media as we count the daily international infection/death figures (Gaurav, 2020). In this post I want to concentrate on how important it is to understand the significance of accurate base-rate (prevalence) reporting  so we can allocate the correct amount of training and resources to the dental profession based of the potential aerosol risk in virus transmission. In a paper by Chambers on base-rates in dental decision-making there is a quote (Chambers, 1999):

‘Base your decisions on either the baseline alone or the evidence alone, depending on which one contains the most information.’

What we are seeing now is a rapid accumulation of both base-line data and evidence,  but policy decisions about the future are being based on data and precautionary principles that were only valid at the start of this pandemic. To highlight this, I would like to explore how we are going to test the real-world effectiveness of the personal protective equipment (PPE) and cross infection protocols that are flooding the profession now.

Methods

I am going to look at three areas here, base-rates (prevalence), numbers of asymptomatic individuals, and power calculations regarding PPE use. For clarity I will use natural frequencies wherever possible.

Firstly, on the 14th May the Office of National Statistics (ONS) in the UK published the results of its coronavirus (Covid-19) infection survey (ONS, 2020). This data was based on 10,705 participants’ swab tests taken over a two-week period from 27th April to 10th May, the sample was drawn from households in which someone has already participated in an ONS survey to ensure the sample was representative of the wider population. From this sample 33 individuals in 30 households tested positive for COVID-19. This equates according to the ONS to 0.27% (95% confidence interval: 0.17% to 0.41%) of the population of England.

Secondly, one of the key problems with Covid-19 is asymptomatic spread (Bai et al., 2020) but there is no reliable data so any calculations here need to be looked on as a Fermi (back of an envelope) problem. To get a best estimate on the proportion of asymptomatic patients I conducted a meta-analysis of the data presented on the Oxford Covid-19 Evidence Service (Heneghan et al., 2020). The meta-analysis was carries out in R using a random effects model (See Figure 1.)

Figure 1. Forest plot of asymptomatic individuals

Asympto

There are two points of note from this forest plot, the summary estimate for asymptomatic individuals is 27% (95% CI: 12 to 45%) and the heterogeneity between studies (variability) is extremely high.

The next stage is to put the base-rate and number of asymptomatics together in the form of a frequency tree. For ease of calculation I have rounded the figures so a base-rate of 0.27% becomes 1 in 400,  and the number of asymptomatics becomes 30% (See Figure 2.).

Figure 2. Frequency tree of asymptomatic vs symptomatic

Asympto_1

The frequency tree illustrates that in this population 1 out of every 1333 people could be an asymptomatic carrier of Covid-19.

How does this relate to dentistry? On the precautionary principle we are operating under at the moment the presumption is that all (100%) the patients are asymptomatic carriers rather than the true figure of 0.075% ( I have assumed symptomatic patients will not be attending a dental surgery or will be triaged out prior to entering the clinical environment). This becomes important when we want to test if our PPE and protocols are effectively protecting both the patients and the staff.  We now need to set up a study comparing PPE that is adequately powered to eliminate the effects of random chance around such a small prevalence statistic (Button et al., 2013). I have created three examples of PPE for aerosol generating procedures:

  • Perfect PPE model (fluid resistant disposable gowns, FFP3 masks, visors, ventilation, long fallow periods etc) with a 99%  chance of reducing viral contamination
  • Realistic expectations of enhanced PPE practice (FFP2, reusable surgical gown, rubber dam etc) at 93%.
  • Standard practice (surgical masks etc) at 80%.

I placed the data into an apriori sample size calculator (G*Power 3.19.2) with a error probability is 0.05 and power (1-b error probability) of 0.8 (See Table 1).

Table 1. Sample sizes for a well powered study into PPE effectiveness

asympto_tableAs we can see even in a simulation study, we are going to have to at least place over a hundred individuals in each arm of the study. To see if the benefit translates into the real-world, we need to go up two orders of magnitude to see if there is a significant difference between perfect and good PPE  based on accurate population base-rate figures.

Discussion

The purpose of this opinion paper was to highlight the potential problems that a precautionary principle can create in healthcare when we work on the assumption that 100% of the patients attending a dental surgery are infectious. Guidelines and protocols need to take into consideration the absolute risk within the population based on data that is accurate and up to date. Simulation studies, and pilot studies rarely carry their full reported success into the real world (Kistin and Silverstein, 2015). Without taking a deep breath and objectively assessing the changing data regarding Covid-19, policy makers, academics, and clinicians can unconsciously fall fowl of the base-rate fallacy and availability biases created by the modern media. High quality PPE and staff training is a vital component of keeping everyone safe from this virus but we must be mindful of the other effects that perfect practice can have on the health economics and affordability of health care to those most vulnerable.

Disclaimer:  The article has not been peer-reviewed; it should not replace individual clinical judgement, and the sources cited should be checked. The views expressed in this commentary represent the views of the author and not necessarily those of the host institution. The views are not a substitute for professional advice.

References

BAI, Y., YAO, L., WEI, T., TIAN, F., JIN, D. Y., CHEN, L. & WANG, M. 2020. Presumed Asymptomatic Carrier Transmission of COVID-19. JAMA.

BASTIAN, H. 2020. A rebuttal to “A fiasco in the making?” [Online]. Available: http://hildabastian.net/index.php/8-secondary/87-a-rebuttal-to-a-fiasco-in-the-making [Accessed].

BUTTON, K. S., IOANNIDIS, J. P., MOKRYSZ, C., NOSEK, B. A., FLINT, J., ROBINSON, E. S. & MUNAFÒ, M. R. 2013. Power failure: why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience, 14, 365-376.

CHAMBERS, D. W. 1999. The roles of evidence and the baseline in dental decision making. J Am Coll Dent, 66, 60-7.

GAURAV, S. 2020. Behavioural Economics in the Fight Against COVID-19: BOMA Framework.

HENEGHAN, C., BRASSEY, C. & JEFFERSON, T. 2020. COVID-19: What proportion are asymptomatic? [Online]. Available: https://www.cebm.net/covid-19/covid-19-what-proportion-are-asymptomatic/ [Accessed].

IOANNIDIS, J. P. 2020. A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data [Online]. STAT. Available: https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/ [Accessed].

KISTIN, C. & SILVERSTEIN, M. 2015. Pilot studies: a critical but potentially misused component of interventional research. Jama, 314, 1561-1562.

ONS. 2020. Coronavirus (COVID-19) Infection Survey pilot: England, 14 May 2020 [Online]. Available: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronaviruscovid19infectionsurveypilot/england14may2020 [Accessed].

TALEB, N. 2020. EVIDENCE BASED is often BS [Online]. Available: https://twitter.com/nntaleb/status/1240641133820207104 [Accessed].

 

 

Opinion: Dentistry, Diagnostic Test Accuracy (DTA) and the Covid-19 Antibody Test

Link to the Dental Elf

Virus Outbreak Germany

As we are acutely aware the whole country has been locked down in an attemp to reduce the spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2 (Covid-19)), while we restructure the healthcare systems and direct research funding into either a treatment and/or a vaccine. The next stage is to rapidly scale up diagnostic testing to collect robust base-rate data at both an individual, and population level for future high quality decision-making (Ioannidis, 2020). The need to know who has been infected will be particularily important for dental care professionals due to the potential risk of Covid-19 spread via aerosol generating procedures (Coulthard, 2020). So what is the current state of affairs regarding quick point-of-care antibody testing? How good is it, how good should it be, and why is this important?

Firstly I am going to use a recently published paper that evaluated 10 antibody tests for SARS-CoV-2 using Enzyme-Linked Immunosorbent Assay (ELISA) and Lateral Flow Immune Assay (LFIA), lateral flow assay utilise the same technology commonly used for pregnancy tests (Crook, 2020). This particular paper has sparked considerable controversy which I will go into later. If you need more detail on how these tests work and their advantages/disadvantages then please take a look at the Oxford COVID-19 Evidence Service (Green et al., 2020). When looking at diagnostic accuracy testing the two main concepts to understand are:

Sensitivity The ability of a diagnostic test to give a positive result when it is supposed to be positive.

Specificity The ability of a diagnostic test to indicate a negative result when it is supposed to be negative.

 Results

After extracting the primary data from the individual tests, meta-analysis was carried out using the ‘mada’ package in R. The summary estimate for sensitivity was 62.7% (95%CI: 57.5 to 67.7) and specificity 96.9% (95%CI: 95.2 to 98.0) The results for the ELISA test and the 9 LFIA tests were plotted below on to a Summary Receiver Operating Characteristic (sROC) curve ( See Figure 1). The y-axis represents the sensitivity (1.0 =100%), and the x-axis represents 1- specificity (0.10 = 10%). For a test to be perfect the summary estimate point should be in the top left corner representing 100% true-positives and 0% false-positives. The blue dot represents the summary estimate for the tests surrounded by a 95% confidence area, the red dot represents the specification target of >98% (95%CI: 96 to 100%)  for sensitivity, and >98% (95%CI: 96 to 100%) for specificity, set by the Medicines & Healthcare Products Regulatory Agency (MHRA., 2020).

Figure 1. sROC curve – Antibody tests

sROC_DT_1_LI (2)

Discussion

So why is so important to set the levels of sensitivity and specificity so high? Imagine we have a small city with a population of 100,000. The prevalence of people who have had the virus and recovered is 6%  and you are using an LFIA test with a sensitivity of 63% and specificity of 97%. Out of 6600 people who test positive for virus antibody only 3780 are true positives which corresponds to 57% having a correct positive diagnosis (See Figure 2.). Another important consideration is that 3% of the population who do not have antibodies test positive (n= 2820) and could lead them to believe they are immune.

Figure 2.Frequency tree of Covid-19 antibody screening

Annotation 2020-05-09 193854

The second point about this paper is that in its present format it cannot be used in any future analysis since the companies that supplied the tests required a commercial confidentiality agreement to be signed with the UK Department of Health making it impossible to discriminate between tests. The current set of results show poor performance, and that is why the MHRA has specifically set its targets high because of the risks that  results could pose if they were used to ease a lockdown, or they become part of an immunity passport system (Mahase., 2020; WHO, 2020). The final point is that whenever we have to deal with diagnostic tests or screening devices in either our professional or private lives we need to be able to identify the products, their comparators, and how accurate they are before making a significant decision to use or purchase the product. False-positive and false-negative results can pose significant harms to both ourselves and the population in general.

Disclaimer:  The article has not been peer-reviewed; it should not replace individual clinical judgement, and the sources cited should be checked. The views expressed in this commentary represent the views of the author and not necessarily those of the host institution. The views are not a substitute for professional advice.

References

COULTHARD, P. 2020. Dentistry and coronavirus (COVID-19) – moral decision-making. Br Dent J, 228, 503-505.

CROOK, D. W. 2020. Evaluation of antibody testing for SARS-CoV-2 using ELISA and lateral flow immunoassays [Online]. Department of Microbiology, John Radcliffe Hospital, Oxford, OX3 9DU, United Kingdom.  [Accessed].

GREEN, K., , A. W., DICKINSON, R., GRAZIADIO, S., ROBERT WOLFF, MALLETT, S. & ALLEN, A. J. 2020. What tests could potentially be used for the screening, diagnosis and monitoring of COVID-19 and what are their advantages and disadvantages? [Online]. Available: https://www.cebm.net/covid-19/what-tests-could-potentially-be-used-for-the-screening-diagnosis-and-monitoring-of-covid-19-and-what-are-their-advantages-and-disadvantages/ [Accessed].

IOANNIDIS, J. P. 2020. A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data [Online]. STAT. Available: https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/ [Accessed].

MAHASE., E. 2020. Covid-19: Confidentiality agreements allow antibody test manufacturers to withhold evaluation results.

MHRA.2020.Target_Product_Profile_antibody_tests_to_help_determine_if_people_have_immunity_to_SARS-CoV-2_ [Online]. Available: https://www.gov.uk/guidance/guidance-on-coronavirus-covid-19-tests-and-testing-kits [Accessed].

WHO.2020. “Immunity passports” in the context of COVID-19 [Online]. Available: https://www.who.int/news-room/commentaries/detail/immunity-passports-in-the-context-of-covid-19 [Accessed].

Disclaimer: the article has not been peer-reviewed; it should not replace individual clinical judgement, and the sources cited should be checked. The views expressed in this commentary represent the views of the author. The views are not a substitute for professional medical advice.

 

What is the most appropriate gown/apron for preventing Covid-19 contaminated fluids transfer in dental practice?

Originally posted on the Dental Elf

L0028811 A nurse and a surgeon, both wearing gown and mask. Etching b

Question:

Which gown/apron combination provides the best protection in the dental practice?

Bottom-line answer:

From this reanalysis of the primary data the reusable cotton surgical gown may be more practical in the dental environment in the long-term than the disposable fluid resistant gown due to its reduce potential for cross contamination during use. The plastic apron creates the most cross contamination and should only be used if there is significant risk of fluid contamination.

Background

This paper is a reanalysis of a recent systematic review (Verbeek et al., 2020) on personal protective equipment (PPE), reframing the question to fit into the new clinical workflow created by Covid-19, and dental aerosol generating procedures (AGPs). I have covered face masks in a previous post.

Much has been written on the epidemiology of Covid-19 and its transmissibility via contact, droplets, aerosols, or faeco-oral route. The main concern within dentistry being the aerosol generated during many routine dental procedures (Coulthard, 2020). To reduce this contamination risk Public Health England’s guidance document for personal protective equipment updated 3 May 2020 Section 10.4 (GOV.UK, 2020) states that:-

‘Disposable fluid repellent coveralls or long-sleeved gowns must be worn when a disposable plastic apron provides inadequate cover of staff uniform or clothes for the procedure or task being performed, and when there is a risk of splashing of body fluids such as during AGPs in higher risk areas or in operative procedures. If non-fluid-resistant gowns are used, a disposable plastic apron should be worn’.

As this advice is generic and the workflow within a critical care unit differs from a dental practice it is important to evaluate the best available evidence from a primary care rather than secondary care perspective.

Method

To save unnecessary duplication of search strategies and risk of bias/quality assessments I utilised the most up to date Cochrane Review of PPE for preventing highly infectious diseases due to exposure to contaminated body fluids in healthcare staff (Verbeek et al., 2020). This systematic review included 17 studies with 1950 participants evaluating 21 interventions. The authors concluded:

‘We found low- to very low-certainty evidence that covering more parts of the body leads to better protection but usually comes at the cost of more difficult donning or doffing and less user comfort, and may therefore even lead to more contamination. More breathable types of PPE may lead to similar contamination but may have greater user satisfaction.’

The authors conclusion helped to focus the next stage of analysis which was based around the levels of contamination and wearability. From the included studies two randomised simulation trials, one of a parallel design (Wong et al., 2004) and a second of a cross-over design (Guo et al., 2014) were selected as they contained sufficient primary data to undertake a meta-analysis. A simulation trial utilises aerosolised fluorescent dye sprayed on the PPE instead of true viral contamination. It was possible to extract data on contamination of  fluid resistant disposable gowns, standard cotton surgical gowns, and plastic aprons. The data was placed in Excel and then transferred to R for meta-analysis using the ’meta’ package. A random effect model was used with a Hartung Knapp conversion due to variability within the studies. Prediction intervals were included to facilitate the estimation for future studies.

Findings

The first meta-analysis compared a fluid resistant disposable gown with a standard cotton gown for both Wong and Guo. In Guo’s study the group tried two methods of doffing PPE: their individual accustomed removal method (IARM), and gown removal methods recommended by the Centers for Disease Control and Prevention (CDC).

The overall result favoured the cotton gown, but the result was non-significant, the mean difference (MD) was 0.91 (95%CI: -0.34 to 2.66) (See Figure 1.). The second meta-analysis showed significantly less contamination of the cotton surgical gowns compared with the plastic apron, the  MD was 8.4 (95%CI: 0.59 to 16.2) (See Figure 2.). The final analysis looked at the  contamination of the clinician post PPE removal showing equal contamination between the different PPE types MD was -0.02 (95% CI: -1.43 to 1.40) (See Figure 3).

Figure 1.Forest plot of disposable fluid resistant gown vs cotton gown

FIGURE 1

Figure 2. Forest plot of plastic apron vs cotton gown

FIGURE 2.

Figure 3. Forest plot of body contamination

 FIGURE 3

Conclusion

From the results of the meta-analysis there is little difference between the disposable fluid resistant gown and the reusable cotton surgical gown in terms of contamination/protection of both the wearer, patient, and clinical environment. The results favour the cotton gown as cotton through its material and properties can absorb droplet contaminants and thereby reduce opportunities for such contaminants to spread to the environment. The plastic apron performed worst and may significantly increase the risk of cross contamination both to the clinician and patient and should only be necessary where there is a risk of serious fluid contamination.

There is an interesting paper recently published by Phan and co-workers (Phan et al., 2019) who observed that ‘90% of observed doffing was incorrect, with respect to the doffing sequence, doffing technique, or use of appropriate PPE. Common errors were doffing gown from the front, removing face shield of the mask, and touching potentially contaminated surfaces and PPE during doffing’.

These results presented are hypothetical and due to the lack of specific studies of virus penetration through gowns in dentistry and are based on surrogate, and composite outcomes. There is an urgent need for specific studies to address PPE performance in the dental surgery environment.

Disclaimer:  The article has not been peer-reviewed; it should not replace individual clinical judgement, and the sources cited should be checked. The views expressed in this commentary represent the views of the author and not necessarily those of the host institution. The views are not a substitute for professional advice.

References

COULTHARD, P. 2020. Dentistry and coronavirus (COVID-19) – moral decision-making. Br Dent J, 228, 503-505.

GOV.UK. 2020. COVID-19 ( personal protective equipment (PPE) [Online]. Available: https://www.gov.uk/government/publications/wuhan-novel-coronavirus-infection-prevention-and-control/covid-19-personal-protective-equipment-ppe [Accessed].

GUO, Y. P., LI, Y. & WONG, P. L. 2014. Environment and body contamination: a comparison of two different removal methods in three types of personal protective clothing. Am J Infect Control, 42, e39-45.

PHAN, L. T., MAITA, D., MORTIZ, D. C., WEBER, R., FRITZEN-PEDICINI, C., BLEASDALE, S. C., JONES, R. M. & PROGRAM, C. P. E. 2019. Personal protective equipment doffing practices of healthcare workers. Journal of occupational and environmental hygiene, 16, 575-581.

VERBEEK, J. H., RAJAMAKI, B., IJAZ, S., SAUNI, R., TOOMEY, E., BLACKWOOD, B., TIKKA, C., RUOTSALAINEN, J. H. & KILINC BALCI, F. S. 2020. Personal protective equipment for preventing highly infectious diseases due to exposure to contaminated body fluids in healthcare staff. Cochrane Database Syst Rev, 4, CD011621.

WONG, T. K., CHUNG, J. W., LI, Y., CHAN, W. F., CHING, P. T., LAM, C. H., CHOW, C. B. & SETO, W. H. 2004. Effective personal protective clothing for health care workers attending patients with severe acute respiratory syndrome. American journal of infection control, 32, 90-96.

Other  links

Personal protective equipment: a commentary for the dental and oral health care team on  Verbeek et al .