The Cochrane Collaboration is the world’s leading evidence-based clinical healthcare organisation. They encode clinical questions of interest in a standardised framework known as PICO (Population, Intervention, Comparator and Outcome) which are translated into search protocols to be run against biomedical literature databases such as PubMed and EmBase.
Search results are then assessed for risk of bias, and the remaining studies are used as the data foundation of a systematic review production process as the result of which a PDF is published which summarises the total known evidence about the clinical question. Consumers of Cochrane systematic reviews include government health policy bodies, clinical guidelines organisations, health researchers and practicing clinicians.
Cochrane’s review production methodology is highly rigorous, as a result of which their output is regarded as the gold standard of scientific objectivity. However this rigour costs time, and historically they have struggled to meet demand for both new questions and updates to existing questions. Also, publishing reviews as semi-structured PDF documents severely impacts their data usability and flexibility.
As a result, Data Language were engaged by Cochrane to explore whether linked data could be used to improve productivity without compromising quality to create a fully structured clinical evidence dataset that could meet the flexibility, extensibility and queryability needs of customers.
Review authors and information specialists are now able to compile PICO questions into SPARQL or ElasticSearch queries, automating the previously time-consuming manual search of primary evidence. In conjunction with data extraction and meta-analysis services, this gives Cochrane the capability to generate baseline systematic review datasets on demand.
In addition, by annotating primary evidence at ingest time, alerting services allow review authors to be notified as soon as new evidence is published which is relevant to systematic reviews they have previously published. PICO micrograph annotations then allow consumers to perform complex searches of primary and aggregated evidence, via APIs or via rich and friendly user interfaces.
Cochrane’s Linked Data Platform has now attracted major investment from US charitable foundations who are keen to leverage its capabilities in their health care initiatives. Also a number of other healthcare organisations have recently approached Cochrane about utilising its linked data capabilities, which is creating new and unexpected revenue opportunities.
If you would like to know more about Data Language or have a project you would like to discuss with us, then please get in touch.