Precision medicine: lessons learned

connie dello buono

10 years ago

By Connie Dello Buono , motherhealth@gmail.com

I learned many care practices/protocols/collaborations and advance health IT and medicines in the area of precision medicine during the last conference.

FDA (Health Insurance companies and FDA are not present in the conference)

Regulations around cloud and internet are onerous.
Regulations and health insurance companies not aligned to ensure pharmacogenetic tests is a standard of care for all doctors at a reasonable costs for doctors, labs, health consumers and health insurance companies
Real problem: International, privacy rules are not the same

Collaborations

Collaboration among bio-pharma, cancer organizations and life science centers, hospitals and clinicians in the area of precision medicine is starting to form during the last two years to better understand cancer related health data in clinical space.

Precision medicine must use health data insights with actionable data points.
Plural anecdotes does not equal good data.
From Clinical outcomes to value outcomes , which is not the present scenario.
We have to scale the big data and derive health data insights. How?
There is a gap in health care practice, in balancing costs, tribal fashion of clinical practice ; ~80% softer, inferential, tribal
Curated health data to clinical decision is not translated.
Data warehouse to data mart to data cube for consumer oriented data, curated and reduced processing and search. New, faster data access: score atomically from one metadata ; one single copy of data and apply link to it for easy access.
Complete DNA sequencing allows discovery of new biomarkers based on output DNA data where utilized data ; 95% , shown in 5 continental population, to shown complex disease, 95% of casual variants identified by GWAS are multi-ethnic

Disease phenotypes

Disease phenotypes in different ethnic population ; statistical significance higher in 5 continental population
Strong evidence for clinical associations, to establish new functionalities
22,483 variants known clinical associations

Individual protocol

Pathogenic annotated GSA variant
Annotate/cross reference clinvar
Confirm clinical lab criteria
Based on evidence report

Population

Application areas: sample stratification and QC, early detection and intervention, participant selection for clinical trials, biomarker discovery (drug target ID)
Sample stratification and QC: population of increasing sample size, use SNP variance, epigenetics, gen or physical traits,
Epidemiology: 1 out of 5 with risk variance has kidney failure
Early detection: 70% present in African americans, a protective mutation against __ disease
Intervention: genetic counseling from test results for lifestyle modifications
Wilson Disease: gene ATP7B; high in Japan and Slovenia; body retains too much copper ; deposited in liver; undiagnosed organ failure to death
Warfarin Treatment: based on genetic profile, tailor dose
Biomarker discovery: 7000 SNPs are associated with 400 disease

Precision medicine with select hospitals for cancer patients, targeted based on DNA sequence and drugs

Contradiction of Precision Medicine Drugs

860 billion interactions in a single cell each day, 210 cells types and 78 organs and 100-300 Trillion microbes.

Issues with clinical data:

Randomized controlled clinical trial: too expensive, too long
Existing data sources: very valuable with many data insights; not structured; not specific to a patient cohort;
Registries; traditional site-based; track patients over time; data mining to search for particular study data; very expensive; longer time to collect; more patient-centered;

Data scientists:

Give patients service, manage their medical records, organize and distribute them among their doctors
Structure and normalize patient data
Give patient a timeline of their health information
Together with other clinicians and doctors, provide data insights and actionable health data

precision

Important considerations for a successful precision medicine standard of care

Institutions: Collaboration between hospitals, doctors, labs for complete DNA sequence and bio-pharma allowed the delivery of precision medicine
Doctors: Alignment of doctors with patient and his/her care team in implementation of precision medicine
Legacy system: Providing care teams transparent access to array of precision medicine from legacy systems
Data Science: Allow over-ride in variance matching with whole genome variants and cancer drugs
Clinical operations have access to wide array of precision medicine
Actionable results: stripping raw data to arrive at actionable results from data , using pattern generation
In genomics: See the patient in context.
Data Fidelity: Clinician reviews raw data from feedback loop. Change workflow routine on data collection, use of natural Language processing, SW solutions around data mapping, human curation of data elements
Consolidate databases:
1. No clear annotated database for easy data mining; we reference back to the original raw data; which one to choose; drill down to clinical database evidence
2. Updating DB and having the latest data

Other notes

Confidentiality of data: use an army of lawyers to maintain privacy of all (institution and consumers); finding balance enough liquidity, shareable while still conforming with regulatory environment
Real problem: International, privacy rules are not the same
Regulations around cloud and internet are onerous.
How do you prioritize data? Laser focus on goals , increasing data capture with focus on data elements that will allow to fire a large number of health data measures as quality loop for clinicians
Identify to make positive change is in the standardization of care.
Achievable scaling costs, level of trusts/granularity/specificity

Big Data

N of millions ; characterize as fast as you can and then create your big data asset ; under the complexity, you derive data insights ; Millions of N = 1S
Make informed clinical decision

How to nurture big data enterprise in clinical cancer space?

There is a coalition building to pool resource knowledge to leverage novel cancer insights for cancer care. Institute of Medicine, of the national academes has a system to set into motion to be able to use emerging science, informatics to improve cancer care and reduce suffering.

What if we can bring all electronic health data into one space ; share health data that is not financially injurious but drives science into creation of clinical knowledge

This is true in clinical science where we can see from genotypes, with proper computing power and scientific rigor
There 7 molecular drivers (HER,etc) – drives cancer

For health care provider: Use knowledge base, oncology database, derive what truth is (pain model, care guideline, population at risk)

Only 3% of the population are entered into a clinical trial
Everyday patients tend to be older, less healthy and more diverse
We are treating a more skewed view of the population

Challenge: how do we disseminate knowledge into actionable points and how to make people keep up with exploding knowledge base

Our ability to analyze disease (to sequence genes) outstrips our ability to know what it means

New diagnostics, new care, cancer biomarkers

We need too learn from patients, take data from backend, not to increase the burden, to rationalize information from various sources into a common understanding as to how care is rendered into clinical outcomes into wide variety of care in the US, use of modern computing skills to mine data and get data insights; fearless collaboration; to change incentives on how to mine data insights; reward institutions and clinicians for contributing to a common knowledge source – health data insights