Co-PI ($624K) Total $950K (PI Megan O'Boyle from PMSF )
PCORI - PPRN http://www.pcori.org/research-results/2013/phelan-mcdermid-syndrome-data-network
The PMS_DN mission is to advance knowledge, care, and treatment of Phelan-McDermid Syndrome (PMS) and related conditions by integrating diverse, complex data sources into a richly populated, high quality, and accessible centralized database to facilitate patient-centered research. The network promotes a culture of transparency and authentic engagement and leadership of families in network vision, governance, and generation of research priorities. The organizations participating in the PMS_DN include: The Phelan-McDermid Syndrome Foundation, the Harvard Medical School (HMS) Department of Biomedical Informatics, and Boston Children’s Hospital.
PMS, a rare genetic condition caused by errors on the end of chromosome 22 (22q13), is associated with autism/autism spectrum disorders, intellectual disabilities, sleep disorders, and seizures. There is no treatment for PMS. There are ~1,100 diagnosed cases worldwide, with at least 520 cases confirmed in the US. However, a conservative estimate, regardless if known/diagnosed, is 8,898 US cases. Since PMS_DN’s launch in July 2014, 281 families have consented to participate in the network and PCORnet studies (~50% of known US diagnosed cases). The PMS_DN’s enrollment target for Phase II is 400 US cases (~75% of known US diagnosed cases). PCORnet offers a vital opportunity to advance knowledge and care for PMS, capitalizing on the consortium to enroll families in observational studies and pragmatic clinical trials, while identifying undiagnosed individuals and connecting families to needed services. The PMS_DN brings unique assets to PCORnet: an established data portal for collecting patient-reported outcomes (PROs), the PMS International Registry (PMSIR), and i2b2/tranSMART, a platform that allows researchers to analyze complex data types harmonized to the PCORnet Common Data Model (CDM) v2 and enables integration of PROs with validated knowledge extracted from EHR clinical notes through Natural Language Processing (NLP).
The PMS_DN integrates PROs with data from the application of NLP to EHR clinical notes submitted by families. The technical platform to achieve this integration and harmonize data domains with the current version of the PCORnet CDM is i2b2/tranSMART. The PMS_DN completed a mapping file for PRO data that has been validated with real patient data and production of an audit trail of the clinical data integration pipeline in tranSMART. An advantage of i2b2/tranSMART is that an investigator can quickly validate outcomes of interest from the output of the NLP with the i2b2/tranSMART interface, viewing all anonymized sentences from which the concept was generated. The network uses cTAKES including a new temporal module that extracts time and event concepts from clinical notes, enabling a complete longitudinal view of a patient. The PMS_DN maps to several ontologies/terminologies: SNOMED CT, ICD9, ICD10, Human Phenotype Ontology, Gene Ontology, and MeSH.
