Chronic Disease Management Mitigates the Relationship between Literacy and Health Outcomes Darren A. DeWalt, MD, MPH RWJ Clinical Scholars Program Division of General Internal Medicine University of North Carolina, Chapel Hill June 7, 2004
Collaborators Michael Pignone, MD, MPH Russell Rothman, MD, MPP Robert Malone, PharmD Morris Weinberger, PhD RWJ Clinical Scholars Program Pfizer Health Literacy Initiative UNC Program on Health Outcomes UNC Center for Research on Chronic Illness Funding Sources
Low Literacy Related to Worse Control of Chronic Illness Diabetes –Worse glycemic control –More long-term sequelae HIV –Higher viral load Depression –More severe disease Hospitalization –Consequence of several diseases
LiteracyHealth understanding/ knowledge Health behavior/ Self-care Health outcome Educational opportunity (sum of general knowledge) Learning potential Exposure to reading Poverty/wealth/SES Poor environment Conceptual Framework Patient-provider communication Self-efficacy Participation in medical encounter Trust in physician/health system
Chronic Disease Management
Disease Management Components Multidisciplinary teams/Care coordination Defined follow-up procedures Treatment algorithms based on best available evidence Information systems for tracking patients Patient education for self-care
Disease Management May Reduce Literacy-related Barriers Careful design of educational strategies may improve self-care, particularly for low literacy patients Disease management programs have not examined the role of literacy in the effectiveness of the interventions
Literacy Health-related knowledge Self-efficacy/ health behavior Access to providers/therapy Quality of care Health Outcome Learning potential Educational opportunity SES Why Would Disease Management Work?
Evidence of Effectiveness Heart failure randomized clinical trial Diabetes randomized clinical trial
Disease Management Intervention 1-hour individualized education session Education booklet < 6th grade level Digital bathroom scale Scheduled follow-up phone calls Easy access to our team
Randomized Controlled Trial
Design of RCT Usual Care Disease Management compared to Included patients with low and high literacy for a pre-specified sub-group analysis 64 patients65 patients
Outcome Measures Primary Outcomes –HF-quality of life (Minnesota Living with Heart Failure Questionnaire) –Hospitalization or death Secondary Outcomes –HF knowledge –HF specific self-efficacy –HF self-care behavior
Follow-up to Date Enrolled and randomized 129 Control 65 Intervention 64 Withdrawal: 2 58 (95%)56 (95%) Withdrawal: 6 50 (85%)56 (93%) 6 month to date 12 month to date Death: 5
Baseline Characteristics Variable Control (n=65) Intervention (n=64) Mean Age, years (SD)62 (10)63 (10) African American, %55%56% Male, %42%58% Education, years Income <15,000/yr, %68%69% Medicaid, %32%36% Medicare, %73%72% Literacy (S-TOFHLA) Inadequate, %40%45%
Baseline HF Characteristics Variable Control (n=65) Intervention (n=64) Time with HF, median years (interquartile range) 3 (1-9)2.5 (1-8) NYHA class, % II III IV Systolic dysfunction, %4836 Medications, % ACEI or ARB6975 -blocker 7155 Digoxin3727
Reduced Hospital Admission or Death Incidence Rate Unadjusted Incidence Rate Ratio (IRR) 0.66 [0.38, 1.12] Adjusted IRR 0.56 [0.32, 0.95] *Adjusted for baseline HFQOL, B-blocker use, digoxin use, systolic dysfunction and hypertension
Results for Patients with Low Literacy 53 patients found to have inadequate literacy based on the TOFHLA (28 intervention, 25 control)
Inadequate Literacy Lower Admission or Death Incidence Rate Unadjusted Incidence Rate Ratio (IRR) 0.69 [0.28, 1.75] Adjusted* IRR 0.38 [0.16, 0.88] *Adjusted for baseline HFQOL, B-blocker use, ACEI or ARB use, and hypertension
Heart Failure Study Summary HF disease management appears to decrease the combined endpoint of hospitalization or death Effect may be greater among those with low literacy
Diabetes Disease Management
Patient registry Treatment and monitoring algorithms Patient education Care coordination
Educational Strategies Patient centered learning Focus on behaviors rather than knowledge Repetition/reinforcement Teach-back method
Care Coordination Call patient at least once a month Review self-care skills Help to navigate health care system Address barriers of medication access, transportation, and communication
Evaluation with RCT Usual Care Disease Management compared to 112 patients 105 patients One Time Management Session
Outcome Measures Primary Measures –A1C –Blood pressure –Aspirin use Secondary Measures –Diabetes knowledge –Treatment satisfaction –Medical visits –Potential harms
Variable Control (n=105) Intervention (n=112) Age, mean (SD), y56.7 (10.8)53.5 (12.5) Female, No. (%)59 (56%)63 (56%) African American, No. (%)62 (59%)78 (70%) Household Income $20,000, No. (%) 78 (76%)77 (69%) Less than HS education, No. (%)46 (44%)40 (36%) REALM Score, median (IQR)57 (32-64)55 (31-62) Low Literacy, No. (%)**34 (32%)49 (44%) Demographics ** Defined as 6th Grade Literacy Level on REALM
Improvement in A1C Worse Control Better Control
Results According to Literacy Status
Diabetes Control: Results for Patients with Literacy Above 6th Grade Level Worse Control Better Control
Diabetes Control: Results for Patients with Literacy at or Below 6 th Grade Level Worse Control Better Control
Diabetes Study Summary Patients with low literacy had greater improvement in glycemic control
Comment Both studies performed in general internal medicine practice at academic center Both studies used multiple components in the intervention to improve outcomes Cannot isolate individual effects of parts of the interventions
Conclusions Chronic disease management appears to mitigate literacy related disparities Structured care appears helpful, especially for patients with low literacy Easy-to-read materials insufficient Need to combine with other systems of care to address other barriers
Future Directions Research on the components of disease management that are critical for patients with low literacy Research on existing disease management programs and whether they reduce disparities Consider disease management as a policy strategy to reduce disparities
Baseline HF Measures Variable Control (n=65) Intervention (n=64) Knowledge, mean percent5756 Self-efficacy, mean score22 Daily weight measurement, %1513 HFQOL score, mean score (range 0-105) 5745
Improved HF Knowledge, Self-Efficacy, and Self-Care Behavior 6 Month OutcomeControlIntervention Difference (CI) P value Knowledge change (4, 19) <0.01 Self-efficacy change (0.5, 3.1) <0.01 Daily weight measurement, % (53, 81) <0.01
No Difference in HFQOL at 6 or 12 Months Time Unadjusted Difference (CI) Unadjusted P value Adjusted Difference* (CI) Adjusted P value 6 Months-6 (-15, 2) (-11, 6) Months -3 (-11, 5) 0.47 (-9, 7) 0.84 *Adjusted for baseline HFQOL, B-blocker use, systolic function, sex, diabetes, hypertension
Follow-up to Date Enrolled and randomized 129 Control 65 Intervention 64 Withdrawal: 2 58 (95%)56 (95%) Withdrawal: 6 50 (85%)56 (93%) 6 month to date 12 month to date Death: 5