Using Relative Humidity Forecasts to Manage Meningitis in the Sahel. A23A-0178. Meningitis epidemics occur frequently in the Sahel. Persistent humidity ends ...
Using Relative Humidity Forecasts to Manage Meningitis in the Sahel
A23A-0178
Rajul Pandya1, Abudulai Adams-Forgor8, Patricia Akweongo5, Timothy Awine7, Maxwell Dalaba7, Vanja Dukic4, Arnaud Dumont3, Mary Hayden3, Abraham Hodgson6, Thomas M Hopson3, Stéphane Hugonnet9, Thomas C Yoksas2 1. Spark UCAR Science Education, UCAR, Boulder, CO 2. Unidata Program, UCAR, Boulder, CO 3. Research Applications Laboratory, NCAR, Boulder, CO, 4. Applied Mathematics, University of Colorado, Boulder, CO 5. School of Public Health, University of Ghana, Accra, Ghana 6. Ghana Health Service, Accra, Ghana 7. Navrongo Health Research Centre, Navrongo, Ghana 8. War Memorial Hospital, Navrongo, Ghana 9. Pandemic and Epidemic Diseases, World Health Organization, Geneva, Switzerland.
Meningitis epidemics occur frequently in the Sahel
Persistent humidity ends epidemics without using scarce vaccine Greenwood, et al. (1976) first documented a relationship between low humidity and meningitis, although people living in Sahel have long known meningitis ends with the start of the rainy season.
The Sahel is home to 60M people and annual per-capita income ranges from $500 to $1500.
Meningitis has a 10% fatality rate, and 10% of survivors are permanently disabled
Data collected by A. Adams-Forgor, Navrongo Health Research Centre
Centers for Disease Control
Epidemics occur periodically: 188,000 people were infected in 1996, 88,000 in 2009
Hayden, Akweongo and Awine found that a single case of meningitis can cost as much as two times the average annual household income
Dukic analyzed monthly meningitis cases and environmental variables from 1998 to 2008 for Navrongo, Ghana, and found that relative humidity showed the strongest correlation with number of cases with an inflection point at 40%.
Vaccine Shortages
Data came from Clément Lingani, and covered the entire belt for the 2008 and 2009 meningitis season.
Photo by Rodrigue Barry/WHO
A vaccine that is effective against all serogroups Neisseria meningitidis can only be used reactively to contain the spread of existing epidemics. • It does not produce an immune response in children under two. • It is only protective for two years. • It does not prevent transmission by the individual vaccinated . A new conjugate vaccine can be used proactively against the most common strain of meningitis, but cases of meningitis have been attributed to other serogroups.
Socio-Economic Factors Provides a more complete picture and allows for a better understanding of intervenable factors to reduce the burden of disease
Routine forecasts from international centers, courtesy of the World Meteorological Organizations’ TIGGE Project, provide an ensemble of relative humidity predictions up to two weeks in advance.
Mera examined links between hindcasted weather variables and cases in regions with meningitis outbreaks. In April 2009, the Kano epidemic threshold ended after relative humidity crossed above 40%. Attack rates fell in D’jamena and Gaya when average relative humidity for the week rose above 40%.
Hopson uses “quantile regression” to correct the ensemble forecast using historic observations. The right panel shows that the ensemble members’ hind-casted relative humidity (colored lines) bracket the observations (thick black line) after quantile regression is applied, but not before (right panel).
Dumont and Yoksas developed a live website that displays relative humidity forecasts (contours) and epidemic status (colored dots, red = epidemic). The contour colors are chosen to highlight the critical transition of humidity above 40% and be convenient to public health officials
Dukic also applied Generalized Additive Models, widely used to tie air pollution to health impacts, to meningitis. This treats timevarying confounding processes (e.g. seasonal migration, new strains) as a background function that varies the intercept by year, and asks how much of the other variability can be related to specific variables which vary more quickly in time. The analysis of the same 11 years of Navrongo reaffirms importance of relative humidity and shows that amount of CO in the air due to burning biomass (fires) is also related to meningitis.
Starting with a differential equation disease model based on known transmission dynamics, and cross-validating the model with only case data meningitis (i.e. every district/week is a unique event), we asked if the model can be improved by including other data (over 90 variables were tried). Hopson and Dukic found including relative humidity two weeks in the past improves accuracy in anticipating an epidemic by ~25%. Air temperature and Northeasterly winds also improved the forecast, but less than relative humidity.
Coupled with a two week forecat of relative humidity, this indicates an improved ability to anticipate a roll-off in epidemic four weeks in advance. Hopson and Dukic also found that the probability of an epidemic decreased significantly for relative humidity above 40%, as shown (right).
External drivers, which range from climate change to human migration which allows for the introduction of new serogroups; Exposure, including environmental factors such as pollutants (e.g., indoor cooking/dust) ; Sensitivity, or susceptibility based on age, poverty and access to health care; and Adaptive capacity, including knowledge of meningitis, behavior and traditional medicine and western medicine-based approaches to disease management.
A new tool for public health decision makers
Knowledge, Attitudes and Practices survey in K-N District in northern Ghana (pop. 150,000) led by Hadyen, Akweongo, and Dalaba: • Survey conducted between May 2010 and May 2011 • Structured questionnaire focused on knowledge of meningitis, health care seeking behaviors and cost of illness • Case control study with a total of 222 participants o 74 laboratory confirmed cases o 148 controls matched by age, gender and geographic location • 48% male, 52% female
During the 2011 Meningitis Season, public health officials from several countries in the Sahel participated in weekly phone calls discussing meningitis risk. This prototype was used in those conversations. We have a Memorandum of Understanding with the African Center for Meteorological Applications for Development to help integrate this system into their routine meningitis-relevant weather forecasting
Acknowledgement This work supported by Google.org
Research result Humidity can be used to predict the end of an epidemic with 2-4 weeks lead time
Potential public health action Allocating vaccine to areas with persistent low relative humidity
Next steps or hurdles Decision support system for regional-scale decision makers Train met services of Africa to provide forecast
CO, humidity, wind, current cases can estimate future cases by district
Earlier vaccination because of knowledge of disease trends
Decision support system for local scale
Early cases of meningitis are often mistaken for less serious disease
Earlier self-diagnosis
Public health outreach
Migration and socioeconomic status influence disease transmission in identifiable, if surprising, ways
Targeted vaccination, education about travel
Public health outreach
