A pilot study led by Dr. Saumya Shekhar Jamuar analyzed over a million health records, uncovering new symptoms and progression patterns in two rare diseases. Their research enhances our understanding of genetic conditions and demonstrates the value of EHRs and innovative population health tools like PopAnalyzer in fostering personalized and proactive approaches to treating patients.
Download.jpg)
Rare genetic diseases are a hidden epidemic, affecting a sizable portion of the population and often going undiagnosed or misdiagnosed.
A team of trailblazing researchers from SingHealth Duke-NUS Institute of Precision Medicine (PRISM), Singapore General Hospital (SGH), KK Women’s and Children’s Hospital (KKH), National Heart Centre Singapore, Rare Care Centre of Western Australia, alongside the SingHealth Office of Insights & Analytics (OIA) and the Curtin University School of Electrical Engineering, Computing, and Mathematical Sciences, are harnessing the power of electronic medical records (EMR) and healthcare data analytics tools to mine vast amounts of healthcare information and improve diagnostic methods for rare conditions.
Dr. Saumya Shekhar Jamuar, MBBS, MRCPHCH, FAMS, FACMG, Senior Consultant of Genetics Service at KKH and PRISM Director, renowned for his expertise in treating rare genetic conditions, led this team in a 3-year pilot study to mine 1.28 million electronic health records to identify previously unknown connections between symptoms, clinical data, and disease progression for two rare conditions: familial hypercholesteremia and Fabry disease.
Researchers estimate that 5-8 percent of the population in Singapore has a rare genetic disease. In addition, according to Rare Diseases International, about 300 million or about 5-6 percent of the world’s population are affected with a rare disease as of 2019.
Implementing technological innovation in a field that is often overlooked allows medical professionals and researchers to make data-driven decisions based on real-world evidence, even when there’s limited availability of patient data, rather than relying solely on individual case studies or conventional medical knowledge.
One of the most significant advantages of data mining in rare genetic disease diagnosis is its ability to integrate multiple sources of healthcare data. Scientists can create a comprehensive profile of individuals affected by these diseases by analyzing electronic records, genetic sequencing data, medical imaging results, and even patient-reported symptoms.
This holistic approach improves diagnostic accuracy as it considers a wide range of factors that are typically overlooked when providers rely on a single source of information.
The team mined data for markers and symptoms associated with familial hypercholesteremia and Fabry disease. However, the process of rendering EMR data usable for mining was nuanced and multifaceted. It involved many steps to ensure the quality, integrity, and confidentiality of health records.
First, data extraction was required to pull relevant clinical data from the EMR system, collected from various sources, including labs, radiology, pathology, diagnoses, and detailed patient chart information—all of which contained identifiable patient records.
To ensure privacy and compliance with security protocols, researchers used internal tools to de-identify the patient data, removing all personal identities before further processing, according to Analysis and visualisation of electronic health records data to identify undiagnosed patients with rare genetic diseases.
Once the data was de-identified, they leveraged Health Catalyst Pop AnalyzerTM to normalize and standardize the structured data. The tool identified or corrected errors, including removing duplicate records, correcting misspelled or inaccurate entries, and resolving any missed values. If done manually, this task is time-consuming and laborious.
One major obstacle to detecting and diagnosing rare conditions is the lack of data and research available to healthcare professionals. Moreover, many rare diseases have complex and varied symptoms that overlap with common ailments. Further complicating matters is that genetic testing for rare diseases can be costly and time-consuming, making such testing less likely to be employed in routine clinical practice.
Fabry disease is a rare genetic, neurological disorder affecting multiple bodily systems. It is caused by a mutation in the GLA gene, which leads to a deficiency of the alpha-galactosidase A enzyme. This enzyme breaks down a fatty substance called globotriaosylceramide (GL-3). Without enough of this enzyme, GL-3 builds up in various cells and tissues throughout the body, leading to a range of symptoms and severity levels. Additionally, the onset and progression of Fabry disease can vary, making it difficult to predict how the condition will impact someone’s life.
Similarly, researchers have found that the diagnosis of familial hypercholesterolemia (FH), characterized by abnormally high serum levels of low-density lipoprotein (LDL) cholesterol, is often delayed. Patients are typically diagnosed after presenting with a “catastrophic event." Meanwhile, individuals with FH inherit the condition from one or both parents with the gene mutation, which highlights the importance of genetic testing and family screening to identify those at risk and intervene with appropriate treatment and strategies early.
Saumya’s study identified two possible patients with Fabry disease, leading to a potential 50 percent increase in the number of patients with Fabry disease within the local healthcare system. His team also discovered over 12,000 patients with suspected FH, which is close to their population’s expected prevalence.
Analyzing extensive clinical documentation from both affected individuals and unaffected populations uncovered previously unknown parameters, including genetic aberrations associated with rare diseases.
Researchers wrote that this study is an early step towards increasing the quality of care for people living with rare diseases by reducing their diagnostic odyssey so that more time and resources can be spent on managing the disease.
Indeed, their approach to these findings could open new avenues for greater diagnostic accuracy, targeted therapies, and personalized medicine.
Would you like to learn more? Check out these related articles:
Standardizing Excellence: Five Pillars for Reducing Unwarranted Care Variation
How One Hospital System Advances Clinical Documentation Improvements with Augmented Intelligence
Three Ways Healthcare Analytics Streamlines Workflows and Enhances Productivity
