Precision Health: The Role of the US National Library of Medicine in Broadening the Conversation Patricia Flatley Brennan, RN, PhD, FAAN Director National Library of Medicine
Abstract It’s useful but limiting to think of precision medicine as a conversation between genes and drugs. However, this narrow conversation belies the great advances in human health arising from genomic discoveries and data-driven science. Integrating knowledge from other health practice disciplines, such as nursing, expands Precision Medicine to Precision health, characterizing people by more than their genes and characterization of the person to one not limited to genes and extending nature of precision interventions beyond medicine to health. Grounded with examples from nursing research, this talk will identify the role of the US National Library of Medicine in transforming precision medicine to precision health.
Precision Precis isio ion n engine neerin ring Cardiology Precisio cision Imaging aging Precision Drug Development Precision agriculture Precision oncology Precision drug design Precision nursing Precision stereotactic surgery Precision stomotology Precision health care Precision Health Economics Precision diabetes Precision psychiatry Precision radiotherapy
Precision Therapeutics Targeting Therapies • Case: 64 year old man diagnosed with familial adenomatous polyposis • Sequence to determine the allele location • Evidence-based decision making – Surgery or medicine? – Insurance coverage – Patient preferences and assets APC gene
Precision Screening Optimal Mammography Protocol • Case: 54-year-old woman, perimenopausal with 20 years use of Birth Control Pills; 3 positive 2 nd -degree relations; sedentary; had 13 mammograms • Three tools: – Data mining to find important relationships – Simulation for policy, capacity, population – Optimization: “smartly” choose among 1,000s of pathways
What might a nurse know about these two people ?
Environment Person Organ Cell/Molecular
Precision Health… it’s about broadening the conversation
What do Nurses know that others don’t know? • Human response • Action directed information • Care between the care • Being present in the intimate moments of life
Characterizing the human response • Inflammation – – Cox – systemic inflammatory process mediates chronic critical illness – Fredrickson and colleagues note that our genes are wired for adversity, but eudamonia (and not hedonia) mitigates CRTA • Fatigue – men who experience fatigue demonstrate differential gene expression NINR Study Identifies Genes Involved in Cancer Treatment- related Fatigue US Navy
Action Directed Information • Isoflavones reduce hot flashes and some help mood and sleep disturbance; amino acids don’t help • Information visualization improves understanding; aids people in knowing what to do
The Care Between the Care Go where the care happens.
Care between the Care • Caring for a spouse with end-stage heart failure through implantation of a left ventricular assist device as destination therapy. (Kitko et al 2017) • Frontloading and intensity of skilled home health visits: a state of the science . Oconnor et al 2014
Being present in the intimate movements of life • Advance Care Planning and End-of-Life Decision Making in Dialysis helps caregivers feel less burdensome, more in control • What happens to women in prison at the end of life? • Comprehending patient vernacular improves understanding of symptoms and fear in heart failure patients approaching end of life
Precision Health Management Optimizing Medication Effectiveness • 19 year old college freshman tackling self-management of cancer recovery • What influences medication absorption? • Data-driven investigations – Metabolomics – Metagenomics – Circadian rhythms
What does the NLM do to support transforming Precision Medicine to Precision Health • Enhances Information Delivery • Promotes Access to Research Data • Fosters Common Data Elements • Conducts Research to Build Methods
Enhance information delivery 01 01 010 101 01 110 1011 0110 s t a n d a r d s
Protocols Code Funding Models Clinical Data Library Fostering a Study Data ecosphere of discovery People Instruments digital research objects Pathways
PubMed Literature is the primary repository of knowledge 23
25
Relevance based ranking & Snippets 26
Data Discovery in PubMed Central & PubMed
Data Citations and Supplementary Data in PMC Examples of what you will find in supplementary material: Computer code • Mathematical or computational • models Audio or video clips • Oversized tables • Intervention protocols • Primary or supplementary data • sets Expanded methodology sections • Additional figures • PMC Search: has suppdata[filter] Source: Publication Manual of APA (6th ed.)
Data Links in PubMed: Secondary Source ID Secondary Source ID data sources: • Publishers • NLM indexers • PMC Challenges: • NLM indexing resources were reduced in 2016, leading NLM to explore alternative options • No incentives for journals/publishers to supply Example PubMed Searches: hasdatabanklist these metadata to NLM genbank[si] OR figshare[
Data Links in PubMed: LinkOut Links to the materials directly supporting the research discussed in the cited article, including data sets from experiments/studies accessory graphics, images, sound, and multimedia files related to the article. loprovfigshare[Filter] OR loprovdryaddb[Filter]
Common Data Elements Make data findable, interoperable ◼ Structured human & machine readable definitions of NIH CDEs allowing ◆ Search for individual CDE or sets per programs ◆ Compare & harmonize similar but distinct CDEs ◆ Select or create CDEs with minimal duplication
NINR CDE Project 1. Anxiety 10.SF-36 2. Cognition 11.Self-Efficacy 3. Demographics 12.Self-regulation 4. Depression 13.Sleep Disturbance 5. Diagnosis 14.Pediatric Global Health Assessment 6. Fatigue 15.Pediatric Parent Proxy 7. Global Health Assessment Fatigue 8. Pain 16.Pediatric Parent Proxy 9. Positive Affect & Well-being Global Health 17.Pediatric Short Form Fatigue
Fosters Research
Video
Deep Learning for Cervical Cancer Screening • Leading cause of cancer mortality in low-resource settings • Limited usefulness of visual inspection with acetic acid (VIA), whether in-person or via telemedicine. Goal: apply deep learning to automate diagnosis
Deep Learning for Cervical Cancer Screening Automated visual evaluation (AVE) algorithm • Highly accurate – Identifies prevalent precancer/ cancer (AUC = 0.95) – Predicts incident cases several years in advance • Outperforms human AUC = 0.9540 interpretation • Requires minimal clinical training and cost • R&D ongoing
CRISPR- Cas
New CRISPR-Cas classification Discovered by computational methods and partially validated experimentally Makarova et al, 2015. Nature rev Microbiol. Koonin, Makarova, Zhang, Curr Opin Microbiol 2017 (2016-2018)
Discovery in Clinical Text Guergana Savova, PhD [U54LM008748 R01LM010090] THYME Project thyme.healthnlp.org Goal: Extract temporal relationships from clinical text • Recognizes disorders • Normalizes clinical narrative • Integrates individual episodes into an aggregate patient timeline • Annotates within a document and across documents • Incorporated into Apache via cTAKES Lorem ipsum dolor sit amet, consectetur adipiscing elit. In mollis nisi et tellus aliquam dignissim. Nam semper tincidunt sem, ac ultrices nunc semper at. ctakes.apache.org Nulla quis tincidunt dui. Etiam in tincidunt arcu, at placerat nulla. Donec volutpat auctor faucibus. Nunc a finibus leo, a lacinia purus. Praesent eu nisl orci. Quisque scelerisquecommodo leo in iaculis. Pellentesque lorem felis, ullamcorper ut orci sit amet, maximus fermentum turpis. Interdum et malesuada fames ac ante ipsum primisin faucibus.
Reaching NLM emey87 / IconArchive / CC BY-NC-ND-4.0 @NLM_NEWS patti.brennan@nih.gov @NLMdirector
Transforming Information into Discovery Reach more people in Accelerate discovery Build a workforce and advance health more ways through for data-driven enhanced through data-driven research and health dissemination and research engagement NLMTownHall@mail.nih.gov
Goal 1 1.1 Connect the resources of a digital research enterprise 1.2 Advance research and development in Accelerate discovery and biomedical informatics and data science advance health 1.3 Foster open science policies and practices through data- 1.4 Create a sustainable institutional, driven research physical, and computational infrastructure NLMTownHall@mail.nih.gov
Recommend
More recommend
