Utah Health Data Committee (HDC) Meeting
Minutes
Tuesday, May 9th, 2017
3:03 pm- Meeting started by Lynette Hansen after establishing quorum in person and via phone.
The meeting took place between 3:03 and 5:05 pm in Room 125 at the Utah Department of Health 288 N 1460 W, Salt Lake City.
Chair Jim VanDerslice was unable to attend the meeting. Lynette Hansen, chair of the Payer Task Force (PTF), conducted the meeting.
Attendees:
Lynette Hansen
Lynda Jeppesen
Hinkley Jones Sanpei
Jim Murray (phone)
Mark Bair (phone)
Alan Ormsby (phone)
Vaughn Holbrook (phone)
Steve Neeleman (phone)
Tanji Northrup (phone)
Excused:
Jim VanDerslice
Sharon Donnelly
Absent:
Michael Hale
Kevin Potts
David Purinton
Guests:
John Barlow
Kimberly M. (HealthInsight)
Aaron N. (Health Catalyst)
1. Approval of March 2017 Meeting Minuets
Lynette Hansen asked if any changes needed to be made to the minutes.
Motion to adopt meeting minutes: Alan Ormsby
Second: Vaughn Holbrook
None Opposed
2. Subcommittee Update and Review
Best Use Task Force
Sharon Donnelly, chair of the Best Use Task Force, was unable to attend the meeting. Norman Thurston, OHCS Director, gave an update from the Best Use Task Force Meeting held on April 12, 2017 on Strategic Prioritization (reviewed handout).
Key Highlights:
Expected Outcome: As a result of strategic prioritization, more of our products and services will be worth purchasing.
· Section 1- How should HDC determine strategic priorities for data use? How can we maximize the value of our data products and services?
1. Demonstrated value
2. Potential impact
3. Sustainability
4. Data Privacy
5. Transparency
· Section 2- Based on the criteria outlined in Section 1, what are some of the most interesting unanswered questions that our data could inform?
1. Opioids – data could be used to help identify the root cause of opioid abuse.
2. Hip & knee replacements- data could be used to help identify areas with higher and lower costs and possible areas with high variability.
3. Vaccinations- data could be used to help identify the potential impact of vaccination or lack thereof on costs for specific areas.
4. Care management for newly enrolled- could use claims data to help identify patterns and needs and trigger early and more intensive interventions, leading to improved care and reduced costs.
5. Aggregation Efficiencies- calculate HEDIS measured at clinic/provider level rather than calculating by health plan.
6. Unwarranted variations- data mining techniques could be used to identify areas where further investigation is needed.
7. Air and water quality- link facilities data (ER encounters and inpatient admissions) with death records and air quality data.
8. Understanding Social Determinants of Health (SDOH)
· Question: Could patient Never or Sentinel Events be captured by our data?
Answer: Hospital inpatient encounter data may have those types of events.
· Question: What about deaths that occur in the operating room, and similar events?
Answer: In-patient data has been used the past for patient’s safety.
· Add an additional item- (#9) Sentinel Events.
· Lynda Jeppsen suggested- Obesity and Orthopedic procedure correlations, as a possible subcategory for item #2.
· Going forward:
o Prioritize- Find the priority. Which item/items from the list do we want to address?
o Partnerships – Who is already working on the item/items? Collaborate with them. If no one is currently working on the items, reach out to possible groups who would be interested in working in collaboration with us.
· Norman asked- Who would be interested in #7- Air Quality, linking facilities data with death records and air quality data? Our data goes back to 1996.
o Lynette suggested- Asthma Task Force and Department of Economic Development, would most likely be interested in this data.
o A lot of people are interested in results of this study. Who will conduct study?
Facilities Task Force
Charles Hawley with the Office of Health Care Statistics gave an update on the Facilities Task Force.
Key Highlights:
Possible changes to Data Submission Guide (DSG) guideline:
· Agreement to give facilities at least 10-12 months notification of changes.
· Contracted with Mercer to do processing- attempt to streamline process
o There is still fair amount of variation. Data doesn’t come in as standard as we would like.
· 1 Single File format for Inpatient, ED, AMB. T
o This would make it easier for smaller facilities
o Include all elements on a UB-04 form that are not currently included- SSN, race, ethnicity, marital status.
§ Question: Why use/collect SSN.
§ Answer: Mainly for linkage. SSN is encrypted, numbers are scrambled- turn the SSN into another number.
o The facilities agreed to continue communication via email
o We will continue to work with the facilities over summer and bring final version to HDC September meeting.
o We want make the DSG changes:
1. Easy for small hospitals and small surgical center.
2. More valuable for the larger systems.
PBM CAHPS Survey
· Received the results. OHCS is not sure if it’s worth going forward, continuing PBM next year. Less than 9% survey response rates. By next September we need to decide if we will continue PBM CAHPS.
Data Use Subcommittee (DUS)
Jim VanDerslice is the chair of the DUS. Chair VanDerslice was excused from the meeting. Norman Thurston, OHCS Director conducted the discussion for three Data Use requests.
Key Highlights:
Motion to allow John Barlow to speak: Norman Thurston
Seconded: Hinkley Jones Sanpei
None opposed
Approved
Data Requests
· Data Requests 1 & 2 are from the University of Utah, regarding requisition cost data- sudo identifier for facility data.
· John Barlow- Director of RGE, from the University of Utah, working with Dr. Mangles. Norman Thurston stated- we have an agreement with the University. They have our data on their systems and University researchers can request the data, but any request that goes to RGE need to go through the Data Use Subcommittee for approval.
1. University of Utah- Dr. Mangle
· Dr. Mangles study is only using the APCD.
· The primary concern the Data Use Subcommittee has with this request is that the facilities could possibly be re-identified.
· Question to HDC- Under what circumstances will data that has the possibility of Facilities being identified, be available for studies?
· Norman Thurston stated- Hospitals are not entitled under any laws to not be identified. Our office has been always been cautious when it comes to possible re identification of any data.
o It comes down to trust; do we trust the researchers will keep the agreement and not identify the facilities?
· Lynette Hansen asked- What questions will access to the Cost Data answer?
Answer: The study is on Continuity of Care in Physical Therapy- Provider and System Continuity of care, focusing on muscular skeletal conditions.
o Literature shows, when a patient sees more than one provider for an episode of care, their cost of care increases.
o If a patient is seen by only one physical therapist vs. being seen by more than one physical therapist, what is the effect/downstream initialization and cost?
o System Continuity- if a patient is seen by a provider and receives treatment outside of health care system, what is the effect/ downstream unitization and cost?
· Individual Healthcare Systems (facilities) would be reported by aggregate. They never would identify a healthy care system (facility) by name ever.
· Dr. Bair stated- The request can be approved, as long as they meet the Data Subcommittee’s requirement of a compliance review prior to publication.
Motion to approve: Dr. Mark Bair
None opposed
Approved
2. University of Utah- Dr. Mac
· This request is very similar to request #1. Hip Arthroplasty - use data to compare condition costs associated with facilities.
· Noman Thurston stated-Data can be used to compare hospitals and prices, but the hospitals cannot be identified (not allowed to use hospital names).
· Norman Thurston and Charles Hawley stated- there was no major concern with this request.
Motion to approve: Steve Neeleman
Seconded: Lynda Jeppsen
None opposed
Approved
· Jim Murray asked: Is the entire data base being accessed?
Norman Thurston: No, access will only be provided for fields that have a justified request.
3. Health Catalysts Data Request
· Charles Hawley stated, Health Catalyst is building a tool- CAFÉ is their data base.
· They want to use the APCD to pool values and compare to expected values.
· They are requesting limited data set access. Variation request- they don’t need Geography but want Service Dates.
· Norman Thurston stated, providing Geography poses the same (equal) risk factor as providing Service Dates.
· Health Catalyst will use the All Payer Claims Database (APCD) to link Service Dates across patients, so they can have a history of dates.
· Norman Thurston stated, this accented by the Data Use Agreement and Contract.
· Charles Hawley stated, Service Date is not considered PHI Data as defined by HIPPA.
General RULE: Swap out Date of Service Data for Geography Data for a Limited Data Set.
Motion to approve: Alan Ormsby
Seconded: Lynda Jeppsen
None opposed
Approved
· Health Catalyst is requesting Provider ID (NPI)
· Norman Thurston stated, the NPI identifies provider of care.
· Question: Why does Health Catalyst need NPI?
Charles Hawley stated, for a Benchmark Health System you have to identify the provider.
· Aaron N. (from Health Catalyst) said they are only interested in the hospital NPI.
· Jim Murray asked- What is end goal of project?
Answer: To create a Health Care system where clients can go to Health Catalyst and compare themselves to the State wide average.
· Arron N. (Health Catalyst) stated, the unique value of this project is that it is a Claims Value Dataset.
o Claims Data provides information that EMR Data cannot and helps provide a better understanding of utilization rates.
o Without NPI analysis they are unable to apply attribution model or benchmarks.
o Only requesting organizational level NPI.
· Contract and overview would include that Health Catalyst is not in violation of contract.
· Jim Murray stated; this is the first time a monitored product has been created using our APCD. What is the return to the State and APCD?
Norman Thurston replied, they are allowed to use the data for their clients.
o We have a flat fee. The Data Fee for Health Catalyst is $15k.
o They are asking for 2 years of data.
o We would receive $30k for 2 years of data.
o Upcoming years may have a profit share requirement in contrast.
o This model has been applied with Facilities Data.
· Jim Murray asked, what publication rules are in place for secondary uses?
Norman Thurston replied, Health Catalyst would have to come back to the State with their clients is request for publication. The State can then approve or deny the request.
· Release of Facility NPI needs to go back to Data Use Subcommittee and release of NPI needs to be determined on a case by case basis.
Lynette Hansen proposed the following monition- Approve release of Facility NPI in this case, as long as the request goes through Data Use Subcommittee review. The Data Use Subcommittee may approve.
Motion to approve: Lynette Hansen
None opposed
Approved
3. DMBA
Norman Thurston, OHCS Director gave a quick update on the possible DMBA 2017 APCD opt-out.
Key Highlights:
Norman Thurston stated- We think we have a solution to DMBA contribution to APCD. More information is to come.
4. The Hitachi Research Question
Norman Thurston, OHCS Director and John Barlow from the University of Utah conducted the discussion on the Hitachi Research question.
Key highlights:
Norman Thurston reviewed the APCD Data Request Executive Summary.
Overall objective: To improve health outcomes for individuals with diabetes and high blood pressure through the use of advanced data science. Somewhat like data mining
· Dr. Kawamoto and his team want to use the data to help build a tool- predictive model. Data scientists want to build mathematical modules to see what medications would be the most effective. Module will be licensed and sold.
o Develop predictive model.
o License predictive model that will be added to Electronic Health Records (EHR).
· John Barlow stated; Dr. Kawamoto would be the only one with access to certain data. Workers would be required to work on the project only on a University of Utah health server. Any legal issues will be resolved in the state of Utah.
· Timeline is dependent on University of Utah receiving All Payer Claims Data.
· How does this benefit the State and community of Utah?
o Supportivecreating a community predictive model that improves overall health of State.
o Commercialization- a new commercialization tool.
· Once the model is built the data is irrelevant.
o Once the model is built, the APCD data can be deleted and the model would still work.
o Dr. Kawamoto is not planning on updating or maintaining the model.
· Jim Murray asked, can you release results of analysis?
Dr. Kawamoto is planning to publish and wants to submit to a conference.
· Jim Murray suggested, this to go back to Data Use Subcommittee- providing everything in APDC- with existing patients/clients.
Note: This had been through IRB review and defined as Strictly Research
· Each time the model is updated it needs to come back for review.
Motion to approve Kawamoto Application Model: Lynda Jeppsen
Second: Lynette Hansen
Opposed: Jim Murray
Abstained: Hinkley Jones Sanpei
Approved
5. Clinic Comparison Methods and Results
Charles Hawley, with the Office of Health Care Statistics, gave an update on Clinic Comparison Methods and reviewed the results.
Key Highlights:
Charles Hawley reviewed draft quality measurers (see handout).
· Re-Occurring Measures
o Comprehensive Diabetes Care: HbA1c Screening
§ Twice as many diabetics were attributed even though some clinics were dropped from last year.
o Avoidance of Antibiotic Treatment for Adults with Acute Bronchitis(AAB)