Disparity in LDLC Testing of Dually Enrolled Patients with Diabetes Patient and Practitioner Factors

1
Disparity in LDL-C Testing of Dually Enrolled
Patients with DiabetesPatient and Practitioner
Factors

• Ruth Medak, MD
• Senior Clinical Coordinator
• OMPRO
• AHQA Technical Conference
• Analytic Methodologies Track
• 1110 am, February 1, 2002

2
Project Goal

• Investigate causes of performance disparities in
  diabetes indicator tests between
• Dually enrolled Oregon patients on Medicare FFS
  Medicaid
• Non-dually enrolled Oregon patients on Medicare
  FFS
• Design and implement interventions based on
  factors

3
Performance Disparity in Diabetes Indicator Tests
for Oregon Medicare FFS Patients

• Significant disparity in LDL-C testing (12.2)
  between dually enrolled and non-dually enrolled
  Oregon Medicare FFS patients with diabetes
• No significant disparity in HbA1c testing
• Significant disparity in dilated eye exams (7.5
  below CMS threshold)

Data source 19971998 Oregon Medicare FFS
diabetes claims
4
Project Design
Interviews baseline medical record abstraction
Hypothesis development
Intervention development
Intervention implementation
Remeasurement(medical record abstraction and
claims)
5
Semistructured Interviewsand Medical Records
Abstraction

• Sampling criteria physicians with gt5 DE and gt10
  NDE patients
• Sample 40 physicians
• Recruited 18 physicians
• Interviewed 16 physicians
• Records abstracted 244 patients of 15
  physicians 5 diabetes specialists
• 6 general internists
• 4 family physicians

1 physician withdrew following the interview
6
Semistructured Interviews

• Fifteen physicians and staff
• use of lipid testing guidelines
• lipid testing practices
• barriers to testing
• characteristics of dually enrolled and non-dually
  enrolled patients with diabetes
• use of diabetes management systems

7
Baseline Measurement Retrospective Medical
Record Abstraction

• 66 DE records and 178 NDE records abstracted
• test dates and results
• blood pressure
• hyperlipidemia treatment
• use of systems
• patient comorbidities
• patient behavior (missed appointments,treatment
  refusal, etc.)

8
Analytic Methods

• Combination of qualitative and quantitative
  methods
• Results from semi-structured interviews analyzed
  using NudIst qualitative software
• Results from medical record data abstraction
  analyzed using MS Access and SPSS
• Statistical analysis Chi-square test was used
  to check for statistically significant
  differences between variables at the a0.05 level
  of significance

9
Results Practitioner Interviews

• LDL-C testing goals compatible with CMS
  quality
• indicators and ADA guidelines
• Reasons not to test advanced terminal
  illness
• normal LDL-C without
  medications
• patient indifference to treatment
• Patient indifference perceived as common
  among to treatment Medicaid patients
• most physicians not discouraged by
  initial patient indifference regarding
  glycemic and lipid treatment

10
Results Practitioner Interviews (continued)

• Delegation no standing order protocol for
  LDL-C testing
• most reported referring patients
  to nurse educator or CDE for education
• Systems approach 7 of 15 reported use of
  flow sheet
• 10 of 15 reported obtaining lab
  prior to visit more often than not
• 10 of 15 reported using flow sheet
  or obtaining lab prior to visit

11
Results Performance in LDL-C and HbA1c Testing
for DE and NDE Patients of Interviewed Physicians

• Significant disparity (23.4) in LDL-C testing
  between dually enrolled and non-dually enrolled
  patients with diabetes
• No significant disparity in HbA1c testing

Data source March 1, 1999February 28, 2001
medical record abstraction
12
Results Factors Associated with Dually Enrolled
Patients

• Medicare patients with Medicaid coverage
• More likely than patients without Medicaid
  coverage to have
• mobility limitation
• nephropathy
• insulin therapy
• psychiatric disorder

13
Results Patient Factors Associated with LDL-C
Testing

• For the aggregate sample
• Less likely to receive testing
• mobility limitation
• nephropathy
• Although some patient factors were significantly
  more common among DE patients, no significant
  association was found with LDL-C testing
  disparity between DE and NDE

14
Results Use of a System (Flow Sheet or Planned
Visit)
Use of a diabetes management system was not
significantly more likely to be found in the
charts of non-dually enrolled patients
Aggregate

• Diabetes specialists were significantly more
  likely to use systems

Data source March 1, 1999February 28, 2001
medical record abstraction
15
Results LDL-C Tests Among Patients Whose Charts
Show Use of a System, by Patient Coverage

• Use of a diabetes management system was
  significantly associated with LDL-C testing
• aggregate
• non-dually enrolled patients

Data source March 1, 1999February 28, 2001
medical record abstraction
16
Results LDL-C Tests Among Patients Whose Charts
Show Use of a System, by Provider Type

• No significant performance difference between
  specialists and nonspecialists

Data source March 1, 1999-February 28, 2001
medical record abstraction
17
Conclusions from Interviews and Chart Abstraction

• The interviews and abstracted chart data did not
  explain the disparity in LDL-C testing between
  dually enrolled and non-dually enrolled patients
  with diabetes.

18
Conclusions from Interviews and Chart
Abstraction (continued)

• Factors associated with receiving a biennial
  LDL-C test
• documented diagnosis of hyperlipidemia
• treatment by a diabetes specialist
• use of a diabetes management system
• Factors associated with not receiving a biennial
  LDL-C test
• mobility limitations
• nephropathy

19
Hypothesis for Intervention

• Implementation of a patient management system
  will lead to increased LDL-C testing of both
  dually enrolled and non-dually enrolled patients
  with diabetes
• Implementation may not reduce LDL-C testing
  disparity between dually enrolled and non-dually
  enrolled patients with diabetes

20
Intervention Tools

• Data support for systems change
• Practitioner-specific performance data
  (automatically generated)
• Flow sheet
• Electronic registry
• Planned visit concept paper
• Systems change concept paper

21
Intervention Participants and Methods

• Participants
• Project participants
• Additional target practitioners treating 25 of
  Oregon Medicare FFS dually enrolled patients
• Methods
• Detailing visits to project participants
• Detailing visits to 1/3 to 1/2 of target
  practitioners
• Mail packets to remaining target practitioners

22
Remeasurement

• Interim measurements
• Medical record abstractions
• laboratory tests and results
• blood pressure
• use of systems
• Final measurement
• Medicare FFS claims data CY 20012002

56 months after interviews and 56 months after
intervention