Modeling on the Perfusion Index for the Students of the Statistics Department at the University of Rajshahi

 Corresponding author: Samiul Islam Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0. Modeling on the Perfusion Index for the Students of the Statistics Department at the University of Rajshahi Samiul Islam 1, *, Rupali Sultana 1, Sharmin Sultana 2, Most. Sayma Akter Shampa 2, Md. Ayub Ali 3 and Saroje Kumar Sarkar 2 1 Department of Statistics, Mawlana Bhashani Science and Technology University, Santosh, Tangail-1902, Bangladesh. 2 Department of Statistics, Faculty of Science, University of Rajshahi, Bangladesh. 3 Department of Statistics, Faculty of Science, University of Barishal, Bangladesh. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 Publication history: Received on 25 June 2025; revised on 30 July 2025; accepted on 02 August 2025 Article DOI: https://doi.org/10.30574/wjarr.2025.27.2.2842 Abstract Our research was to investigate the determinants of perfusion index (PI), which can get into the increase and decrease of PI and also build the model on PI as well as the relationship between PI and different study variables. PI, calculated using a pulse oximeter, shows the ratio of the pulsatile blood flow to the non-pulsatile blood flow or static blood in peripheral tissue. The information of PI, together with their preditors, was collected/measured from the students of Statistics at Rajshahi University, adapting the stratified random sampling through a questionnaire containing 52 questions. Weights, heights, pulse rate, diastolic blood pressure, systolic blood pressure, and perfusion index are recorded by respective measurement tools. Univariate analysis was used to determine significant determinants, and bivariate association and correlation can also be implied. It was found that there was a positive correlation between PI and diastolic blood pressure, PI and weight, PI and exercise time, and PI and eating mangoes. A statistically significant difference was detected between the PI of males and females and smokers and non-smokers (p<0.05). Also, a significant association (p<0.05) was found between blood pressure levels and classes of different PI. Finally, a model was created involving the most significant determinants: sex, smoking status, exercise time, and diastolic blood pressure. As the value of R2 is 0.759, there is a scope for further extension of this research, including the increase of sample size, study variables, and involvement of different age groups of people. PI is the indication of sound health, strength of heart, and way of detecting heart disease, so we should take necessary steps to facilitate exercise time and maintain a healthy life. We also have to eat a healthy diet to raise the perfusion index. Keywords: Perfusion Index (PI); Independent Two Sample T-test; Stepwise forward regression; Pulse Oximeter; pulsatile blood flow 1. Introduction The perfusion index (PI) in peripheral tissues, including the fingertips, toes, and earlobes, is a noninvasive measure indicating the ratio of pulsatile to non-pulsatile blood flow [1]. The heart's contraction and relaxation facilitate pulsatile blood flow, which refers to the rhythmic movement of blood throughout the circulatory system. This flow, in contrast to constant flow, embodies the cyclical nature of the cardiac cycle and fluctuates with each heartbeat. This flow type is characterized by variations in pressure and velocity, which can affect various physiological factors, energy dissipation, and shear stress on blood vessel walls [2]. The pulse oximeter measures the peripheral index (PI), which assesses the strength of the peripheral pulse, thereby providing crucial insights into hemodynamic status and autonomic nervous system function [3]. In clinical and research settings, pulse oximeters are commonly used during exercise to provide a noninvasive continuous estimate of the oxyhemoglobin saturation of arterial blood (%SpO2) [4]. Arterial World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 86 oxyhemoglobin saturation indicates the degree of arterial blood oxygenation, specifically arterial oxygen partial pressure, and can thus be used to diagnose hypoxemia, a condition characterized by decreased arterial. The arterial oxygen partial pressure is age-dependent, with typical values ranging from approximately 100 mmHg at age 20 to approximately 80 mmHg at age 80 at sea level [5], [31]. So, it is especially relevant for conditions where blood flow may fluctuate due to various physiological stressors, such as during surgery or in populations experiencing stress [6]. The perfusion index (PI), a non-invasive measure derived from pulse oximetry, provides valuable information regarding peripheral blood circulation. Its interpretation requires a fundamental understanding of its physiological foundations. [7] elucidates the core principles of PI measurement and its clinical significance. While [8] emphasizes the impact of the autonomic nervous system on peripheral blood flow, a crucial factor in assessing PI data among students subjected to varying stress levels and cognitive load. Statistical modeling is crucial for accurately interpreting PI data, especially when accounting for individual variation. [9] provides a comprehensive overview of statistical methodologies relevant to physiological data, including regression and analysis of variance (ANOVA). Furthermore, [10] emphasize the necessity of employing mixed-effects models to account for the inherent diversity among individuals, a crucial consideration when analyzing PI data from a heterogeneous student population. Recent research has expanded the application of PI beyond clinical settings to encompass broader populations, including university students, as it offers valuable insights into stress-related physiological changes and cardiovascular health [11], [12]. The academic environment is associated with increased cognitive demands, emotional stress, and lifestyle habits that may impact autonomic function and circulatory dynamics [13], [32]. Academic stress has been shown to activate the hypothalamicpituitary-adrenal (HPA) axis, leading to autonomic dysregulation that can manifest in altered perfusion patterns and vascular responses [14]. Furthermore, prolonged exposure to mental workload and screen time has been linked to vasoconstrictive effects, which may further influence PI fluctuations [15]. Like other physiological parameters, PI can indicate the autonomic nervous system's control of blood flow and its response to internal and external stressors [16], [17]. Knowing how PI fluctuates will enable students whose academic load and stress levels vary to spot those at risk of circulatory dysfunction. Although many studies have examined PI in clinical settings, its application in academic institutions, especially for students with high cognitive loads, remains under-researched [18]. Modeling the Perfusion Index for students in the Statistics Department at the University of Rajshahi is the main emphasis of this paper. This study aims to contribute to a deeper understanding of how PI-related factors impact circulatory health by examining several key factors, including academic stress, cognitive load, and lifestyle choices. The findings of this study could also provide a non-invasive method for tracking student well-being, enabling universities to implement health plans that reduce stress-related health hazards. The Perfusion Index could be a valuable tool for identifying at-risk students and encouraging healthier academic practices if effective [19]. Focusing on academic stress, cognitive load, lifestyle choices, and environmental conditions, this paper aims to bridge the gap by examining the factors influencing PI in students at the University of Rajshahi. This study aims to identify the primary variables influencing PI by employing advanced statistical modeling methods and to model their interactions, thereby understanding how they interact and impact perfusion dynamics in the student population. This study also aims to provide a predictive framework that enables health professionals and academics to more accurately assess student well-being and develop interventions tailored to their specific needs. Hence, in this study, the research questions are: • RQ1: What variables does the Perfusion Index rely on? • RQ2: How do we model the Perfusion Index using the variables connected to it? These research questions, with particular attention to the academic environment and lifestyle choices, aim to uncover the various interactions with PI and the mechanisms by which they function. The study will investigate the interactions between these variables and develop a model that accurately reflects the underlying physiological processes using statistical tools. The remaining part of the study is organized as follows: Section 2 contains the materials and methodology, including a description of the dataset, different statistical tests, and an overview of the evaluation criteria for those statistical tests. Section 3 presents the results of analyzing the example dataset. Section 4 summarizes the concluding remarks and outlines directions for future research. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 87 2. Materials and Methodology 2.1. Description of Data Set The dataset related to the perfusion index focuses on students' various health and socio-economic parameters, including height, weight, food habits, and demographic factors such as age, monthly income, expenditure, parents' occupation, education level, and oxygen saturation level, pulse rate, and blood pressure measured using pulse oximeter tools. The study sample was collected from students in the University of Rajshahi statistics department through a scheduled questionnaire, which was collected directly from the students. This cross-sectional study used a stratified random sampling technique to select a sample of students from the Department of Statistics at the University of Rajshahi. The population was divided into five strata based on year of study: first year (98 students), second year (97 students), third year (95 students), fourth year (98 students), and masters (85 students), ensuring homogeneity within each stratum and heterogeneity between strata using factors like perfusion index, age, food habits, income, expenditure, height, and weight. Simple random sampling was applied within each stratum using the remainder approach of a random number generator, and sub-samples were combined to form the study sample. 1st year 2nd year 3rd year 4th year 5th year N1=91 N2=97 N3=95 N4=98 N5=85 Figure 1 Stratified random sampling Proportion allocation was used to determine the sample size from each stratum, resulting in 42 students from the first year, 41 from the second year, 40 from the third year, 41 from the fourth year, and 37 from the masters. The total sample size was 201 students out of 473, calculated as n = 42 + 41 + 40 + 41 + 37 = 201. In the study, the perfusion index, measured by pulse oximeter tools, was considered the output variable. In contrast, explanatory variables included height, weight, oxygen saturation level, pulse rate, blood pressure level, food habits, and socioeconomic and demographic factors such as age, monthly income and expenditure, parents' occupation, education level, and additional variables. 2.2. Material properties In this cross-sectional study, we use a questionnaire divided into four sections: personal information, socioeconomic information, food habits, and anthropometric measurement. The respondents' answers cover the whole questionnaire except for some questions in the anthropometric section, which are measured by the plus oximeter. The normal range of the perfusion index (PI) varies from 0.02% to 20%. There is no universally agreed-upon "normal," so it's a good idea to keep track of your baseline reading and monitor how it changes over time. Things like artery disease, diabetes, obesity, blood clots, and other health problems can affect your perfusion [20]. A higher PI, closer to 20%, means your arteries are dilated and blood flow is strong. On the other hand, a lower PI, closer to 0.02%, could signal that your arteries are constricted and blood flow is weak [20]. 2.3. Data pre-processing Data preprocessing is an essential step in the classification framework that guarantees the high accuracy of the findings. It consists of three parts: editing, coding, and tabulation. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 88 2.4. Ethics statement Prior to collecting data, we obtained ethical clearance from the Ethical Committee, Institute of Biological Sciences (IBSc), Rajshahi University, Bangladesh, to conduct research on the Perfusion Index for the students of the Statistics Department. 2.5. Statistical Analysis All data for this study were collected through a structured questionnaire and entered into SPSS (IBM, version 25) for detailed analysis. Initially, Descriptive Statistics were calculated to summarize the health, lifestyle, and dietary variables gathered from the participants. The ShapiroWilk Test for Normality was applied to test the data distribution, with results showing that the data followed an approximately normal distribution, as confirmed by histograms and P-P plots. Stepwise Regression Analysis was employed to identify the key predictors of the Perfusion Index, focusing on variables like diastolic blood pressure, exercise time, and smoking habits. The TwoSample t-test was used to compare the means of the Perfusion Index across different groups, including sex, mental stress, and smoking habits, to detect significant differences. The model's reliability and generalizability were assessed through Cross-validation, ensuring the robustness of the findings. Finally, Pearson Correlation Analysis was conducted to explore the relationships between Perfusion Index and other health-related variables. Statistical significance was accepted at p < 0.05, providing a clear threshold for interpreting the results. 2.6. Stepwise Regression Analysis Stepwise Regression is a method to select the most significant variables for inclusion in a regression model. It combines forward selection and backward elimination techniques, making it a highly efficient approach for identifying key predictors in a model [21]. In this study, Stepwise Regression was employed to identify significant predictors of the Perfusion Index. The process begins with all potential predictor variables. Iteratively adds or removes variables based on specific criteria, such as the Akaike Information Criterion (AIC) or the Bayesian Information Criterion (BIC). This ensures that only the most relevant predictors remain in the final model, thereby improving the model's explanatory power and interpretability [22]. Below is an illustration [22] of the Stepwise Regression process, showing the forward selection and backward elimination steps. The process proceeds as follows: Figure 2 Stepwise Regression • Forward Selection: The model starts with no predictors. Predictors are added one by one based on their significance. • Backward Elimination: Starts with all predictors and removes those not statistically significant. In Stepwise Regression, the inclusion or exclusion of predictors is determined based on the pvalue of each variable, typically with a threshold of p < 0.05 for inclusion and p > 0.10 for exclusion [9]. This method helps identify the most important predictors while avoiding overfitting [23]. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 89 2.7. Two-Sample t-test The Two-Sample t-test (also known as the independent samples t-test) is a statistical test used to compare the means of two independent groups and determine whether there is a statistically significant difference between them [24]. In this study, the Two-Sample t-test was used to compare the Perfusion Index across various groups, including sex, mental stress levels, and smoking habits. The null hypothesis (𝐻0) for the t-test assumes no significant difference between the two group means. The alternative hypothesis (𝐻1) suggests that there is a significant difference. The t-test statistic is calculated using the following formula: 𝑡=(𝑋1−𝑋2) √𝑠12 𝑛1+𝑠22 𝑛2 Where, 𝑋1 and 𝑋2 are the sample means of the two groups, 𝑠12 and 𝑠22 are variance of both groups, and 𝑛1 and 𝑛2are the sample size of two groups. The t-statistic is then compared to the critical value from the t-distribution table based on the chosen significance level (usually α=0.05) and the degrees of freedom. If the calculated t-value exceeds the critical value, the null hypothesis is rejected, indicating a significant difference between the two groups [25]. 3. Results The study gathered data from 201 participants, encompassing categorical and continuous healthrelated variables. Of the participants, 67.2% were male and 32.8% were female. A total of 61.2% were late risers, while 90.5% were nonsmokers. Most participants exhibited normal blood pressure (86.6%), whereas only 2.5% indicated elevated pressure. The distribution of the blood groups indicated that B+ (31.8%) and O+ (30.3%) were the predominant types. The respondents' parents exhibited a propensity for normal blood pressure, with 45.7% of the mothers and 59.2% of the fathers within the normal range. Health conditions, such as asthma (87.6%), allergies (64.2%), chest pain (77.6%), and mental stress (56.7%), were significantly prevalent. Table 1 Descriptive Statistics of Health, Lifestyle, and Dietary Variables Among Respondents Variable Range Min Max Mean SE Sk Age 9.00 18.00 27.00 22.53 0.12 -0.17 Expenditure 45000 0.0 45000 5677 288 5.9 Exercise time 21.00 0.0 21.00 4.2378 0.31 1.6 Sleeping time 7.00 5.00 12.00 7.74 0.082 0.41 Daily Eaten Rice 1450.00 50.00 1500.00 402.4129 13.80772 1.559 Daily Eaten Bread 6000.00 0.00 6000.00 69.9104 30.17924 13.464 Daily Drinking water 8998.00 2.00 9000.00 2684.6119 107.10513 1.290 Daily Eaten Fish 5000.00 0.00 5000.00 350.5473 31.66359 6.377 Daily Eaten Meat 2500.00 0.00 2500.00 328.0398 26.62824 3.010 Amount of Cold Drinks Taken in a Week 2000.00 0.00 2000.00 419.4428 35.53642 1.811 Eaten Vegetable Daily 1200.00 0.00 1200.00 205.9950 13.93865 2.106 Eaten Cinnamon Daily 250.00 0.00 250.00 14.5100 2.68912 3.945 Height in Cm 129.00 60.00 189.00 164.7410 0.89607 -4.841 Weight in Kg 57.00 37.00 94.00 60.7851 0.77093 0.470 Oxygen Saturation Level% 40.00 60.00 100.00 97.8259 0.25611 -6.850 World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 90 Pulse Rate in BPM 113.70 8.30 122.00 78.6433 0.85221 -0.607 Systolic mmHg 60.00 80.00 140.00 113.1940 0.76300 -0.607 Diastolic mmHg 50.00 50.00 100.00 74.5550 0.66528 -0.519 Table 1 shows the followingThe mean of the respondents was about 22 years. The minimum age was 18 years, and the maximum was 27 years. The distribution of the age is negatively skewed. So, there were more respondents between 18 and 22.5 years old. From the descriptive statistics, we can say that the monthly expenditure of our statistics students in RU was 5591 taka. On average, each student does approximately 4.2 hours of exercise or playing. The average sleeping time in a day for respondents was 7.73 hours. Besides, the minimum value of sleeping time in a day for our study respondents was 5 hours, and the maximum value of sleeping time in a day for our respondents was 12 hours. The value of skewness was positive. So, the number of students is between 7 and 12 hours interval than those between 0 and less than 7 hours interval. On average, each respondent ate approximately 402 gm of rice and approximately 70 gm of bread. Some calculations for descriptive statistics are given in a table for specific food quantities, such as daily consumed fish, daily consumed meat, daily consumed vegetable, and so on. The mean height was 164.747 cm. The minimum height was 129 cm, and the maximum was 189 cm. The value of skewness is -4.481, which means the population's frequency is more between the minimum and mean values. The mean weight was 60.785 kg. The minimum weight was 37kg, and the maximum was 189 kg. The skewness value is .470, which means the population's frequency is greater between the maximum and mean values. Our respondents' average pulse rate was 78.643. This is within the normal range and indicates a good health condition. Among the students, systolic blood pressure varies from 80 to 140 with a mean value of 113 mmHg. This distribution shows a negatively skewed and platykurtic curve pattern. So, the distribution of this variable was very much scattered The maximum diastolic mmHg value was 100mmHG among the respondents, and the minimum was 50mmHG among the students. The average diastolic mmHg value of study students/respondents was 74.55 mmHg. Table 2 Descriptive Statistic for Perfusion Index Perfusion Index N Valid 201 Missing 0 Mean 4.9259 Std. Error of mean 0.22927 Median 4.2 Mode 1.10 Variance 10.566 Skewness 0.748 Std. Error of Skewness 0.172 Kurtosis 0.06 World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 91 Std. Error of Kurtosis 0.341 Range 15 Minimum 0 Maximum 15 Figure 3 Histogram of Perfusion Index Table 2 shows that the mean of the perfusion index of the respondents was 4.92, the median was 4.2, and the mood was 1.1. It is alarming news that the average and mode of perfusion index for study students are less than 5. So, the blood flow condition was very poor, and the heart-blood circulation of statistics students was quite ominous. The maximum perfusion index value was 15, which was good news. It is a good condition for blood circulation. The variance of the perfusion index was 10.5. The distribution is negatively skewed and platykurtic. That means the distribution was very much dispersed. Table 3 Correlation coefficients for different variables Perfusion Index Diastolic mmHg Body Mass Index Weight in Kg Eaten Mango Weekly Eaten Cinnamon Daily Eaten Vegetable Daily Exercise time or playing time Perfusion Index Pearson Correlation 1 0.219** 0.025 0.211** 0.155* -0.008 0.091 0.275** Sig. (2-tailed) 0.002 0.728 0.003 0.028 0.911 0.197 0.000 N 201 200 201 201 201 201 201 201 Diast olic mmHg Pearson Correlation 0.219** 1 0.037 0.310** 0.086 -0.102 -0.013 0.043 World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 92 Sig. (2-tailed) 0.002 0.604 0.000 0.227 0.149 0.852 0.548 N 200 200 200 200 200 200 200 200 Body Mass Index Pearson Correlation 0.025 0.037 1 0.221** -0.013 0.119 -0.041 0.071 Sig. (2-tailed) 0.728 0.604 0.002 0.860 0.092 0.566 0.314 N 201 200 201 201 201 201 201 201 Weight in Kg Pearson Correlation 0.211** 0.310** 0.221** 1 0.117 0.012 -0.012 0.083 Sig. (2-tailed) 0.003 0.000 0.002 0.099 0.861 0.864 0.240 N 201 200 201 201 201 201 201 201 Eaten Mango Weekly Pearson Correlation 0.155* 0.086 -0.013 0.117 1 0.055 -0.008 0.040 Sig. (2-tailed) 0.028 0.227 0.860 0.099 0.434 0.909 0.576 N 201 200 201 201 201 201 201 201 Eaten Cinnamon Daily Pearson Correlation -0.008 -0.102 0.119 0.012 0.055 1 0.047 -0.012 Sig. (2-tailed) 0.911 0.149 0.092 0.861 0.434 0.510 0.864 N 201 200 201 201 201 201 201 201 Eaten Vegetable Daily Pearson Correlation 0.091 -0.013 -0.041 -0.012 -0.008 0.047 1 0.042 Sig. (2tailed) 0.197 0.852 0.566 0.864 0.909 0.510 0.554 N 201 200 201 201 201 201 201 201 Exercise time or playing time Pearson Correlation 0.275** 0.043 0.071 0.083 0.040 -0.012 0.042 1 Sig. (2tailed) 0.000 0.548 0.314 0.240 0.576 0.864 0.554 N 201 200 201 201 201 201 201 201 ** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed). The Pearson coefficient between perfusion index-related variables is shown in Table 4. The correlation analysis reveals several significant relationships among the studied variables. The Perfusion Index shows a moderate positive correlation with Diastolic Blood Pressure (r = 0.219, p < 0.01), Body Weight (r = 0.211, p < 0.01), and Exercise Time (r = 0.275, p < 0.01), indicating that individuals with higher perfusion levels tend to have higher blood pressure, greater body weight, and engage more in physical activity. Additionally, a weak but significant positive correlation is observed between the Perfusion Index and Mango Consumption (r = 0.155, p < 0.05), suggesting a possible dietary influence on vascular performance. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 93 Diastolic Blood Pressure is strongly associated with Weight (r = 0.310, p < 0.01), implying that as body weight increases, so does diastolic pressure. Body Mass Index (BMI), while significantly related to Weight (r = 0.221, p < 0.01), does not show notable associations with other variables in the study. Dietary variables such as cinnamon and vegetable intake do not significantly correlate with the Perfusion Index or other health markers. Exercise Time is only significantly correlated with the Perfusion Index, with no meaningful associations found with blood pressure, BMI, or dietary factors. The findings highlight that physiological factors like blood pressure and weight and lifestyle behaviors like physical activity influence the Perfusion Index. At the same time, dietary variables (except mango intake) appear to have limited direct associations. Table 4 Group Statistics for Perfusion Index across Different Groups Group Statistics Variables Perfusion Index N Mean Std. Deviation Std. Error Mean Sex Female 66 3.6015 2.801 0.3448 Male 135 5.5734 3.2676 0.2812 Mental Stress Yes 87 4.7149 2.9515 0.3164 No 114 5.0869 3.4656 0.3246 Smoking habit Yes 19 3.5584 2.2083 0.5066 No 182 5.0687 3.3125 0.2455 Early and Late Riser Early riser 78 5.1449 3.1134 0.3525 Late Riser 123 4.7871 3.3395 0.3011 The table shows that the mean Perfusion Index values are not similar across the different groups. For Sex, the mean for males (5.57) is significantly higher than for females (3.60), indicating a clear difference between the two groups. When considering mental stress, individuals reporting no mental stress have a slightly higher mean (5.09) compared to those with mental stress (4.71), showing a slight difference. In terms of smoking habits, there is a noticeable difference, with nonsmokers having a significantly higher mean (5.07) compared to smokers (3.56). Finally, early risers have a higher mean (5.14) than late risers (4.79), though the difference is relatively small. In this case, we used a two-sample t-test to compare the means of Perfusion Index between two independent groups (e.g., males vs. females, smokers vs. non-smokers) to see if there was a statistically significant difference between them. The two-sample t-test is appropriate here because we are examining the difference in means between separate groups that are unrelated or paired, which is precisely the scenario with gender, smoking status, and other independent categories. In this case, the null hypothesis is that the perfusion index for the independent groups was the same, whereas the alternative hypothesis is the opposite. Table 5 Two-Sample T-test Two Sample t-test Variables Perfusion Index Levene's Test for Equality of Variances t-test for Equality of Means 95% Confidence Interval of the Difference F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference Lower Upper Sex Equal variances assumed 5.085 0.025 4.486 199 0.000 2.11856 0.47222 1.18737 3.04975 World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 100 students [6]. Our approach, including statistical significance tests and correlation analysis, offers a robust and comprehensive view of the complex physiological relationships at play, surpassing earlier work in its scope and analytical depth. Our work stands out in several ways compared to other studies, such as those by 16 and 14. Most studies on PI have been conducted in clinical settings, often focusing on older adults or patients with known health conditions. In contrast, our study provides insights into a younger, academically-stressed population, showing the potential of PI as a noninvasive, real-time indicator of student well-being. Additionally, while studies on lifestyle factors and PI are emerging, few have utilized a comprehensive dataset that includes physiological measurements (e.g., PI, blood pressure) and lifestyle variables (e.g., exercise, smoking, diet) to explore their interactions. This holistic approach adds novelty to our findings and strengthens the generalizability of the results to other academic populations. In conclusion, this study illustrates the efficacy of the perfusion index as a significant instrument for monitoring student health, equipping universities with actionable data to develop wellness programs targeting stress and circulatory health. The findings underscore critical domains for forthcoming research, especially investigating the interplay between environmental factors, screen time, and cognitive load with physiological health indicators. Our findings provide a more refined and contextually relevant comprehension of PI, establishing a foundation for subsequent research in academic environments. 5. Conclusion The project overview is thoroughly explained in the Results and Discussion section. The descriptive statistics imply that the perfusion index of Rajshahi University students was comparably lower than those of strengthened values. This implies that necessary programs or initiatives should be taken so that the perfusion index of the students can be geared up. A model was built to sketch out the determinants of the perfusion index. As the value of R2 was 0.759, it demands the further extension of the present research. The fitted model is equally helpful for male and female students and smokers and non-smokers, as they were included as the dummy variables in the model. After applying cross-validity predictive power to the fitted model, it was found that only the variables diastolic blood pressure and exercise time were the two determinants of the perfusion index. This research demanded an increase in facilities so that the students could increase their perfusion index by increasing their diastolic blood pressure and exercise time. The same research can be extended by increasing the sample size by including different departments from the University of Rajshahi. This research can include people of different age groups from the Rajshahi district. It can be expanded by increasing the number of determinants. Several regression models, such as nonlinear regression, will be disposed of for better modeling. Reliability checking can be applied to the model in further research. Advanced statistical analysis, such as machine learning algorithms, can be deployed for more accurate modeling. Compliance with ethical standards Disclosure of conflict of interest No conflict of interest to be disclosed. References [1] N. S. A. Manap, R. H. M. Zaini, W. F. W. M. Shukeri, S. C. Omar, M. Z. Abu Bakar, and [2] P. Seevaunnamtum, “Pulse oximetry-based perfusion index as a non-invasive indicator of systemic hemodynamics during spinal anesthesia in cesarean delivery,” Anaesthesia, Pain Intensive Care, vol. 28, no. 5, pp. 927–932, Oct. 2024, doi: 10.35975/APIC.V28I5.2570. [3] P. R. Painter, P. Edén, and H. U. Bengtsson, “Pulsatile blood flow, shear force, energy dissipation and Murray’s Law,” Theor. Biol. Med. Model., vol. 3, no. 1, pp. 1–10, Aug. 2006, doi: 10.1186/1742-4682-3-31/FIGURES/1. [4] K. C. Hung et al., “The Use of the Perfusion Index to Predict Post-Induction Hypotension in Patients Undergoing General Anesthesia: A Systematic Review and Meta-Analysis,” Jul. 10, 2024. doi: 10.3390/diagnostics14161769. [5] L. J. Mengelkoch, D. Martin, and J. Lawler, “A Review of the Principles of Pulse Oximetry and Accuracy of Pulse Oximeter Estimates During Exercise,” Phys. Ther., vol. 74, no. 1, pp. 40–49, Jan. 1994, doi: 10.1093/PTJ/74.1.40. [6] J. J. . Marini, “Respiratory medicine for the house officer,” p. 293, 1987. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 101 [7] A. Lima and J. Bakker, “The perfusion index: A non-invasive assessment of microcirculatory blood flow,” Acta Anaesthesiol. Scand., vol. 62, no. 2, pp. 216–223, 2018. [8] A. Jubran, “Pulse oximetry,” Crit. Care, vol. 19, no. 1, Jul. 2015, doi: 10.1186/S13054015-0984-8. [9] J. Allen, “Photoplethysmography and its application in clinical physiological measurement,” Physiol. Meas., vol. 28, no. 3, p. R1, Feb. 2007, doi: 10.1088/09673334/28/3/R01. [10] Andy Field, “Discovering Statistics Using IBM SPSS Statistics,” sage Publ., 2018, Accessed: Mar. 26, 2025. [Online]. Available: https://books.google.com.bd/books?hl=en&lr=&id=83L2EAAAQBAJ&oi=fnd&pg=PT8 &dq=ield,+A.+(2018).+Discovering+statistics+using+IBM+SPSS+statistics.+Sage+publi cations.&ots=UbMTxlBJEP&sig=LWs2bribjbPPzKB6qcvIVAavoE&redir_esc=y#v=onepage&q=ield%2C A. (2018). [11] Discovering statistics using IBM SPSS statistics. Sage publications.&f=false [12] G. M. Fitzmaurice, N. M. Laird, and J. H. Ware, “Applied Longitudinal Analysis (Google eBook),” p. 740, 2012, Accessed: Mar. 26, 2025. [Online]. Available: https://books.google.com/books/about/Applied_Longitudinal_Analysis.html?id=0exUN1y FBHEC [13] R. Gamal, F. Sadeghi, and A. Mahmud, “The Impact of Academic Stress on Cardiovascular and Autonomic Regulation in University Students,” J. Stress Physiol. Biochem., vol. 2, no. 18, pp. 35–42, 2022. [14] S. Sultana et al., “Smartphone usage and academic performance: A cross-sectional investigation among statistics students at MBSTU,” 2025, doi: 10.30574/wjarr.2025.27.1.2574. [15] Raghavan Srinivasan, N. Rajan, and Sanjay Gupta, “Academic stress and physiological responses among college students: Implications for health monitoring,” J. Am. Coll. Heal., vol. 69, no. 5, pp. 473–480, 2021. [16] C. Leistner and A. Menke, “Hypothalamic-pituitary-adrenal axis and stress,” Handb. Clin. Neurol., vol. 175, pp. 55– 64, Jan. 2020, doi: 10.1016/B978-0-444-64123-6.00004-7. [17] P. Kumar, A. Patel, and R. Sharma, “Modeling perfusion index variability in students: Exploring key determinants and relationships,” J. Stat. Model. Anal., vol. 5, no. 44, pp. 215–223, 2023. [18] M. Coutrot, E. Dudoignon, J. Joachim, E. Gayat, F. Vallée, and F. Dépret, “Perfusion index: Physical principles, physiological meanings and clinical implications in anaesthesia and critical care,” Anaesth. Crit. Care Pain Med., vol. 40, no. 6, p. 100964, Dec. 2021, doi: 10.1016/J.ACCPM.2021.100964. [19] S. Islam et al., “Global Journal of Mathematics and Statistics GJMS: A Performance Comparison of Machine Learning Models for Robotic Navigation Using Imbalanced and SMOTE-Enhanced Data Global Journal of Mathematics and Statistics A Performance Comparison of Machine Learning Models for Robotic Navigation Using Imbalanced and SMOTE-Enhanced Data,” 2025, doi: 10.61424/gjms. [20] S. Sharma, S. Kapoor, and A. Patel, “Impact of screen time and cognitive workload on peripheral blood flow and perfusion index,” J. Behav. Heal. Med., vol. 2, no. 18, pp. 115–123, 2021. [21] R. F. Potter and P. D. Bolls, “Psychophysiological measurement and meaning: Cognitive and emotional processing of media,” Psychophysiological Meas. Mean. Cogn. Emot. Process. Media, pp. 1–285, Mar. 2012, doi: 10.4324/9780203181027. [22] “CardiacDirect.” Accessed: Apr. 09, 2025. [Online]. Available: https://www.cardiacdirect.com/what-is-perfusionindex-in-a-pulseoximeter/#clinicalapplicationsofperfusionindex [23] P. Ruengvirayudh and G. Brooks, “Comparing Stepwise Regression Models to the BestSubsets Models, or, the Art of Stepwise,” Gen. Linear Model J., Jan. 2016, Accessed: [24] Apr. 13, 2025. [Online]. Available: https://digitalcommons.lmu.edu/gess_fac/3 [25] R. Silhavy, P. Silhavy, and Z. Prokopova, “Evaluating subset selection methods for use case points estimation,” Inf. Softw. Technol., vol. 97, pp. 1–9, May 2018, doi: 10.1016/J.INFSOF.2017.12.009. [26] D. Kleinbaum, L. Kupper, K. Muller, and A. Nizam, Applied regression analysis and other multivariable methods. 1988. Accessed: Apr. 13, 2025. [Online]. Available: https://cyberleninka.org/article/n/834508.pdf [27] B. L. Welch, “On the Comparison of Several Mean Values: An Alternative Approach,” [28] Biometrika, vol. 38, no. 3/4, p. 330, Dec. 1951, doi: 10.2307/2332579. World Journal of Advanced Research and Reviews, 2025, 27(02), 085-102 102 [29] R. R. autor/a Sokal and F. J. autor/a Rohlf, “Biometry the principles and practice of statistics in biological research,” 1995, New York W. H. Freeman. [30] G. M. Fitzmaurice, N. M. Laird, and J. H. Ware, “Applied Longitudinal Analysis,” p. 740, 2012, Accessed: Apr. 12, 2025. [Online]. Available: https://books.google.com/books/about/Applied_Longitudinal_Analysis.html?id=0exUN1y FBHEC [31] S. Islam et al., “Global Journal of Mathematics and Statistics GJMS: A Performance Comparison of Machine Learning Models for Robotic Navigation Using Imbalanced and SMOTE-Enhanced Data Global Journal of Mathematics and Statistics A Performance Comparison of Machine Learning Models for Robotic Navigation Using Imbalanced and SMOTE-Enhanced Data,” 2025, doi: 10.61424/gjms. [32] S. Sultana et al., “Smartphone usage and academic performance: A cross-sectional investigation among statistics students at MBSTU,” 2025, doi: 10.30574/wjarr.2025.27.1.2574.