The study’s results indicate a high frequency of diagnostic and initial management errors in the pre-hospital and hospital phases. In both triage and the healthcare center, 8/10 treated patients experienced at least one error or omission during the process; in observation/Uviclin, this error occurred in 5/10 patients. Similarly, the healthcare center and triage incorrectly classified 1/3 of the patients based on their severity, which improved under observation. Only 3/10 patients in triage and 2/10 patients in the healthcare center, respectively, received adequate hydration.

The results may show bias in healthcare staff documentation because they may not have enough time to record every detail. Another possibility is a lack of standardized formats set up by the new National Technical Standard (5). This reality is often evident in scenarios of patient overload, epidemics, and care provided by front-line personnel with limited experience. A recent example was the COVID-19 pandemic, which made this type of inconvenience clear (15). On the other hand, a lack of adherence to the CPG recommendations is common not only in our country but in other realities and infectious pathologies such as COVID-19 (16).

Evidence suggests that the region has conducted training. Experts conducted four training sessions between 2023 and 2024: the first for 150 primary care health professionals (17), and the second for personnel with similar characteristics in the Morropon-Huancabamba Health Sub-Region (18). The Regional Directorate of Health Piura trained more than 1,200 nurses from the Ministry of Health (MINSA), Social Security (Essalud), Police Forces Health Facilities (Sanidad), and other healthcare establishments in a 12-hour workshop course in October 2023 (19, 20). However, it is crucial for any strategy to not only provide training for the personnel in charge but also to measure the baseline (knowledge, attitudes, and practices) (8) (10) and determine the impact afterwards. In this scenario, quasi-experimental studies are an attractive model and an especially useful tool (21). In an 18-month study in India by Balkrishnan et al., in patients with DWS, pre- and post-intervention measurements were carried out, determining that the educational program had an impact on improving the admission and classification criteria, monitoring of vital signs, correct use of bolus fluids, fluid therapy in titration, hematocrit monitoring, platelet transfusion when indicated, and discharge criteria. Also, mortality decreased from 7.1% to zero. The median hospital stay decreased by one day (22). An attractive possibility is to describe the before and after evaluations conducted on doctors in Piura in 2023 (20). Qualitative research is another unexplored scenario that is potentially helpful in exploring the perceptions, attitudes, and limitations of human resources. Quantitative measurement of adherence to a guideline is insufficient to explore these scenarios (23).

The Peruvian Health Ministry Directorate for the Prevention and Control of Metaxenic and Zoonotic Diseases developed a strategy to measure adherence to the National Technical Standard’s recommendations (5). According to their findings, in Peru, adherence was 70.5% (249 evaluations), with 51.27% in treatment and 73.76% in diagnosis. After reviewing 33 records, global adherence was 68.22% in Piura, 73.74% in diagnosis, and 49.19% in treatment, ranking ninth out of 13 regions (24). These figures are more encouraging; however, the evaluation of compliance should focus on the correct completion of forms and sociodemographic data rather than clinical aspects. We believe that our results can provide input for metaxenic strategies and other managers looking to improve training, infrastructure, and political commitment to addressing dengue. In the case of the Peruvian Social Security Health System (ESSALUD), we audited 42 clinical records as of March 2024 from cities Piura, Lambayeque, La Libertad, Ancash, and Ica. We found that the frequencies of notification to epidemiology surveillance, recording of the date of onset of symptoms, observation/hospitalization, parenteral hydration, recording of vital functions, and request for at least one blood count were 62%, 83%, 83%, 76%, 71%, and 91%, respectively (25). The National Health Ministry Health System made similar observations.

Even in patients with DWS, the high mortality rate of 18% is striking. Other probable causes include related comorbidities (which were not measured in this study) and limitations in diagnosis and case management. In a recent systematic review of our country, out of 17 publications between 1993 and 2023, including 10 regions of Peru and 2310 patients, mortality was 1.73% (26). The significant variation in mortality could potentially be attributed to our study’s bias, as we exclusively included patients from Piura, our non-random sampling method, and the severity of our patients (20.7% of whom had severe dengue). However, in the exploratory multivariate analysis, the presence of at least one error in the hospital, in the entire process, and in each of the areas, as well as other management problems, was not ultimately associated with mortality. This could be attributed to biases in the sample size (141/158) and sample design (consecutive), as well as the challenges of applying a retrospective checklist, which, despite training, may have measurement biases. Only dengue severity was associated with higher mortality.

There has been serious criticism of the efficiency of the reference and counter-reference systems in Peru. The referral process took more than 2 hours for more than 60% of patients (27). According to the regional government, the hospital where the study took place has approved the formation of a counter-referral team, but its implementation is still under development (28). Meanwhile, over 30% stayed in triage for over four hours. Understaffing and insufficient areas of attention are possible explanations. The current National Technical Standard mandates that in DWS, medical care should begin in less than 30 minutes, with immediate intravenous hydration; in SD, immediate attention is needed (5). This study has found many challenges in implementing these measures.

Regarding the observation/Uviclin area’s findings (55% of “at least one error”), we could attribute these findings to time constraints, patient overload, burnout, unnecessary form filling (National Insurance/discharge procedures), and possibly a lack of organization. This area requires monitoring with practical protocols. It is also imperative to verify the proper operation of equipment and supplies.

Theoretically, hospitalized first-level health care facilities (I-3 and I-4) can handle dengue diagnosis and alarm signs by providing an immediate response based on hydration and vital sign monitoring (blood pressure, heart rate, respiratory rate, temperature, oxygen saturation, water and electrolyte balance, etc.), which forms the foundation of effective dengue management. These facilities typically have a basic laboratory that works between 12 and 24 hours for monitoring with a complete blood count, although dengue management is still clinical in the Americas, according to the WHO (29). Again, human resource training (3, 4), as well as a lack of clear and socialized processes, could explain this reality.

One of the study’s limitations was that it did not meet the required sample size (n = 159 medical records). Some of the reasons for this were the inability to find the clinical history, incomplete data in the follow-up, and a lack of digitization of data and processes. However, the limitations prevented a random sampling process. Likewise, the 2023 cases were limited to weeks 18–25 (the peak of the outbreak). However, we believe that these results can serve to support any effort in the implementation of the electronic health record to improve information management (30). The Ministry of Health and the regional governments still have unresolved issues.

In conclusion, we found a high frequency of errors in the diagnosis of severity and in the treatment of patients with DWS and SD in the pre-hospital and hospital phases, but the sample size was insufficient to determine whether this finding had an impact on mortality.