226 International Journal of Statistics in Medical Research, 2023, 12, 226-232 E-ISSN: 1929-6029/23 A Novel Algorithm for Predicting Antimicrobial Resistance in Unequal Groups of Bacterial Isolates Tareef Fadhil Raham 1,* , Haider Hussain Ali Al. Zubaidi 1 , Abbas Oweid Olewi 2 , Aya Ahmed Abddul-Fatah Al-Aboosi 3 , Nassera Attia 3 , Senaa Jaleel 3 and Abdulkhaleq Abduljabbar Ali Ghalib Al-Naqeeb 4 1 Department of Pediatrics, Al-Alwyia Teaching Pediatric Hospital, Baghdad, Iraq 2 Department of Pediatrics, AL-Kindy College of Medicine, University of Baghdad, Iraq 3 Department of Laboratory, Al-Alwyia Teaching Pediatric Hospital, Baghdad, Iraq 4 Medical & Health Technology College, Baghdad-Iraq Abstract: Choosing antimicrobials is a common dilemma when the expected rate of bacterial resistance is high. The observed resistance values in unequal groups of isolates tested for different antimicrobials can be misleading. This can affect the decision to recommend one antibiotic over the other. We analyzed recalled data with the statistical consideration of unequal sample groups. Data was collected concerning children suspected to have typhoid fever at Al Alwyia Pediatric Teaching Hospital in Baghdad, Iraq. The study period extended from September 2021 to September 2022. A novel algorithm was developed to compare the drug sensitivity among unequal numbers of Salmonella typhi (S. Typhi) isolates tested with different antibacterials. According to the proposed algorithm, the predicted resistance values were more valid than the observed values. This proposed algorithm is expected to help the hospital antibiotic policy committee recommend the proper antibacterial agents for S. Typhi and further bacterial isolates. Keywords: Unequal groups, ranking, salmonella drug resistant, predicted value, observed value. HIGHLIGHTS • Isolates with different sample sizes of groups tested for antimicrobial sensitivity are difficult to analyze. • A novel algorithmic method is developed to analyze the retrieved data of a sample of unequal groups of bacterial isolates tested for different antibacterial agents in order to rank these groups according to their degree of resistance. • This method can be used in analyzing different unequal group sizes in general. 1. INTRODUCTION Antimicrobial resistance (AMR) has emerged as a major public health problem all over the world. By 2018 Salmonella typhi (S. Typhi) had drug resistance to multiple antimicrobials including fluoroquinolones (FQ) [1]. Furthermore, there were many internationally transmitted reported cases of Extensively drug- resistant (XDR) S. Typhi strain from the United States, *Address correspondence to this author at the Department of Pediatrics, Al- Alwyia Teaching Pediatric Hospital, Baghdad, Iraq; E-mail: tareeffadhil@yahoo.com United Kingdom, Denmark, Germany, Canada; and Australia [2]. S. Typhi, a gram-negative bacilli that belongs to the species S. enterica, caused typhoid fever for an estimated 21.7 million illnesses and 216,000 deaths globally in 2000 [3-5]. Although the first outbreak of third-generation cephalosporin resistance was identified in Pakistan, resistant S. typhi strains have later been reported from India, Bangladesh, Philippines, Guatemala, Italy, and Iraq [6]. The first reported extended resistant S. typhi strain reported in Iraqi patients was back in 2008 [7]. Between 2002 and 2007, Multidrug -resistant (MDR) strains prevalence were 83% In Iraq [8]. Previous studies on Salmonella-resistant typhoid cases in Iraq showed a diversity of resistance patterns S. typhi among different institutions and geographical locations. Furthermore, these studies were not limited to a pediatric age group [8-10]. In Iraq and many other countries, there are limitations in using the national antimicrobial policy which include the local resistance problems which