Diabetes is a leading cause of disparities in differences in diabetes outcomes persist, Adams et al, Diabetes Care (2005) Racial Differences in Mean …

Diabetes Disparities in an HMO: Challenges and Opportunities for Improvement?
Alyce S Adams May 31, 2008

Outline
Background and motivation
Focus on medication management

Longitudinal study of medication management and black-white gap in A1c control
Methods Key findings Limitations and conclusions

Implications for interventions in the primary care setting

Background and Motivation
Diabetes is a leading cause of disparities in mortality in the US Racial differences in diabetes outcomes persist, even among insured populations Self-care is key to diabetes control
Influenced by non-health care factors Challenging to monitor most self-care behavior Clinicians may face resource and other barriers to provide the intensive self-care support recommended for all patients
Source: Wong 2002; Albright et all 2001; Karter et al 2002; Sequist 2006 Grant et al 2007; Adams et al 2008

Summary of Previous Findings
Persistent gap in A1c between black and white patients over time Decreasing disparities in the quality of care Persistent differences in self-monitoring of blood glucose and medication adherence
Provision of free blood glucose monitors resulted in a significant, but temporary increase
in SMBG among black patients Gap in medication adherence evident at time of initial treatment
Source: Sequist, Adams, et al 2006; Trinacty 2006; Trinacty, Adams, et al Manuscript

Racial Differences in Mean HbA1c 1994-2001 among Continuously Enrolled Diabetes Patients
MEN
95 9

WOMEN
95 9

black

Avg HbA1c

85 8 75 7 65 6

black

85 8 75

white

7 65 6

94 95 96 97 98 99 00 01

represent 95 Confidence Intervals Adams et al, Diabetes Care 2005

Racial Differences in Mean HbA1c 1994-2001 among Continuously Enrolled Newly Diagnosed Diabetes Patients
95 9

MEN

95 9 85

WOMEN

Avg HbA1c

85 8 75 7 65

black

8 75 7 65

black

white
6 1 2 3 4
6

white
1 2 3 4

X-Axis Years Since 1st Diabetes Diagnosis represent 95 Confidence Intervals Adams et al, Diabetes Care 2005

Initiation of Treatment as a Point of Intervention
Most physician-patient encounters result in a prescription Medication use can be monitored and adjusted to individual patient needs Efficacious strategies for improving medication management are available Few studies have examined the contribution of medication management to racial disparities
Diabetes Care 2008: Vol 31:
916-921

Research Questions
What is the contribution of medication management to the black-white gap in A1c in this setting?
Medication adherence Therapy intensification Self-monitoring of blood glucose Can improving medication management reduce disparities in diabetes control?

Setting and Data Sources
Multi-specialty group practice in Boston, MA Data sources
Electronic medical record Insurance claims Prescribing notes Patient self-reported race information

Outcome measure
Lab assessed HbA1c

Methods
Retrospective, Longitudinal Repeated Measures Design 1,806 Adult newly diagnosed diabetes patients identified between 1994 and 2000
467 black 1339 white

Followed for 12 months following initiation of oral hypoglycemic therapy

Medication Management Measures
Medication Adherence
Average mgs dispensed/mgs prescribed during 3 months prior to each lab test Monotherapy or multiple oral medications

Intensification of Therapy
Increase in mgs or added medication during the last six month

Self Monitoring of Blood Glucose
Average number of test strips dispensed during the last 3 months

Analysis
Hierarchical Models
Stratified and combined models With and without clustering by clinic
Multiple Imputation Methods used for missing covariates

Baseline Differences by Race
Female Age 45 46-64 65 SBP130mm Hg Mean A1c Black 26 52 White 74 41

29 62 9 61 98 24

14 54 32 67 89 21

Differences in Self-Care Behaviors Over Time
Following Therapy Initiation 6 Months Black Medication Adherence SMBG Therapy Intensification
p00001

12 Months Black 717 155 382 White 776 171 378

White 783 207 272

727 203 286

Association between Medication Management and A1c
25 increase in medication adherence from 50 to 85 of days covered was associated with
005 lower A1c among black 007 lower A1c among whites

An additional 20 test strips per month avg 70 to 90 was associated with
02 lower A1c among blacks 01 lower A1c among whites

Therapy Intensification was not a statistically significant predictor

Other Important Predictors
1 additional physician visit was associated with
022 lower A1c for blacks 014 lower A1c for whites

1 year older at first diagnosis
003 lower A1c for blacks 002 lower A1c for whites

Results of Combined Model
Black and white patient benefited from improved medication adherence and improved self-monitoring of blood glucose Medication management contributed
marginally 2 to the black-white gap, more than 50 of which remained unexplained

Limitations
Single HMO, non-representative of typical patient Potential for measurement error
Adherence may overestimate actual adherence Therapy intensification may be poorly measured

Exclusion of patients using insulin Lack of data on diet, exercise, SES and other contributing factors

Conclusions
We found persistent disparities between black and white patients that were present at diagnosis Higher medication adherence was associated with lower A1c for all patients However, improving medication adherence, as a single strategy, is unlikely to reduce the blackwhite gap in A1c in this setting

What can clinicians do?
An early intervention approach that emphasizes medication adherence, SMBG and more frequent physician visits Advocate for System-Level Strategies to Improve Secondary Prevention
Co-pay waivers for diabetes medication, test strips and physician visits during the first year Use of automated telephone reminders to reduce missed appointments

Acknowledgements
Research Team National Institute for Diabetes, Digestive and Kidney Diseases Harvard Pilgrim Health Care Foundation Harvard Pilgrim
Health Care Patients and Staff at Harvard Vanguard Medical Associates

Thank You
Contact Information: Alyce S Adams Research Scientist Division of Research Kaiser Permanente alycesadams@kporg

Results of Hierarchical Models
Blacks
Medication Adherence SMBG

Whites

Combined

-0002 -0004, -0001 -0003 -0004, -0002 -0003 -0004, -0002

-001 -001, -001

-0005 -001, -0004

-001 -001, 0005

Therapy -006 -026, 014 Intensification Black Race –

0002 -008, 009

-002 -010, 007

–