Leveraging graph topology and semantic context for pharmacovigilance through twitter-streams

BMC Bioinformatics

Table 2 ADEs Discovered through DEG Analysis

Drug	Effect	Ref	Drug Tweet	Effect Tweet
Pravastatin	Eczema	[38]	Sertraline and Propranolol in the morning low dose Propranolol throughout the day Trazodone, Loxapine and pravastatin at night	then theres the Clobetasol Propionate for my Eczema.
Atenolol	Cramps	[39]	I’m only way to changing my BP medication. The Atenolol I have been taking hasn’t been effective enough. I was in hospital last weekend.	I’m just here trying to deal with cramps while watching Jesus documentaries on TV.
Vyvanse	Alopecia (Balding)	[40]	Starting to take Vyvanse again tomorrow	Heck yeah I went bald
Digoxin	Exhaustion	[41]	digoxin why u gotta be so complicated?	Why am I so exhausted 24/7 I’m so over this ughhhhh
Tamoxifen	Stroke	[42]	I learned about tamoxifen and strokes after my aunt died of a massive bleed after unnecessary long tamoxifen treatment. Nurse told me!	[see Drug tweet]

Table 2 This table contains ADEs that were implied by the topology of DEG_Sider. Evidence for each of these ADEs was found both in the Twitter-stream as well as in medical literature. Twitter text is provided with the medical entities in bold (text is nominally altered/censored where appropriate). This process of ADE hypothesis has potential use in helping fill gaps in the databases like SIDER, when evidence from the literature is present. It may also be used to hypothesize new and unreported ADEs for further investigation

ISSN: 1471-2105