Dear Elisabeth,

I work as a biostatistician/data analyst in SaniPath group. I hope I can answer your questions about QMRA.

The QMRA wiki website qmrawiki.canr.msu.edu/index.php?title=Qu...sessment_(QMRA)_Wiki summarizes some concepts about QMRA. On this website, QMRA was defined as a framework and approach that brings information and data together with mathematical models to address the spread of microbial agents through environmental exposures and to characterize the nature of the adverse outcomes. As far as I understand, using quantitative data (microbial and behavioral) and mathematical models distinguishes QMRA from traditional risk assessments . As for the mathematical models and calculations, they don’t have to be as complex as Bayesian Analyses or Monte Carlo Simulations. They could be as simple as multiplying frequency of behavior and dose per contact together.

SaniPath is an application of QMRA. However, since our goal for SaniPath is to provide overall fecal microbe exposure information instead of information specific to one pathogen, the SaniPath methodology doesn't use a dose response model, which must correspond to one specific pathogen. Therefore, the SaniPath methodology functions as an exposure assessment rather than an estimate of the probability of infection.

An exposure assessment is a two part process that includes 1) evaluation of pathways that allow microorganisms to be transported from the source to the point of contact with human beings and 2) estimation of the amount of exposure that is possible between humans and contaminants. Conducting a realistic exposure assessment is important for good risk characterization as well as risk management. (ref: qmrawiki.canr.msu.edu/images/Introduction.pdf)

The SaniPath exposure assessment for the Phase 1 in-depth analysis tracks back the source of fecal microbes that enters the human body and estimates its amount using a huge Monte Carlo Simulation. The microbial and behavioral data are all based on distributions, which are estimated using Bayesian Analysis. We can provide more detailed information about our exposure assessment after our exposure paper is finished.

Thanks,
Yuke (Andrew) Wang

Yuke Wang, MSPH
Data Analyst
Emory University
Rollins School of Public Health–Hubert Department of Global Health
Center for Global Safe Water
1518 Clifton Road, NE
MS: 002-7BB CNR6040B
Atlanta, GA 30322
Phone: 404-727-2238
Fax: 404-727-4590

Dear Suraja,

This might be a stupid question but could you please explain to me how QMRA* relates to your Sanipath tool, i.e. is Sanipath an application of QMRA or an extension (or simplification) of it?
You mentioned QMRA once in your project description under objectives:

o Conduct an exposure assessment and quantitative microbial risk assessment (QMRA).

Thanks,
Elisabeth

* QMRA = Quantitative Microbiological Risk Assessment, see also: en.wikipedia.org/wiki/Quantitative_Micro...ical_Risk_Assessment

Yes, in the example graph that I posed with the three pathways, it would be reasonable to expect that a user of the tool may choose to prioritize exposure from produce over that of drains. You can see from the graph that what is driving this greater relative risk is really the frequency of ingestion of produce (in Ghana where this data is from, consuming raw produce is common) as opposed to frequency of exposure to drains (which may be accidental and relatively rare).

Variability is definitely an issue- particularly when we deal with limited samples and survey sizes, and as you mentioned sampling variability/chance effects. In an effort to stay simple, this first version of our tool's analysis calculated risk of exposure using point estimates, which do not take into account variability. However, we recognize that variability is a concern and we are currently working on utilizing the distribution of sample and survey data so that variability can be better accounted for.

Thanks, that's helpful. In the example of the "risk of exposure from three pathways in one neighbourhood for adults" graph in your first post, might a government use the tool to prioritise exposure from the produce rather than the drains?

I think your worker in the video I posted was mentioning that the soil around the school and even the school desks were highly polluted - so I'm guessing a similar analysis had been used to show that these were areas to address to reduce infection rates.

One thought I had was around variability of the sampling - and chance effects. If a municipality used the tool and found that the school soil (for example) was an area to prioritise, it might set in place a policy to reduce that risk. If it then reanalysed using this tool, is it possible that a sample with low E.coli rates could be seen as a 'success' - when in reality it was just accidental? Or alternatively maybe the particular desk where a single child was sitting was particularly contaminated, whereas the others were cleaner. How robust is your tool to these kinds of variations?

Hi Joe, thanks for your message. That is an excellent question, and I apologize that it wasn't clear in the original post. After samples and surveys are collected, the data are combined along with other parameters from the literature, such as intake volumes and duration of exposure to characterize risk of exposure (see diagram below). The output of the tool allows people to compare the relative risks of exposure for different pathways. For example, if you are looking at open drains, produce, and piped water, the tool would allow you to be able to rank the risks relative to each other. The output also allows users to understand what is contributing more to the risk- frequency of exposure or dose (a combination of contamination level and intake values from the literature). Our goal is that the tool can be used to prioritize investments/programs based upon which pathways pose the highest environmental risk. We want to encourage users to think of upstream interventions that can help intercept the exposure (for example, covering open drains) or decrease contamination (chlorinating piped water at the municipal level). Currently, we are focused on pathways in the public domain that we believe to be most actionable by municipal governments and policy makers. I hope that helps to answer your questions. Let me know if you have any other questions or comments. We are always looking for ways to improve upon the tool.

-Suraja

Dear Suraja, can you give us a bit of context on how the SaniPath tool will be used? They collect samples to analyse for E.coli, but what does the tool tell them? Is it to be used to identify the most important health or hygiene risks in the community?

I have been watching this video from your website, which is very interesting but does not seem to explain much about using the tool

I would like to tell you about our project, SaniPath, which we have been working on since 2010 with funding from the Bill and Melinda Gates Foundation. The lead organization is Emory University in the US and I am one of the research project coordinators on the study.




Title of grant: SaniPath - Assessment of Fecal Exposure Pathways in Low-Income Urban Settings

• Name of lead organization: Center for Global Safe Water at Emory University
• Primary contact at lead organization: Christine L. Moe, PhD, (Principal Investigator); Suraja Raj, MPH (Research Project Coordinator)
• Grantee location: Atlanta, GA (USA)
• Developing country where the research is being or will be tested: Accra, Ghana; Vellore, India; Maputo, Mozambique, other locations to be determined
• Start and end date: 09 November 2010 to 31 December 2015 (extended to December 2017)
• Grant type: Global Development Grant
• Grant size in USD: $3,721,284 as per the BMGF grant database: www.gatesfoundation.org/How-We-Work/Quic...s/2010/11/OPP1016151

Short description of the project:

Rapid urbanization has led to a growing sanitation crisis in urban and peri-urban areas of low-income countries. Currently, over half of the global population reside in urban areas, with cities in Asia and Africa projected to experience the greatest increase in population. Rapid population growth in developing countries has outpaced existing infrastructure, including water and sanitation systems, leading increasing numbers of urban dwellers, especially those in informal settlements, to live in very polluted environments. Despite the considerable sanitation needs of urban and peri-urban communities, there are little data to inform strategies to mitigate risks of fecal exposure in developing countries. Consequently there is a need for site-specific evidence to help make decisions about sanitation investments.

Phase 1 on the study involved in-depth data collection on behavior and environmental contamination along multiple pathways of exposure in Accra, Ghana. Examples of pathways of exposure to fecal contamination include drinking water, recreational water, waste-water irrigated produce, surfaces in public latrines, drain water, etc. Based on lessons from Accra, we have been developing the SaniPath Rapid Assessment Tool to assess key pathways of exposure in urban, low-income settings. The tool guides users in collecting much needed data on exposure to fecal contamination in low income, urban neighborhoods, and synthesizes these data for community, government, and service providers’ decision-making. The pathways are characterized by taking into consideration the frequency of exposure to a particular pathway along with the dose (the amount of E.coli ingested per ingestion event). The tool, which consists of training material, protocols, as well as data entry/analysis software, is publicly available on our website or on a USB flash drive and is intended for non-governmental organizations, funders, and government institutions.




Goal(s):

To improve our understanding of risks from fecal contamination in urban and peri-urban settings in developing countries and ultimately develop effective strategies and inform investments and policies to reduce these risks.


Objectives:

The objectives of the SaniPath Study are to :

• Conduct an exposure assessment and quantitative microbial risk assessment (QMRA) with Phase 1 behavioral and environmental data from Accra, Ghana to understand exposure to fecal contamination in low-income, urban neighborhoods.
• Use lessons learned from Phase 1 in developing the SaniPath Rapid Risk Assessment Tool, which utilizes a mixed methods approach to understanding exposure to fecal contamination (i.e. environmental sample collection, individual and group surveys, key informant interviews, GPS mapping, etc.)
• Synthesize these data for community, government, and service providers’ decision-making
• Create a tool whose results will be used to characterize a neighborhood or city according to a matrix of fecal exposure pathways. The outputs will serve as a simplified, but still informative, means of identifying priorities for sanitation investments or interventions.

Research or implementation partners:
o Noguchi Memorial Institute for Medical Research (Accra, Ghana)
o The London School of Hygiene & Tropical Medicine (London, UK)
o The International Water Management Institute (IWMI) (Accra, Ghana)
o Water Research Institute (WRI) (Accra, Ghana)
o TREND Group (Accra, Ghana)
o Research Triangle Institute (RTI) (Raleigh, NC, USA)
o Christian Medical College, Vellore (Vellore, India)
o Georgia Institute of Technology (Atlanta, GA, USA)

Links, further readings – results to date:
Our project website: www.Sanipath.com
Documents in SuSanA library (includes my recent presentation at the FSM3 conference in Hanoi): www.susana.org/en/resources/library/details/2192

See my presentation on video here (it starts at 1 min. 33 s):



Direct link to where my part starts





Current state of affairs:

Phase 1 was completed in 2012 as a 15-month in-depth interdisciplinary exposure assessment in Accra, Ghana in order to identify, characterize and prioritize fecal exposure pathways in both the public and private domains using environmental sampling to quantify contamination and social science methods to investigate behaviors that lead to exposure. Phase 1.5 was completed in 2013, and involved the development and piloting of the first version of the SaniPath Rapid Assessment Tool. Phase 2 included further development of the tool (including softwar e), piloting of the tool in Vellore, India, as well as additional data collection through collaboration with the Mal-ED study in Vellore (mal-ed.fnih.org/) in order to understand the relationship between environmental exposure risk and enteric infection rates.

Data from Phase 1 and 2 are currently being analyzed and prepared into publications . We recently finished baseline data collection in Maputo, Mozambique through a collaboraion with the Georgia Institute of Technology’s MapSan study (tinyurl.com/labgubb)Through this collaboration we hope to understand the applications of the SaniPath tool in monitoring sanitation interventions. Over the next few months, we will be refining components of the SaniPath Rapid Assessment Tool based on lessons learned from tool deployments and input from target user groups. Improvements we are exploring include enhanced flexibility, improved data visualization and analysis components of the software, and adaptation for mobile data collection (with mobile phones).

Biggest successes so far:

• Development of simple protocols for behavioral and environmental data collection (www.sanipath.com/manual/)
• Identification of relevant pathways of fecal contamination in the public domain (i.e. shared spaces and areas actionable by government policies and programs)
• Collection of data from field sites in Ghana, India, and Mozambique on exposure to fecal contamination
• Development a pilot version of a software that is freely available to the public, see: www.sanipath.com/download-sanipath/

Main challenges / frustrations:

• Fecal exposure pathways may be very complex and thus difficult to fully characterize. To mitigate this risk our interdisciplinary exposure assessment investigates a range of potential exposure pathways. The mixed methods approach (i.e. environmental sample collection, individual and group surveys, key informant interviews, mapping, etc.) gives us more opportunities to characterize complex and intersecting exposure pathways.
• Data collection and microbial quantification methods can be challenging to perform in a setting with limited resources. To facilitate successful data collection we have partnered with institutions in Ghana, India, and Mozambique with established track-records of maintaining dependable infrastructure and results.
• Developing a rapid risk assessment tool that provides relevant and useful information, but maintains a level of simplicity is challenging. Identifying audiences (such as NGOs/CNGOs, city planners and government officials, contractors, or funding agencies) who would be most likely to use the tool as well as the degree of flexibility that is (or is not) needed has been key to improving the usability of our tools and protocols. Lastly, the long term sustainability of maintaining such a tool is an important consideration as we move forward.

Please feel free to post with any questions or comments!

Regards,
Suraja