Evaluation of a Novel Web-Based Prior Approval Application for Palivizumab Prophylaxis of Respiratory Syncytial Virus in a State Medicaid Program

AUTHORS: Kristin Lundeen, Trista Pfeiffenberger, Julie Jacobson Vann, Timothy O’Brien, Charlene Sampson, Steven Wegner



BACKGROUND: Recent disproportionate increases in use of specialty medications, such as palivizumab (Synagis), compared with steady utilization of traditional medication use, have prompted complex utilization management strategies that require frequent evaluation to facilitate cost-effectiveness while preserving patient access. Clinical criteria utilized by North Carolina (NC) Medicaid for use of palivizumab for respiratory syncytial virus (RSV) prophylaxis are consistent with the most recent guidelines published in the Red Book: Report of the Committee on Infectious Diseases. Prior to the 2011-2012 RSV season, prior approval (PA) requests were submitted by facsimile using the NC Medicaid Synagis PA form. A web-based PA application, which includes automatic approval capability, monthly dose prompts to providers, and a standardized dose projection formula, was developed for the 2011-2012 RSV season.
OBJECTIVES: To evaluate the timeliness of palivizumab coverage determination, compliance with palivizumab prophylaxis regimen, and the accuracy of the dose projection formula achieved with this novel web-based PA application for palivizumab prophylaxis in NC Medicaid recipients.

METHODS: A historically controlled retrospective cohort study was conducted in which all palivizumab PA submissions and supporting documentation from the 2010-2011 and 2011-2012 RSV seasons were retrospectively reviewed for date and time of original submission and final coverage determination. Submissions from the 2011-2012 season were also retrospectively reviewed for number of doses approved, number of doses administered, date of administration of each dose, and actual dosage administered. These data were used to evaluate compliance and the projected versus actual beneficiary weight and dose to assess the accuracy of the dose projection formula. Submissions lacking required information were excluded. Time from PA submission to coverage determination was compared between seasons using a 2-sample t-test. The proportion of compliant recipients was calculated based on number of doses received and dosing interval of no more than 35 days. Accuracy of the dose projection formula was evaluated using a paired Student’s t-test.

RESULTS: Time to coverage determination decreased overall, on average, by 3.7 days (mean [SD] 8.5 [15.4] vs. 4.8 [9.3]; P < 0.001) for the 2011-2012 season using the electronic web-based PA application compared with the traditional facsimile-based system used in the 2010-2011 season. Decreased time to coverage determination was observed in both PA requests that required medical review and those that did not. Of all palivizumab recipients who were eligible to receive at least 2 doses (n = 1,233), 61.1% were fully compliant with all doses, and 86.9% received all but one documentable dose. Of those who received at least 2 documented doses (n = 1,091), 62.8% received all doses within 35 days of the previous dose. When both definitions of compliance were applied concurrently, 39.3% of all palivizumab recipients were considered compliant; the mean difference between projected and actual doses was 7.1 mg (95% CI: 6.8-7.5; P = 0.001) or 8.6% (95% CI: 8.0-10.0). Projected and actual doses did not vary significantly in the sensitivity analysis when excluding entries with ≥ 50% difference.

CONCLUSIONS: The 2011-2012 web-based PA application improved the timeliness of palivizumab coverage determination compared with the 2010-2011 facsimile-based system. Observed compliance rates for NC Medicaid recipients were slightly lower than those reported in the literature when defined by number of doses received but were higher when defined by interval between doses. The dose projection formula used for the web-based application appears to be accurate for infants 0-2 years of age.

Content for class "break" Goes Here