The Time Trend of Callssuspected of Having A Stroke with The Emergency Medical Services and Its Forecasting for The Next 5 Years in Iran


 Backgrounds: Stroke is known as the second leading cause of death and the first leading cause of disability in developing and underdeveloped countries and the annual absolute numbers of people with stroke, stroke survivors, stroke-related deaths, and overall stroke burden have shown an increasing trend. This study was aimed to investigate the time trend of calls for suspected stroke received by the Emergency Medical Servicesand to predict this trend for the next 5 years inShiraz, Iran.Methods: This was a descriptive-analytical ecological study with time seriesdata. The required data were extracted from Shiraz Emergency Medical Center between2013and2019. ITSM 2000, Excel 2013 and SPSS, version 19.0 were used to analyze and predict the data.Results: During 2013-2019, a total of 844004 calls were received by EMS call center in Shiraz city, of which 6620 calls were suspected of strokes. The number of suspected cases of stroke between 2013and2016demonstrated a steady trend in a certain range, while between 2016 and 2017, it has significantly increased and after that the number of stroke cases has reached a relative stability.In a short time period, like monthly or seasonally, the highest incidence of stroke occurs in February, the coldest month of the year in shiraz city, and the lowest incidence of stroke occurs in July and August, the warm times of the year in Shiraz city, and this trend is repeated every year. Also, the predict of stroke incidence for the next 5 years indicated that the trend is almost constant with a slightdecline, that is the total number of stroke caseswill diminish by 6% in 2024 compared to 2019. Conclusion:The results of study demonstrated that the long-term trend of reported suspected calls cases of stroke in EMSis currently stable following an increase in Iran, and itis expected to decline in the upcoming years. Besides, in short-term seasonal trend, the number of reported suspected calls cases of stroke in EMS is higher in cold seasons than in warm seasons.


Introduction
Stroke is the second leading cause of death and disability in developing and underdeveloped countries (1)(2)(3). In the United States, stroke is the 7th leading cause of death (4). As the World Health Organization (WHO), 6.1 million deaths occurred due to stroke in 2019 (5). Although,mortality rates globally have dropped over the past two decades, the absolute number of people with stroke each year, stroke survivors, stroke-related deaths, and overall stroke burden (measured by the disability-adjusted life year (DALYs)) have shown an increasing trend. Temporal changes in stroke incidence as well as mortality from cerebrovascular disease have been investigated in several studies in different countries with various climatic conditions, but contradictory ndings have obscured the conclusions (6-11). A systematic review of a population-based stroke study on 28 countries showed an increasing trend in stroke incidence in low-income and middle-income countries. However, there has been a 42 percent reductionin high-income countries over the last four decades (12). In a short-term trend review in Finland, in winter season, the trend of acute ischemic and hemorrhagic strokeswas shown to increaseseasonally (13).
Since the management of acute stroke is time-dependent, early detection of symptoms and quick transportation to hospital can reduce related mortality and disability (5). EMS personnel should be sophisticated enough tobe able to diagnose, examine, manage, treat, triageand transport stroke patients in a timely manner (14). In Iran, the system "724", which provide speci c services for acute stroke patients seven days a week and 24 hours a day and the code SAMA (Emergent Stroke) aimedat rapid identi cation and screening of patients suspected of acute stroke, andestablish better communication and coordination between the pre-hospital system and stroke centers for faster transmissionby either air or land, have been established since July 2015 (15). It seems necessary to assess factors associated with pre-hospital and hospital delay, the quality of provided care and individual factors associated with timely treatment in any society (16). In Iran, there are challenges and potentials for improvement of stroke care that can be improved by increasing public awareness and creating an organized program in the healthcare system (17).
Fars province located in the south of Iran was established in 1316 as the center of Shiraz metropolis. According to 2016 Census data, the population of Shiraz city is 1,565,572 (29).
The short-term and long-term trend of stroke in Iran is not well-known and also due to the important roles of diagnosis, treatment and early transfer of stroke patients by EMS, this work was carried out to investigate the seasonal trend of the number of calls received byEMScall center in Shiraz city during 2013-2019, and forecasting the frequency of calls for the next upcoming months to help o cials and authorities to establish training intervention regardingin order to improve awareness of public and the staff towards system 724 instructions.

Methods
Thecurrent study is a descriptive-analytical ecologicusing time series analysis. A time series is a set of statistical data that are collected at regular intervalsand the statistical methods that use these statistical data is called time series analysis (18). The data of this work were extracted from 2013 to 2019 from the emergency medical call center of Shiraz city, which was registered by the software in this center, by month and year.In this investigation, all contacts that were established with the Shiraz Emergency Medical Service from 2013 to 2019 and were diagnosed with a suspected stroke according to the diagnosis of emergency medical technicians or the physician of the EMS were participatedin our study.
ITSM 2000, Excel 2013 and SPSS version 19.0 were used to analyze and predict the data. Initially, the data were extracted in the form of Excel les from the EMScallrecording software at the Shiraz EMScalling center. Then, these data were entered into SPSS and the frequency of suspected cases of stroke was evaluated by month and year, and the number of accidents in different months was determined. Then the data trend was examined separately by year and month in different seasons. Data were entered into the ITSM program and staticdataanalysis was performed. The data tting was performed to best t different models. Thus, we usedself-correlated seasonal Auto Regressive Integrated Moving Average (SARIMA) to predict the data. In this model, S is the seasonal period, which is selected for the trend of callssuspected of stroke between2013and2019. The autoregressive (AR)isthe self-correlated model with the order P, and MA is the average mobility with the order Q. The nal model for SARIMA is as follows: In thecurrent study, the tted model is as follows: Two tests were applied to ensure the accuracy of modeling and prediction in this work; 1) data randomization following model tting and prediction using Turning Points method, i.e. the p-value was equal to 1 at a signi cance level of 0.05 in this method, thus the data are random, indicating the good reliability of the model, 2) Using ACF Residual and PACF Residual in the data prediction model.
To review and analyze the data with time series, the data oncalls for suspected stroke received by EMSin Shiraz city were entered into ITSM. Figure 1 shows the Preliminary image of the data,following entering the ITSM software.
For static data as well as eliminating the seasonal trend of data and estimating the trend of these events, the BOX-COX transformationwas rst set to zero and DIFFERENCE 12 and 1 were usedfor the next 5 years. Figure 2 shows the data after the mentionedmeasures.
Then, in order to determine the order of q and p in AR, MA and ARMA models, autocorrelation function (ACF) and partial autocorrelation function (PACF) plots were used. Therefore, the appropriate values of p and q were equal to 14 and 11, respectively. Figure 3 shows the ACF/PACF plot.
In the next step, different models were tted to the data and among the evaluated models, MA model (3) based on innovations method and search for the smallest AICC value by ITSM had the lowest value of Akai index and was selected as the best model. Based on this model, the AICC coe cient was calculated as -24.428596.
Finally, two tests were used to evaluate the reliability of the data prediction model. The P-value of this method was equal to 1 and was statistically signi cant at the level of 0.05. Therefore, the data were random, indicating the good reliability of the model. Using ACF and PACF in the data prediction model. Given that in both cases, the number of lag times outside the zero range is less than 0% of the total number of lag times, it can be concluded that the time series model is best ttedand the predictions are reliable. The results of prediction of trend of suspected stroke calls over the next ve years on a monthly basis according to Figure 4 and Table 2 with a 95% con dence interval are as follows: As shown in gure 4, the overall long-term time trend of stroke incidence after a signi cant increase between 2016 and 2017, has reached a relative stability and continued to be stable for the next upcoming years. These results are also presented in more detail in Table 2. The results presented in table 2indicate that the prediction of stroke incidence is almost constant with a slight decrease, so that the total number of stroke cases in 2024 compared to 2019 will decrease by 21 percent. It was also found that in the short-term trend (e.g. monthly or seasonal) the highest incidence of stroke occurred in February, the coldest month of the year in Shiraz city, whereas the lowest incidence of stroke occurred in July and August, the warm seasons of the year.

Discussion
The results of our study revealed that the time trend of incidents leading to contact with the EMS call center, suspected of stroke after a raise is currently at a steady annual level, while in seasonal and monthly surveys, this prevalence is more in winter, especially in February compared with other seasons. These ndingswerewell-grounded by other studies. As such, Farhang and colleagues investigated the seasonal prevalence of thrombotic stroke in two seasons, winter and summer of 1998-1999 in Loghman Hakim medical training center anddemonstrated that 63% of patients with thrombotic stroke in the neurology ward of Loghman hospital occurred in winter and 37% in summer (19). As well as, Jalkorevic and colleaguesassessed seasonal variations in the prevalence of stroke in the Finnish adult population between 1982 and 1992, and reported the prevalence of stroke as 12% higher in winter than in summer (20). However, another crosssectional study aimed at assessing the frequency of stroke patients who referred to the emergency department in summer and winter in 2020 in Tehran indicated that the prevalence of stroke was higher in hot seasons than in cold seasons (21). These contradictory results may be attributed to the classi cation of types of stroke. Overall, the prevalence of stroke was shown to be higher in winter than in summer, which may be due to the increased prevalence of respiratory infections in winter season. Respiratory infections can also increase plasma brinogen and anti-cardiolipin antibodies and reduce protein C levels (22). Cold climate is also associated with high blood pressure, so cold weather, especially in winter, raises blood pressure, especially in the elderly, and increase the risk of stroke (23).
The results of our study proved that the trend of suspected stroke events is currently in an almost constant annual condition and this stability initiated almost from 2017 and continued until 2019 with a slightdecline (e.g. 4 percent in Year), which is expected to continue. This nding was consistent with the results of previous studies. A study on stroke incidence and 10-year survival in Sweden from 1975 to 2001 found that stroke rates and short-term mortality followed a stable pattern in Sweden between 1975 and 1990 (24). The results of other studies have also shown that the incidence of stroke in most industrialized countries was stable from the second half of the 1970s to the end of the 1980s . Another study conducted between 1990 and 2010, indicated that the number of strokes decreased by approximately 10% in developed countries and increased by 10% in developing countries (28). It seems that in Shiraz city as the representative of Iran, the increasing trend of stroke has ended and the decreasing trend has begun, which demonstrates the end of the silent peak of stroke. This change is similar to the change in disease trends in developed countries where the rate of reduction in the incidence of stroke began many years ago, which may be due to the increased public awareness about risk factors for stroke or lifestyle changes in the community.

Conclusion
The results of our study disclosed that the long-term trend of reported suspected cases of stroke in EMSafter an increase, is currently stable and is expected to decline in the next upcoming years in Iran. Besides, we found that in the short-term seasonal trend, the number of suspected cases of stroke in EMSwas higher in cold seasons compared to warm seasons.

Limitations
One of the limitations of this paperinclude the Lack ofdemographic information of callers with EMS for suspected cases of stroke. If this information was available, a more e cient prediction with demographic characteristics could be made.

Recommendations
It is suggested to conductfuture studies in other populations with demographic characteristics and compare the results. Also, we recommend to determine the actual number of calls, suspected of stroke between 2020 and 2024 on the study population and compare the results with the projected results in the present study.   prediction of calls for suspected stroke on a monthly basis over the next 5 years