Carrillo et al. aimed to distinguish features of dengue, chikungunya, and Zika to differentiate their diagnosis, noting non-specific symptoms like fevers and rashes as well as age, sex, and time-course. The single-centre prospective cohort study used data from the ongoing Pediatric Dengue Cohort Study (PDCS) in Nicaragua, including 1980 unique participants from ages 2 to 18 between January 2006 and December 2023 with laboratory confirmed dengue, chikungunya, and Zika. Patients visited the local health center at symptom onset and consented to the collection of multiple blood samples at the discretion of the physician. Those with fever and two other symptoms (headache, retro-orbital pain, myalgia, arthralgia, rash, haemorrhagic manifestation, or leukopenia), fever without cause/other symptoms, or afebrile rash without cause (added when Zika was introduced to the study in 2016) were confirmed with inhouse RT-PCR, IgM capture, and Inhibition ELISA assays. The study selected 30 clinical features during the first 10 days of illness and compared these across the three diseases using Kruskal–Wallis and Dunn tests along with Hochberg corrections to adjust for multiple comparisons. Due to the limited number of Zika cases, an N-1 Pearson chi-squared test was used. 1321 dengue cases (many caused by different disease variants), 517 chikungunya, and 522 Zika cases were confirmed. Basophilia, monocytopenia, abdominal pain, and leukopenia best distinguished dengue; arthralgia, lack of rash, and conjunctival injection best distinguished chikungunya; and rash, absence of fever, headache, myalgia, and lymphocytopenia best distinguished Zika. Several machine learning models (MLMs) based on each disease were also utilized to differentiate these clinical features creating boosted regression tree models for accuracy in predictions, with various restrictions on the number of clinical features included and days of illness. The MLMs had high specificity (89-96%) and fair sensitivity (68-86%). They correctly identified 72.5% of chikungunya cases, 86.1% of dengue (performing better with febrile dengue compared to afebrile dengue), and 68.2% of zika cases (performing well on afebrile Zika compared to febrile). Limitations of the study included the singularity in the patients’ race as well as decreased observation of acute stages of illnesses due to the outpatient nature of the health clinic. Overall, the study was able to determine several distinguishing characteristics for each disease and test a MLM but noted difficulty differentiating afebrile dengue from afebrile Zika cases.