On September 20, 2017 Category 4 Hurricane Maria made landfall in Puerto Rico (PR), causing widespread flooding, power outages, and lack of water service. Given the potential for infectious disease outbreaks, the Department of Veterans Affairs (VA) and Centers for Disease Control and Prevention established enhanced surveillance to actively monitor priority infections at VA facilities.
We queried VA data sources from August 27, 2017 to February 3, 2018 (pre-storm dates included to establish baselines). Trends in infectious disease ICD-10 syndrome groupings (respiratory illness/pneumonia, Influenza-like illness (ILI), gastrointestinal illness, conjunctivitis, rash-like Illness, jaundice) as a percent of total emergency department (ED) visits were tracked. The total number of laboratory tests performed, and percent positive per week, for influenza, hepatitis A, dengue (DENV), zika (ZIKV), leptospirosis, and chikungunya (CHIKV) were calculated.
ILI increased from 9.3% to 12.6% during the surveillance period (peak epi week 52: 15.7%) (Figure 1), while other ICD-10–based syndromes remained relatively stable. Weekly influenza testing increased shortly after landfall averaging 105 rapid influenza tests per week (epi weeks 41–4) (Figure 2). Influenza positivity increased in epi weeks 41 and 42 (7%), dropping the following weeks, and peaked at 15% in epi week 2 (Figure 3). Four acute infections were detected: 2 + leptospirosis DNA, 1 + CHIKV RNA, and 1 + Hepatitis A IgM. The remaining 34 positive tests were ZIKV, CHIKV, or DENV IgM positive or equivocal, awaiting confirmation (Figure 3).
We quickly established a simple surveillance system to monitor trends in priority infectious diseases. Increases in ILI, weekly influenza testing volume, and percent positive of influenza tests coincided with onset of influenza season. Diseases of public health importance were identified through laboratory-based surveillance. The impact of Maria on VA healthcare operations, including clinic closures, power outages, and disrupted care seeking patterns limited this system. However, the timeliness and flexibility of this surveillance system provides a model for disease monitoring following future natural disasters.
Figure 1.
Figure 2.
Figure 3.