Interesting update on the GEMS study:
www.cawst.org/blog/bydate/2016/12/diarrh...se-the-top-culprits/

It seems like more thought has to be put into preventing the spread of viruses especially regarding pitlatrines as viruses can travell quite far in groundwater compared to the typical ecoli indicators.

The short answer is I agree that in most cases microbial fingerprinting would not be practical for many public safety monitoring purposes. In our project we are using fingerprinting data to address methodological questions about the performance of general fecal indicators, individual pathogens, and combinations of pathogens at accurately detecting enteropathogens. Using Kisumu as a model, we are quantifying uncertainty associated with using any of these strategies to assess enteropathogens in publically-shared environments. As an example, let’s say Cryptosporidium has a 0.6 probability of accurately predicting whether any other pathogen is present or not present, Crypto and the ST toxin gene of ETEC achieves 0.75 accuracy, and Crypto, ST, and Adenovirus achieves 0.91 probability. Stakeholders like you mentioned can weigh the cost-benefits of progressively more expensive approaches and make informed decisions about how much uncertainty they are willing to accept.

The longer answer requires more space than reasonable for a blog, but is basically that no one really knows just yet whether the interactions between humans, the environment, and human pathogens or the indicators used as proxies can be easily reduced to simple relationships. I think that (mostly) everyone acknowledges that it is necessary - the WASH sector needs to get better about objectively evaluating policy and interventions/programs. Do shared latrines pose higher risks of exposure, and if so how? Are improvements in sanitation coverage reducing open defecation? How much impact will latrines have given all of the other sources of feces and pathways to infection?

Unfortunately, all the evidence suggests that in places like Kisumu, the story may be more complex than we hope. And the easiest to use, lowest cost indicators are not appropriate for many of these questions although they are quite useful for drinking water/food/hands where tolerance for any contamination is low. The starting point begins with using rigorous science like microbial fingerprinting to understand how humans, animals, etc. affect safety of the environment and to figure out which indicators best measure risks from different sources. Second is to validate those patterns across sites with similar developmental challenges. Then we will are equipped to have an evidence based discussion about observable factors or indicator assays for risk evaluation by utilities and others.

I hope that doesn't come across as a non-answer, but I think we have to be transparent about current knowledge gaps.

Kelly

so reluctance to 'change behaviour' from open defecation to using shared toilets is perfectly understandable.
In general, people might quite rationally choose open defecation, especially in the bush, as more hygienic for the individual defecator even though it's less hygienic for everyone else.

This sounds definitely very interesting from an academic point of view, but I have a bit of an hard time imagining how microbial fingerprinting ( en.wikipedia.org/wiki/Community_Fingerprinting ) could become part of a risk-evaluation tool set of a urban utility company responsible for sanitation services for example.

Maybe you have already thought about the possible practical applications in more detail and would be willing to share some thoughts on it here?

Hi Elisabeth,
I knew about Susana, but a colleague that works for an NGO involved in EcoSan projects reminded me that this is a great forum where researchers like myself can understand the priorities and perceptions of practioners, and maintain awareness of projects that may not necessarily wind up in journals or text books. As a researcher hoping to advance the science of global development evaluation - this is important!

We are working on the Social Microbes manuscripts now. But some details from a UNC conference presentation last fall are here: www.researchgate.net/publication/2834211...urban_slums_in_Kenya

So far our major take-away observations (from Kisumu, Kenya) are:

1. The microbial diversity of fecal pathogens detected in neighborhoods with low sanitation coverage and limited domestic animal management is extraordinary. We found ~14 - 19 (depending upon neighborhood) different fecally-transmitted diarrhea or helminth pathogens in soil and surface water in Kisumu in just one 1-week rapid assessment during a dry to wet transitional season. There is a great deal of overlap between the top 5 pathogens detected in <5 children's stool in GEMS and what we are detecting in the environment. In comparison, at a "control" site in Iowa, areas impacted by cattle and pig farms, we typically detected only 1 pathogen (Crypto) at far lower frequencies.

2. We can tell the difference between sites impacted by domestic animals or open defecation. Animals and OD increase the risk of detecting any fecal pathogen in communal areas, and more importantly they increase the risk of simultaneously detecting multiple fecal pathogens. We are also detecting greater fecal pathogen diversity at the neighborhood level as well, although we are still working on disentangling whether those differences are due to differences in sanitation coverage vs other factors like infrastructure, tenancy, wealth, etc. The fecal fingerprint (diversity) for animal impact on environmental contamination versus OD are different, suggesting different potential indicators for risk evaluation.

3. Children playing in communal areas can be ingest multiple potentially-infectious pathogens per day from relatively infrequent contact behaviors.

This was for all practical purposes a pilot project, and we are hoping to expand on this study soon to address many of the questions relevant to this forum.
Kelly

Dear Kelly,

Great to see you here on the forum, as one of the authors whose paper is being discussed here. I am really happy about that (did someone alert you to this thread or did you just happen to come across it?).

You wrote:

We have found extraordinarily complex patterns of enteropathogen exposure in urban child play areas (see Social Microbes study)

Which Social Microbes study do you mean?

Regards,
Elisabeth

This a good study and required to be strengthen around the world in order to make sure sanitation education spread to urban and rural areas,

In May 2016 issue of PLOS Medicine, an interesting paper, titled: Sanitation and Hygiene-Specific Risk Factors for Moderate-to-Severe Diarrhea in Young Children in the Global Enteric Multicenter Study, 2007–2011: Case-Control Study (available at journals.plos.org/plosmedicine/article/a...nal.pmed.1002010.PDF) was published, in which Kelly Baker et al. examined sanitation and hygiene exposures, including shared sanitation access, as risk factors for moderate-to severe diarrhea (MSD) in children less than 5 years of age.

Kelly Baker and colleagues’ Global Enteric Multicenter Study (GEMS) collected data on MSD among children reporting to health centers in seven sites in seven countries from 2007 to 2011, with cases matched to controls by village and homes visited within 90 days to observe sanitation and hygiene conditions.

The authors conclude that sharing a sanitation facility with just one to two other households can increase the risk of MSD in young children, compared to using a private facility. Interventions aimed at increasing access to private household sanitation facilities may reduce the burden of MSD in children. The World Health Organization/ United Nations Children's Emergency Fund categorize shared sanitation as unimproved.

Jonny Crocker and Jamie Bartram have discussed the paper of Kelly Baker and colleagues (Interpreting the Global Enteric Multicenter Study (GEMS) Findings on Sanitation, Hygiene, and Diarrhea, (blogs.plos.org/everyone/files/2016/04/journal.pmed_.1002011.pdf). The authors say that there are a number of limitations to the sanitation and hygiene indicators that suggest caution in interpreting the findings. The authors say:

“GEMS sanitation and hygiene indicators are at the household, not individual, level and are
indicators of access, not behavior (except child feces disposal). Access does not equate to use,
and behaviors within a household often vary, for example, by age and gender. Survey best
practice is to inquire about individual behaviors both at and away from home, in addition to
observing sanitation facilities. Likewise, the link between available handwashing materials
and behaviors is not a given.”

More than 761 million people rely on shared sanitation facilities (Shared Sanitation versus Individual Household Latrines: A Systematic Review of Health Outcomes, by Marieke Heijnen, Oliver Cumming, Rachel Peletz, Gabrielle Ka-Seen Chan, Joe Brown, Kelly Baker,and Thomas Clasen).

A blog in Sci Dev by Munyaradzi Makoni (Shared toilets increase diarrhoea risk for children - www.scidev.net/global/children/news/toil...3A%2016%20May%202016), says that sizable risk to under-fives results, when two or three households share toilet; and such facilities are only marginally safer for kids than public latrines.

Kelly Baker and colleagues have produced useful paper, collecting data on MSD among children reporting to health centers in seven sites in seven countries from 2007 to 2011, including Pakistan. Discussion by Jonny and Jamie is also interesting.

Despite the fact that 761 million people resort to shared sanitation, my own experience in Sindh province, Pakistan, shows that that sharing toilets among households in Sindh is almost negligible. This is, in part, due to the religion – predominantly Muslim population.

Some might contradicts this, as according to a WSUP report (Can behaviour change approaches improve the cleanliness and functionality of shared toilets? A randomised control trial in Dhaka, Bangladesh – available at
r4d.dfid.gov.uk/pdf/outputs/Wsup/DP009-E...f-shared-toilets.pdf), “Households living in densely populated urban slums often lack the space for their own toilet, making shared sanitation the only viable solution. This is the situation in Dhaka, where many of the city’s low-income residents depend on one of the city’s enormous number of shared compound toilets: a recent study by the International Centre for Diarrhoeal Disease Research Bangladesh estimated that 4.3 million people in Dhaka use such facilities.”

A recent news item in Dawn of 10 May 2016 (www.dawn.com/news/1257403/recruitment-of...ness-at-kp-hospitals) gives an insight as to how religion affect sanitation.

Bangladesh is a Muslim country. However, in case of Sindh, another factor that comes into play is our Sindhi traditions. It is not possible for a Sindhi woman to use a toilet that is used by male from another house.

Shared toilets in Sindh (e.g., school toilets, toilets in offices, toilets in highway restrooms, etc) are most pathetic and, are sure to cause major diseases, let alone diarrhea in children. No wonder, joint monitoring program of WHO and UNICEF calls shared sanitation as unimproved.

F H Mughal

I'd like to add some background information that could not be included into our paper, but provides interesting food for thought for this discussion.

First, the original goal was to use open defecation (OD) as the reference for comparison, but OD was too low at some sites so we could not do this systematically for all sites. However, for the sites where open defecation could serve as a reference, we observed a two-fold or more INCREASE in odds of MSD * for sharing at any level in Pakistan (semi-urban) and Mozambique (rural), with no significant difference in MSD elsewhere.

Second, pan-site model adjustment was limited by sample sizes in some sanitation access categories for some sites. Two indicators of wealth were chosen since as common higher-level drivers of access to water, soap, food, etc., they were the best options to control for a broad scope of latent variables. However, there was correlation between sanitation access level and other sanitation and hygiene practices at some sites. For the sites where sample sizes permitted more extensive adjustment, the relationship between sharing sanitation and MSD was remarkably stable. Adjusting for handwashing with soap, improved drinking water access, open child feces disposal, and conditions in latrines did not change our conclusions.

Third, in addition to age and gender GEMS cases and controls were matched by village or neighborhood. In theory, this matching process could have controlled for macro-scale community conditions, which would mean that cases and controls might have similar potential for exposure to public domain contamination. However, this matching might not have captured micro-scale differences in conditions, like wealth pockets or living near OD sites. But it does at least hint at the possibility that if effects from shared sanitation were a proxy for some sort of latent factor, it is likely to be in the private domain.

Alternatively, we will soon explore a hypothesis in an unrelated study in Kenya that maternal practices related to where children are allowed to play outside the home is a key mediator of child exposure. We have found extraordinarily complex patterns of enteropathogen exposure in urban child play areas (see Social Microbes study). Maybe mothers who keep a tighter leash on their children buffer their interaction with highly contaminated public areas. Translation to the GEMS study, maybe mothers who prioritized watchfulness of child behavior also prioritized private sanitation, or even latrine cleanliness. The nuances of the GEMS study design make interpretation of our results even more intriguing.

K. Baker


* MSD = Moderate-to-severe diarrhea (added by moderator)

I found this part (p.12) from the discussion interesting:

GEMS findings suggest access to private household latrines can provide protective benefits against MSD, even in communities like rural western Kenya, where open defecation and open child feces disposal was common. It is unlikely that private latrine access influences whether children play outside the home or not, so this protective effect may reflect protective benefits of private household latrine access on private (domestic) exposure pathways such as contaminated drinking water, food, household play areas, or hands.

This is interesting because it would go against common wisdom that the protective effect of toilets is mainly a community wide effect and thus full ODF should be reached. However I think they have cause and effect slightly confused in this paper with their narrow focus on household toilets Vs. shared facilities.
But a bit further down they seem to at least acknowledge that possibility:

The reverse explanation for this relationship could also be true: households that prioritize safe hygiene practices are more likely to invest in private sanitation facilities than those that do not prioritize hygiene.

And I would add: households that can afford private sanitation facilities are much more likely to be also able to afford other things that have a big effect on occurrences of child MSD.

Edit: they did some social-economic adjustments, but it seems to have been rather limited:

We selected two sociodemographic indicators, a wealth quintile index and having two parents in the home. Although wealth indices are widely used in WASH research, it may not be a robust way of adjusting for sociodemographic confounding.

Thanks Eddy. I took another closer look and realise that there is only one research paper here (from Kelly Baker and many "et al"s) whilst the other is in fact a comment on the paper by Baker et al by Crocker & Bartram. Nonetheless, both the paper and the associated article are highly insightful and thought provoking and I too would also encourage the SuSanA community to read, reflect and consider what this means on policy and practice.

As stated by Baker et al "This is a controversial topic, as communal facilities are the most economical and feasible solution for providing sanitation access to the 2.5 billion people without a private facility". Indeed this is the case if we compare the cost of installing one communal toilet compared with the cost of each family installing private facilities. But this does not take into account important issues of space and the fact that many families lack security of tenure and lack incentive to invest in improved latrines. Additionally, private landlords may raise the rent when there is access to a private facility. So, it's certainly a tricky one.

It is apparent why many organisations choose to use funds to construct communal facilities as well-constructed household toilets according to standards are so expensive. It is harder to adopt a CLTS approach and not offer any form of support to build the toilets.

Generally I think it is true to say that donors prefer to go for communal or private but not something in the middle. However container-based sanitation may turn this on it's head as it provides the privacy associated with private on-plot sanitation and is clearly much more affordable. Key factors are usage and quality of service. A poorly maintained facility is likely to become a loci of disease transmission whereas a well-maintained one, even if used by a large number of people, can be an effective means to reduce transmission of diarrhoeal disease provided of course that users practice good hygiene behaviour.

I do agree that with the conclusion that "shared facilities may still have a role in addressing open defecation in challenging settings" but I also agree that "For reasons beyond just health such as dignity and gender equity, we should advocate for private access whenever possible." And if Crocker & Bartram state that this is based on the best dataset yet on diarrheal disease associated with sanitation and hygiene, then that's a compelling case indeed !

p.s. I've attached the articles to this posting to make it even easier for you to find the paper, download and digest.

Colleagues:

This is a very important topic. I encourage all to take a few minutes to read the discussion and conclusion section of the actual paper ( not the review of the paper). It is only a couple of pages and written in accessible language and provides ( in my opinion) a well balanced and nuanced discussion of the issues of private vs. shared and also important to understand the articulated limitations of what the data was able to look at what is was not able to look at.

The discussion in the paper acknowledges that the private household vs shared latrine debate is not helpful in situations were private household latrines are not viable ( such as in extremely densely populated informal urban settlements where over 1 billion people live - see recent Habitat data on slums) and where shared sanitation is the only option - often as public latrines managed by local governments, NGOs or private sector operators. These are indeed often not safely managed and provide inadequate protection from fecal contamination. But given the lack of alternatives, reducing the discussion to HH v Shared often results in a disincentive by governments and development partners to invest in shared sanitation to make more hygienic and reduce public health risks.