Role of patient descriptors in predicting antimicrobial resistance in urinary tract infections using a decision tree approach: A retrospective cohort study

Highlights

• Antimicrobials are prescribed empirically for urinary tract infections without microbiological confirmation (pathogen) in general practice.

• Antimicrobial resistance is driven by antimicrobial prescribing, of which 80% in general practice.

• A decision tree model based on patient’s descriptors was developed to predict culture positivity and antimicrobial susceptibility.

• The model shows 68%–92% accuracy in predicting culture and resistance based on patients descriptors.

• Extending the application of decision tree models could support general practitioners to improve antibiotic prescribing.

Abstract

Background

In general practice, many infections are treated empirically prior to or without microbiological confirmation. Prediction of antimicrobial susceptibility could optimise prescribing thus improving patient outcomes. Decision tree models are a novel idea to predict AMR at the time of clinical presentation. This study aims to apply a prediction model using a decision tree approach to predict the antimicrobial resistance (AMR) of pathogens causing urinary tract infections (UTI) for patients over 65 years based on pre-existing routine laboratory data.

Methods

Data were extracted from the database of the microbiological laboratory of the University Hospitals Galway (UHG). All urine results from patients over 65 years, their microbiological analysis and susceptibility (AST) results from January 2011 to December 2015 were included. The primary endpoint was culture result and resistance to antimicrobials (nitrofurantoin, trimethoprim, ciprofloxacin, co-amoxiclav, and amoxicillin) commonly used to treat UTI. A non-parametric regression tree analysis i.e. a decision tree model was generated with the 75% of the dataset (training set) and validated with the remaining 25% (test set). The model performance was evaluated measuring Area Under the Curve Receiver Operating Characteristic (AUC_ROC) curve.

Results

A total of 99,101 urine samples of patients over 65 years were submitted for culture over the five years and 27% had significant bacteriuria (≥10⁴ cfu/ml) and AST. The most common identified causative organisms were E.coli, Klebsiella spp. and Proteus spp. E.coli was more often resistant to amoxicillin (66%) followed by Proteus spp. (41%). Klebsiella spp. and Proteus spp. were more often resistant to trimethoprim (78% and 54% respectively). E. coli resistance to nitrofurantoin is low (<10%). The decision tree model showed an AUC-ROC score of 0.68 for culture and in between 0.60 to 0.97 for antimicrobial resistance of the pathogens, with the inclusion of patient’s descriptors only. Including the uropathogen in the model did not change model performance.

Conclusions

The decision tree models using patient descriptors available at the time of presentation showed fair to excellent performance in predicting culture and antimicrobial resistance. The presented models provide an alternative approach to decision making on antimicrobial prescribing for UTIs. Increasing more predictors in the model could improve the model performance. Prospective data collection, validation and feasibility testing of the model including data from other laboratories will progress the practical implementation of similar models.

Introduction

Urinary tract infection (UTI) is the most frequently occurring bacterial infection [1]. Both in the UK and the Netherlands 1% of all general practitioners (GPs) consultation are for UTIs [2,3]. In the United States alone 7.9 million physician visits per year are for UTI treatment [4]. The prevalence of UTI increases with age and UTIs occur more often in females than in males [5,6]. Elderly in the community (over 65 years) as well as nursing homes (NHs) residents more often have UTIs [7]. UTI is the number one infection in long-term care facilities (LTCFs) in Europe with 31% of residents reported as having a UTI (10% confirmed and 21% probable UTI) on the day of the survey [8]. UTI is caused by various pathogens, but gram-negative bacteria such as E. coli, Klebsiella species, and Proteus mirabilis species are the most commonly identified pathogens [9,10].The rapid increase in the resistance of these bacterial pathogens to commonly used antimicrobials is a global concern. Due to increased resistance, the number of alternative options of antimicrobials to treat infections is decreasing [11]. The National Institute for Health and Care Excellence (NICE), UK and the Health Service Executive (HSE) Ireland have developed guidelines for appropriate use of antimicrobials for patients over 65 years [12,13]. According to the HSE, trimethoprim and nitrofurantoin (except when renal impairment) are the recommended first-line antimicrobials to treat UTI in patients over 65 years [13]. Co-amoxiclav, ciprofloxacin, or cephalexin are recommended when acute pyelonephritis develops [13,14]. Empiric treatment should take account of local resistance patterns where applicable to ensure effective treatment of the condition in question [13,15]. In the western region of Ireland, it has been shown that 50% of E. coli are resistant to ampicillin and 30% to trimethoprim, making these agents unsuitable for empiric UTI treatment [15,16]. Various guidelines outline about the appropriate prescribing of antimicrobial, but still a high level of inappropriate and overprescribing of antimicrobials has been observed in the community [17]. This has negative consequences including the development and further spread of multidrug resistance (MDR) [18,19]. Studies conducted in NHs estimate that 25–75% of the systemic antimicrobials and up to 60% of topical antimicrobials are prescribed inappropriately [20].

Urine samples constitute the largest category of specimens sent from the community to medical microbiology laboratories [21,22]. Guidelines recommend urine culture in certain conditions such as symptomatic UTI, recurrent UTI, infections of upper urinary tract, during pregnancy or where firstline antimicrobials failed [13,14]. However, in a previous study it was shown that general practitioners routinely submit samples on from all patients suspected to have a UTI [23]. Generally, in patients with typical clinical features of UTI empiric prescribing is initiated prior to microbiological confirmation of bacteriuria and antimicrobial susceptibility testing. Predicting culture results and possible resistance of pathogens to antimicrobials based on patient descriptive characteristics at General Practice could improve prescribing, assist in reducing AMR and ultimately the cost of care. This motivated us to apply a decision tree approach to develop a prediction model on AMR. The model predicts culture results (positive/negative) and resistance of pathogens associated with UTI to commonly prescribed antimicrobials (nitrofurantoin, trimethoprim, ciprofloxacin, amoxicillin, and co-amoxiclav) using basic patient descriptive features.