AI in Drug Dosing: Improving Patient Outcomes and Reducing Cost of Care

By | HealthTech

Artificial intelligence (AI) is poised to transform chronic disease management by empowering physicians to personalize drug dosing at scale. Dosis, a personalized dosing platform used by dialysis providers to manage chronically administered drugs, has harnessed the power of artificial intelligence to disrupt traditional one-size-fits all dosing approaches, and has introduced precision dosing that is personalized, based on patients’ actual observed responses to their medication. This personalization significantly improves dosing efficiency–achieving equivalent or improved patient outcomes while reducing the cost of care.

Why is now the right time to use AI for dosing?

There are several developments that have come together to create the conditions necessary for the growth of AI-powered drug dosing. The first is technological advancement – modern-day computing power allows for efficient processing of large, complex datasets, making AI solutions practical to implement.

The second is the general public’s familiarity with artificial intelligence as a tool that can solve complex problems. Public familiarity with a diverse array of AI applications in imaging, consumer electronics, and logistics lend popular credibility to the technology, and allow healthcare providers to be comfortable incorporating such tools in clinical settings.

The third is the availability of reliable data in a clinical setting – electronic medical records codify and standardize data in a manner that is much more ingestible by algorithms than free-form paper medical records. The accessibility of dosing and outcomes data in electronic medical records and the ability to effectively analyze this data has made applying artificial intelligence and control algorithms to dosing much more practical and efficient. As a result, it is now possible to draw on data from millions of drug doses and their associated outcomes, a vast improvement on previously siloed and inaccessible datasets.

The final development that cements the need for more innovative dosing approaches is the fact that drugs are now more complex and have more potential to impact basic physiologic processes than at any point in the past. Drugs that impact multiple physiologic processes and have a narrow therapeutic window – the “sweet spot” between toxicity and ineffective therapy – provide the greatest opportunity for  AI-powered dosing.

Dialysis patients and providers significantly benefit from AI-based drug dosing

There are more than 550,000 dialysis patients with End-Stage Kidney Disease (ESKD) in the United States. Dialysis patients tend to have elevated risk for adverse outcomes, and are often on multiple medications. Almost 90% of dialysis patients experience chronic anemia and are treated with Erythropoiesis Stimulating Agents (ESAs). However, exposure to high doses of ESAs is associated with an increase in serious adverse cardiovascular events, like heart attacks and strokes, so the primary clinical intent is to use the minimum amount of medication necessary to prevent patients from requiring blood transfusions.

Dosis’s flagship tool, Strategic Anemia Advisor (SAA), offers dialysis providers clinically validated decision support, empowering them to make optimal dose adjustments based on individual patients’ needs and deliver patient-centered care. It allows clinicians to maintain or improve their hemoglobin outcomes while significantly reducing patients’ exposure to ESAs. This has benefits for both patients and providers: patients are exposed to 25% less drug on average, which means a lower likelihood of serious adverse events like heart attacks or strokes, while providers lower their drug utilization, saving an average $1250 per patient per year.

 The Future – Chronic Dosing To Be AI-Powered

Dosis has proven that precision dosing at scale is possible, and has established a path to success for future applications of AI within the kidney space, including precision dosing for mineral bone disorder (MBD), expansion of value-based chronic kidney disease (CKD) models, and more.

Over the next decade, partially as a result of Dosis’s innovations, AI-driven dosing models will likely be the standard of care across the chronic care spectrum. In addition, as more tools are developed and more opportunities to use those tools are identified, there will be exponential growth in the use of AI to drive innovative therapies.

Analyzing the Healthcare Industry’s Approach to Innovative Technology

By | HealthTech

Originally used as a metric to characterize the treatment of hypertension, therapeutic inertia is a concept that’s recently been reappropriated to describe the healthcare industry’s often slower-than-expected uptake of new tools and technologies. Often described as “the measurement of initiating or changing therapy in a timely manner according to evidence-based clinical guidelines,” therapeutic inertia is a useful lens through which to view innovation in healthcare.

Using Therapeutic Inertia As a Barometer for Innovating Care

As a treatment metric, therapeutic inertia is now standardly used to analyze physicians’ behavior during treatment courses for hypertension, diabetes, and hyperlipidemia. Typically calculated as the percentage of provider visits that resulted in a change in the treatment course, a higher percentage indicates a slow treatment of a condition by a healthcare provider, while a low percentage showing a quick response in providing a new treatment for a medical condition.

Studies have shown that reducing therapeutic inertia number results in improved patient outcomes and better-managed care, and we can see this measurement of treatment being applied to all chronic conditions as a whole. In fact, studies have found that reducing therapeutic inertia is crucial in order to more effectively control hypertension in over 50% of patients. This also is observed to be a significant issue in treating type 2 diabetes. However, this aversion to changing therapies and trying new medicines or treatments is broader than the metabolic conditions in which it has been directly studied.

The Risk May Be in Sticking with the Status Quo

The healthcare industry is risk-averse for understandable reasons: when patient health is at stake, the risks of new tools, treatments, and operating workflows can seem to outweigh the benefits. However, many new tools are aimed at reducing medical error, which is currently listed as the third leading cause of death in the United States, just behind heart disease and cancer. Death due to a medical error can be caused by simple mistakes, errors in judgment, or system defects, including computer breakdowns. With medical errors causing up to 440,000 deaths per year, technical innovations are needed to help limit human and system errors that could be fatal.

Using Technology for Early Intervention

It’s known that earlier intervention and advancements are better for patients’ overall outcomes. And, while healthcare practitioners naturally turn to their own clinical experience to make decisions on how to treat their patients, turning to artificial intelligence-based technology that synthesizes a broader pool of experience to help them in their work has proven to be quite effective.

A recent study presented at the 2019 ASCO Annual Meeting showed that physicians changed their treatment decisions in 13.6% of cases after being presented treatment recommendations from IBM’s Watson for Oncology tool. Diving deeper, of those that changed their decisions, more than half did so because the innovative technology provided more up-to-date, evidence-based information on newer treatments compared to their own individual knowledge. Having artificial intelligence to help reduce healthcare practitioners’ errors in judgment leads them to perform their duties to the best of their abilities. This is a game-changer in treating all types of ailments effectively and efficiently.

Preparing for Value-Based Care

Investment in innovative technologies gives modern healthcare organizations an edge as they prepare for value-based care models. Although most clinics run on a pay-per-visit model currently, the transition to value-based care will require maximum efficiency and effectiveness.

Tracking and working to reduce therapeutic inertia will also push healthcare providers to adapt faster and make iterative improvements to patients’ treatments plan.

With the stakes so high, providers have a good reason to be risk-averse. However, if this risk aversion is taken too far, it can impede innovation, ultimately resulting in harm to the patients that healthcare providers are trying to protect.