Reflections

Category: AI Clinical Discovery

  • Outlier analysis for clinical discovery: An augmented intelligence framework and a systematic review

    Outlier analysis for clinical discovery: An augmented intelligence framework and a systematic review

    Outlier analysis for accelerating clinical discovery: An augmented intelligence framework and a systematic review

    Introduction

    Clinical research often hinges on identifying unusual or outlier cases, which can lead to groundbreaking discoveries. Traditionally, these outliers are identified through manual methods like case reports, which are time-consuming and limited by human observation. However, a recent study published in PLOS Digital Health proposes a novel approach using an augmented intelligence framework for outlier analysis. This method systematically scans large datasets to identify unique cases, offering a more efficient and comprehensive way to accelerate clinical discovery.

    Methods

    The study employed two primary outlier analysis techniques: the extreme misclassification contextual outlier method and the isolation forest point outlier method. The extreme misclassification approach used a random forest predictive model to identify misclassified data points, with a confidence level exceeding 90%. On the other hand, the isolation forest method identified outliers based on path length z-scores in the dataset. These methods were applied to two large datasets: the folic acid clinical trial (FACT) and the Ottawa and Kingston birth cohort (OaK). Clinical experts then reviewed the identified outliers to determine their potential as novel discoveries.

    Outlier analysis diagram

    Discussion

    The study found that the extreme misclassification method was more effective in identifying potential novel discoveries compared to the isolation forest method. Specifically, in the FACT dataset, 76.9% of the identified outliers were considered potential novelties, whereas in the OaK dataset, 32.7% were deemed novel. This highlights the importance of using advanced machine-learning techniques to enhance the process of clinical discovery. The researchers suggest that this method could be generalized across various medical disciplines, potentially transforming how we approach clinical research.

    Conclusion

    The study concludes that augmented intelligence, through systematic outlier analysis, offers a feasible and effective way to accelerate clinical discoveries. By identifying unique cases that might be overlooked by traditional methods, this approach has the potential to uncover novel insights and improve patient outcomes. The researchers advocate for broader adoption of this framework in clinical research to enhance the efficiency and impact of medical discoveries.

    Read full paper
    Request a demo
    Subscribe

    FAQs

    What is the purpose of using augmented intelligence in clinical discovery?

    Augmented intelligence is used to systematically identify outlier cases within large datasets, accelerating the discovery of novel clinical insights that may be missed by traditional methods.

    What are the key methods used in the outlier analysis described in the study?

    The study used two methods: extreme misclassification contextual outlier analysis and isolation forest point outlier analysis to identify unique cases in clinical datasets.

    How does the extreme misclassification method compare to the isolation forest method?

    The extreme misclassification method was found to be more effective, identifying a higher percentage of potential novel discoveries compared to the isolation forest method.

    What datasets were analyzed in this study?

    The study analyzed data from the folic acid clinical trial (FACT) and the Ottawa and Kingston birth cohort (OaK).

    What potential impact does this augmented intelligence approach have on clinical research?

    This approach could significantly accelerate clinical discovery across various medical fields by identifying unique cases more efficiently, leading to faster and more impactful medical breakthroughs.

    How can this method be applied in other areas of clinical research?

    The study suggests that the augmented intelligence framework could be generalized across different medical disciplines to enhance the efficiency of clinical discovery processes.

  • Augmented Intelligence for Clinical Discovery in Hypertensive Disorders of Pregnancy Using Outlier Analysis

    Augmented Intelligence for Clinical Discovery in Hypertensive Disorders of Pregnancy Using Outlier Analysis

    Clinical discoveries are heralded by observing unique and unusual clinical cases. The effort of identifying such cases rests on the shoulders of busy clinicians. We assess the feasibility and applicability of an augmented intelligence framework to accelerate the rate of clinical discovery in preeclampsia and hypertensive disorders of pregnancy-an area that has seen little change in its clinical management.

    Methods


    We conducted a retrospective exploratory outlier analysis of participants enrolled in the folic acid clinical trial (FACT, N=2,301) and the Ottawa and Kingston birth cohort (OaK, N=8,085). We applied two outlier analysis methods: extreme misclassification contextual outlier and isolation forest point outlier. The extreme misclassification contextual outlier is based on a random forest predictive model for the outcome of preeclampsia in FACT and hypertensive disorder of pregnancy in OaK. We defined outliers in the extreme misclassification approach as mislabelled observations with a confidence level of more than 90%. Within the isolation forest approach, we defined outliers as observations with an average path length z score less or equal to -3, or more or equal to 3. Content experts reviewed the identified outliers and determined if they represented a potential novelty that could conceivably lead to a clinical discovery.

    Results


    In the FACT study, we identified 19 outliers using the isolation forest algorithm and 13 outliers using the random forest extreme misclassification approach. We determined that three (15.8%) and 10 (76.9%) were potential novelties, respectively. Out of 8,085 participants in the OaK study, we identified 172 outliers using the isolation forest algorithm and 98 outliers using the random forest extreme misclassification approach; four (2.3%) and 32 (32.7%), respectively, were potential novelties. Overall, the outlier analysis part of the augmented intelligence framework identified a total of 302 outliers. These were subsequently reviewed by content experts, representing the human part of the augmented intelligence framework. The clinical review determined that 49 of the 302 outliers represented potential novelties.

    Conclusions


    Augmented intelligence using extreme misclassification outlier analysis is a feasible and applicable approach for accelerating the rate of clinical discoveries. The use of an extreme misclassification contextual outlier analysis approach has resulted in a higher proportion of potential novelties than using the more traditional point outlier isolation forest approach. This finding was consistent in both the clinical trial and real-world cohort study data. Using augmented intelligence through outlier analysis has the potential to speed up the process of identifying potential clinical discoveries. This approach can be replicated across clinical disciplines and could exist within electronic medical records systems to automatically identify outliers within clinical notes to clinical experts.

    Keywords: augmented intelligence; clinical discovery; clinical trials; hdp; hypertensive disorders of pregnancy; preeclampsia treatment; preeclampsia-eclampsia; real-world data; research methods and design.

    Read full paper
    Request a demo
    Subscribe

    FAQs

    What is augmented intelligence and how is it used in clinical discovery?

    Augmented intelligence refers to the collaboration between human expertise and artificial intelligence (AI) to enhance decision-making and discovery processes. In this study, it was used to accelerate clinical discoveries in hypertensive disorders of pregnancy by identifying unique and unusual cases through outlier analysis.

    What methods were used to identify potential clinical discoveries in this study?

    The study employed two outlier analysis methods: extreme misclassification contextual outlier analysis and isolation forest point outlier analysis. These methods were applied to data from the folic acid clinical trial (FACT) and the Ottawa and Kingston birth cohort (OaK) to identify cases that could represent potential clinical discoveries.

    What were the key findings from the outlier analysis in this study?

    The study found that the extreme misclassification contextual outlier analysis identified a higher proportion of potential novelties compared to the isolation forest approach. Specifically, 76.9% of outliers identified using the extreme misclassification method in the FACT study and 32.7% in the OaK study were considered potential novelties by clinical experts.

    How can augmented intelligence benefit clinical research in hypertensive disorders of pregnancy?

    Augmented intelligence can significantly speed up the process of identifying potential clinical discoveries by automating the identification of unusual and unique cases. This approach allows for faster and more efficient discovery of novel clinical insights that could improve the management and treatment of hypertensive disorders of pregnancy.

    Can the outlier analysis approach used in this study be applied to other clinical disciplines?

    Yes, the outlier analysis approach demonstrated in this study can be replicated across various clinical disciplines. It has the potential to be integrated into electronic medical records systems, where it could automatically flag outliers in clinical notes for further review by experts, thereby facilitating faster clinical discoveries.

    What is the significance of identifying outliers in clinical trials and real-world cohort studies?

    Identifying outliers in clinical trials and real-world cohort studies is crucial as these outliers may represent novel or rare cases that could lead to significant clinical discoveries. By focusing on these outliers, researchers can uncover new insights that may not be apparent in the general data, potentially leading to advancements in medical knowledge and patient care.