JMIR Publications

ABSTRACT

Background:

Evidence-based medicine combines scientific research, clinical expertise, and patient preferences to enhance patient outcomes and improve healthcare quality. Clinical data is crucial in aligning medical decisions with evidence-based practices, whether derived from systematic research or real-world data sources. Quality assurance of clinical data, mainly through predictive quality algorithms and machine learning, is essential to mitigate risks such as misdiagnosis, inappropriate treatment, bias, and compromised patient safety.

Objective:

This study aims to demonstrate the varying quality of medical data in clinical primary source systems and provide researchers with insights into data reliability through predictive quality algorithms utilizing machine learning techniques.

Methods:

A literature review was conducted to evaluate existing approaches to automated quality prediction. Additionally, metadata relevant to clinical data was stored in a relational database, taking into account factors such as data granularity and quality metrics. A predictive quality algorithm was developed using machine learning, focusing on preprocessing the dataset, training machine learning algorithms on echocardiographic, laboratory, and medication data, and assessing various prediction models to identify the most effective algorithms for quality classification.

Results:

Classifiers were used to predict the quality of medical data, and the performance of the algorithms was assessed based on accuracy, precision, recall, and scoring. Extreme Gradient Boosting (XGB) demonstrated the highest performance with an accuracy of 84.7%, AUC ROC of 84.6%, F1-score of 84.0%, and precision of 83.9%.

Conclusions:

This proposal presents a template for predicting data quality and incorporating the resulting quality information into the metadata of a data integration center, a concept not previously implemented. The model was deployed for data inspection using a hybrid approach that combines the trained model with conventional inspection methods.