LSU Health Services Research Program (LSU HSRP) (Text Version)

On September 14, 2009, Ron Horswell made this presentation at the 2009 Annual Conference. Select to access the PowerPoint® presentation (720 KB).

---

Slide 1

LSU Health Services Research Program (LSU HSRP)

Ron Horswell, PhD

Slide 2

LSU HSRP

• Mission: Accelerating translation of evidence into practice, emphasizing topics of importance to the LSU Health Care Services Division (LSU HCSD), a provider system serving low-income, predominantly uninsured patients.
• Funded by AHRQ from 2001 to 2006.
• Original P.I. was Fred Cerise, MD.

Slide 3

LSU HCSD System

An image of the state of Louisiana is shown. On the map, the LSU HCSD System is marked.

Slide 4

LSU HCSD Disease Management and Population Health Programs

An image of the LSU HCSD Disease Management and Population Health Programs is shown.

Slide 5

LSU HSRP

• Methods and Program Elements (initially):
• Recruit (largely clinical faculty) participants.
• Pair participants with research mentors.
• Core Program Office.
• Works in Progress meetings.
• Affiliated with the Harvard Pediatric Health Services Research Program
  • Sharon Muret-Wagstaff, PhD.
  • Donald Goldman, MD.

Slide 6

LSU HSRP

• Results (by Sept 2006):
• 25 projects.
• $680,000 in additional funding.
• Published 11 papers, probably approximately 15 by now, with some still in progress.

Slide 7

LSU HSRP

• Core Program Office:
• Most successful of the original program elements.
• Shannon McNabb, MA MPH.
• Managed IRB relationships.
• Managed the Works in Progress meeting process.
• Participated in research oversight committees at HCSD sites.
• Helped instigate much higher LSU HCSD participation in "external" research.

Slide 8

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 9

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 10

LSU HSRP themes: Disparities and Underserved Patient Groups

• Conclusions:
  • "Self pay" does worst.
  • Small racial disparities on processes.
  • More disparities on outcomes.
  • Variation across clinics within sites is greater than variation across sites.
  • Focus on patient subgroups with persistent problems.
  • Need to get patients to "in-care" status and keep them there.

Slide 11

Eye Exam Within Past Year

An image of a graph that shows GenRace, Insured, Site, and SiteClinic versus Fraction of Unique Patients is shown on the screen.

Slide 12

Eye Exam Within Past Year

• Pc sustain: .5
• Pc new: .4
• Pc 1 year: .3
• Pc 2 year: .15
• Pc none: .12

Slide 13

Fraction with HbA1c under 7%

An image of a graph that shows GenRace, Insured, Site, and SiteClinic versus Fraction of Unique Patients is shown on the screen.

Slide 14

Fraction with HbA1c under 7%

• Pc sustain: .5
• Pc new: .45
• Pc 1 year: .40
• Pc 2 year: .33
• Pc none: .31

Slide 15

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 16

CONCENTRATION ANALYSIS FOR GENERAL EVENTS
Events Saved Per Patient Treated

An image of graph that shows Accumulative Average Event per Patient Vs. Fractiles by Descending Predicted Prob. is shown on the screen.

Average event per patient = 413/1883 = .22

Slide 17

HbA1c over time

A graph of "HbA1c over time" is shown on the screen.

Slide 18

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 19

Relative cost effectiveness a function of:

• Cost of BNP.
• Cost of Echocardiography.
• Prevalence of undiagnosed HF.
• Sensitivity of status quo.
• Specificity of status quo.

A graph of BNP Sensitivity Vs. BNP Specificity is shown on the screen.

Slide 20

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 21

HEDIS® Measure BCS

NOTE: Graph shows only providers with 10+ recipients.
Symbol size proportional to # of recipients.
Green line (40%) is State Average.

Slide 22

HEDIS® Measure BCS

BCS - Breast Cancer Screening measures the percentage of women 40 - 69 years of age who had a mammogram to screen for breast cancer.

Provider Group	Denominator	Numerator	HEDIS
FQHC	6,439	2,603	40.40%
LSU	12,709	6,790	53.40%
RHC	7,981	3,100	38.80%
Other	61,910	23,179	37.40%
Total	89,039	35,672	40.06%

Slide 23

HEDIS® Measure CDC - Eye Exam

NOTE: Graph shows only providers with 10+ recipients.
Symbol size proportional to # of recipients.
Green line (38%) is State Average.

Slide 24

HEDIS® Measure CDC - Eye Exam

CDC - Comprehensive Diabetes Care is a set of measures for members 18-75 years of age with type 1 or type 2 diabetes. The measures include: eye exam, HbA1c, LDL-C and screening for nephropathy. Each measure is calculated individually. Presented in this chart is the CDC-Eye Exam measure.

Provider Group	Denominator	Numerator	HEDIS
FQHC	3,388	1,198	35.40%
LSU	7,296	4,333	59.40%
RHC	3,894	1,308	33.60%
Other	32,197	10,418	32.40%
Total	46,775	17,257	36.89%

Slide 25

Selected Projects

• DIABETES: Evaluation of a telemedicine diabetes foot program.
• DIABETES: Barriers to diabetic eye care.
• DIABETES: Effect of indigent pharmacy medication program.
• HF: Mortality benefit of a HF disease management program.
• HF patient registry development.
• HF: Cost effectiveness of using BNP as a screener for heart failure.
• HF: Heart failure acute event prediction model.
• TOBACCO: Survey assessing patterns of tobacco use among HCSD patients.
• HTN: Hypertension management using group visits.
• NEONATES: IT network supporting preventive services for at-risk infants.
• DISPARITIES: Racial disparities and qualify of care in disease management programs.
• DISPARITIES: Decomposition of sources of variation in disease management performance measures.
• CANCER: Comparison of mammography screening levels across three population (Medicare, Medicaid, and the Uninsured.)
• CANCER: Relationship of screening mammography to tumor stage at diagnosis.
• CANCER: Relationship of tumor stage at diagnosis to survival.

Slide 26

Distribution of Stage at Diagnosis

Insitu

• HCSD Patients: 14
• Fly-in: 7

Localized

• HCSD Patients: 49
• Fly-in: 40

Regional

• HCSD Patients: 35
• Fly-in: 42

Distant

• HCSD Patients: 4
• Fly-in: 11

Slide 27

Survivor functions

A graph of Survivor functions (Unscreened Probability) adjusted for age=55 and race=Afm is shown on the screen.

Slide 28

Screening and Tumor Stage

	2-year Screening history	In Situ	Localize	Regional	Distant
In Situ cases included	screened	0.144	0.470	0.353	0.033
	unscreened	0.081	0.465	0.386	0.068
	Risk ratio (screened/unscreened)	1.78	1.01	0.91	0.49
In Situ cases excluded	screened		0.549	0.412	0.039
	unscreened		0.506	0.420	0.074
	Risk ratio (screened/unscreened)		1.08	0.98	0.53

Slide 29

Survivor functions

A graph of Survivor functions (Probability of Survival) adjusted for age=55 and race=Afm is shown on the screen.

Slide 30

LSU HSRP

• Our recommendations to the LSU HCSD when the LSU HSRP program concluded:
• Maintain Core Office (partially done).
• Abandon mentoring concept (done).
• Create an analysis department (partially done).
• Become programmatic and purposeful (not done).

Slide 31

LSU HSRP

• Observation:
• "…A distinction exists between:
  • A research program designed to mentor individual researchers and move them along their career paths, and
  • A research program designed to address major needs within client organizations' patient populations."

Slide 32

Purposeful and Programmatic

Area	QI	Retrospective	Prospective	Summary
Diabetes	(1) Accu-Chek	(1) Chiou, (2) pharmacy paper(3) eye camera	(1) Depression & diabetes	(1) HbA1c trends
CHF	(1) CHF registry	(1) predict admit, (2) predict mortality, (3) IV lasix	None	(1)Beta Blocker trend (2) IP and ER trends
Asthma	None	(1) EKL asthma intercept program	None	(1) IP and ER trends
HIV	None	None	None	None
Cancer		(1) screening and stage, (2) stage and survival	None	(1) mammography rate history (2) comparison to Medicaid
Tobacco		(1) Baseline patient survey	None	None
General		(1) population model (2) capacity project (3) sources of variation/disparities project		(1) disparities summary

Slide 33

LSU HSRP

• Our recommendations to the LSU HCSD when the LSU HSRP program concluded:
• Maintain Core Office.
• Abandon mentoring concept.
• Create an analysis department.
• Become programmatic and purposeful.
• Maintain chronic disease patient registries.
• Systematic analysis to identify sources of variation (potential opportunities).
• Create a practice-based research network.
• Identify patient subgroups with persistent problems and develop interventions.
• Develop and test means of patient empowerment.

Slide 34

Practice-based Research Network

• We suggested:
• LSU HSCD "medical home"-like clinics as participants.
• Both QIPs and research projects.
• Would provide the platform for mutually beneficial collaboration with other organizations and with various faculty.
• This would enhance funding prospects.
• Flow:
  • Idea.
  • Retrospective.
  • Pilot.
  • Larger project.
  • Funding.

Slide 35

Reasons for LSU PBRN

• Tele-monitoring projects
  • HF patients with history of ED/IP events.
  • Diabetes patients with persistently high HbA1c levels.
• Medical home development.

Slide 36

Medical Home Patients

An image of a flowchart of "Medical Home Patients" is shown on the screen.

Slide 37

Medical Home Patients

An image of a revised flowchart of "Medical Home Patients" is shown on the screen.

Slide 38

Reasons for LSU PBRN

• Tele-monitoring projects
  • HF patients with history of ED/IP events.
  • Diabetes patients with persistently high HbA1c levels.
• Medical home development.
• Need to link quality improvement and access improvement to business model.

Slide 39

LSU HCSD Screening Colonoscopy Capacity Requirements

Simulation Results: Long-run Needed Capacity for combinations of Assumptions
	Policy: Screen Every 5 Years		Policy: Screen Every 10 Years
% of incoming patients already screened	Screen Uninsured only	Screen All	Screen Uninsured only	Screen All
0%	8,223 (98)	15,463 (184)	7,261 (86)	14,340 (171)
25%	7,173 (85)	13,397 (159)	6,075 (72)	11,538 (137)
50%	6,120 (73)	11,324 (135)	4,679 (56)	8,918 (106)
75%	5,125 (61)	9,123 (109)	3,376 (40)	6,322 (75)
100%	4,100 (49)	7,043 (84)	2,096 (25)	3,632 (43)

Slide 40

END