(P79) REFERENCE POINT PERCENT CHANGE TREND ANALYSIS IN HEALTH DISPARITY MEASURES

Background:
Percent change analysis is widely used to assess temporal trends. However, conventional percent change (CPC) usually ignores reference points (e.g., rate ratio=1), potentially incorrectly calculating percent change when measures of association cross the reference point between time-points. We propose a new reference point percent change (RPPC) estimator incorporating the reference point and compare the performance of RPPC to CPC analysis in the context of mpox health disparities.

Methods:
We examined early (May22-Aug31) and late (Sep01-Dec31) trends in mpox diagnoses and vaccination from the 2022 mpox outbreak. We calculate rate ratios (RRs) and associated trends of racial and ethnic disparities, using RPPC and CPC, highlighting where estimated trends were different.

Results:
Black and Hispanic people experienced disproportionate case burden throughout (Black:White RRs=1.9, 6.9, and 6.5 at the early, peak, and late outbreak times and Hispanic:White RRs=2.2, 4.1, and 3.6). Because the reference point (RR=1) was never crossed, both approaches estimated identical trends, indicating initial disparity increases compared to Whites. Vaccination disparities were more complicated with RRs crossing the reference point over time (Black:White RRs=0.5, 1.4, and 1.2, and Hispanic:White RRs=0.9, 1.4 and 1.5), which led to diverging trends. Comparing Blacks to Whites, initially the RPPC estimated a 65% decrease while the CPC a 170% increase; both estimated a 10% decrease later. Comparing Hispanics to Whites, initially the RPPC estimated a 25% increase while the CPC estimated a 64% increase; both estimated a 9% increase later.

Conclusion:
When RRs cross a reference point, CPC only considers either an uptrend or downtrend and ignores a reducing trend toward the reference point in determining percent changes, while RPPC accounts for the reducing trend toward the reference point, resulted in different percent change values for the temporal trends of the vaccination disparities. Findings from this analysis show the utility of RPPC in assessing temporal trends in health disparities.