Mining and analysis
Logical data and Domain modelling
Patient data can be hard to analyse because it is collected by a range of people, for a variety of specific purposes, using a range of data collection technologies, form paper forms, to electronic patient records. As a result data is frequently a mish-mash of duplicated and confusing records. When linking this data for analysis a common model for understanding the data from disparate systems is needed. Our clinical and domain modelling expertise enables the integration of complex clinical data.
Clinical Terminology mapping and analysis
Clinical terminologies represent a special problem in data analysis. Data may need to be mapped from one terminology to another, or analysis may need to account for complex parent-child hierarchies in the clinical terminologies. Our detailed understanding of terminologies and tools such as the UMLS is essential for analysis of coded clinical data.
Emerging patterns from clinical data is often as simple as looking at the data in the right way. That might be achieved through a traditional exploration of the facets of relational data, or by exploring the clustering of complex network graphs. Every data set has it’s own story to tell and we approach each with an open mind.
Data cleansing and transformation
As clinical data people we can help you set up data transformation pipelines to cleanse and normalise your clinical data. Whether this be anonymisation, terminology mapping, or natural language processing, we can help wrangle your clinical data into the best possible state for analysis.