Effect of electronic prescribing with formulary decision support on medication tier, copayments, and adherence

Table 3 Logistic regression evaluating the relationship between formulary decision support and preferred medication tier (n = 14660) ^*

PCP or Claim characteristic^†	Odds ratio	95% CIs	P value
Medication class – inhaled steroid	4.1	3.4-5.0	<0.0001
E-prescribing with formulary decision support (FDS) usage
Low users (<30% of time)	0.9	0.8-1.1	0.35
High users (>30% of time)	0.8	0.6-1.1	0.16
Time periods
Non-interruptive FDS time period	1.1	0.9-1.3	0.42
Interruptive and Non-interruptive FDS time period	1.0	0.9-1.1	0.59
Interactions
Low users during time period with Non-interruptive FDS only	0.9	0.6-1.5	0.74
Low users during time period with Interruptive and Non-interruptive FDS	0.8	0.6-1.1	0.13
High users during time period with Non-interruptive FDS only	0.9	0.4-2.3	0.83
High users during time period with Interruptive and Non-interruptive FDS^‡	1.9	1.0-3.4	0.04

^*Variables without association, and therefore not used as predictors in this model, are: date dispensed, PCP prescribing volume with this insurer, patients’ average pharmaceutical claims per month, and patient race and income (as estimated from zip code data).
^†The referent categories were as follows: angiotensin receptor blocker medication class, non-users, and claims from the time period prior to e-prescribing activation.
^‡As noted in the text, this odds ratio remained in the range of 1.6-2.0 in five other models where we: (1) excluded pediatricians completely, (2) controlled for physician specialty, (3) controlled for practice size, (4) restricted to ARB only, and (5) restricted to IS only.

ISSN: 1472-6947