To enable comparison of periods before and during the conflict, we gathered open-source data from EPIWATCH for Ukraine for 2 periods before (November 1, 2021–February 23, 2022, and February–July 2021) and during (February 24–July 31, 2022 the conflict. EPIWATCH collects outbreak reports on specific infectious diseases and clinical syndromes, including undiagnosed syndromes such as severe acute respiratory infection, pneumonia, rash and fever, and encephalitis.

The before-conflict period served as a baseline for comparing disease reports from the during-conflict period. To capture a relevant snapshot of epidemics in Ukraine around the time the conflict began, which could influence epidemics occurring during the conflict, we selected the 3 months before the conflict as a baseline for the immediate before-conflict period. To account for seasonal variations of disease incidence, we also collected data for the same period during the previous year (February 24–July 31, 2021) as a second comparator.

EPIWATCH searches for ≈200 specific disease and syndrome terms in 46 languages, and ≈70% of all intelligence gathered is from non-English sources (17). EPIWATCH gathered data for all areas within the internationally recognized borders of Ukraine in the Ukrainian and Russian languages (because Russian is spoken in some parts of Ukraine). The Russian language had already been in use in EPIWATCH since September 2019. After Russia invaded Ukraine on February 24, 2022, the EPIWATCH team added the Ukrainian language to the EPIWATCH search engine. To reduce ascertainment bias caused by adding a new language, identical EPIWATCH search terms were applied retrospectively in the Ukrainian language by using the Google search engine for the 2 before-conflict control periods in 2021–2022; those terms were added to the dataset to compare trends before and during the conflict. Because the retrospective search used the same search terms and search engine (Google) for prospective data collection, the results should be comparable.

We downloaded EPIWATCH outbreak report data for Ukraine for the periods of interest, combined with retrospective data in the Ukrainian language, and screened the data for eligibility. Inclusion criteria for the final analysis were reports focusing on infectious diseases and syndromes among humans; zoonotic diseases (diseases circulating among animals that can be transmitted to humans); and having data regarding confirmed, probable, or suspected cases. We excluded reports not meeting those criteria.

To conduct descriptive epidemiologic analysis of outbreak reports, we used Microsoft Excel (https://www.microsoft.com). We extracted data on diseases, syndromes, populations affected, and location and time and analyzed the data by reporting periods. We compared numbers of reports from the before- and during-conflict periods as well as from the previous year and constructed graphs by using Prism (GraphPad, https://www.graphpad.com).

We next extracted case numbers from the reports. We extracted from EPIWATCH reports the number of cases of the 8 most reported diseases and created a line list for further analysis. We included cases in the line list only if reports had definitive numbers. If case numbers were vague or cumulative, we excluded them. We also removed duplicate case numbers. Then, we searched for formal case numbers published by organizations engaged in formal surveillance (e.g., the World Health Organization and the European Centre for Disease Control and Prevention). We then compared during-conflict case numbers from EPIWATCH with case numbers of the same diseases published by formal surveillance during the same period, when available.

Figure 1

Figure 1. Daily EPIWATCH (https://www.epiwatch.org) reports of outbreaks in Ukraine before and during Russia’s invasion, November 1, 2021–July 31, 2022.

Figure 2

Figure 2. Comparison of daily EPIWATCH (https://www.epiwatch.org) reports of outbreaks in Ukraine during Russia’s invasion (February 24, 2022–July 31, 2022) and in the same period of the previous year (February...

We identified 805 outbreak reports in EPIWATCH during February 24, 2022–July 31, 2022, for an average of 5 reports per day. In comparison, 259 reports were collected in the 3 months before the conflict (November 1, 2021–February 23, 2022), an average of 2 reports per day (Figure 1). Reports initially peaked (n = 14) on April 7, 2022, largely regarding a diphtheria outbreak in the Ternopil Oblast of western Ukraine. The largest peak was in June, when 37 reports were attributed to an outbreak of cholera in Mariupol in eastern Ukraine. For the same period in the previous year, we found 180 reports, indicating a 447% increase during the conflict (Figure 2).

Figure 3

Figure 3. Number of reports of clinical syndromes per day before and during Russia’s invasion, Ukraine, November 1, 2021—July 31, 2022.

Before the invasion, there were a total of 4 reports of clinical syndromes, all of which were acute gastroenteritis (Figure 3). After the invasion, syndromic reports increased; acute gastroenteritis (87.1% of reports) was the most common, followed by influenza-like illness (6.5%), fever of unknown origin (3.2%), and meningitis (3.2%).

According to individual reports from June 2022 of acute gastroenteritis, many illnesses were identified as dysentery or suspected cholera. Among other reports of acute gastroenteritis, a report from May 26, 2022, cited an outbreak of unknown cause resulting in acute gastroenteritis signs/symptoms among 17 internally displaced persons in the Lviv region. In addition, starting on July 20, 2022, an outbreak affecting 91 persons within 1 week was reported in the Odessa region. A report of meningitis recorded on July 31, 2022, mentioned an overall increase in viral/bacterial meningitis cases in Ukraine since before the conflict.

Before the conflict, the most frequently reported outbreaks were COVID-19, influenza, and poliomyelitis, accounting for 70% of the total reports during that period (Table 1). In the during-conflict period, the most frequently reported diseases were COVID-19 (increased from 106 to 187 reports), cholera (from 0 to 157 reports), botulism (from 22 to 122 reports), TB (from 11 to 75 reports), HIV/AIDS (from 6 to 61 reports), rabies (from 12 to 40 reports), salmonellosis (from 4 to 39 reports), and diphtheria (from 2 to 29 reports) (Table 2). Only cholera was not reported before the conflict. Frequency of reporting decreased for poliomyelitis (from 37 to 19 reports), influenza-unspecified (from 39 to 10 reports), and influenza A(H3N2) (0 reports) during the conflict (Table 1).

With regard to our comparison of case numbers from EPIWATCH and from formal surveillance, most diseases were not reported by formal surveillance, except for botulism, TB, and diphtheria (Table 3). For botulism, the number of cases identified through EPIWATCH (32 cases) was lower than that identified through formal surveillance (51 cases) (22). For TB, the only statistic found by formal surveillance was for the Rivne Oblast for the first quarter of 2022 (113 cases), whereas EPIWATCH was able to collect case number data for the entire reporting period and across more locations (5,647 cases) (23). Both EPIWATCH and formal surveillance identified 2 cases of diphtheria (22).

The ongoing escalation of the conflict in Ukraine is one of the world’s fastest growing humanitarian crises and has disrupted health systems and reduced outbreak detection and response capabilities in the country (24). We have demonstrated the value of using open-source epidemic intelligence to gain information about unfolding epidemics and public health priorities in a conflict zone where formal surveillance is reduced or lacking. We were able to identify new epidemics that occurred during the conflict, such as cholera, botulism, and human cases of rabies (which were presumably exacerbated by an increased number of displaced domestic dogs). The increases in cholera and gastroenteritis reflect declining hygiene and sanitation during the conflict, including lack of access to safe drinking water and toilets and subsequent improper disposal of fecal waste (25). The increase may also be exacerbated by improper handling and disposal of dead bodies (26), such as burial in shallow graves, which increases the risk for transmission of some diseases (26,27). We also identified increased HIV/AIDS and TB cases. HIV clinics were shut down during the conflict (4), most likely affecting testing and surveillance. In addition, medications such as antiretroviral drugs have become scarce or been misused, which may increase the risk for drug-resistant HIV and may subsequently limit treatment options and further increase transmission (6). Overall, the disruption of transportation networks during the conflict has decreased access to medical supplies and treatment for infectious diseases (4). Lack of access to testing and treatments has also resulted in loss of continuity of care, poorer outcomes, and increased community transmission (6).

Using EPIWATCH, we were able to extract more complete case data for the 8 most reported infectious diseases compared with formal surveillance. During the conflict, formal reporting of infectious diseases such as COVID-19 notably decreased, most likely because of lack of testing. The collapse of formal surveillance systems in Ukraine during the conflict resulted from a variety of factors (e.g., high levels of displacement, attacks on healthcare facilities, lack of routine data collection, limited testing and treatment, reduced diagnostic capabilities, and changes in disease testing policies) (28). After the invasion, case numbers from the Ukraine government or other formal surveillance sources for many diseases either ceased to be reported or were not up to date. Some formal surveillance systems continued to report cases of botulism, TB, and diphtheria, but even for those, we showed that numbers are probably underestimated.

Weak or absent formal surveillance during wartime hampers timely and targeted interventions such as vaccination programs (29). Solely relying on formal surveillance may result in missed early warning signals from other sources, heightening the risk that diseases will spread internationally, particularly if refugees migrate to other countries. In that context, open-source epidemic intelligence can provide early warnings of epidemics.

Among the limitations of our study, open-source data are not subject to validation compared with formal surveillance data. However, EPIWATCH uses 3 AI subsystems to improve data quality and exclude irrelevant material (30). All data are then further curated by human analysts, who follow standardized operating procedures. EPIWATCH provides reports of outbreaks, rather than case numbers; thus, monitoring report trends reflects signal strength, and case numbers extracted from reports may be less precise. However, for most diseases there was no formal surveillance or case reporting during the conflict, whereas we were able to extract case numbers from the open-source reports. Another limitation is ascertainment bias because media reporting may be increased in regions of Ukraine with larger populations. As a result, data from smaller regions were less frequently obtained, and thus, infectious diseases may be underestimated. Last, because we added the Ukrainian language in February 2022, ascertainment bias could have contributed to the observed increase in reports from that time. However, we accounted for that difference by using a control period in the previous year, for which we retrospectively collected data in the Ukrainian language. Historical open-source reports of outbreaks from Ukraine might have been removed or censored by 2022, which may have reduced the potential for data capture; however, there is no evidence to suggest this was the case. A longer control period may have been more informative of longer-term trends and is planned in a follow-up study.

Our study provides an overview of epidemic activity in Ukraine during the Russian escalation of the Russo-Ukrainian war in 2022, demonstrating that EPIWATCH was able to capture a breadth of data that was not captured by other formal sources. Open-source digital disease surveillance therefore provides a useful way to gather real-time health intelligence in a conflict zone when formal surveillance is absent or reduced. Using EPIWATCH and other open-source health intelligence systems can be valuable for real-time decision support in disaster contexts, including conflict or natural disasters.