This study examined the epidemiology, outcomes, and associated factors of total diagnostic delay in P. vivax malaria in an urban Thai setting. As Thailand progresses towards malaria elimination, the number of imported P. vivax cases increased from 586 cases in 2021 to 7,246 in 2023 before declining to 4,516 in 2025 [13]. In urban areas, particularly Bangkok, most P. vivax cases were imported, either internationally from neighboring countries or domestically from provinces bordering Myanmar and Cambodia [13].

The diagnostic challenges observed in Thailand reflect many countries approaching malaria elimination. Under the WHO E-2025 initiative, 25 countries, including Thailand, Bhutan, Timor-Leste, Nepal, and Malaysia, were targeted for malaria elimination by 2025 [1]. Despite achieving zero indigenous transmission in some settings, such as Bhutan and Timor-Leste, timely recognition of imported malaria remains a key challenge in sustaining malaria-free status [1, 22]. In Kuala Lumpur, for example, imported P. vivax accounted for a substantial proportion of reported cases, underscoring the shared challenge of managing imported malaria in urbanized, near-elimination settings [23].

In this cohort, almost 60% of P. vivax patients experienced total diagnostic delay, a proportion higher than that reported in some previous studies [20, 24], but comparable to findings from France, Sri Lanka, and China where malaria cases were predominantly imported [11, 25,26,27]. Care-seeking delay was the primary contributor, accounting for 73% of cases, while medical diagnostic delay accounted for 14.1%. The median time from symptom onset to first medical visit was four days, exceeding the three-day care-seeking threshold and suggesting limited patient awareness of P. vivax malaria. Similar findings have been reported, with care-seeking delay accounting for most diagnostic delays [28]. Furthermore, misdiagnosis at the initial healthcare encounter is common in imported malaria (46–57%) and substantially contributes to prolonged diagnostic delays [16, 17].

Four factors were independently associated with total diagnostic delay in multivariable analysis: age 45–54 years, employment as an office worker or business owner, absence of prior malaria history, and initial misdiagnosis. These findings underscore the contribution of both patient and healthcare-system-level barriers. Urban workers and malaria-naïve individuals may have low perceived malaria risk and limited symptom awareness, resulting in delayed healthcare seeking [20, 28, 29]. In particular, office workers and business owners, whose predominantly urban lifestyles are associated with infrequent malaria exposure, may be less likely to suspect malaria after travel to endemic areas [20, 28]. Likewise, individuals without previous malaria experience may lack the illness familiarity that facilitates early symptom recognition and prompt healthcare seeking [30, 31].

At the healthcare system level, nearly one-quarter of patients were misdiagnosed at their initial healthcare encounter, reflecting insufficient clinician awareness of malaria in non-specialist settings [20, 28, 29, 32]. Notably, initial misdiagnosis in our cohort occurred predominantly in public hospitals, which were largely secondary or tertiary care centers, contrasting with previous reports in which misdiagnosis was more frequent in primary healthcare emergency settings [16]. As malaria transmission declines and becomes increasingly concentrated in limited geographical areas, clinicians in non-endemic settings may have fewer opportunities to encounter malaria cases. This reduced clinical exposure may contribute to decreased clinical suspicion and increase the risk of delayed diagnosis, particularly among patients with relevant travel histories [33,34,35]. Thus, routine assessment of travel and exposure history at triage remains a critical diagnostic step, even in metropolitan settings.

Although Thai patients experienced higher rates of total diagnostic delay than non-Thai patients on crude analysis (p = 0.036), this difference was not significant after adjustment for other covariates. The crude difference likely reflects the substantially different demographic and exposure profiles of the two groups rather than nationality per se. For instance, patients from neighboring countries in our cohort more frequently sought care directly at HTD, a specialized referral center with rapid diagnostic capacity, whereas Thai patients more often visited public or private hospitals first, likely influenced by insurance scheme differences. Additionally, patients from neighboring countries in this study may benefit from prior malaria exposure that enhances symptom recognition and prompts earlier care-seeking, and clinicians may maintain a higher index of suspicion for malaria in this group [34, 35]. Similar patterns have been reported in Canada and Spain, where presentation to specialized referral centers was associated with more accurate and timely diagnosis [16, 36]. A stratified analysis by nationality was not performed given the relatively small Thai subgroup (n = 136); larger future studies are warranted to characterize nationality-specific predictors of diagnostic delay.

Age 45–54 years was independently associated with total diagnostic delay, consistent with findings from China [20]. By contrast, age ≥ 55 years did not reach significance in the multivariable model, likely reflecting the small number of patients in this stratum (n = 25) and consequent wide confidence intervals. Previous studies have suggested that older adults, particularly those from endemic areas, may underestimate malaria risk, prefer watchful waiting, or attempt self-medication with antipyretics prior to seeking care [20, 37].

Different studies have used varying cut-offs (e.g. 3 or 4 days) and timepoints (e.g. from symptom onset to first healthcare consultation, hospital admission, or malaria diagnosis) to classify delay, showing that estimates and predictors can shift depending on the chosen threshold [16, 20, 25]. Our secondary analysis revealed that the 4-day and 6-day thresholds did not yield consistent findings in the multivariable model, with the exception of initial misdiagnosis at first healthcare facility visited. This finding further supports that initial misdiagnosis is a robust predictor of total diagnostic delay, particularly medical diagnostic delay, consistent with a previous study [20]. Clinically, this finding may be expected because failure to initially recognize malaria may lead to continued evaluation for alternative causes of fever, thereby prolonging the time to correct diagnosis. In contrast, the effects of other factors, such as age, occupation, and previous malaria history, may become less apparent as the illness progresses and additional clinical evaluations are performed. The differences observed between the 4-day and 6-day cut-offs may also partly reflect methodological factors, including the smaller number of delayed cases identify using the 6-day definition (29.7% versus 58.3%, respectively) which may have reduced statistical power and limited the ability to detect significant associations.

For future studies, a 4-day threshold is recommended when the aim is early identification of at-risk patients and cross-study comparability, while a 6-day threshold may be more appropriate when the focus is on prolonged diagnostic failure or prediction of severe clinical outcomes. Earlier studies on imported malaria encompassing all species have reported that P. vivax is associated with longer delays in care-seeking and diagnosis than P. falciparum infection [20, 24, 38]. Although the selected threshold may vary across studies, future prospective studies are needed to derive and validate a disease-specific definition of diagnostic delay for P. vivax, particularly in non-endemic settings. Thresholds established for P. falciparum may not be directly applicable to P. vivax, given differences in disease progression and the risk of severe outcomes. A P. vivax-specific threshold, validated against clinically meaningful outcomes, would facilitate comparisons across studies and identify patients who may benefit from targeted interventions.

Despite the high frequency of delayed diagnosis at the 4-day cut-off, severe malaria on admission and intensive care requirements did not differ significantly between delayed and non-delayed groups. Nevertheless, severe malaria according to WHO criteria was more common among patients with delayed diagnosis (17.0% versus 10.0%; p = 0.073) [21], suggesting a potentially clinically relevant association but the study may have been underpowered to detect it. Previous studies demonstrated an association between delayed diagnosis and severe imported malaria; [16, 20] therefore, these findings should be interpreted with caution. The generally less severe clinical course of P. vivax and the capacity of a specialized referral center to mitigate adverse outcomes after admission may also have attenuated observable differences between groups [20]. On admission, delayed cases more frequently exhibited schizonts, anemia, and transaminitis, reflecting more advanced parasite maturation after a longer duration of untreated infection [39].

Several limitations should be acknowledged. The retrospective design resulted in some missing data from referring facilities, and important socioeconomic determinants of care-seeking behavior could not be assessed. Moreover, variables reported by patients, such as the date of symptom onset and prior healthcare-seeking history, may also be subject to recall bias. In addition, because the study relied on documented travel histories, the exact country of malaria acquisition could not be definitively confirmed. Therefore, the reported country of exposure should be interpreted as the most likely location based on the available epidemiological information. Furthermore, relapse, recrudescence, and reinfection could not be reliably distinguished because molecular genotyping data were unavailable. This limitation may have resulted in misclassification of the presumed site of malaria acquisition, particularly for P. vivax infections that can recur due to activation of dormant hypnozoite stages.

Microscopy-based diagnosis may also have missed very low-density P. vivax infections below the detection limit, although all included patients had microscopically confirmed infection. Finally, as a national referral center, HTD may attract disproportionately complex or delayed cases, resulting in referral bias and potentially overestimating the prevalence of diagnostic delay compared with broader population. Nevertheless, the referral setting of HTD enabled comprehensive characterization of patients’ diagnostic pathways before presentation, which was essential for evaluating factors associated with delayed diagnosis.

Despite these limitations, this study provides insights into diagnostic delay patterns of P. vivax malaria in a non-endemic metropolitan setting over a 15-year period, contributing evidence relevant to malaria elimination efforts in urbanized, near-elimination contexts. Clinicians should maintain high clinical suspicion for malaria in any febrile patient with recent travel to endemic regions, particularly working-age employees, office workers, and business owners without previous malaria experience. Routine travel history should remain standard practice even in metropolitan healthcare settings. Moreover, given that diagnostic delays were most pronounced among Thai travelers, targeted public health messaging delivered through digital and social media platforms may help raise awareness of malaria risk associated with travel. Future multicenter studies including primary care and private sector settings are needed to better characterize the full spectrum of diagnostic delay across diverse healthcare environments. Qualitative studies exploring health-seeking behaviors including barriers to healthcare access and factors contributing to initial misdiagnosis, across patient-related, diagnostic and health-system domains are warranted to inform targeted interventions.