Slide Presentation from the AHRQ 2007 Annual Conference

Targeted Intervention

• 75 "bottom left" provider
  • Control: providers were mailed baseline indicator levels
  • Education only: providers were mailed baseline levels and nurse educators visited providers
  • Education + Reminder System: providers were mailed baseline levels, nurse educators visited providers, and providers had access to the reminder system
• All analyses are "intention to treat"

Notes:

75 lower left hand corner practices selected—evenly divided among 3 groups


COMMENTS ON INTERPRETATION

All results are for CC only, with all people with any third-party insurance dropped
The UNISYS project included only the lowest quartile of PCPs (based on HbA1c testing rates.)
The comparisons of interest are group comparisons on changes in rates from baseline to follow-up; e.g., did the "education only" group change more than the "control group"?
The graphs show raw rates; however, the p-values included on the graphs are for comparisons of adjusted rates.
The baseline data are HEDIS measures calculated as of the 12/31/2005; the follow-up data are HEDIS measures calculated as of 12/31/2006.

COMMENTS, DIABETES MEASURES

The major finding is the upward movement in all the diabetes measures, with the movement quite strong in most of the measures.
This upward movement occurred in all three treatment groups, and it also occurred in the CC population as a whole, not just the UNISYS study PCPs.
In some measures, however, there are groups differences (see graphs), with the "education + reminder system" group generally showing greater improvement.
On the graphs, there are six p-values.
The first three p-values are for changes from baseline to follow-up within each of the three treatment groups
The last three p-values are for the group comparisons