Blood tests are used to help doctors identify and diagnose illnesses. Doctors often perform blood tests during routine health check-ups. To diagnose diabetes, for example, blood glucose (sugar) levels in blood are measured. If they exceed a certain predefined limit, the patient may have diabetes. Standard reference values are the basis for the interpretation of test results. However, lab values are subject to complex control mechanisms within the body and can differ by individual, despite known normal values. Thus, it is not clear whether the "one size fits all" approach is always appropriate. Age, sex and underlying diseases, for example, can have an influence on test results. Nonetheless, most information on reference ranges only looks at one test at a time and does not take into account the manifold interactions within the human body. As doctors perform and record many blood tests during routine health check-ups, the CPRD GOLD provides an ideal setting to start looking at this bigger picture (i.e. combination of multiple tests per patient) and to assess and describe subgroup-specific blood tests performance. This will provide an overview of the distribution of lab values clusters in middle-aged and older persons in a real-world primary care setting.
Most routine lab diagnostics were developed to detect a specific disease, but now are often performed in test panels during routine health checks or in patients with ill-defined symptoms. The interpretation of these test results poses a challenge in patient care, as normal values and acceptable deviations from the norm are not well defined or understood, particularly when it comes to the interpretation of multiple test results at the same time. We plan to study all CPRD GOLD patients aged 40-79 years present in the data between January 1, 2004 and December 31, 2017 to describe the performance of major routine blood tests such as lipid and glucose levels, liver or thyroid function tests and blood cell counts. Using unsupervised clustering we will describe cluster of patients with similar routine diagnostic test performance in this adult primary care population. We will provide this information for the whole study population, as well as for subgroups stratified by age, sex and metabolic or cardiovascular diseases.
Cluster of patients with similar routine diagnostic test performance.
Susan Jick - Chief Investigator - BCDSP - Boston Collaborative Drug Surveillance Program
Cornelia Schneider - Corresponding Applicant - University of Basel
Christoph Meier - Collaborator - University of Basel
Christoph Meier - Collaborator - University of Basel