Limits...
Mapping snakebite epidemiology in Nicaragua--pitfalls and possible solutions.

Hansson E, Cuadra S, Oudin A, de Jong K, Stroh E, Torén K, Albin M - PLoS Negl Trop Dis (2010)

Bottom Line: To avoid this error, we try to identify where underreporting is likely based on available information.The Nicaraguan municipalities are categorized by precipitation, altitude and geographical location into regions of assumed homogenous snake prevalence.The effects of the case detection bias on the distribution of resources against snakebites could decrease.

View Article: PubMed Central - PubMed

Affiliation: Division of Occupational and Environmental Medicine, Lund University, Lund, Sweden. erik.hansson@med.lu.se

ABSTRACT

Background: Snakebites are a public health problem in Nicaragua: it is a tropical developing country, venomous snakes are present and there are reports of snakebites treated both in the formal and informal health care system. We aimed to produce an incidence map using data reported by the health care system that would be used to allocate resources. However, this map may suffer from case detection bias and decisions based on this map will neglect snakebite victims who do not receive healthcare. To avoid this error, we try to identify where underreporting is likely based on available information.

Method and findings: The Nicaraguan municipalities are categorized by precipitation, altitude and geographical location into regions of assumed homogenous snake prevalence. Socio-economic and healthcare variables hypothesized to be related to underreporting of snakebites are aggregated into an index. The environmental region variable, the underreporting index and three demographic variables (rurality, sex and age distribution) are entered in a Poisson regression model of municipality-level snakebite incidence. In this model, the underreporting index is non-linearly associated with snakebite incidence, a finding we attribute to underreporting in the most deprived municipalities. The municipalities with the worst scoring on the underreporting index and those with combined low reported incidence and large rural population are identified as likely underreporting. 3,286 snakebite cases were reported in 2005-2009, corresponding to a 5-year incidence of 56 bites per 100,000 inhabitants (municipality range: 0-600 cases per 100,000 inhabitants).

Conclusions: Using publicly available data, we identified areas likely to be underreporting snakebites and highlighted these areas instead of leaving them "white" on the incidence map. The effects of the case detection bias on the distribution of resources against snakebites could decrease. Although not yet verified empirically, our study provides an example of how snake bite epidemiology may be investigated in similar settings worldwide at a low cost.

Show MeSH

Related in: MedlinePlus

Spatial distribution of municipalities suspected to be underreporting.The 10 municipalities with the worst underreporting indexes and the 24 municipalities identified as “low-reporters” (see Methods section: “Identification of underreporting”).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2990701&req=5

pntd-0000896-g006: Spatial distribution of municipalities suspected to be underreporting.The 10 municipalities with the worst underreporting indexes and the 24 municipalities identified as “low-reporters” (see Methods section: “Identification of underreporting”).

Mentions: 24 municipalities were identified as “low-reporters”. These were distributed all over the country, whereas the municipalities with the 10 worst scores on the underreporting index were all situated in the north-eastern part of the country (Figure 6). Six out of 24 “low-reporters” were among the 10 municipalities with the worst underreporting index scores.


Mapping snakebite epidemiology in Nicaragua--pitfalls and possible solutions.

Hansson E, Cuadra S, Oudin A, de Jong K, Stroh E, Torén K, Albin M - PLoS Negl Trop Dis (2010)

Spatial distribution of municipalities suspected to be underreporting.The 10 municipalities with the worst underreporting indexes and the 24 municipalities identified as “low-reporters” (see Methods section: “Identification of underreporting”).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2990701&req=5

pntd-0000896-g006: Spatial distribution of municipalities suspected to be underreporting.The 10 municipalities with the worst underreporting indexes and the 24 municipalities identified as “low-reporters” (see Methods section: “Identification of underreporting”).
Mentions: 24 municipalities were identified as “low-reporters”. These were distributed all over the country, whereas the municipalities with the 10 worst scores on the underreporting index were all situated in the north-eastern part of the country (Figure 6). Six out of 24 “low-reporters” were among the 10 municipalities with the worst underreporting index scores.

Bottom Line: To avoid this error, we try to identify where underreporting is likely based on available information.The Nicaraguan municipalities are categorized by precipitation, altitude and geographical location into regions of assumed homogenous snake prevalence.The effects of the case detection bias on the distribution of resources against snakebites could decrease.

View Article: PubMed Central - PubMed

Affiliation: Division of Occupational and Environmental Medicine, Lund University, Lund, Sweden. erik.hansson@med.lu.se

ABSTRACT

Background: Snakebites are a public health problem in Nicaragua: it is a tropical developing country, venomous snakes are present and there are reports of snakebites treated both in the formal and informal health care system. We aimed to produce an incidence map using data reported by the health care system that would be used to allocate resources. However, this map may suffer from case detection bias and decisions based on this map will neglect snakebite victims who do not receive healthcare. To avoid this error, we try to identify where underreporting is likely based on available information.

Method and findings: The Nicaraguan municipalities are categorized by precipitation, altitude and geographical location into regions of assumed homogenous snake prevalence. Socio-economic and healthcare variables hypothesized to be related to underreporting of snakebites are aggregated into an index. The environmental region variable, the underreporting index and three demographic variables (rurality, sex and age distribution) are entered in a Poisson regression model of municipality-level snakebite incidence. In this model, the underreporting index is non-linearly associated with snakebite incidence, a finding we attribute to underreporting in the most deprived municipalities. The municipalities with the worst scoring on the underreporting index and those with combined low reported incidence and large rural population are identified as likely underreporting. 3,286 snakebite cases were reported in 2005-2009, corresponding to a 5-year incidence of 56 bites per 100,000 inhabitants (municipality range: 0-600 cases per 100,000 inhabitants).

Conclusions: Using publicly available data, we identified areas likely to be underreporting snakebites and highlighted these areas instead of leaving them "white" on the incidence map. The effects of the case detection bias on the distribution of resources against snakebites could decrease. Although not yet verified empirically, our study provides an example of how snake bite epidemiology may be investigated in similar settings worldwide at a low cost.

Show MeSH
Related in: MedlinePlus