An early-warning system based on statistical modeling that maps occurrences of the most common symptom of polio could identify large outbreaks of the disease at an earlier stage and help limit their spread, researchers said.
In a study published in Emerging Infectious Diseases, Isobel M. Blake, PhD, a research fellow at Imperial College London, and colleagues said an algorithm based on identifying clusters of acute flaccid paralysis (AFP) by location and time could have led to alarms being raised days — even weeks — before two recent polio outbreaks were officially confirmed. The algorithm, however, lagged behind for two other large outbreaks and for several smaller outbreaks, according to the researchers.
“Maintaining high-quality surveillance for polio outbreaks is essential to achieve global eradication of poliomyelitis,” Blake and colleagues wrote. “The longer the delay between the start of a polio outbreak and its detection (and subsequent response), the higher the chance of wide-scale spread and re-established transmission.”
Aiming to improve the current surveillance system for polio, which largely relies upon health care providers reporting AFP cases in children aged younger than 15 years, the researchers focused on 67,218 cases of AFP with clinical onset from 2003 to 2013 in 20 countries. These included Tajikistan, Democratic Republic of the Congo and Somalia, each with more than 150 confirmed cases of polio annually since 2005. They also looked at African countries at high risk for an outbreak of wild poliovirus. For each case, the investigators recorded the patient’s age, sex and residence, the dates of AFP onset, case notification and stool sample collection. Stool samples confirmed 3,089 cases of wild poliovirus and 70 cases of circulating vaccine-derived poliovirus (cVDPV). Among the cases without adequate stool samples, 1,436 were determined by a panel of experts as polio-compatible.
The study used population estimates for each country and studied delays in the reporting of AFP cases in Africa from 2010 to 2013. A mixed-effect regression model predicted background rates of nonpolio AFP at the district level.
The investigators created a real-time AFP database for each Monday from 2003 to 2013 by assuming cases entered the database on the date the case was noted by local health care providers. The researchers used the Kulldorff scan statistic to build space-time cylinders for every district in each country. The surveillance was considered successful by its ability to detect confirmed outbreaks of serotype 1 and 3 wild poliovirus and cVDPV. The results were classified as being detected by the algorithm if a warning alarm was raised within the outbreak period and if the location of the alarm occurred in at least one district containing reported outbreak-associated cases.
Algorithm detects large outbreaks sooner
The algorithm beat official confirmation of large 2010 polio outbreaks in Tajikistan by 39 days and the Democratic Republic of the Congo by 11 days, the researchers said. Two other large outbreaks in Somalia in 2005 and 2013, however, were detected 11 and 27 days after official confirmation. The algorithm also did not perform as well in detecting smaller outbreaks in sub-Saharan Africa. Few false alarms were recorded except in the Democratic Republic of the Congo. Further, the researchers said the algorithm would have generated alarms during cVDPV outbreaks in four African nations, showing its functionality during the transition to remove serotype 2 OPV from routine immunization in April.
“Polio outbreaks that are detected late will threaten the progress of the [Global Polio Eradication Initiative (GPEI)], and consequently there is a need to strengthen ongoing surveillance,” Blake and colleagues wrote. “Although future work is required to test our algorithm in other settings, we have shown that integrating an automated early-warning system based on detection of AFP clusters into the polio information system could be of value to the GPEI.” – by Gerard Gallagher
Disclosure: The researchers report no relevant financial disclosures.