Descriptive analysis of Fiona's retrospective database

[docsteveharris]

• [x] governance: service evaluation
• [ ] Descriptive analysis of Fiona's retrospective database No new data
• [x] submit request for ADS legacy systems for list of patients seen in FW clinic 2001 - 2021
• [x] Compare legacy list to db list
• [ ] Produce a paper

Fiona's data (2002-2019) 2002-2012 600 women 2012- 2019 1000+ denominator validation - not missing visits to her clinic up until Epic implementation (i.e. early 2019)

So you can count women

But outcomes are all problematic

• obstetric outcomes
• fetal outcomes
• cardiac outcomes

Strategies to overcome this missingness

• search against legacy systems maybe back to 2012

[docsteveharris]

2021-08-24 Supervision SH/CA good quality data until the service moves to Barts data quality good for diagnosis/condition lots of missingness for maternal and foetal outcomes

data is likely to be available

• tomcat will hold echo data
• birth and maternal outcomes also likely to exist on legacy data systems at UCLH

SH suggest planning the descriptive study as follows

• focus just on the UCLH period before the move to Barts
• pull in the linked data from other legacy systems (I think the obs database might be called Medway)
• neonates and neonatal outcomes might also be captured by the national audit for neonates but not 100% sure if this is all kids or just kids admitted to NICU
• work toward a single 15 y case series report

Think of this as building the priors for your informatics consult And also building the data spec for future data collection

[docsteveharris]

3 sources of data

• echo notes: done; 1 row per echo; all patients who have been Fiona's clinic from approx 2008
• opera: stuck b/c we don't seem to be able to filter against MRN; Cheryl will follow-up and loop me in if needed
• midwifery notes for outcomes: regulatory stuff in process but prob need to build a relationship with midwife/maternal research teams