Language Specific Gaps in Identifying Early Epidemic Signals – A Case Study of the Malay Language

Background: Internet-based surveillance systems have become invaluable tools to complement traditional surveillance in the early detection of infectious disease outbreaks. Search results limited to only English-language sources may lead to missed opportunities in global outbreak monitoring efforts. The Malay language has 290 million native speakers across the Southeast Asian region (Indonesia, Malaysia, Brunei, Singapore and the southern parts of Thailand and the Philippines), which is an important region for outbreaks. The aim of this research is to determine the completeness of reports of Malay language outbreak detection by commonly used open source surveillance platforms. Methods: We searched ProMED-mail, HealthMap, and Google News to investigate outbreak events in Malaysia and Indonesia between 1st August 2016 and 31st August 2017. We also cross-checked published epidemic reports in Medline to determine the proportion of outbreaks that are published. Results: A total of 371 entries from ProMED-mail and HealthMap were included. A Google News search query using Malay language keywords resulted in 453 news outbreak reports, 98 of which were missed by either or both ProMED-mail and HealthMap. During the study time frame, only 40 published epidemic reports were found in MEDLINE, showing the importance of informal data sources to identify outbreaks. Some diseases such as mumps, dengue and avian influenza had different local terminology in Indonesia and Malaysia. Conclusions: Search results limited to only English-language sources may lead to reduced outbreak detection. The inclusion of keywords in the Malay language improves epidemic intelligence in the region, but needs to be nuanced for local terminology, which may differ between countries. Informal Internet-based surveillance can improve compliance with International Health Regulations (IHR) for low income countries, because it is less resource-intensive than formal surveillance. However, there is a need to include local language search terms to improve outbreak detection, especially in regions at high risk of emerging infectious diseases.