User Tools

Site Tools


groups:healthcare:connecting_health_data_across_silos_for_diagnostics_and_treatment

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision
Previous revision
Last revision Both sides next revision
groups:healthcare:connecting_health_data_across_silos_for_diagnostics_and_treatment [2018/07/17 13:20]
Marissa Iannarone created
groups:healthcare:connecting_health_data_across_silos_for_diagnostics_and_treatment [2018/07/21 21:22]
Marissa Iannarone Added link to the Donor Milk Use Case page
Line 1: Line 1:
 **Section 1 - Problem Statement** **Section 1 - Problem Statement**
  
-To link longitudinal user data identified across silos to help inform diagnostics and treatment - giving ​the individual ownership of their electronic medical records and health data+The Patient Subgroup has submitted an abstract ​to the Hyperledger Global Forum December 12th-15th, 2018:​[[https://​events.linuxfoundation.org/​events/​hyperledger-global-forum-2018/​]]Abstract submitted is as follows:
  
-While aggregated ​healthcare data can be a powerful tool for informing ​algorithms for patient diagnostics and treatmentCurrently data are across various silos in Electronic Medical Records (EMR), laboratories,​ etc., residing with various stakeholders that form a center of trust around the individual. We need a way to connect these data across silos for a specific individual in a de-identified manner, while keeping the individual at the center of all data flow and access+Aggregated ​healthcare data is a powerful tool. Informing ​algorithms for patient diagnostics and treatmentCurrentlydata are across various silos in electronic medical records, laboratories,​ etc., residing with various stakeholders that are disconnected from one another.
  
-When thinking of this individual-owned health data recordwe have identified 2 major use cases to explore (each of these with endless user journey and "sub use cases":​ +With the patient at the centerour product will connect the data across ​silos in a de-identified manner, forming a circle of trust around the patient ​with the patient ​in control.
-* Clinical trials +
-* Sharing ​data across ​healthcare entities +
-  * Emergency room visit how can an individual bring any relevant data with you to an emergency visit +
-  * Acute incident that requires both in-network and out-network provider visits/​interventions+
  
-From here we need to deconstruct these use cases to make the work more actionable (we can't boil the ocean) - journey mapping seems like logical next step - identifying ​the steps that a user would take along the way to get the expected results ​then at each step of the way we can dig into the challenge areas below.+Working with Hyperledger Fabric and Composer, ​we will establish ​network of patients, health care entities, and researchers. Patients will be at the center of their data and researchers will have the ability ​to query the network for de-identified data that will inform diagnostic algorithms and treatment ​at the individual and population-level. As patients move from one doctor to another their data follows securely, where ever they may travel. Solving problems in data access, continuity, security, and medication abuse while lowering costs.
  
-Areas that we'll need to dig into for each "sub use case" include: +**Leveraging Composer Proof of Concept**
-Interoperability +
-Provenance +
-Data standards +
-Data rights +
-* Flow of consent and data+
  
-Initial thoughts on journeys to map: +After reviewing the prototype for Composer found here: [[https://​youtu.be/​cNvOQp8r0xo]] we drew a lot of connections between the entities in their car ordering/​manufacturing ecosystem, and the data sharing ecosystem that we are trying ​to create ​in healthcare. If we view the template ​in the video as the vehicle being the patient ​and the outside interactions being the doctor, hospital, lab, researcher etc., this is a good start.
-* someone signs up to engage ​in our platform ​in order to manage data from and engage in a clinical trail +
-* someone ​is looking to manage data from treatment protocol that requires visits/​interventions with multiple physicians +
-* Opt-in for laboratory sample results to be leveraged to inform patient treatment protocols or diagnosis+
  
-_Please explain ​the business problem ​or opportunity in business terms without technical jargon, and without mention ​of a distributed ledger._+**Use Case Breakdown** 
 + 
 +The MVP use case with this functionality could be with 4 players: 
 +  * Patient 
 +  * Healthcare entity A (hospital, lab, etc) 
 +  * Healthcare entity B 
 +  * Researcher 
 + 
 +If we make the assumption that both healthcare entities use the same data standards (FHIR or other)this is how the work flow could be laid out: 
 + 
 +  - Patient signs up to engage with our product via a IOS app and agrees for pointers to their health data to live on the blockchain 
 +  - Healthcare entity A and B place pointers to all of the patients health data (ideally both retrospectively and forward moving) 
 +  - Researcher wanting to leverage the data (for all patients - not just one) to help inform research or diagnostic algorithm could query the pointers for standard variables that would pull de-identified health data into data set to use 
 +  - Researcher takes data out of the product to do analysis, etc. 
 + 
 +Initial analysis of the Composer code and identified mappings for names and categories:​ 
 + 
 +**Participants:​** 
 +  * Company--->​Doctor 
 +  * Manufacturer>​Hospital 
 +  * Auction house----->​Researcher 
 +  * Regulator----->​Insurance 
 +  * Private owner----->??​data private owner. Same as patient 
 +  * Scrap merchant----->??​deceased data base 
 + 
 +**Assets:​** 
 +  * Order----->​Lab/​ procedure order 
 +  * Vehicle----->​Patient 
 + 
 +**Transactions:​** 
 +  * Place order----->​ Order Lab/ procedure 
 +  * UpdateOrderStatus------->​ Update Lab/​procedure status 
 +  * Application for vehicle registration certificate----->​Insurance application(healthcare) 
 +  * Private Vehicle transfer----->​Patient transfer to new doctor 
 +  * Scrap vehicle----->​patient deceased 
 +  * Update suspicious----->​ ? Fraud or medication abuse  
 +  * Scrap all vehicles by color------>​group patients by disease or category 
 + 
 +**Events:​** 
 +  * Place order event----->​ order lab/ procedure event 
 +  * Update order status event------>​updates lab/ procedure event 
 +  * Scrap vehicle event------>​ deceased 
 + 
 +We've confirmed that we will be working on a open source build. Tony Little suggested looking at the Davinci project from FHIR that that provides an implementation framework for others to follow: [[http://​www.hl7.org/​about/​davinci/​index.cfm?​ref=common]] 
 + 
 +We will also want to work with the Hyperledger ID team to understand better what they have thought of/built already. Once we have a journey map completed, we can loop them in and walk through it with them to gain insights
  
 **Current Solution** **Current Solution**
Line 31: Line 64:
 https://​doc.ai/​product - has been highlighted,​ but doesn'​t address data standardization and GitHub repo is here: https://​github.com/​HD2i/​biomedical-blockchain#​doc.ai https://​doc.ai/​product - has been highlighted,​ but doesn'​t address data standardization and GitHub repo is here: https://​github.com/​HD2i/​biomedical-blockchain#​doc.ai
  
-We will also want to work with the Hyperledger ID team to understand better what they have thought of/built alreadyOnce we have journey map completed, we can loop them in and walk through ​it with them to gain insights+Take a look at how UK NHS is solving this same problem: https://​www.engadget.com/​2018/​06/​29/​britain-nhs-anonymous-health-data-privitar/​ 
 + 
 +We have reviewed [[https://​medicalchain.com/​en/​|Medicalchain]] (Hyperledger project) and believe that they are targeting a very similar solution to what is laid out below. After reviewing the white paper, we do see possible opportunities for a POC to be built that would support Medicalchain work in the following areas: 
 +  * AI - machine learning could be inserted into the network - Anton recommended looking at [[https://​ada.com/​|ADA]] 
 +  * Chain of custody (which appears to be a gap according to the white paper 
 +  * Researchers able to engage with patient data via the '​market place' component laid out in the white paper 
 + 
 +We are waiting to hear back from Medicalchain on the following questions ​and are hopeful that a representative will be on the 7/27 HCWG call so that we can determine whether ​it makes sense to move forward ​with the use case stated above. 
 +  * If Medicalchain has implemented their code onto the main block chain at this time?  
 +  * Where their struggles, if any, may lie in implementation of the project.(This may be a question for them directly.)
  
 _If there are systems in place today which automate the above business problem/​opportunity,​ please explain what exists._ _If there are systems in place today which automate the above business problem/​opportunity,​ please explain what exists._
Line 56: Line 98:
  
 _Provide a list of acronyms, their expansions, and what they actually mean in general language here. Define any terms that are specific to your problem domain. If there are devices, appliances, or software stacks that you expect to interact with Hyperledger,​ list them here._ _Provide a list of acronyms, their expansions, and what they actually mean in general language here. Define any terms that are specific to your problem domain. If there are devices, appliances, or software stacks that you expect to interact with Hyperledger,​ list them here._
 +
 +**Section 5 - Additional Thoughts/​Notes**
 +
 +When thinking of this individual-owned health data record, we have identified 2 major use cases to explore (each of these with endless user journey and "sub use cases":​
 +  * Clinical trials
 +  * Sharing data across healthcare entities: Emergency room visit - how can an individual bring any relevant data with you to an emergency visit; Acute incident that requires both in-network and out-network provider visits/​interventions  ​
 +
 +From here we need to deconstruct these use cases to make the work more actionable (we can't boil the ocean) - journey mapping seems like a logical next step - identifying the steps that a user would take along the way to get the expected results - then at each step of the way we can dig into the challenge areas below.
 +
 +Areas that we'll need to dig into for each "sub use case" include:
 +* Interoperability
 +* Provenance
 +* Data standards
 +* Data rights
 +* Flow of consent and data
 +
 +Initial thoughts on journeys to map:
 +
 +* someone signs up to engage in our platform in order to manage data from and engage in a clinical trail
 +* someone is looking to manage data from a treatment protocol that requires visits/​interventions with multiple physicians
 +* Opt-in for laboratory sample results to be leveraged to inform patient treatment protocols or diagnosis
 +
 +Patient Centered Data mapping provided by Tony Hussain on 7/6: {{ :​groups:​healthcare:​screen_shot_2018-07-06_at_12.17.54_pm.png?​200 |}}
 +
 +**Other Use Cases Being Considered by Patient/​Member Subgroup**
 +[[groups:​healthcare:​donor_milk_transparency_and_traceability|Donor Milk Transparency and Traceability]]
 +
groups/healthcare/connecting_health_data_across_silos_for_diagnostics_and_treatment.txt · Last modified: 2018/07/24 23:53 by Marissa Iannarone