31st January 2022

To celebrate our 50th anniversary, we’re inviting a select number of experts to share what they see as being the 21st-century solutions to one of the biggest and most complex challenges of our generation: antimicrobial resistance (AMR). In this post, Dr Amy Chang and Jonathan H. Chen MD PhD, explore the role artificial intelligence can play in tackling AMR…

Antibiotics are one of the most important advances in modern medicine, increasing average human lifespan by over 20 years. Yet, increasing bacterial resistance is a natural and inevitable consequence of antibiotic use, requiring essential strategies to sustain these gains in health. The driving concept of antibiotic stewardship is using the right antibiotic, at the right dose, at the right time. This collides with data suggesting that up to half of all antibiotic prescriptions are either unnecessary or inappropriate.

Why is appropriate use of antibiotics so hard?

The choice of antibiotics when first encountering a patient, known as “empiric” antibiotic prescribing, is essentially educated guesswork.  Antibiotic appropriateness is determined by the infectious pathogen and its susceptibility to various antibiotics. Unfortunately, the key tests providing this data, microbial cultures, often take days to result. This is far too long to delay empiric treatment that could minimize the chance of patient death from sepsis. If doctors don’t know what we’re treating, but we know we need something that works immediately or else the patient might die, then we often give an antibiotic that treats everything possible. This mentality leads to physicians picking a “shotgun” broad-spectrum antibiotic, when a narrow spectrum precision “scalpel” would have been more effective.  Moreover, many non-infectious diseases can look similar to infection, making decisions even harder as to whether any antibiotic is needed at all.

National guidelines on specific syndromes (e.g. pneumonia, skin/soft tissue infection) provide general suggestions on when to treat infection and what antibiotics to use, but they can take years to develop and can only provide overly general one-size-fits all guidance that providers are often not even aware of. Local hospital antibiograms document resistance patterns within a specific healthcare facility, but still not specific to particular patients. We need better tools at the point of care to guide precision antibiotic prescribing, personalized and rapidly updated to new streams of data.

Artificial intelligence (AI) and machine learning for antibiotic stewardship?

Existing tools and standards to support antibiotic prescribing provide important guidance but are too broad for personalized recommendations that must balance immediate risks of undertreatment against the nebulous risks of overtreatment. This actionable, arbitrary, and ascertainable process where an important decision (empiric antibiotic prescribing) depends on our variable human ability to predict a verifiable result (diagnostic culture susceptibilities) is ideally suited for innovative machine learning methods.

Artificial intelligence (AI) is often depicted as science fiction of a distant future, yet modern algorithms are already shaping our lives every day. Do internet advertisements seem uncannily specific to your interests?  This is the power of predictive analytics: Using machine learning predictive models on large scale data to generate individualized predictions and suggestions. Imagine if we were to use this same power, except in choosing the right antibiotic, individualized towards a single patient.  With similar technology, we can use the vast amounts of data provided by electronic medical records to create predictive models to optimize the accuracy and consistency of the current educated guesswork of empiric antibiotic prescribing.

Electronic patient charts provide ample information that can be utilized by models from the history of past infections and antibiotic susceptibility data to patient specific presenting symptoms, past medical history, laboratory results, and imaging. We currently use this data in large scientific studies to better determine which variables independently predict the need for specific types of antibiotics in general. Using the right tools, AI models can rapidly sift through similar data, and with statistical training, they can be taught to predict the answers to specific, patient oriented questions. Does this particular patient have an infection or not?  If so, is it safer to use the shotgun or the scalpel?  If we give our AI the right variables and the right training, then models can estimate probabilities of infection with drug-resistant bacteria for specific patients quantitatively in a manner that humans can only intuit qualitatively. The output can be real-time, giving providers the guidance needed at a critical time when it is needed the most. Just as we would expect from competent clinicians, AI models should not be stagnant. Continuously learning algorithms should adapt to incoming streams of new patient and epidemiologic data.  With additional training, predictions can become better and more relevant, allowing for a more dynamic guidance tool than static guidelines or yearly antibiograms.

There have already been breakthroughs in this area. For instance, at Stanford we are pitting a machine learning model’s suggestions for empiric antibiotics against the choices already made by our own providers.  We compared the predictions made for antibiotic selection at the beginning of an infection to the resistance profiles from culture information obtained at the end. Such models can choose a correct antibiotic based on resistance profile as well, or better, than the clinician’s choice while choosing narrower spectrum antibiotics. This indicates the potential to improve both safety and stewardship, simultaneously. Similar work is emerging as locally trained and optimized tools in multiple sites.

Impact on population scales will require advances in exchanging data across diverse health systems, the key fuel to power modern AI systems. While substantial inertia tends to lock such data into local health system silos, the shock of the COVID pandemic finally motivated many to pool their data, resources, and effort to face a larger threat. Laws on improving interoperability between health systems as well as international standards like Fast Healthcare Interoperability Resources (FHIR) are advancing efforts in this area.  Hopefully we will collectively recognize the even greater, but slowly boiling, problem of antimicrobial resistance to take action. Within the next decade, we envision providers routinely using AI models as their own personal guidance tools, ones which provide recommendations that are not just effective for treating immediate infections, but also in combating the otherwise relentless development of antimicrobial resistance.

Amy Chang

Dr Amy Chang is a Clinical Assistant Professor within Stanford University School of Medicine’s Department of Infectious Disease & Geographic Medicine.  She also acts as Medical Informatics Director for the Infectious Disease, Infection Prevention, and Antimicrobial Stewardship within Stanford Health Care.   Her training in pharmacy school, then medical school, and finally in fellowship for Infectious Disease married together to develop her keen interest in antimicrobial stewardship.  She pursued a third year in Infectious Disease fellowship training at Stanford University School of Medicine to focus specifically on antimicrobial stewardship, at which time she was awarded with the Society for Healthcare Epidemiology of America’s “Race Against Resistance” Education Scholarship recognizing her work and interest in antimicrobial stewardship.  Her current research interest is in leveraging informatics tools to enhance antimicrobial stewardship, quality improvement, and wellness of infectious disease providers.  She is also active clinically, focusing on management of infectious diseases in orthopedics, critical care, and the general population.

Jonathan H Chen

Jonathan H Chen MD, PhD, is a physician-scientist with professional software development experience and graduate training in computer science. He continues to practice Internal Medicine for the concrete rewards of caring for real people and to inspire his research focused on mining clinical data sources to inform medical decision making.

Chen co-founded a company to translate his Computer Science graduate work into an expert system to solve organic chemistry problems, with applications from drug discovery to a practical education tool distributed to students across the world. To gain first-hand perspective in tackling the greater societal problems in health care, he completed medical training in Internal Medicine and a VA Research Fellowship in Medical Informatics. He has published influential work in venues including the New England Journal of Medicine, JAMA, JAMA Internal Medicine, Bioinformatics, Journal of Chemical Information and Modeling, and the Journal of the American Medical Informatics Association, with research awards and recognition from the NIH Big Data 2 Knowledge initiative, National Library of Medicine, American Medical Informatics Association, Yearbook of Medical Informatics, and American College of Physicians, among others.

In the face of ever escalating complexity in medicine, integrating informatics solutions is the only credible approach to systematically address challenges in healthcare. Tapping into real-world clinical data streams like electronic medical records with machine learning and data analytics will reveal the community’s latent knowledge in a reproducible form. Delivering this back to clinicians, patients, and healthcare systems as clinical decision support will uniquely close the loop on a continuously learning health system. Dr Chen’s group seeks to empower individuals with the collective experience of the many, combining human and artificial intelligence approaches to medicine that will deliver better care than what either can do alone.

Refer to Dr Chen’s web-page for additional in-depth bio information, publication lists, CV, etc.

http://web.stanford.edu/~jonc101

*

Additional reading

*

BSAC’s Initial Response

Artificial intelligence feels like a glimpse into the future, until we realise just how much is possible already.

Clinical algorithms are in routine use in many settings and prescribers routinely work through these to make patient management decisions. These diagnostic processes often rely on clinical scoring tools and data extraction for many of these can be automated. This is happening now, but advances in the scale and speed of data analysis and the ability to link large datasets from multiple different sources make AI very powerful.

Machine learning will make it better and better over time. A key benefit is that it facilitates a much more individualised patient treatment. Relating to antimicrobial stewardship, current empirical prescribing tends to be overly cautious and the identification of individuals who are at low risk and do not need immediate treatment or who can be safely treated with narrow spectrum agents is an immediate benefit.

GP? Pharmacist? Researcher? Social Scientist? Whatever field you work in, if you’re committed to fighting infection then we want you to join us as a BSAC Member.
Become a member
I have read and agree to the Privacy Policy

Having trouble finding what you're after?

Top