Table 3. Regression Results Estimating Days’ Supply.
| Component | Estimated days’ supply (95% CI) | |
|---|---|---|
| Michigan (n=2915 person-y)a | New York(n=4101 person-y)b | |
| Logistic component (estimating any vs no hydroxyurea use measures) | ||
| Years since 2010 | 0.00 (−0.07 to 0.06) | NA |
| Age | 0.09 (0.07 to −0.11) | NA |
| Postguideline changes periodc | 0.42 (0.07 to 0.76) | NA |
| Negative binomial component (estimating count of hydroxyurea use measure) | ||
| Years since 2010 | −0.13 (−0.30 to 0.05) | 0.20 (−0.26 to 0.66) |
| (Years since 2010)2 | 0.02 (−0.03 to 0.06) | −0.03 (−0.25 to 0.20) |
| Postguideline changes periodc | −3.66 (−6.30 to −1.02) | 0.11 (−1.53 to 1.75) |
| Years since 2010 × postguideline changes periodc | 1.14 (0.30 to 1.99) | −0.09 (−0.98 to 0.79) |
| (Years since 2010)2 × postguideline changes periodc | −0.09 (−0.16 to −0.01) | 0.02 (−0.22 to 0.26) |
Abbreviation: NA, not applicable.
a
Michigan data were modeled with zero-inflated negative binomial regression.
b
New York State models include years since 2012. New York data were modeled with negative binomial regression.
c
Reference = preguideline changes period.