Research gaps and priorities for quantitative microbial risk assessment (QMRA).

The coronavirus disease 2019 pandemic highlighted the need for more rapid and routine application of modeling approaches such as quantitative microbial risk assessment (QMRA) for protecting public health. QMRA is a transdisciplinary science dedicated to understanding, predicting, and mitigating infectious disease risks. To better equip QMRA researchers to inform policy and public health management, an Advances in Research for QMRA workshop was held to synthesize a path forward for QMRA research. We summarize insights from 41 QMRA researchers and experts to clarify the role of QMRA in risk analysis by (1) identifying key research needs, (2) highlighting emerging applications of QMRA; and (3) describing data needs and key scientific efforts to improve the science of QMRA. Key identified research priorities included using molecular tools in QMRA, advancing dose-response methodology, addressing needed exposure assessments, harmonizing environmental monitoring for QMRA, unifying a divide between disease transmission and QMRA models, calibrating and/or validating QMRA models, modeling co-exposures and mixtures, and standardizing practices for incorporating variability and uncertainty throughout the source-to-outcome continuum. Cross-cutting needs identified were to: develop a community of research and practice, integrate QMRA with other scientific approaches, increase QMRA translation and impacts, build communication strategies, and encourage sustainable funding mechanisms. Ultimately, a vision for advancing the science of QMRA is outlined for informing national to global health assessments, controls, and policies.

Quantitative Microbial Risk Assessment and Molecular Biology: Paths to Integration.

Quantitative microbial risk assessment (QMRA) has now been in use for over 35 years and has formed the basis for developing criteria for ensuring public health related to water, food, and remediation, to name a few areas. The initial data for QMRA (both in exposure assessment and in dose response assessment) came from measurements using assays for viability, such as plate counts, plaque assays, or animal infectivity. With the increasing use of molecular methods for the measurement of microorganisms in the environment, it has become important to assess how to use such data to estimate infectious disease risks. The limitations to the use of such data and needs to resolve the limitations will be addressed.

A Case Study Evaluating the Risk of Infection from Middle Eastern Respiratory Syndrome Coronavirus (MERS-CoV) in a Hospital Setting Through Bioaerosols.

Middle Eastern respiratory syndrome, an emerging viral infection with a global case fatality rate of 35.5%, caused major outbreaks first in 2012 and 2015, though new cases are continuously reported around the world. Transmission is believed to mainly occur in healthcare settings through aerosolized particles. This study uses Quantitative Microbial Risk Assessment to develop a generalizable model that can assist with interpreting reported outbreak data or predict risk of infection with or without the recommended strategies. The exposure scenario includes a single index patient emitting virus-containing aerosols into the air by coughing, leading to short- and long-range airborne exposures for other patients in the same room, nurses, healthcare workers, and family visitors. Aerosol transport modeling was coupled with Monte Carlo simulation to evaluate the risk of MERS illness for the exposed population. Results from a typical scenario show the daily mean risk of infection to be the highest for the nurses and healthcare workers (8.49 × 10-4 and 7.91 × 10-4 , respectively), and the lowest for family visitors and patients staying in the same room (3.12 × 10-4 and 1.29 × 10-4 , respectively). Sensitivity analysis indicates that more than 90% of the uncertainty in the risk characterization is due to the viral concentration in saliva. Assessment of risk interventions showed that respiratory masks were found to have a greater effect in reducing the risks for all the groups evaluated (>90% risk reduction), while increasing the air exchange was effective for the other patients in the same room only (up to 58% risk reduction).

A quantitative risk assessment method for synthetic biology products in the environment.

The need to prevent possible adverse environmental health impacts resulting from synthetic biology (SynBio) products is widely acknowledged in both the SynBio risk literature and the global regulatory community. To-date, however, discussions of potential risks of SynBio products have been largely speculative, and the limited attempts to characterize the risks of SynBio products have been non-uniform and entirely qualitative. As the SynBio discipline continues to accelerate and bring forth novel, highly-engineered life forms, a standardized risk assessment framework will become critical for ensuring that the environmental risks of these products are characterized in a consistent, reliable, and objective manner that incorporates all SynBio-unique risk factors. In their current forms, established risk assessment frameworks - including those that address traditional genetically modified organisms - fall short of the features required of this standard framework. To address this gap, we propose the Quantitative Risk Assessment Method for Synthetic Biology Products (QRA-SynBio) - an incremental build on established risk assessment methodologies that supplements traditional paradigms with the SynBio risk factors that are currently absent, and necessitates quantitative analysis for more transparent and objective risk characterizations. We demonstrate through a hypothetical case study that the proposed framework facilitates defensible quantification of the environmental risks of SynBio products in both foreseeable and hypothetical use scenarios. Additionally, we show how the quantitative nature of the proposed method can promote increased experimental investigation into the true likelihood of hazard and exposure parameters and highlight the most sensitive parameters where uncertainty should be reduced, ultimately leading to more targeted SynBio risk research and yielding more precise characterizations of risk.

Seasonal Assessment of Opportunistic Premise Plumbing Pathogens in Roof-Harvested Rainwater Tanks.

A seasonal study on the occurrence of six opportunistic premise plumbing pathogens (OPPPs) in 24 roof-harvested rainwater (RHRW) tanks repeatedly sampled over six monthly sampling events (n = 144) from August 2015 to March 2016 was conducted using quantitative qPCR. Fecal indicator bacteria (FIB) Escherichia coli (E. coli) and Enterococcus spp. were enumerated using culture-based methods. All tank water samples over the six events were positive for at least one OPPP (Legionella spp., Legionella pneumophila, Mycobacterium avium, Mycobacterium intracellulare, Pseudmonas aeruginosa, or Acanthamoeba spp.) during the entire course of the study. FIB were positively but weakly correlated with P. aeruginosa (E. coli vs P. aeruginosa τ = 0.090, p = 0.027; Enterococcus spp. vs P. aeruginosa τ = 0.126, p = 0.002), but not the other OPPPs. FIBs were more prevalent during the wet season than the dry season, and L. pneumophila was only observed during the wet season. However, concentrations of Legionella spp., M. intracellulare, Acanthamoeba spp., and M. avium peaked during the dry season. Correlations were assessed between FIB and OPPPs with meteorological variables, and it was determined that P. aeruginosa was the only OPPP positively associated with an increased antecedent dry period, suggesting stagnation time may play a role for the occurrence of this OPPP in tank water. Infection risks may exceed commonly cited benchmarks for uses reported in the rainwater usage survey such as pool top-up, and warrant further exploration through quantitative microbial risk assessment (QMRA).

Dose response models and a quantitative microbial risk assessment framework for the Mycobacterium avium complex that account for recent developments in molecular biology, taxonomy, and epidemiology.

Mycobacterium avium complex (MAC) is a group of environmentally-transmitted pathogens of great public health importance. This group is known to be harbored, amplified, and selected for more human-virulent characteristics by amoeba species in aquatic biofilms. However, a quantitative microbial risk assessment (QMRA) has not been performed due to the lack of dose response models resulting from significant heterogeneity within even a single species or subspecies of MAC, as well as the range of human susceptibilities to mycobacterial disease. The primary human-relevant species and subspecies responsible for the majority of the human disease burden and present in drinking water, biofilms, and soil are M. avium subsp. hominissuis, M. intracellulare, and M. chimaera. A critical review of the published literature identified important health endpoints, exposure routes, and susceptible populations for MAC risk assessment. In addition, data sets for quantitative dose-response functions were extracted from published in vivo animal dosing experiments. As a result, seven new exponential dose response models for human-relevant species of MAC with endpoints of lung lesions, death, disseminated infection, liver infection, and lymph node lesions are proposed. Although current physical and biochemical tests used in clinical settings do not differentiate between M. avium and M. intracellulare, differentiating between environmental species and subspecies of the MAC can aid in the assessment of health risks and control of MAC sources. A framework is proposed for incorporating the proposed dose response models into susceptible population- and exposure route-specific QMRA models.

Reproducible Risk Assessment.

Reproducible research is a concept that has emerged in data and computationally intensive sciences in which the code used to conduct all analyses, including generation of publication quality figures, is directly available, and preferably in open source manner. This perspective outlines the processes and attributes, and illustrates the execution of reproducible research via a simple exposure assessment of air pollutants in metropolitan Philadelphia.

Quantitative microbial risk assessment for recreational exposure to water bodies in Philadelphia.

A quantitative microbial risk assessment was conducted to estimate risk of gastrointestinal (GI) illnesses associated with recreational exposure to Philadelphia waterways, under dry and wet weather conditions. Using maximum likelihood estimation, stochastic exposure models were generated for each exposure scenario and Monte Carlo techniques were applied to characterize uncertainty in final risk estimates. The dry-weather risk estimates were found significantly lower than those predicted for wet-weather conditions. Moreover, the predicted risk, calculated in proportion of the frequency of use, was elevated at 6 out of 10 sites (ranging from 9 to 52 illnesses/1000 users/day). Activities contributing most to the risk of GI illness at creeks were identified as wading and playing (81%), while fishing was the potential risk contributor (65%) at rivers. The quantitative measure of risk contribution from each type of water activity obtained from this study can be useful for policy makers in prioritizing the future interventions.

Application of quantitative microbial risk assessment for selection of microbial reduction targets for hard surface disinfectants.

BACKGROUND: This quantitative microbial risk assessment (QMRA) included problem formulation for fomites and hazard identification for 7 microorganisms, including pathogenic Escherichia coli and E coli 0157:H7, Listeria monocytogenes, norovirus, Pseudomonas spp, Salmonella spp, and Staphylococcus aureus. The goal was to address a risk-based process for choosing the log10 reduction recommendations, in contrast to the current US Environmental Protection Agency requirements. METHOD: For each microbe evaluated, the QMRA model included specific dose-response models, occurrence determination of aerobic bacteria and specific organisms on fomites, exposure assessment, risk characterization, and risk reduction. Risk estimates were determined for a simple scenario using a single touch of a contaminated surface and self-inoculation. A comparative analysis of log10 reductions, as suggested by the US Environmental Protection Agency, and the risks based on this QMRA approach was also undertaken. RESULTS: The literature review and meta-analysis showed that aerobic bacteria were the most commonly studied on fomites, averaging 100 colony-forming units (CFU)/cm(2). Pseudomonas aeruginosa was found at a level of 3.3 × 10(-1) CFU/cm(2); methicillin-resistant S aureus (MRSA), at 6.4 × 10(-1) CFU/cm(2). Risk estimates per contact event ranged from a high of 10(-3) for norovirus to a low of 10(-9) for S aureus. CONCLUSION: This QMRA analysis suggests that a reduction in bacterial numbers on a fomite by 99% (2 logs) most often will reduce the risk of infection from a single contact to less than 1 in 1 million.

Recreational use assessment of water-based activities, using time-lapse construction cameras.

Recreational exposure to surface waters during periods of increased pathogen concentration may lead to a significantly higher risk of illness. However, estimates of elementary exposure factors necessary to evaluate health risk (i.e., usage distributions and exposure durations) are not available for many non-swimming water-related activities. No prior studies have assessed non-swimming water exposure with respect to factors leading to impaired water quality from increased pathogen concentration, such as weather condition (rain events produce increased runoff and sewer overflows) and type of day (heavy recreational periods). We measured usage patterns and evaluated the effect of weather and type of day at eight water sites located within Philadelphia, by using a novel "time lapse photography" technology during three peak recreational seasons (May-September) 2008-2010. Camera observations validated with simultaneous in-person surveys exhibited a strong correlation (R(2)=0.81 to 0.96) between the two survey techniques, indicating that the application of remote photography in collecting human exposure data was appropriate. Recreational activities usage varied more on a temporal basis than due to inclement weather. Only 14% (6 out of 44) of the site-specific activity combinations showed dry weather preference, whereas 41.5% (17 out of 41) of the combinations indicated greater usage on weekends as compared with weekday. In general, the log normal distribution described the playing and wading duration distribution, while the gamma distribution was the best fit for fishing durations. Remote photography provided unbiased, real-time human exposure data and was less personnel intensive compared with traditional survey methods. However, there are potential limitations associated with remote surveillance data related to its limited view. This is the first study to report that time lapse cameras can be successfully applied to assess water-based human recreational patterns and can provide precise exposure statistics for non-swimming recreational exposures.

Dose-response assessment for influenza A virus based on data sets of infection with its live attenuated reassortants.

Reported data sets on infection of volunteers challenged with wild-type influenza A virus at graded doses are few. Alternatively, we aimed at developing a dose-response assessment for this virus based on the data sets for its live attenuated reassortants. Eleven data sets for live attenuated reassortants that were fit to beta-Poisson and exponential dose-response models. Dose-response relationships for those reassortants were characterized by pooling analysis of the data sets with respect to virus subtype (H1N1 or H3N2), attenuation method (cold-adapted or avian-human gene reassortment), and human age (adults or children). Furthermore, by comparing the above data sets to a limited number of reported data sets for wild-type virus, we quantified the degree of attenuation of wild-type virus with gene reassortment and estimated its infectivity. As a result, dose-response relationships of all reassortants were best described by a beta-Poisson model. Virus subtype and human age were significant factors determining the dose-response relationship, whereas attenuation method affected only the relationship of H1N1 virus infection to adults. The data sets for H3N2 wild-type virus could be pooled with those for its reassortants on the assumption that the gene reassortment attenuates wild-type virus by at least 63 times and most likely 1,070 times. Considering this most likely degree of attenuation, 10% infectious dose of H3N2 wild-type virus for adults was estimated at 18 TCID50 (95% CI = 8.8-35 TCID50). The infectivity of wild-type H1N1 virus remains unknown as the data set pooling was unsuccessful.

Time-dose-response models for microbial risk assessment.

While microbial risk assessment (MRA) has been used for over 25 years, traditional dose-response analysis has only predicted the overall risk of adverse consequences from exposure to a given dose. An important issue for consequence assessment from bioterrorist and other microbiological exposure is the distribution of cases over time due to the initial exposure. In this study, the classical exponential and beta-Poisson dose-response models were modified to include exponential-power dependency of time post inoculation (TPI) or its simplified form, exponential-reciprocal dependency of TPI, to quantify the time of onset of an effect presumably associated with the kinetics of in vivo bacterial growth. Using the maximum likelihood estimation approach, the resulting time-dose-response models were found capable of providing statistically acceptable fits to all tested pooled animal survival dose-response data. These new models can consequently describe the development of animal infectious response over time and represent observed responses fairly accurately. This is the first study showing that a time-dose-response model can be developed for describing infections initiated by various pathogens. It provides an advanced approach for future MRA frameworks.

Characterizing the risk of infection from Mycobacterium tuberculosis in commercial passenger aircraft using quantitative microbial risk assessment.

Quantitative microbial risk assessment was used to predict the likelihood and spatial organization of Mycobacterium tuberculosis (Mtb) transmission in a commercial aircraft. Passenger exposure was predicted via a multizone Markov model in four scenarios: seated or moving infectious passengers and with or without filtration of recirculated cabin air. The traditional exponential (k = 1) and a new exponential (k = 0.0218) dose-response function were used to compute infection risk. Emission variability was included by Monte Carlo simulation. Infection risks were higher nearer and aft of the source; steady state airborne concentration levels were not attained. Expected incidence was low to moderate, with the central 95% ranging from 10(-6) to 10(-1) per 169 passengers in the four scenarios. Emission rates used were low compared to measurements from active TB patients in wards, thus a "superspreader" emitting 44 quanta/h could produce 6.2 cases or more under these scenarios. Use of respiratory protection by the infectious source and/or susceptible passengers reduced infection incidence up to one order of magnitude.

Legionnaires' disease: evaluation of a quantitative microbial risk assessment model.

BACKGROUND: The quantities of Legionella vary considerably from natural waters to water in contaminated domestic hot water supplies, whirlpool spas and cooling towers, with the risk for LD rising as the Legionella counts grow. We currently report the results from our Quantitative Microbial Risk Assessment (QMRA) model evaluation. We developed the LD QMRA model to better understand Legionella exposure risks. METHODS: Using an animal data derived model for LD, we calculated risks from estimated exposures for a whirlpool spa outbreak, two hot spring spa outbreaks and compared the results to the reported LD risks. RESULTS: The QMRA model shows agreement (generally less than an order of magnitude discrepancy) with the reported Legionnaires' disease sub-clinical severity infection, clinical severity infection, and mortality risks. CONCLUSIONS: The LD QMRA model may lead to risk based limits to supplement the current guidance on Legionella control in cooling towers, whirlpool spas and other potential exposure sources. The verification of QMRA for LD also suggests the techniques, given suitable animal model data, may be useful in quantifying human response to other airborne pathogens.

Wastewater disinfection by peracetic acid: assessment of models for tracking residual measurements and inactivation.

With its potential for low (if any) disinfection byproduct formation and easy retrofit for chlorine contactors, peracetic acid (PAA) or use of PAA in combination with other disinfectant technologies may be an attractive alternative to chlorine-based disinfection. Examples of systems that might benefit from use of PAA are water reuse schemes or plants discharging to sensitive receiving water bodies. Though PAA is in use in numerous wastewater treatment plants in Europe, its chemical kinetics, microbial inactivation rates, and mode of action against microorganisms are not thoroughly understood. This paper presents results from experimental studies of PAA demand, PAA decay, and microbial inactivation, with a complementary modeling analysis. Model results are used to evaluate techniques for measurement of PAA concentration and to develop hypotheses regarding the mode of action of PAA in bacterial inactivation. Kinetic and microbial inactivation rate data were collected for typical wastewaters and may be useful for engineers in evaluating whether to convert from chlorine to PAA disinfection.

Quantitative microbial risk assessment model for Legionnaires' disease: assessment of human exposures for selected spa outbreaks.

Evaluation of a quantitative microbial risk assessment (QMRA) model for Legionnaires' disease (LD) required Legionella exposure estimates for several well-documented LD outbreaks. Reports for a whirlpool spa and two natural spring spa outbreaks provided data for the exposure assessment, as well as rates of infection and mortality. Exposure estimates for the whirlpool spa outbreak employed aerosol generation, water composition, exposure duration data, and building ventilation parameters with a two-zone model. Estimates for the natural hot springs outbreaks used bacterial water to air partitioning coefficients and exposure duration information. The air concentration and dose calculations used input parameter distributions with Monte Carlo simulations to estimate exposures as probability distributions. The assessment considered two sets of assumptions about the transfer of Legionella from the water phase to the aerosol emitted from the whirlpool spa. The estimated air concentration near the whirlpool spa was 5 to 18 colony forming units per cubic meter (CFU/m(3)) and 50 to 180 CFU/m(3) for each of the alternate assumptions. The estimated 95th percentile ranges of Legionella dose for workers within 15 m of the whirlpool spa were 0.13-3.4 CFU and 1.3-34.5 CFU, respectively. The modeling for hot springs Spas 1 and 2 resulted in estimated arithmetic mean air concentrations of 360 and 17 CFU/m(3), respectively, and 95 percentile ranges for Legionella dose of 28 to 67 CFU and 1.1 to 3.7 CFU, respectively. The Legionella air concentration estimates fall in the range of limited reports on air concentrations of Legionella (0.33 to 190 CFU/m(3)) near showers, aerated faucets, and baths during filling with Legionella-contaminated water. These measurements may provide some indication that the estimates are of a reasonable magnitude, but they do not clarify the exposure estimates accuracy, since they were not obtained during LD outbreaks. Further research to improve the data used for the Legionella exposure assessment would strengthen the results. Several of the primary additional data needs include improved data for bacterial water to air partitioning coefficients, better accounting of time-activity-distance patterns and exposure potential in outbreak reports, and data for Legionella-containing aerosol viability decay instead of loss of capability for growth in culture.

Assessment of benefits from use of antimicrobial hand products: reduction in risk from handling ground beef.

Quantitative microbial risk assessment (QMRA) has been used to estimate the benefits resulting from the use of hand cleansing products (e.g., soaps) containing anti-microbial ingredients. This was done by developing a model for the scenario of hand contact with ground beef during food preparation, considering transference of bacteria to the hands, removal and inactivation by handwashing, and subsequent transference from the hands to the mouth. Organisms of interest in this case study were pathogenic Escherichia coli and the particular strain E. coli O157:H7. It was found that QMRA could be applied to this problem, and that the antimicrobials provided some quantifiable benefit (i.e., reduced the risk of infection and illness). Benefits from the use of triclosan-containing products were less than from the use of products in which alcohols or chlorhexidine were active ingredients.

The Milwaukee Cryptosporidium outbreak: assessment of incubation time and daily attack rate.

The time course of reported illnesses (epidemic curve) in the 1993 Milwaukee outbreak of cryptosporidiosis was analysed using a dynamic model considering time variant force of infection and incubation time distributions. Different functional forms for the force of infection and incubation time distribution were tested. The resulting model is a coupled integro-differential equation system. These models gave a good fit to the data, although depending upon the functional forms of the underlying distributions, different incubation time and force of infection curves were obtained. However there was reasonable agreement with respect to a baseline illness rate that existed. This demonstrates that useful information may be obtained in this manner, although it should be supplemented with other data (e.g. serology) for a precise assessment of dynamics of disease occurrence during waterborne epidemic conditions.

Comparison of tissue culture and animal models for assessment of Cryptospridium parvum infection.

The current increased interest for using tissue culture as a surrogate for mouse infection to assess Cryptospridium viability suggests that a comparison of the two models is essential for data interpretation. Therefore, a need remains for a statistical comparison that can demonstrate if infection and inactivation predicted by new tissue culture models are comparable with those predicted by animal models. Data from a total of 31 dose-response trials using both tissue culture and mouse models to assess C. parvum infectivity were compared. The dose needed to infect 50% of the tissue cultures (ID(50)) was also compared to each ID(50) in mice. Average ID(50)s developed using the logit dose-response method for tissue culture and mice were 8 and 107, respectively, suggesting that tissue culture was more sensitive to infection. However, correlation (r) between tissue culture and mouse infectivity was statistically significant (0.9167 [95% CI=0.8428 to 0.9594, p<0.0001]). Comparison of oocyst disinfection by UV and chlorine dioxide showed no significant difference between inactivation predicted by tissue culture and mouse models (p=0.8893; t=0.0141; n=21). These results demonstrate that tissue culture can successfully be used to measure C. parvum infection and can be used for determining inactivation in disinfection studies.