Develop antimicrobial stewardship programs using AI models to ensure adequate antibiotic selection
Roughly 12,000 audits conducted at SGH in 2018 alone revealed that 20 to 30 % of these prescriptions were inappropriate.
Overuse or misuse of antibiotics in hospital settings is a global issue which fuel one of the most important threats to public health: antimicrobial resistance. An estimated 20 to 50 per cent of acute care hospitals worldwide do not adequately prescribe antibiotics. As a result, micro-organisms, including antibiotics, are immune to antimicrobials and can become superbugs that don't react to any medication.
In Singapore, up to 30% of infections in hospitals are resistant to cephalosporins of the third generation (a large group of antibiotics derived from mold acremonium) that are widely used in the wide range of antibiotics.
One way to tackle the antimicrobial resistance (AMR) issue is through the implementation of antimicrobial stewardship programs (ASP).
These initiatives aim to avoid overuse of antimicrobial agents and ensure optimum antibiotic selection. Setting a consistent prescribing protocol is a key priority at acute care hospitals in Singapore, led by ASP teams and backed by accurate , detailed data. However, the development of an effective ASP brings with it many challenges and considerations, as Dr Vinod Seetharaman, Asia's Chief Medical Officer at DXC Technology shared with SGH from their recent collaboration.
Trajectories of the disease can vary widely because of a variety of factors such as patient genotype, phenotype, pathogen, co-morbidity, prescribed medication, and other social factors. Therefore, it is important to realize that all these factors can be of varying relevance to different groups of diseases and that resolving for a group of diseases at a time would be a more prudent strategy that allows for the gradual refinement of techniques with a well-defined scope, according to Dr Seetharaman.
One of the biggest problems was having a line of sight on all factors that affected them. Furthermore, the clinical condition chosen would also be one in which the lifecycle of the patient's path from diagnosis to resolution should include a data trail, for example, inpatient conditions usually occurring in the hospital setting and often most likely to be resolved in a hospital environment.
