Diabetes Disease Management Results in Hispanic Medicaid Patients

Esteban R. López, MD, MBA, FAAPProgram Director and Medical Director, McKesson Health SolutionsNational Hispanic Medical AssociationMarch 20, 2011Washington, DC

Objective of Presentation

• To present medical service utilization from a telephonic nursing disease-management program for Medicaid patients with diabetes residing in Puerto Rico

• Published in Journal of Health Care for the Poor and Underservedo May 2009

Road Map of Presentation

• Diabetes Disease Management• Methods• Results• Discussion

Disease Management

• Care Continuum Alliance (CCA) disease management definition:o Supports the physician/practitioner patient relationship

and plan of careo Emphasizes prevention of exacerbation and

complications utilizing evidence-based practice guidelines and patient empowerment strategies; and

o Evaluates clinical, humanistic, and economic outcomes on an on-going basis with the goal of improving overall health

Components of Disease Management

• Disease management components include:o Population identification processo Evidence-based practice guidelineso Collaborative practice models (physicians and others)o Patient self management educationo Process and outcomes measurement, evaluation, and

managemento Routine reporting/feedback loop

Diabetes Disease Management

Previous research shows that an increased number of patients can be managed in an ambulatory care setting through• Increased understanding of pathophysiology of

diabetes• Pharmacological interventions• Non-pharmacological intervention

Goals are:• Increased quality of life and less expensive health care

 

Diabetes Disease Management

Centers for Medicare and Medicaid Services (CMS)• Recognize burden of chronic diseases• Will pilot differing disease management strategies• Previous research finds a multidisciplinary approach

has increased QOL and reduce overall medical costs through

• Disease management nurses• Frequent physician office visits• Telephone contact systems

Diabetes Disease Management

Limitations of Previous Research• Pre/Post Evaluations

o Least rigorous study methodology• Clinic-based interventions

o Not representative of community of diabetes patients• Costs are not clearly delineated

This Research• Uses a more rigorous study methodology• More representative of community of diabetes patients

Methods: Study Population

Diagnosed with Diabetes through administrative claims

Medicaid plan in Puerto Rico

Age 1-64

Methods: Study Population

Excluded people:• Those engaged in a local formal diabetes program.• Members age 65 or over• Members with less then three months eligibility prior to their study

start date or less than three months eligibility after their study start date.

• ESRD, Dialysis, Transplants, HIV/AIDS• Hospice • SNF• Intervention group members with less than three months

participation in the disease management program.

 

Methods: Study Population

Sample of 490 diabetes participants and 490 matched non-participants

• Matched non-participants drawn from sample of 7,966

Methods: Intervention

Created a customized self-management intervention plan o Risk stratificationo Formal scheduled nurse education sessionso 24 hour access to nurse counseling and symptom adviceo Printed action planso Workbookso Individualized assessment letterso Medication compliance reminders and vaccine reminderso Physician alerts

Methods: Intervention

Guidelines used:• The American Diabetes Association

Methods: Intervention

Changes in medical service utilization is expected to result from improvements in patients’

Knowledge Behavior Health status

Methods: Study Design

Alternatives include• Randomized control trial• Matched two-group cohort• Population based pre/post• Participant only pre/post• Others

We used a 12 month, matched-cohort study.

Results

Medical service utilization (annualized rate per 1000)

Study group

Control group P-value

Change (%)

Inpatient admits 174 268.4 0.112 -35.2

Inpatient bed days 920.3 1,770.00 0.021 -48

Emergency Department visits 773.6 758.3 0.778 2Physician evaluation & management visit 5,153 4,651.80 0.649 10.8

Pharmacy scripts 39,530.40 40,932.90 0.704 -3.4

Diabetes inpatient admits 39.8 14.90 0.437 167.2

Diabetes inpatient bed days 148.8 108.60 0.699 37Diabetes Emergency Department visits 81.8 95.80 0.603 -14.7

Cardiac inpatient admits 25.2 98.00 0.001 -74.3

Cardiac inpatient bed days 134.2 528.20 0.002 -74.6

Cardiac Emergency Department visits 16.8 12.80 0.591 31.2

Inpatient 30 day readmits 29.4 42.60 0.635 -31.1

Results

Prescription drugs (% of people who have)

Study group

Control group P-value

Change (%)

ACE inhibitor (%) 31.6 25.7 0.041 23

Beta blocker (%) 27.6 25.7 0.516 7.1

Antihypertensives (%) 54.9 49.8 0.11 10.2

Diuretics (%) 45.3 36.1 0.004 25.4

Cardiac glycosides (%) 5.1 6.3 0.409 -19.4

Blood glucose regulators (%) 90.4 90.4 1,000 0

Results

Procedures performed (% of people who have)

Study group

Control group P-value

Change (%)

Hemoglobin A1c 21.2 16.5 0.061 28.4

Lipid panel 28 23.7 0.126 18.1

Eye examination 16.3 13.9 0.285 17.6

Maculopathy 3.9 3.5 0.734 11.8

Microalbumin 1.4 1.2 0.78 16.7

Echocardiography 4.9 7.6 0.086 -35.1

Cardiac catheterization 1.2 5.7 0 -78.6

Myocardial imaging/ perfusion 1.8 2 0.817 -7.1

Influenza immunization 7.1 2.4 0.001 191.7

Pneumococcal immunization 2.9 1 0.037 180

Results

Average costsStudy group

Control group

P-value

Change (%)

Monthly medical costs ($) 74.5 154.66 0.001 -51.8

Monthly pharmacy costs ($) 79.25 80.11 0.848 -1.1

Monthly total costs ($) 153.75 234.78 0.002 34.5

Discussion

Drugs and device manufactures often subject their products to clinical research to determine• Safety• Efficacy

Healthcare services are rarely subject to similar levels of clinical research • Some exceptions• CMS randomized pilot• HealthDialog has a randomized trial published in NEJM

Discussion

75% of managed care plans report having comprehensive disease management programs as defined by CCAIndustry growth likely due to:• Frustration with pace of guideline adoptions• Guaranteed financial savings by DM companies• High patient satisfaction• Other reasons

Discussion

Although Propensity Scores balance observable variables, unobservable variables may not be balanced• Motivation • Psycho-social factors• No drug information for this study

Selection Bias?• Is selection determined by observable or unobservable

variables?• If by unobservable, then bias may exist