Introduction
Clinical reasoning is a process used for analyzing knowledge relative to a clinical situation or specific patient. This skill is introduced in medical schools as a methodology for representing and relating clinical problems; however, it is truly learned through repeat exposure to real case examples that illustrate the complexity of patients’ conditions. The fast pace of healthcare leaves little room for the development of this skill. Medical science is advancing every day, and it has become increasingly difficult for healthcare providers (HCPs) to integrate emerging evidence into their clinical reasoning and decision making processes.
The evidence-to-practice gap
Evidence-based medicine aims to integrate clinical expertise with the best available evidence coming out of clinical trials and disease registries. Still, the majority of patients in the United States are not receiving care that aligns with current evidence. This disparity is partially attributed to the fact that it takes an average of 17 years before medical evidence is integrated into routine patient care. This delay not only impacts patient outcomes but also burdens healthcare systems with inefficiencies and elevated costs; it also compromises life science companies’ ability to recover the costs of researching and developing a new therapy before its patent cliff expires.
Challenges in clinical practice
HCPs face numerous challenges in assimilating the ever-expanding medical knowledge into their clinical decisions. While clinical practice guidelines (CPGs) are developed to help guide them, more than 8,000 CPGs exist and periodic updates make it difficult to stay current on the latest best practices. Current reference tools like UpToDate, Epocrates, MDCalc, and PubMed provide access to CPGs but still require HCPs to interpret and apply these guidelines in the context of their patients' unique needs, a task that is both time-consuming and prone to natural variations in interpretation.
Clinical prediction models: A double-edged sword
In addition to clinical reference tools, HCPs often rely on risk estimators, or clinical risk prediction models, to inform their decision-making process. Clinical prediction models incorporate statistical methods to estimate the likelihood of a specific outcome, such as the presence of disease or the probability of an event occurring in the future. These models can tend to oversimplify risk categories, however, leading to imprecise estimates and potentially leading to unnecessary or inappropriate interventions. For example, in the illustration below, we demonstrate how we developed a scoring algorithm that is more accurate than the current standard for predicting an individual's likelihood of experiencing a cardiovascular event.
The Clint model identified 39,404 patients as high-risk (represented in orange) compared to the ASCVD Risk Estimator, which identified 39,548 patients as high-risk (represented in green).
An innovative approach
Clint addresses the current limitations of clinical reference tools and prediction models with an innovative approach that we term ‘clinical concept mapping.’ This approach enables us to uncover patterns and relationships in patient data and provide personalized recommendations for treatment.
Advantages of clinical concept mapping
- Personalization: Clint automates patient assessment by extracting concepts from patient health records, a method of natural language processing (NLP), and mapping clinical concepts to CPGs (as well as clinical trial criteria) to determine a diagnosis and make personalized recommendations for care — whether it be a test, prescription, or referral into a clinical trial.
- Quality: Automating the assessment process reduces the time HCPs spend on patient assessment, allowing for more patient-focused care.
- Accuracy: Our algorithms provide more precise risk predictions, improving the reliability of clinical interventions
Scenario: Cardiovascular disease risk assessment
|
Without Clint |
With Clint |
|
|
Conclusion
Clint offers a sophisticated yet practical tool for bridging the evidence-to-practice gap. We are supporting HCPs to deliver personalized, evidence-based care, ultimately driving improvements in patient outcomes and healthcare system-wide performance. As this technology continues to evolve and integrate into clinical workflows, it promises to reshape the practice of evidence-based medicine, making care more precise and patient-centered than ever before.
Authored by
Cassandra Broadwin, MPH, Rajesh Dash, MD, PhD
