Ag in the Aftermath: A Framework for One-Health Monitoring in Haiti

¶ It is now three weeks following the devastating earthquake in Haiti, with more than 150,000 estimated deaths, that number again in injuries, and 1.5 million Haitians displaced.  As we move forward past the initial response, public health and infrastructure disruption continue to grow as primary concerns (when I use the term ‘public health’ I generally mean the overall interaction of human, animal and crop health).  Housing conditions, post-trauma injuries, mental health, and infectious disease all rank high as human public health threats, while the agricultural infrastructure faces challenges associated with mass population migration, food security, infectious disease, and production-related dysfunction.  The condition of Haiti’s public health infrastructure was degraded before the earthquake, which will exacerbate the health and infrastructure challenges to come.

¶ There is an urgent need for increased and continuing public health monitoring throughout the country.  What mechanism will be used to track disease and public health infrastructure in the coming days and months?  How will international organizations, the Haitian government and relief agencies track and respond to the inevitable public health emergencies that are likely to crop up as we move forward in rebuilding Haitian infrastructure?  Non-traditional methodologies and technologies can play a significant role in monitoring and responding to such important infrastructure risks, it’s just a matter of deciding the best way forward since the tools are available.  Continuing to break through bureaucratic and organizational barriers is the key. 

¶ Urban centers such as Port-au-Prince, Jacmel, and Leogane have been the primary locations for public health response and intervention to date.  However, upwards of 500,000 Haitians have already relocated to more rural areas of the country following the earthquake.  This influx of population in rural areas risks straining infrastructure in those locations, to include increasing both human health risks and strain to agricultural systems.  As described in the previous post, Ag in the Aftermath: Dearth of Haitian Agricultural Baseline Information, due to the nature of the rural Haitian agriculture industry, there was a significant lack of socio-economic and animal demographic information prior to the earthquake.  Now, the population dynamics and societal infrastructure has been even more disrupted, thereby further confounding baseline information for human and animal populations throughout the country.  Population movements and societal disruption can have profound effects on disease transmission, food security, and infrastructure strain – the ability to monitor these populations and infrastructure components will become ever more important as relief efforts move from response to recovery.

¶ So, how would one go about collecting demographic data on human/animal populations, identifying disease outbreaks and other public health concerns, while also providing useful information back to the local Haitian population?  As this press release points out, the CDC plans to collect reporting from approximately 50 sentinel surveillance sites.  This system represents the ‘gold standard’ for haiti disease surveillance and public health response; however, due to the nature of the reporting mechanism and requirements for validation, it can tend to be somewhat slow.  Additionally, it tends to be associated with one-way communication (collection) and dependent upon local Haitian use of the sentinel sites’ services.

¶ Using a multi-tiered approach to these challenges provides the most cost-effective, timely, and overall comprehensive mechanism.  The CDC approach of high fidelity, validated and structured information on current disease, injuries, and treatments is one tier.  Couple this with less validated, but more mobile and dispersed collection mechanisms, which provide a broader reach, near real-time reporting, and inclusion of cultural components.  True, such less validated mechanisms might not provide a fully accurate picture of the specific disease situation, but it provides a tip-off to a potentially disruptive public health or agricultural event that might warrant more detailed follow-up.

¶ Let’s explore the integrated components included in such a system:

System Components:

• Tier 1: Validated, Structured

Established hospitals and treatment facilities, NGOs, and international organizations (e.g. Centers for Disease Control and Prevention‘s proposed surveillance mechanism) provide high fidelity, validated, and structured information on current disease, injuries, and treatments.  They are dependent on physical presence at sentinel sites, as well as collection and validation protocols that have the risk of being fairly slow, especially with increased workloads at the various sites.  The strength is that such a system provides detailed information that is generally validated by health care professionals, relief workers, etc.  In Haiti, initial presence of such sites focused on urban centers, leaving public health and agricultural infrastructure in rural areas without coverage.  Further, since many such mechanisms rely on professionals not native to the location, some loss of cultural context is expected.

• Tier 2: Unvalidated, Structured:

Using various lightweight technical means, demographic, public health, and agriculture reporting can be brought to the level of the local population in a distributed manner.  Systems such as EpiSurveyor, FrontlineSMS, Geochat, and Ushahidi, and Sahana have already been successfully implemented and validated in various public health and disaster response roles.  The integration methodology and application of such technology on the ground in Haiti is where the true value lies.

In Africa, mobile technology has been used successfully to report local market prices to agricultural producers, in order to provide them with near real-time pricing information used by the producers to determine the best market at which to sell their products.  A co-op of producers from a community hire mobile-enabled individuals, referred to locally as ‘market spies‘, to troll the various regional markets and return current prices to the co-op.  These market spies generally also come from the region in which they serve, thereby adding valuable local knowledge and cultural considerations to their reporting capacity.

The United Nations plans to hire up to 200,000 Haitians by March to assist with post-earthquake clean up and reconstruction efforts.  I would propose hiring Haitians from various locations throughout the country to report demographic, public health and agriculture information via mobile devices.  Basic training on the reporting format, physical signs, etc., would be provided to enhance easy processing and increase specificity of reporting.  Selection of literate individuals who potentially also have specific local/regional knowledge, specialized skill sets, etc., would also increase the fidelity of reporting.  As an example, EpiSurveyor allows a user to develop specific forms to conduct surveys from mobile devices.  As the surveys are conducted, it sends the results wirelessly back to the central server.  Individual public health and agricultural surveyors could have multiple, standardized forms on their mobile devices and report remotely from regions throughout Haiti.  The standardization of questionnaires would provide more validation to the reporting; however, this would not preclude the surveyors from sending free-form or minimally-formatted SMS messages back to the central aggregator.

• Tier 3: Unvalidated, Unstructured:

The least validated and most unstructured reporting tier would be collection of unstructured SMS reporting from the public at large.  The adoption of public reporting was exemplified by the response to the Ushahidi platform following the earthquake.  The Haitian public can use a SMS short code to send requests for emergency response assistance, food, and shelter, as well as to report searches for family members.  GeoChat, by InSTEDD, is also currently being used with a shortcode to coordinate efforts between citizens, NGOs, and various government agencies.  Providing the general population the ability to report disease conditions, agricultural requirements, etc., would provide an additional source of information, adding to the overall demographic and public health picture.  The limitation in current systems is enabling (and enacting) a conversation between the aid providers, government organizations, etc., and the Haitian people.  While much previous work has focused on receiving reports from the population, we now need to transition to more of a two-way information flow.  The question now becomes, are aid organizations and government entities willing and interested to engage in this sort of conversation with the Haitian citizenry?

¶ Adoption/Incentive: How would this system incentivize the local population to report information and cooperate with public health and infrastructure reconstruction efforts?  This is one of the biggest barriers to direct mobile reporting systems and is quite dependent upon the characteristics associated with the current situation, population and economic demographics, and cultural/societal practices.  In the case of Haiti, Tier 1 reporting is sustained through bureaucratic mandate of reporting requirements at sentinel sites and adopted/supported through buy-in by health and relief professionals staffing those locations.  Separately, Tier 2 collection incentivizes the surveyors to report by maintaining their employment and salary, providing basic levels of training, and through the selection process to hire individuals who are suitable and motivated.  Finally, Tier 3 reporting raises the most difficult challenges to system adoption and incentive to report.  As was observed with Ushadihi, emergency situations can quickly lead to adoption; however, on a non-emergency basis, the same motivation to use the Ushahidi platform will greatly decrease.  Therefore, providing information back to the user population, proper advertising of the system, and providing access to services quickly and easily could all be used as incentives – it’s just a matter of identifying the correct ones.  BarCamp-like community meetings could provide an interesting method to determining the key interests and needs of the potential user population, based on their own assessment, not those of an external source.

¶ Services and Information: Successful outcomes for all three surveillance tiers ultimately rely upon the provision of services and information to the Haitian population.  Whether these services and information are direct medical care, food distribution locations, shelter and employment programs, or market prices and livestock production assistance, depends upon the target population and local requirements for assistance.  There are many examples of mobile systems being used to both collect information and provide information back to populations.  An example of this is very well described by Grameen Foundation’s AppLab in Uganda:

Farmer’s Friend is an agricultural information service based on text messages. It was launched in June by MTN (Ugandas largest mobile network operator), Google and the Grameen Foundation’s “Application Laboratory”, or AppLab. The service accepts queries such as “rice aphids”,”tomato blight”or “how to plant bananas” and retrieves advice from a database. More complicated questions are forwarded to human experts. The query “pineapple disease” elicits the answer “Copper deficiency in pineapples leads to fruit rot. Cut affected fruit as soon as noticed and dispose of where they will not contaminate other fruits or burn.”

¶ In the end, the overall time and financial investment to implement such a system would be fairly minimal in relation to the benefit of receiving demographic, public health, and infrastructure information from across Haiti.  Each tier acts in concert (and complement) to the others, augmenting reporting while mitigating the particular weaknesses of each individual component.  The technology is already available, motivated people have shown a willingness to donate time and expertise (as seen with CrisisCommons), and the need for such information is critical.  The primary challenges would be successful integration to ongoing efforts and successful adoption, by the Haitian population as well as the various government, international organizations, and NGO stakeholders currently operating in the country.

Tagged as: Agriculture, analysis, animal health, Biosurveillance, CrisisMappers, developing world, disease surveillance, Haiti, mobile technology, one health, public health, surveillance, technology