200 likes | 517 Views
Association of Care Coordination with Diabetes Outcome Measures among Adults with Diabetes. David M. Mosen, PhD, MPH Carol L. Remmers, MPH Erin M. Dirks, MS Jim Bellows, PhD Richard Mularski, MD, MSHS, MCR. Background. Diabetes Mellitus (DM) is a major cause for morbidity and mortality.
E N D
Association of Care Coordination with Diabetes Outcome Measures among Adults with Diabetes David M. Mosen, PhD, MPH Carol L. Remmers, MPH Erin M. Dirks, MS Jim Bellows, PhD Richard Mularski, MD, MSHS, MCR
Background • Diabetes Mellitus (DM) is a major cause for morbidity and mortality. • Patients with DM often have complex care needs and are in need of care coordination services. • Enhanced care coordination may: • improve overall control of DM • reduce hospital admissions and emergency department (ED) utilization • improve satisfaction with care • Little research has examined association of care coordination with DM.
Study Objectives • Among adults with DM, to identify the association of care coordination with: • DM control measures • HgA1c control • LDL-c control • DM-specific outcome utilization measures • ED utilization visits • hospital admissions • Patient satisfaction • satisfaction with chronic illness care • satisfaction with personal doctor
Study Population: Diabetes within Kaiser Permanente (KP) • An estimated 550,000 adults throughout the KP system have DM. • Patients were identified via KP automated databases: • pharmacy • inpatient and outpatient utilization • laboratory information • Total system-wide diabetes prevalence is estimated around 10%. • Patients receive care in a group model HMO setting—8.7 million members.
Overview of KP’s Care Management Institute (CMI) – Study Sponsor Mission: • To improve quality and efficacy of care delivered to members with chronic health conditions Goals: • To ensure that care processes known to be effective are uniformly delivered to KP members • To improve knowledge inside and outside KP of effective healthcare delivery, including new approaches to physician and staff learning • To advance understanding of how patient-provider interactions affect quality of care and outcomes • Collect health-related quality-of-life (HRQOL) information on patients with chronic conditions
2006 CMI Patient Care Survey • Survey assessed several domains across four chronic condition cohorts: • care coordination • health status • demographics • patient satisfaction • asthma, DM, coronary artery disease (CAD), and heart failure • Information from survey intended to be used for quality-improvement purposes • Survey representative of KP regions
Methods • Survey of 1,542 persons with DM was completed in late Fall 2006. • Mode of administration = mail/telephone survey • Survey sample identified using HEDIS inclusion criteria: • > 1 inpatient admission in 2004 or 2005 with DM diagnosis • > 1 ED visits in 2004 or 2005 with DM diagnosis • > 2 outpatient visits in 2004 or 2005 with DM diagnosis • > 1 dispensings of insulin or oral hypoglycemics/ antihyperglycemics in 2004 or 2005 • Survey data linked with 2006 DM controland utilization measures: • DM control and utilization measures obtained from CMI’s Clinical Outcomes, Reporting and Evaluation (CORE) Database • These measures routinely used for performance reporting.
Survey Process Flow Initial sample N = 3,137 Contact rate N = 2,242 (71.5%) Response rate N = 1,542 (68.7%) Analytic sample N = 1,399 (90.7%)
Care Coordination Measure (Independent Variable) • Care coordination survey items adapted from the 2005 Commonwealth Fund Survey of Sicker Adults • Respondents were asked about medical care they received in the past 24 months and whether they: • were ever given conflicting advice from different providers other than their regular doctor; • had an appointment where test results were not available; • had an appointment where tests were ordered that should have already been done; • had an appointment where their provider asked questions that should have already been known. • Each yes/no response (0=yes, 1=no) was summarized into a composite (0=lowest care coordination, 4=highest care coordination).
Covariate Measures • Age (continuous) • Gender: Male vs. female • Race/ethnicity: • White vs. African-American, Hispanic, Asian-American, other • Educational attainment: • Less than high school vs. high school graduate, some college/tech school, college graduate or higher • Self-reported health status: • Fair/poor vs. excellent/very good/good • Comorbidities: DM only vs. >= 1 comorbidities • Geographic location: • California regions vs. non-California regions
Outcome Measures • DM control measures • Good HgA1c control (< 7% vs. > 7%) • Good LDL-c control (< 100 mg/Dl vs. > 100 mg/Dl) • DM-specific utilization measures • > 1 ED visits (vs. < 1) • > 1 Hospital admissions (vs. < 1) • Patient satisfaction • Overall satisfaction with chronic illness care • Overall satisfaction with personal doctor • Each measure scored on 1-10 scale (1=lowest satisfaction, 10=highest satisfaction) • Both measures dichotomized (> 9 vs. < 9)
Analysis • Descriptive statistics • Bivariate analysis of care coordination with each outcome measure • Care coordination measure analyzed as three-level variable: • Low rating (0-2) • Moderate rating (3) • High rating (4) • Logistic Regression Models constructed to assess independent association of care coordination (low vs. high, moderate) with outcome measures • Models constructed for outcome measures that had significant associations with care coordination (p < .10) in bivariate analysis • Models adjusted for age, gender, race/ethnicity, educational attainment, health status, comorbidity status and geographic location
Descriptive Statistics: Care Coordination Ratings and Outcome Measures
Association of Care Coordination Scores with Outcome Measures
Logistic Regression Results: Association of Care Coordination Scores with DM-specific Hospital Admissions 1Models adjusted for age, gender, educational attainment, race/ethnicity, health status, comorbidity status, and geographic location.
Logistic Regression Results: Association of Care Coordination Scores with Satisfaction Measures 1Models adjusted for age, gender, educational attainment, race/ethnicity, health status, comorbidity status, and geographic location.
Conclusions • Highest care coordination ratings (after adjusting for demographics, health status, comorbidities, and geographic location) independently associated with: • lower DM-specific hospital admissions • higher satisfaction with chronic illness care and personal doctor • Care coordination ratings not associated with: • HgA1c and LDL-c control • DM-specific ED utilization
Limitations • Cross-sectional study design • DM-specific utilization measures not followed prospectively • Limited measures of care coordination studied • No data available on care management enrollment • Results cannot be generalized beyond group model HMO setting
Implications for Policy and Practice • Further studies are needed to confirm the prospective relationship of care coordination with DM-specific process and utilization measures. • If prospective relationship can be confirmed, identification of patients with low self-reported care coordination scores can be used to develop targeted quality-improvement programs. • Prospective interventional studies needed to determine whether incremental improvements in self-reported care coordination scores result in improved DM-specific outcomes.