KEYWORD1:
NIS, Provider Response, State Comparisons

BACKGROUND:
Published rates from CDC’s National Immunization Survey (NIS) are commonly used by immunization programs and other local entities for benchmarking and state comparisons. Provider response can vary by area, especially where children are vaccinated by more than one provider. Such variations can contribute to the lack of state-to-state comparability.

OBJECTIVE:
We analyzed NIS data to determine if state multi-provider response rates (MPRR) and immunization rates are associated. We then developed a model to adjust for MPRR variations and assessed its effect on state rates.

METHOD:
State MPRRs for 2000 were calculated from NIS public release data. For children with more than one identified provider, the MPRR is the total of provider responses divided by the total number of providers. States’ NIS 4:3:1 and 4:3:1:3 rates were correlated with state MPRRs and tested for significance. The effect of controlling for MPRR on state rates was also modeled.

RESULT:
For the 4:3:1:3 series, the correlation between state rates and MPRRs in 2000 was .34 (p<.01). Adjusting 4:3:1:3 rates for state MPRRs changed state rates by an average of 3%. Also the range from highest to lowest immunized states decreased for the 4:3:1:3 series, from 18.7% to 10.2%. In addition, the percentage of multiple providers by state also correlates with state rates; for a 4:3:1:3 series, r = -.39. After controlling for the percentage of multiple providers by state, the correlation between 4:3:1:3 rates and MPRRs increased to .51, (p<.01).

CONCLUSION:
State immunization rates from the NIS are influenced by MPRRs. MPRRs appear to have a separate effect on state rates than do differential state proportions of multiple providers. Programs should apply caution in interpreting state comparisons.
LEARNINGOBJECTIVES:
Participants will learn about issues around MPRRs. The comparability of NIS results on a state-by-state basis will be discussed.