Urban: Now how better to screen for decision support? Who to test, and why, will realize cost savings. #PMWC17
12:27pm January 24th 2017 via Hootsuite
Q: Addiction medicine, genetics, self-funded, where does research go? Urban: Very expensive cp to genetics for prognosis or Dx a win #PMWC17
Urban: For the ave pediatrician? What about later at 6mos or 12mos? A 20-pg report, what to do next w/o academic ctr specialists? #PMWC17
12:26pm January 24th 2017 via Hootsuite
Urban: Negative biopsy, sent out WES, rare disease assoc'd w/gypsies, from a caucasian family. #PMWC17
12:24pm January 24th 2017 via Hootsuite
Q: Speed of info to PoC? Urban: What test to send? (Was something missed?) Liver failure, unk cause, common in neonates. #PMWC17
12:23pm January 24th 2017 via Hootsuite
Mattison: Ontology diversity: dozens of overlapping, conflicting ones. May need perhaps 6; want 'patients like mine' #PMWC17
12:19pm January 24th 2017 via Hootsuite
Q: Standards? Mattison: 2 reasons why there are many stds: ontology based upon spec design goals (drug res vs. clin care, others) #PMWC17
12:18pm January 24th 2017 via Hootsuite
Dilks: With Genospace tool, can bring in criteria to filter pts as they come in. #PMWC17
12:17pm January 24th 2017 via Hootsuite
Dilks: HCA does early-phase trials. Molecular requirement is only growing. How to educate? How to bring EMR data in? Genomic data? #PMWC17
12:16pm January 24th 2017 via Hootsuite
Diamond: One pt with 'suitable for hospice' by family doctor, and then looked up clinical trials options... #PMWC17
12:15pm January 24th 2017 via Hootsuite
Urban: And how to explain results back to the family? Next steps for treatment? #PMWC17
12:14pm January 24th 2017 via Hootsuite
Urban: Neonatal: can suspect something, send it out, get back a 'clunky PDF'. Challenge: what beyond first-level testing? #PMWC17
Mattison: Getting to behavioral info - for the individual to be better-informed. #PMWC17
12:13pm January 24th 2017 via Hootsuite
Mattison: The necessity of convergence - interpretable, actionable info is 'huge'. And decision-support Direct to Consumer #PMWC17
12:12pm January 24th 2017 via Hootsuite
Mattison: Similar to Kaiser's problem - physicians complain about unstructured PDF data, which they've solved. But training... #PMWC17
Dilks: Very hard for these physicians to determine what to do next. Level of training, sophistication, determines what they do next #PMWC17
12:11pm January 24th 2017 via Hootsuite
Q: Where are clinicians today? Dilks: Have >500 clin trials, tests are being ordered from community hospitals, figuring it out #PMWC17
12:10pm January 24th 2017 via Hootsuite
Panelists (con't): Amy Urban (UPitt MC) - neonatal med, new technology for UPMC, Joel Diamond (2bPrecise) #PMWC17
12:08pm January 24th 2017 via Hootsuite
Panelists: Holli Dilks (Sarah Cannon Res Inst/HCA) - biomarker strategies; John Mattison (Kaiser) - health IT, integrative care; #PMWC17
12:07pm January 24th 2017 via Hootsuite
Halevy: Showed work they are doing with the NCI/NIH both somatic and germline. #PMWC17
12:05pm January 24th 2017 via Hootsuite
Halevy: Bringing actionable machine-readable object for treatment, at POC (2bPrecise is a division of Allscripts) #PMWC17
12:04pm January 24th 2017 via Hootsuite
Assaf Halevy (2bPrecise) The Last Clinical Mile of Precision Medicine: Challenges at the Point of Care #PMWC17
12:03pm January 24th 2017 via Hootsuite
Komatsoulis: EMR interfaces (about a dozen), 1M pts w/ >800M reported events. '17 goals: reporting, first delivery of 'Discovery db' #PMW
11:48am January 24th 2017 via Hootsuite
Komatsoulis: Implemented in HANA (SAP cloud), showed dashboard. 1.6M pts total, 76 participating institutions; >2K oncologists #PMWC17
11:46am January 24th 2017 via Hootsuite
Komatsoulis: They extract from EMR, to clinical DB, to de-ID analytical DB, human curation to a 'Discovery db' w/partner requests #PMWC17
11:45am January 24th 2017 via Hootsuite
Komatsoulis: IOM, ASCO launched CancerLinQ: to learn from every pt, to harness clinical practice data, to support practicioners #PMWC17
11:42am January 24th 2017 via Hootsuite
Komatsoulis: And 90% of NCI trials are white, vs 23% of the US population is non-white (i.e. 77%) #PMWC17
11:40am January 24th 2017 via Hootsuite
Komatsoulis: Older - 25% are >65% vs 61% of the 'normal' cancer population. 40% of kidney cancer pts too sick for trials #PMWC17
Komatsoulis: 1.7M diagnosed/y w/cancer, 3%, 50K enroll in trials. The 97%: older, less healthy, more diverse. #PMWC17
11:39am January 24th 2017 via Hootsuite
Komatsoulis: Shows example of types and distribution of genes mutated in B-cell, T-cell lineage across adults, youth #PMWC17
11:38am January 24th 2017 via Hootsuite
Komatsoulis: Complex circuitry means multiple genes. 'Cancer is a set of diseases in the molecular circuits that control cells' #PMWC17
11:37am January 24th 2017 via Hootsuite
Komatsoulis: Points to Hanahan Hallmarks figure (Cell 2000), evading apoptosis, angiogenesis; invasion & metastasis. #PMWC17
11:36am January 24th 2017 via Hootsuite
Komatsoulis: Evidence of cancerous tumor in fossil of a 1.7M year old human CNN story https://t.co/I05N5bEFEb #PMWC17
11:35am January 24th 2017 via Hootsuite
George Komatsoulis (CancerLinQ, ASCO) Big Data in Cancer Care #PMWC17
11:33am January 24th 2017 via Hootsuite
Q (con't) Abernethy: A high-trust dataset is returned back, with academic partner's own controls. #PMWC17
Q: How is data protected? Gross: De-identified, minimal amount of info is distributed back to academic partners. Full charts go in #PMWC17
11:31am January 24th 2017 via Hootsuite
Gross: Also survival analysis, if NGS improved outcomes. Lessons: shared governance impt. Master Services Agreement, priorities #PMWC17
11:26am January 24th 2017 via Hootsuite
Gross: NGS: tend to be younger, less minority and wealthier and showed data from their dataset. Next: looking at findings #PMWC17
11:25am January 24th 2017 via Hootsuite
Gross: Pointed to this NYT piece on minority hurdles to access immunotherapy. https://t.co/3U3CDmuzSE #PMWC17
Gross: Data on NSCLC, melanoma, kidney cancers. Preface: doctors typically very slow to adopt new Rx. Showed impact of approval #PMWC17
11:24am January 24th 2017 via Hootsuite
Gross: Two projects - immunoRx and NGS. Looked at 2 PD1/PDL1 agents. Showed adoption - 0.4% to 48% in 2y for 2nd-line Rx #PMWC17
11:23am January 24th 2017 via Hootsuite
Gross: Other guiding principles: using science to guide; questions w/clinical relevance; team emphasizing common goals #PMWC17
11:19am January 24th 2017 via Hootsuite
Gross: Yale sharing data with Flatiron Health first had a guiding principle: putting patients first; it is a stewardship role #PMWC17
11:18am January 24th 2017 via Hootsuite
Cary Gross (Yale SOM) Early insights from an Academic-Industry collaboration #PMWC17 Unusual - in that it is a data-sharing collaboration
11:17am January 24th 2017 via Hootsuite
Abernethy: Able to overlay Overall Survival (OS) compared to TCGA datasets and their similarity. #PMWC17
11:16am January 24th 2017 via Hootsuite
Abernethy: Via FMI datashet with Flatiron - shows comparative data to TCGA dataset, able to follow it across time #PMWC17
11:15am January 24th 2017 via Hootsuite
Abernethy: Shows metrics around quality of their data curation; 1.5M+ records, always on. TCGA ref: https://t.co/ze44Xj0q4s #PMWC17
Abernethy: Essentially, high efficiency chart abstraction, highly parallelized by pathology, radiology, physician notes... #PMWC17
11:10am January 24th 2017 via Hootsuite
Abernethy: And shows a scan of a KRAS report, how it needs a human to boil that down. Interpretation at scale as structured data #PMWC17
11:09am January 24th 2017 via Hootsuite
Abernethy: Shows all the varieties of Albumin data - dozens of codes, descriptions, etc from sources boiled down to 1 record #PMWC17
