DSS Program Office: Trends in Diabetes Monitoring by Merging Five DSS NDE

1
Trends in Diabetes Monitoring by Merging Five DSS
NDE
by Elisabeth McSherry, M.D., MPH DSS Program
Office Director, Database Development
(175F) March 10, 2004
This acknowledges the key contributors in
developing the merged data sets for reporting
Dr. George Busch, John Paris, RN, John Knox,
Nancy Coughlin, RN. For the mini reports Nancy
Coughlin, RN.
2
Trends in Diabetes MonitoringTodays Discussion

• Method to Merge
• Deriving Mini Reports
• Findings in Two VISNs
• Opportunities for Care Enhancements

3
Merging

• To find all cases receiving HbA1C tests (from
  VHA), we used the LAB National Data Extract
  (NDE).
• To find all cases, tested with HbA1C, who also
  had the diagnosis of Diabetes (1º or other) in
  the last three years to NPCD.
• To find the results of HbA1C (glycohemoglobin)
  and LDL (Low Density Lipo Protein) levels, we
  used the LAB Results NDE.
• To find all coded Diabetics who had HbA1C tests
  and prescribed hypoglycemic medications, we used
  the Pharmacy NDE.
• To find the costs of the subsets of these cases
  per year, we used the NDE TRT Inpatient and NDE
  Outpatient.

4
Merging Issues

• The major issue was that some VAMCs did not map
  their local lab tests (LMIPs) to the National
  (DSS) Lab Results Table.
• This resulted in missing data for those VAMCs.
• A new National LAB Results (NDE LAR) Audit Table
  was developed using the new National VHA
  Financial Clinical Data Mart (FCDM) at AAC and a
  Data Cube now reported on the ProClarity Cube
  website https//10.224.151.59/pas.

5
(No Transcript)
6
VSSC KLF/DSS Website
7
FCDM ProClarity WebsiteLab Test Hemoglobin A1C
(Glycohemoglobin)
4 sites did not setup the A1C test on
the VistA-DSS extract
?
8
FY04 Lab Results Tests in the LAR Extract
9
Caveats

• National DSS NDE Clinical Data, 100
  patient-specific samples represent
• Trend data
• Not real-time intervention (see VistA Clinical
  Indicators)
• Not VHA Performance Measures (see OPQ/EPRP 1-3
  samples)
• NDE data shows
• Actual trends
• Data quality issues, if present
• A focus on cases without VHA Lab tests need to
  check for external (Non-VA) Lab tests

10
National FY03 Coded Diabetics Tested
11
National DSS Data

• Finds percent of all coded Diabetics tested (and
  the Diabetes cases not yet tested)
• Finds percent of all non-coded patients on
  hypoglycemics (gives estimate of
  not-recently-coded-Diabetics for data quality
  improvement)
• Finds number of unique SSNs/tests on veterans
  neither on hypoglycemic medications or with coded
  diagnosis of Diabetes for process
  improvement/resource conservation

12
National PatientsUncoded for Diabetes with A1C
Hypoglycemic Medications
Coding Issue?
13
National Cost to Test Patients without Diabetes
Code
Fiscal Year-to-Date National Cost for
Testing Patients w/o code for Diabetes
5,135,144
Coding? Inappropriate Use?
14
VISN 22 Coded Diabetes Mellitus Patientswith A1C
Tests in Last Two Years
15
VISN 23 Coded Diabetes Mellitus Patientswith A1C
Tests in Last Two Years
16
VISN 22 A1C Levels in CodedDiabetes Mellitus
Patients
17
VISN 22 A1C Levels in CodedDiabetes Mellitus
Patients
18
VISN 23 A1C Levels in CodedDiabetes Mellitus
Patients
19
VISN 23 A1C Levels in CodedDiabetes Mellitus
Patients
20
VISN 22 Coded Diabetes Mellitus Patientswith
LDLC Tests in Last Two Years
21
VISN 22 Coded Diabetes Mellitus Patientswith
LDLC Tests in Last Two Years
22
VISN 23 Coded Diabetes Mellitus Patientswith
LDLC Tests in Last Two Years
23
VISN 23 Coded Diabetes Mellitus Patientswith
LDLC Tests in Last Two Years
24
VISN 22 LDLC Levels in CodedDiabetes Mellitus
Patients
25
VISN 22 LDLC Levels in CodedDiabetes Mellitus
Patients
26
VISN 23 LDLC Levels in CodedDiabetes Mellitus
Patients
27
VISN 23 LDLC Levels in CodedDiabetes Mellitus
Patients
28
Summary

• Merging five NDE (three clinical) enables
  national trending.
• DSS NDEs show 100 samples of patient groups, not
  1-3 like EPRP.
• VAMCs which failed to link LAB Results tests,
  show no results for the monitor.
• This trend data is not real-time intervention
  data. When a process at the VAMC changes to
  improve real-time intervention, it will be shown
  in the next NDE data pull.

29
Summary(contd)

• Mini reports by graphs and data cubes allow a
  wider group of VAMC/VISN executives to analyze
  their trends and benchmark vs. other facilities.
• The limitation to this data is that it is derived
  from VHA Lab and Pharmacy extracts and does not
  reflect testing or medications from external
  sources.