Impact of a Step Therapy for Guanfacine Extended-Release on Medication Utilization and Health Care Expenditures Among Individuals Receiving Treatment for ADHD.

backgroundWhile step therapy (ST) policies are generally effective at reducing cost through the managed utilization of targeted medications, the clinical implications of ST policies are not clear and may vary across therapeutic areas. Guanfacine extended-release (GXR) is approved by the FDA for the treatment of attention-deficit/hyperactivity disorder (ADHD) as both monotherapy and adjunctive to stimulant treatment. At the introduction of GXR to the market, Humana implemented an ST policy on GXR requiring the documentation of previous treatment, intolerance, or contraindication to generic clonidine or guanfacine.

objectiveTo examine the impact of a GXR ST coverage determination (i.e., approved vs. denied) on medication utilization and health care costs among members of a commercial health plan with an ST policy for GXR.

methodsThis study was a retrospective cohort study of administrative claims data. Humana commercial members prescribed GXR who had an ST coverage determination review were identified. All members included in this analysis were required to be aged 6-17 years, have a diagnosis of ADHD or be receiving stimulant medication, have an ST coverage determination (index event) between September 1, 2009, and May 30, 2012, and have 6 months of pre- and post-index continuous enrollment. Members were assigned to either the approved or denied group based on the outcome of the ST coverage determination. Medical and pharmacy claims data were used to measure baseline demographic and clinical characteristics and to measure medication utilization and health care costs. Outcomes assessed during follow-up included ADHD medication use, proportion of days covered (PDC) with any ADHD medication treatment, time to first observed post-index ADHD treatment, and all-cause and mental health (MH)-related health care costs. Administrative costs associated with the coverage determination process were also estimated. Bivariate and multivariable adjusted analyses were conducted to compare medication utilization and health care costs between the approved and denied groups.

resultsA total of 642 members were included in the analysis (denied group n = 395 [61.5%], approved group n = 247 [38.5%]). The approved and denied groups were similar in terms of baseline demographics, provider characteristics, and baseline MH diagnoses, with the exception of anxiety disorders being more prevalent in the approved group compared with the denied group (18.2% vs. 10.6%, P = 0.006). A denied GXR coverage determination was associated with a greater percentage of members receiving no ADHD treatment post-index (13.9% vs. 3.2%, P  less than  0.001), greater mean [SD] number of days between index and first observed post-index ADHD medication claim (44.5 [59.6] vs. 17.6 [33.4], P  less than  0.001), and lower mean [SD] PDC with any ADHD medication post-index (0.59 [0.33] vs. 0.75 [0.26], P  less than  0.001). These findings remained statistically significant in multivariable regression models. Unadjusted pre-index median total health care costs and MH-related costs were greater among the approved group compared with the denied group (total health care: $1,582 vs. $1,465, P = 0.033; MH-related: $993 vs. $981, P = 0.020). Likewise, post-index median total health care and MH-related costs were greater among the approved group compared with the denied group (total: $2,056 vs. $1,420, P  less than  0.001; MH-related: $1,543 vs. $946, P  less than  0.001). After adjustment for potentially confounding covariates including pre-index costs, there were no statistically significant differences between the approved and denied groups in all-cause total health care (P = 0.393) or MH-related health care costs (P = 0.054).

conclusionsThe current study found that GXR coverage denial was associated with lower rate of ADHD medication utilization, greater delay in receiving ADHD medication, and lower PDC with ADHD medication. There were no differences observed between the approved and denied group in terms of all-cause total health care or MH-related total health care costs after controlling for potentially confounding variables. Prior to implementation in the ADHD therapeutic area and others, payers should consider the potentially unintended consequences of ST policies, including delay in treatment and undertreatment.