Introduction Bluetongue disease (BTD) is a vector-borne, viral disease of domestic and wild ruminants associated with considerable impact on the animal health and consequently significant economic losses in the agricultural production sector. Circulation of the BTD virus until 1998 had been localized within the tropical and sub-tropical climate zones while after this period the virus spread to the north, and by 2006 had spread to western and central Europe encompassing the belt of temperate and sub-polar climate zones. Over the past twenty years, BTD outbreaks caused by a variety of serotypes have occurred and spread several times across different regions of Europe. Although the affected countries have largely adopted standardized control measures, the approaches used has not resulted in elimination of disease. The occurrence of BTD during this period included unexpected epidemics in areas where the disease had never been confirmed before, low level circulation of certain serotypes, some of which have an unclear origin, incursions of new serotypes, vaccine incidents and disease reoccurrence. Implementation of appropriate surveillance, control and preventive measures are further complicated by the incomplete understanding of the ecology of the vectors, viral transmission dynamics and maintenance in a given area. In the Balkan Peninsula region, where BTD had not occurred for more than 10 years, an unexpected epidemic of serotype 4 occurred in 2014. The first detection of this serotype in the region was followed by rapid spread to 11 south-eastern European countries over a very short time period. Although many studies have investigated BTD in Europe, no nationwide study on the prevalence of BTD serotype 4, influence of climatological factors, vectors and overwintering mechanism or efficiency of applied vaccination programs from Croatia have been carried out. In 2015, the Croatian Ministry of Agriculture implemented a comprehensive mass vaccination campaign against BTD for all ruminants in Croatia. Vaccination campaign followed the recommendations of the EFSA 2017 Scientific report on Bluetongue: control, surveillance and safe movement of animals until the end of implementation in 2019. Despite implemented measures, only 18 months upon last vaccination campaign BTD re-occurred in 2020 when serotype 4 was re-confirmed in sentinel animals in Istarska, Dubrovačko-neretvanska and Požeško-slavonska counties. Objectives The recent BTD outbreaks in Croatia suggest that the implementation of measures cannot be generalized. Planning must take into account the specific conditions of the area, such as population density and demographics, climatic factors, vector ecology and availability of administrative resources, to ensure that a strategy for BTD disease control and potential eradication is based on a case-by-case approach. In a view of the above, the main objective of this research was to develop a new approach to the implementation of BTD control and prevention measures that is evidence-based and includes forward strategy planning on the basis of relevant data, taking into account all the characteristics of the disease epidemiology in Croatia. For that purpose, certain specific objectives of the research included calculation of disease frequency measures, assessment of the magnitude of seroprevalence variability, analysis of risk factors, spatial-temporal mapping and evaluation of the impact of vaccination programs. Material and methods This research covered the areas of all Croatian counties where measures for the control and monitoring of BTD where implemented in the period 2014 to 2019. In order to do a comprehensive epidemiological analysis, data on laboratory testing, dynamics of ruminant population, movement of live animals, clinical signs detected, vaccination of cattle, sheep and goats, entomological studies of Culicoides species and climatic factors were collected. All data were analysed after being separated into a clinical phase that lasted from 27 October 2014 till 01 July 2015 and a phase of active control from 15 August 2015 till 31 December 2019. Descriptive statistics included frequency tables and epidemiological curves. Hierarchical logistic regression with binary outcome of serological testing in sheep was used to establish relative variability in serological positive response at the county, municipality and herd level, where the latter administrative and organizational units were used as random effects in a hierarchical logistic regression model. Risk factors for the number of BTD cases in the county over time during the clinical phase were determined through negative binomial regression with county as a random effect on intercept. Univariable analysis was followed by building the final multivariable regression model. During the phase of active control, likelihood of detecting a positive sentinel animal was determined by fitting the logistic regression model with county as a random effect on the intercept. The main exposures of interest were the percentage of vaccinated cattle, sheep, goats and ruminants in general. Regression models containing individual exposures of interest, and adjusted for the fixed effect of year, as well as random effect of county were used to inform development of the final multivariable models. Results Analysis of the ruminant population dynamics from 2014 to 2019 showed that mixed-type herds with more than one ruminant species predominated in Croatia and that more than 83% of herds kept less than 50 animals. Bovine population is continuously declining while sheep and goats are slightly increasing. In the territory of Croatia, two serotypes of the BTD virus were circulating; serotype 4 circulated throughout the area of the entire country whereas serotype 1 did not spread beyond the boundaries of the location of the island of Lastovo. During the clinical phase of epidemic, the morbidity and mortality in affected herds were 19.33% and 3.31%, respectively. Severe clinical signs described in the literature were not be observed and disease had a mild course. In the four southern counties where clinical disease was observed during clinical phase, herd-level seroprevalence confirmed by ELISA in the study population was 89.83% and animal-level seroprevalence was 29.38%. The overall herd-level prevalence in the study population confirmed by RT-PCR was 68.08% with the animal-level prevalence 23.66%. All animals laboratory tested due to clinical suspicion were positive by RTPCR. The highest number of seropositive herds was located near the first confirmed case in the affected area. Results of the variance component analysis which included three hierarchical levels county, municipality and herd revealed that the highest level of variability was detected at county level, regardless of the type of model and the approach used. This suggested that implementation of disease control and prevention measures at the county level would indeed be the most effective way to implement disease control measures. Univariabile analysis of climatological factors detected a statistically significant association with the number of new cases in the county only for temperature. The increase in mean daily temperature to 18⁰ C increases the number of cases in the county by 4.18 times (95% CI 1,18 -14,73). For precipitation, humidity and wind speed no statistically significant association was found with the occurrence and spread of BTD. Seropositive sentinel animals confirmed by ELISA were detected during the entire phase of active control of epidemic, with the highest animal-level seroprevalence (23%) in 2015 and the lowest (9%) in 2019. At the species level, the highest seroprevalence was found in cattle (10.62%), followed by sheep (5.64%) and then goats (0.13%) which actually corresponded to the number of sentinels used per ruminant species. Overall 5.06% sentinel samples tested by RT-PCR were positive with the highest prevalence during October, a tendency to decrease in November and December and no recorded BTD positive cases from February to August. The highest overall prevalence (15.53%) was in 2016, and the lowest in 2017 (0.52%), while in 2018 and 2019 BTD was not confirmed by RT-PCR. Such findings are directly related to the results of the vaccination. Cumulative herd vaccination coverage was 88.93% and 84.55% at the animal level. Statistical analysis of the vaccination impact has shown that as the percentage of vaccinated animals increase, the probability of detecting seropositive sentinel animals decreases. This type of association was found to be statistically significant for all variables (year, animals and herds of cattle and goats) except for the percentage of vaccinated sheep at herd and animal level, which was an unexpected finding. The dominant presence of palearctic Culicoides species from Obsoletus and Pulicaris complexes was confirmed, as well as the complete absence of C. imicola. Circulation of vectors in Croatia has two peaks, the first (larger) during May-June and the second (smaller) during October. A smaller number of vectors was isolated during December and the period JanuaryMarch which could indicate that the specified period could be considered as a vector-free period. Discussion and Conclusions The last introduction of BTD in Croatia is likely a consequence of passive transport of infected vectors by wind. In the overall sample morbidity and mortality were low and the circulation of serotypes detected was not evenly distributed throughout the country. The analysis of variance at three hierarchical administrative levels (county, municipality, herd) in all models indicated that the highest proportion of total variance resided at the county level. Using univariable and multivariable regression models among all the analysed climatological factors that may affect BTD incidence, statistically significant association was found only for temperature. Precipitation, relative humidity, and wind speed could not be identified as statistically significant factors. A comparison of the vaccination results and the overall prevalence suggested that vaccination leaded to a decrease in prevalence and that the percentage of seropositive herds declined, but without complete elimination of infection. Evidence of seropositive sentinel animals during all five years of vaccination clearly indicated that vaccination coverage less than 90% of total population did not ensure complete elimination of the virus and it continued to circulate at low levels throughout the whole period. The reasons for the lack in the implementation of the previous policies for the control of BTD, such as limited data collection and storage as well as inadequate, insufficient and incomplete implementation of control and prevention measures, were identified. Based on the obtained results, key factors in design of a new strategy were defined and the most significant includes: development of disease monitoring models that would ensure early response of the system and implementation of necessary measures before the disease is introduced into the territory of Croatia; continuous renewal of knowledge and raising the level of awareness among veterinarians and animal owners about early detection of the disease; adjusted targeting of measures to the level of the highest variability - county; planning of the entomological research to enable the analysis of the influence of climatological and ecological factors on the occurrence and distribution of different types of vectors and to compare the same with the occurrence of cases in ruminants; research of possible role of wild ruminants in maintaining infection and overwintering. In addition to prospective planning, a detailed method for data collection, evaluation and integration as well as verification of procedures during the implementation of the program must be developed. The approach applied in this research and the obtained results represent the basis for the development of future strategy for control and prevention of BTD in Croatia, and the same could be applied, as a scientific base, for planning the national control strategies of other vector borne animal diseases.