Demonstration Project Updates: Identifying and Predicting Patients - - PowerPoint PPT Presentation

demonstration project updates identifying and predicting
SMART_READER_LITE
LIVE PREVIEW

Demonstration Project Updates: Identifying and Predicting Patients - - PowerPoint PPT Presentation

Demonstration Project Updates: Identifying and Predicting Patients with Preventable High Utilization Rainu Kaushal, M.D., M.P.H. Professor and Chair, Healthcare Policy & Research @rainukaushal Rainu Kaushal, MD MPH Has nothing to


slide-1
SLIDE 1

Demonstration Project Updates: Identifying and Predicting Patients with Preventable High Utilization

Rainu Kaushal, M.D., M.P.H.

Professor and Chair, Healthcare Policy & Research @rainukaushal

slide-2
SLIDE 2

2

Rainu Kaushal, MD MPH

  • Has nothing to disclose.
slide-3
SLIDE 3

3

Background

  • Five percent of “high-need high-cost” patients account for a disproportionate 50% of health

care utilization.

  • A considerable proportion of their utilization is preventable with appropriate interventions.
  • We have developed a fluid taxonomy (‘computable phenotypes’ [CPs]) to identify and predict

preventable utilization among high need high cost patients by:

  • analyzing clinical, Medicare and Medicaid claims, and social determinant data;
  • grouping patients with shared characteristics and needs;
  • conceptualizing preventable utilization so that interventions can be targeted;
  • incorporating perspectives of patients, clinicians, and health system leaders.
slide-4
SLIDE 4

4

Results

slide-5
SLIDE 5

5

Manuscripts

3 Published Manuscripts

  • "Precision health" for high-need, high-cost patients (The American

Journal of Managed Care)

  • High-Need, High-Cost Patients Offer Solutions for Improving Their Care

and Reducing Costs (NEJM Catalyst)

  • Drivers of preventable high healthcare utilization: A qualitative study of

patient, physician and health system leader perspectives (Journal of Health Services Research & Policy)

6 manuscripts under review

slide-6
SLIDE 6

6

Qualitative: Stakeholder Engagement for Phase I & II

Aims: To identify drivers of preventable high healthcare utilization from the perspectives of health system leaders, physicians and HNHC patients

  • What factors contribute to preventable high healthcare utilization among HNHC patients in urban and

suburban settings? Analyze stakeholder needs and perceptions about the application of computable phenotypes in a learning health system.

  • Would the computable phenotypes (CPs) be useful in proactively identifying high-need, high-cost

patients?

  • Who (what roles) would use CPs?
  • What would facilitate successful implementation and long-term use of CPs?
  • What barriers or potential unintended consequences are there related to implementation of CPs?
slide-7
SLIDE 7

7

Phase I Participants

  • 1. We completed qualitative interviews with health system leaders from NYC-CDRN,

OneFlorida and CAPriCORN (n=11, 5 health systems across 3 states)

  • 2. We completed focus group discussions with physicians and patients from NYC-CDRN

and OneFlorida (n=21 for each group, 1 health system from each state)

slide-8
SLIDE 8

8

Results – Themes: drivers of preventable high healthcare utilization

  • 1. Unmet behavioral health needs
  • Feeling depressed, inadequate access, stigma
  • 2. Socio-economic determinants of health
  • Social: Poor health literacy, unstable housing conditions, lack of social support
  • Economic: Being uninsured, having Medicaid, burden of copays, meds and

devices

  • 3. Challenges associated with accessing health delivery systems
  • Transportation challenges, long wait times, administrative pressures for

physicians (15-minute visit time)

slide-9
SLIDE 9

9

Results – Exemplar Quotes

Theme Subtheme Quotation Unmet behavioral health needs Serious mental illness Health system leader: “Mental health is obviously huge, and the average primary care or specialist physician has very poor access for their patients to any kind of mental health services in any kind of timely way.” Socio-economic determinants of health Inadequate health literacy Patient: “Learn more about your condition. Get the proper medications, get the proper things that you need to be able to take care of yourself at home.” Unstable housing conditions Patient: “They put you in a bed. They take care of you, take your blood...your vitals laying on the bed with a nice warm pillow and blanket. They even offer you food.” Limited social support Health system leader: “[Workable solutions] probably have to do with...social support and reduction of isolation and helping them negotiate the complexities.”

slide-10
SLIDE 10

10

Results – Exemplar Quotes (Continued)

Theme Subtheme Quotation Socio-economic determinants of health Insurance challenges Patient: “I have Medicaid and some of the doctors don't take [it], and so you say to yourself, well, even if I get an appointment are they going to take me? So just go to the emergency room and let them handle it from there.” Financial burden Patient: “A lot of times [we’re] on a fixed income and you need to see a specialist...You may have a copay with your specialist. It could add up if you go

  • excessively. It’s easier to go to the [ED] and get what you need.”

Challenges associated with accessing health delivery systems Transportation Physician: “I often hear patients voice that they can only come to the ED because that was when they got a ride...by ambulance [and] that’s the only way they can get a ride.” Administrative pressures Physician: “When your hospital is basically saying…here is 15 minutes for a repeat visit for another patient, I mean how are you gonna be able to actually provide the kind of care they need?”

slide-11
SLIDE 11

11

Discussion

  • Need for increased access to and engagement in mental health care.
  • Social programs to improve housing instability, health literacy, transportation and

social support are imperative.

  • Need to incentivize physicians and health systems to accept Medicaid, especially in

light of Medicaid expansion.

  • Need to make medications and med devices more affordable.
  • Revisit 15-minute visit time reimbursement model as it is inadequate for HNHC

patient health needs.

slide-12
SLIDE 12

12

Phase II Participants

Stakeholder group Included roles Florida CDRN New York City CDRN End users (of the computable phenotypes) PCPs (including practice leads), hospitalist, NPs (primary care), care manager (RN), health coach (RN), case manager/social worker 6 9 Informatics CIOs, CMIOs, Chief Analytics Officer 3 4 Operations ACO Director, Chief Transformation Officer, Dir./Ass. Dir. of Population Health, Chief Q&S Officer, Chief/Associate Medical Directors, VP Care Integration, Dir. of Care Management, Dir. of Community Health, Chief Scientific Officer, Clinical Department Chairs 9 18 Total 18 31

slide-13
SLIDE 13

13

Results – Emerging Themes: stakeholder needs and perceptions about the application of computable phenotypes in a learning health system.

Emerging themes:

  • 1. Perception that CPs would be useful for specific user groups
  • 2. Implementation success will depend upon:
  • A. Operational resources
  • B. Credibility and perceptions
  • C. Technical challenges
slide-14
SLIDE 14

14

Results – Operational Resources

Emerging Category Emerging Theme Exemplar Quotes Operational resources The predictive model is irrelevant if there is no infrastructure (process, resources, and band-with) to support necessary actions “We have to face into the reality institutionally of; if we’re going to surface this information, are there going to be resources to actually do something about it?” (Informatics) “Unfortunately, because time is so limited that even if they see the score and even if it is useful… getting anyone to act on it in a meaningful way at least on the provider level might be hard.” (Operational) Arranging people and processes is more challenging than technical aspects of implementation “From a technology point of view, I think it’s almost – very little. I mean, it’s probably the easiest thing. It’s the ‘Who does what with it?’ That’s the big concern to me.” (Informatics)

slide-15
SLIDE 15

15

Results – Credibility and Perceptions

Emerging Category Emerging Theme Exemplar Quotes Credibility and perceptions It is important to differentiate from other predictive models “One has to be very careful to only bring forward new innovations into and organization that have clear value because there are so many companies out there pitching particularly IT-based tools at this point in time. There really is a huge cacophony.” (Operational) Need to overcome skepticism related to predictive models “There’s sort of a baseline level of skepticism … that you would have to overcome and somehow demonstrate…that this is somehow different, better, more valuable…” (Operational) Transparency and comprehensibility are key for trust and use “We’ve seen so many of them and one of the big questions that comes up is like how did you get to this number? What went into it? If there’s detail on what went into it, for sure it’d be a confidence booster.” (End user)

slide-16
SLIDE 16

16

Results – Credibility and Perceptions

Emerging Category Emerging Theme Exemplar Quotes Technical challenges Different sources of data have different cost-benefit tradeoffs “Well, there would be huge problems there because the [claims] data is…as much as three months [behind]. So I think that with that then the scores aren't contemporary anymore.” (Informatics) Local customizations of electronic health record systems “There are a couple ways. There are some where we actually put in it on the banner...there are other places we put…something in the doc flow sheet. And then there are some that are richer where we actually put it on somebody's dashboard.” (Informatics)

slide-17
SLIDE 17

17

Discussion

Strong interest in CPs Keys to success for implementing the CPs:

  • Ensure the proper resources are in place to support action related to the CP
  • High-level information related to what variables are driving the score should be

accessible

  • This multi-site contract offers a unique opportunity to determine if/how CPs and

their implementation can be generalized across systems

slide-18
SLIDE 18

18

Questions?

slide-19
SLIDE 19

19

Thank You!

Rainu Kaushal, M.D., M.P.H.

Professor and Chair, Healthcare Policy & Research @rainukaushal