Cost-effectiveness analysis of palivizumab among pre-term infant populations covered by Medicaid in the United States

Abstract

Objective:

Medicaid infants are at high risk of severe respiratory syncytial virus (RSV) disease. The study objective was to estimate the cost-effectiveness of palivizumab in a Medicaid population.

Methods:

A societal cost-utility analysis was conducted of prophylaxis with palivizumab vs no prophylaxis among four groups of premature infants: (1) <32 weeks gestational age (wGA) and ≤6 months chronologic age (CA); (2) 32–34 wGA, ≤3 months CA with 2009 American Academy of Pediatrics (AAP) risk factors (RF); (3) 32–35 wGA, ≤6 months CA with 2006 AAP RF; and (4) 32–35 wGA, ≤6 months CA with ≤1 RF. Full dosing of palivizumab was assumed throughout the RSV season (consistent with the FDA-approved label). All costs were in 2010 US dollars. The societal public payer spend for palivizumab was estimated using Medicaid reimbursement methodologies for the top 10 palivizumab-using states in 2010 minus mandatory manufacturer rebates. This study reports the incremental cost-effectiveness ratios (ICERs) in cost per quality-adjusted life-year (QALY) gained. Sensitivity and probabilistic analyses were also conducted.

Results:

Palivizumab saved costs and improved QALYs among infants <32 wGA. Palivizumab was cost-effective in infants 32–34 wGA with 2009 AAP RF ($16,037 per QALY) and in infants 32–35 wGA with 2006 AAP RF ($38,244 per QALY). The ICER for infants 32–35 wGA with ≤1 RF was $281,892 per QALY. Influential variables in the sensitivity analysis included the background rate of RSV hospitalization, the cost of palivizumab, and the efficacy of palivizumab.

Key limitations:

These results are not generalizable to commercially insured infants or infants outside of the US.

Conclusions:

This is the first cost-utility analysis of palivizumab in a Medicaid population. Palivizumab, when dosed consistent with the FDA-approved labeling, was either cost-saving or cost-effective among current guideline-eligible infants in the Medicaid population. Palivizumab did not demonstrate cost-effectiveness in 32–35 wGA infants with ≤1 RF.

Keywords::

• Respiratory syncytial virus
• Palivizumab
• Prophylaxis
• Cost-effectiveness
• Medicaid

Introduction

Respiratory syncytial virus (RSV) is the leading cause of infant hospitalization in the US1. Infants at highest risk for an RSV-related hospitalization are infants with chronic lung disease of prematurity (CLDP), children with hemodynamically significant congenital heart disease (CHD), or premature birth at less than 36 weeks gestational age (wGA)2,3. Several studies have suggested that the risk of RSV hospitalization is higher among ethnic minorities, including African Americans, Hispanics, and Native Americans1,4–10. Several studies have reported disproportionate rates of prematurity and low birth weight among ethnic/minority populations11–13, thereby increasing the risk of severe RSV disease for this population. Furthermore, the rate of ethnic/racial groups in Medicaid in several states is higher compared with the overall ethnic/racial rate in the US population14. Observational studies also have indicated high rates of severe RSV disease among infants insured by Medicaid1,15,16.

Palivizumab was approved by the Food and Drug Administration (FDA) in 1998. Palivizumab is indicated for the prevention of serious lower respiratory tract disease caused by RSV in pediatric patients at high risk for RSV disease17. However, immunoprophylaxis for all infants born 32–35 wGA is not recommended by the American Academy of Pediatrics (AAP) due to ‘cost considerations’. According to the 2009 AAP guidelines, ‘economic analyses fail to demonstrate overall savings in healthcare dollars because of the high cost if all at risk infants receive prophylaxis’18. Therefore, the 2009 AAP guidelines recommend prophylaxis with palivizumab in a sub-population of infants 32–34 wGA with risk factor criteria (i.e., being of young chronologic age CA [≤3 months] and the presence of ≥1 of the following two risk factors: day care attendance and/or having a sibling younger than 5 years). Approximately 1.6 million Medicaid births were financed in the US between 2002 and 2003, representing ∼40.8% of total births annually in the US19. We estimate that ∼95,000 Medicaid births a year are premature infants of 32–35 wGA and that only 15% of these infants will meet current 2009 AAP eligibility criteria for RSV prophylaxis based on gestational age and risk factor criteria.20

Few economic analyses examined RSV prophylaxis in Medicaid populations21–23. An earlier study by Shireman et al.22 examined the cost-benefit ratio of RSV immune globulin and palivizumab in premature infants or infants with CLDP (10 months or younger) covered by the Kansas Medicaid program during the 1999–2000 RSV season, whereas Wegner et al.23 examined the direct costs of palivizumab for infants born at 32–35 wGA in North Carolina Medicaid’s program during the 2002–2003 RSV season. Recently, Hampp et al.21 examined the cost to avoid an RSV hospitalization among various high-risk cohorts using Florida Medicaid claims data during 2004–2005. Although these studies concluded that the costs of palivizumab exceeded the benefits, these studies did not measure quality-adjusted life-years (QALY), the unit of the effectiveness of a drug or therapy that incorporates the quality and quantity of life gained compared with no therapy in a cost-utility analysis. Furthermore, these studies did not factor in all direct medical costs after the initial RSV episode24 or the significant federally-mandated rebates public payers receive from the manufacturer25^–27. The US Panel of Cost-Effectiveness in Health and Medicine and the 2000 Institute of Medicine report Vaccines for the Twenty-First Century recommended dollars per QALY using the societal perspective as the standard for US cost-effectiveness evaluations28,29. A cost-effectiveness analysis based on cost per QALY gained, instead of cost per event avoided as previously reported, would allow comparisons of the value of drug therapies across multiple disease areas.

The previous economic studies of palivizumab across various Medicaid programs did not account for the effect of risk factors on severe RSV disease in their economic evaluations. Although prematurity alone can significantly increase the risk for serious RSV disease, several studies have shown that the presence of other risk factors increases the risk of RSV hospitalization. Some risk factors include infant exposure to tobacco smoke, day care attendance, and crowded residential living conditions30–33. Multivariate regression techniques and risk-scoring tools from Canada and Spain have recently been developed to identify high-risk populations for serious RSV disease31,32,34–39. These studies have shown an independent effect of various risk factors on RSV hospitalization. In addition, the prospective 2-cohort study in infants of 32–35 wGA by Figueras-Aloy et al.32 demonstrated that the presence of ≥2 risk factors increases the risk of RSV hospitalization in an additive to multiplicative fashion depending on the risk factors. Infants with school-aged siblings, a mother smoking any number of cigarettes daily during pregnancy, and young CA (i.e., ≤10 weeks at the start of the RSV season) had an overall odds ratio for RSV hospitalization of 9.89 (3.96–24.67) compared with an infant without any risk factors.

The incorporation of risk factors in cost-effectiveness analyses (CEAs) could significantly affect the results. A recent CEA conducted in Canada included risk factor data from a validated risk-scoring tool using the Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) dataset35,37. Infants classified as ‘moderate risk’ and ‘high risk’ from the risk-scoring tool had RSV hospitalization rates of 7.1% and 18.7%, respectively37 compared with low-risk infants who had a hospitalization rate of 1.7%. The results of the Canadian cost-effectiveness model indicated that the incremental cost-effectiveness ratios (ICERs) for RSV prophylaxis in infants of 33–35 wGA classified as moderate or high risk were acceptable under their particular healthcare system (moderate-risk group: C$34,215 per quality adjusted life year [QALY]; high-risk group: C$5,765 per QALY)35. Lanctot et al.35 also demonstrated that the ICERs became more compelling as the number of risk factors increased. Similarly, other studies have demonstrated that palivizumab is cost-effective across several high-risk populations outside the US40–43.

In light of these recent findings, the objective of this analysis was to assess the cost-effectiveness of palivizumab prophylaxis in premature infants in Medicaid programs in the US. To our knowledge, this is the first cost-utility analysis of palivizumab conducted in a Medicaid population using Medicaid-specific data. Our economic analysis includes healthcare costs associated with severe RSV disease in Medicaid infants and children, utilities associated with the RSV-related hospitalization, and an estimate of the societal spend for palivizumab that incorporates manufacturer rebates. A Medicaid-specific cost-effectiveness analysis would help decision-makers in setting RSV prophylaxis policies that target patients who would benefit the most from prophylaxis against severe RSV disease44.

Methods

Study population

We conducted a cost-effectiveness analysis of prophylaxis with palivizumab vs no prophylaxis. The cost-utility model used the societal perspective and, therefore, includes direct and indirect costs. A decision tree model (Figure 1) was constructed to compare the cost and outcomes of RSV prophylaxis among four Medicaid premature infant populations: (1) <32 wGA and ≤6 months CA; (2) 32–34 wGA, ≤3 months CA with 2009 AAP risk factors (i.e., having siblings <5 years of age and/or attending day care); (3) 32–35 wGA, ≤6 months CA with 2006 AAP risk factors (any two of the following: exposure to environmental air pollutants, congenital abnormalities of the airways, severe neuromuscular disease, school-age siblings, day care attendance); and (4) 32–35 wGA, ≤6 months CA with ≤1 risk factor as described by Sampalis et al.37. We examined full dosing of palivizumab (≤5 doses depending on month of birth with respect to the RSV season start, consistent with the FDA-approved labeling). Our model did not examine the 2009 AAP recommendation of a maximum of three doses for infants of 32–34 wGA because the clinical efficacy associated with this dosing administration is unknown. We excluded infants and children who were pre-term and who had CLDP or hemodynamically significant CHD from our analysis. Although premature infants with CLDP and children with CHD are included in the labeled indication for palivizumab, we only examined premature infants without these conditions to control for the additional effect of comorbidity on the risk of RSV hospitalization. CA was determined at the start of the RSV season (November 1).

Figure 1.  Structure of simple decision tree model. ICU = intensive care unit; RSV = respiratory syncytial virus.

Model structure

As shown in Figure 1, the infants of interest received either prophylaxis with palivizumab or no prophylaxis. The clinical pathway simulated various clinical events associated with serious RSV disease. Based on a Medicaid study of healthcare utilization and costs among RSV premature infants conducted by Palmer et al.45 and Shi et al.24, the RSV hospitalization was estimated to occur on average 2.4 months after birth. Among those infants with an RSV hospitalization, there is a probability that a proportion of them will require care in the pediatric intensive care unit (ICU). Of those infants in the pediatric ICU, the outcomes are either death during the hospital stay or survival. In our model, mortality is incorporated for pediatric RSV ICU patients only. In accordance with guidance from the US Panel on Cost-Effectiveness in Health and Medicine29, we analyzed a lifetime horizon. The model rolls back the life expectancy (77.45 years46), after discounting quality-of-life adjustments at 3% annually. We varied the discount rate from 0–5% in sensitivity analyses.

Model assumptions

We made several assumptions in our economic analysis of palivizumab. We assumed risk of RSV infection only in the first RSV season, defined in our model as November 1 to March 31, although the start and length of the RSV season can vary across geographic regions in the US. Infants received ≤5 monthly doses of palivizumab throughout the RSV season, consistent with the dosing regimen in the registration trials for palivizumab. Infants were assumed to have received prophylaxis in the outpatient setting during a periodic preventive medicine visit. We assumed that the first dose of prophylaxis for infants born during the RSV season was administered during the birth hospitalization before discharge. We assumed 100% compliance with palivizumab and 5% drug wastage47. Similar to economic models for vaccines, our model uses a baseline utility of 1.0 for infants29,48–52. Life expectancy after surviving RSV hospitalization for premature infants was assumed to be equivalent to the life expectancy for the general US population and unchanged across infant gestational age categories. Three caregiver hours were lost owing to administration of palivizumab; caregiver time lost owing to RSV hospitalization per day was assumed to be 8 h. We did not examine prophylaxis in home health settings, which may be associated with lower costs53,54. Because the IMpact-RSV Clinical Study did not show a statistically significant difference in reported adverse events, we did not incorporate adverse events–related costs in the analysis. Some cases of anaphylaxis, including fatal cases, have been reported in post-marketing reporting.

Respiratory syncytial virus and its complications

The background rates for RSV hospitalization are expected to vary according to the risk profiles of each of the four populations of interest (Table 1). We first estimated the overall background rates of serious RSV disease in the general population and then adjusted these rates based on the data that suggest that the rates of RSV hospitalization are high among Medicaid and ethnic/minority populations1,4–10. For infants <32 wGA and ≤6 months CA, we used RSV hospitalization data from a recent meta-analysis of 10 studies examining prophylaxis with palivizumab vs no prophylaxis55. Checchia et al.55 found a 10.6% RSV hospitalization background rate for pre-term infants ≤32 wGA. The meta-analysis, however, did not specify rates for the other cohorts in our model. Therefore, we used the estimate of 7.1% for 32–34 wGA infants with 2009 AAP risk factors and 32–35 wGA infants with 2006 AAP risk factors, and a 1.7% RSV hospitalization rate was used for 32–35 wGA infants with ≤1 risk factor based on the moderate- and low-risk groups from the PICNIC study in Canada37.

Table 1.  Probabilities used in the model by palivizumab prophylaxis and no prophylaxis.

Variable	Infant population	Base case input for variable	Source for base case inputs	Sensitivity analysis range for base case input	Source for sensitivity analysis range
Placebo
Background risk of RSV hospitalization	<32 wGA, ≤6 mo CA	18.5%	*	9.4%–34.2%	9.4%: Boyce et al.15 34.2%: Boyce et al.15
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	13.4%	*	8%–34.2%	8%: Boyce et al.15 34.2%: Boyce et al.15
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF	3.4%	*	1.7%–6.6%	1.7%: Sampalis et al.37 6.6%: Wegner et al.23
Background risk of ICU admission among infants with RSV hospitalization	<32 wGA, ≤6 mo CA	19.5%	Forbes et al.58	17.8%–48.4%	17.8%: Figueras-Aloy et al.32 48.4%: Horn et al.59
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	21.9%	Forbes et al.58	17.8%–48.4%	17.8%: Figueras-Aloy et al.32 48.4%: Horn et al.59
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF
Mean wGA	<32 wGA, ≤6 mo CA	28.6	US Natality data64	–	–
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	33.3		–	–
	32–35 wGA, ≤6 mo CA with 2006 AAP RF	34.1		–	–
	32–35 wGA, ≤6 mo CA with ≤1 RF
Mean length of stay of RSV hospitalization, d	<32 wGA, ≤6 mo CA	7.0	Forbes et al.58	6.2–7.8	Forbes et al.58
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	10.4	Forbes et al.58	9.6–11.2	Forbes et al.58
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF
Case fatality rate among infants admitted to pediatric ICU with RSV	<32 wGA, ≤6 mo CA	4.3%	Gunville et al.60	3%–8.11%	3%: Yount et al.75 8.11%: Sampalis et al.76
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	6.1%	Gunville et al.60	3%–8.11%	3%: Yount et al.75 8.11%: Sampalis et al.76
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF
Number of hours lost – RSV hospitalization (per day)	All groups	8	Assumption	6.4–9.6	Assumption (±20%)
Palivizumab
RRR of RSV hospitalization	<32 wGA, ≤6 mo CA	75.1%	Post hoc analysis of IMpact-RSV Clinical Study20	27.2%–91.5%	Post hoc analysis of IMpact- RSV Clinical Study20
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	82.2%	Post hoc analysis of IMpact-RSV Clinical Study20	45.4%–94.2%	Post hoc analysis of IMpact- RSV Clinical Study20
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF
RRR of ICU admission among hospitalized infants	All groups	63.0%	Post hoc analysis of IMpact-RSV Clinical Study20	0%–94.8%	Post hoc analysis of IMpact- RSV Clinical Study20

AAP = American Academy of Pediatrics; CA = chronologic age; ICU = intensive care unit; LRI = lower respiratory tract infection; RF = risk factor; RRR = relative risk reduction, RSV = respiratory syncytial virus; wGA = weeks gestational age.

*We first estimated the background RSV hospitalization rates for <32 wGA, ≤6 mo CA (10.7%); 32–34 wGA, ≤3 mo CA, 2009 AAP RF (7.1%); 32–35 wGA, ≤6 mo CA, 2006 AAP RF (7.1%); and 32–35 wGA, ≤6 mo CA, ≤1 RF (1.7%) from Checchia et al.55 (10.7%) and Sampalis et al.37 (7.1% and 1.7%). To obtain rates of RSV hospitalization for these cohorts covered by Medicaid, we converted these probabilities to odds (O = P/[1 − P], where O = odds and P = probability). We then multiplied the baseline odds by the odds ratio (OR) of 2.03 (95% CI, 1.99–2.06), the increased odds of RSV hospitalization among infants from the Medicaid California population compared with the non-Medicaid California population.9 Finally, we converted the transformed OR to probabilities of RSV hospitalization for Medicaid infants <32 wGA, ≤6 mo CA (18.5%); 32–34 wGA, ≤3 mo CA with 2009 AAP RF (13.43%); 32–35 wGA, ≤6 mo CA with 2006 AAP RF (13.43%); and 32–35 wGA, ≤6 mo CA with ≤1 RF (3.4%).

Because Medicaid infants use hospitalization services frequently56,57, we adjusted the expected background RSV hospitalization rates for the four groups of premature infants by converting the original background RSV hospitalization rates (10.7%, 7.1%, 1.7%) to odds (O = P/[1 − P], where O = odds and P = probability). We then multiplied the baseline odds by 2.03 (95% CI = 1.99–2.06), which represented the increased odds of RSV hospitalization among Medicaid compared with non-Medicaid infants and children in California9. Few studies have compared the rates of RSV hospitalization among Medicaid-insured infants compared with commercially-insured infants. Frogel et al.16 also found increased odds of RSV hospitalization among Medicaid vs non-Medicaid infants (odds ratio = 1.76; p < 0.0001), although this particular study was a prospective cohort registry enrolling those who received ≥1 dose of palivizumab during any RSV season (2000–2004) in the US. These revised odds ratios were then transformed back to probabilities of RSV hospitalization for Medicaid infants using the formula P = O/(1 + O). The odds of RSV hospitalization were 18.5% (<32 wGA), 13.4% (32–34 wGA with 2009 AAP risk factors and 32–35 wGA with 2006 AAP risk factors), and 3.4% (32–35 wGA with ≤1 risk factor) in our Medicaid population.

Intensive care unit admission

We used data from a recently published study that examined the characteristics of RSV hospitalizations among high-risk infants58 to inform our model on the rates of ICU admission among those hospitalized with RSV. Specifically, Forbes et al.58 examined the average cost of RSV hospitalization across several RSV seasons and characterized the RSV hospitalization (length of stay, ICU admission, readmissions within 30 days) among premature infants compared with full-term infants. Approximately 20% of premature infants hospitalized for RSV required an ICU admission58. For infants <32 wGA, we used a 19.5% ICU admission rate based on the rates among early–pre-term infants (<33 wGA) in the Forbes et al.58 study. We used a 21.9% ICU admission for infants of 32–34 wGA with 2009 AAP risk factors, 32–35 wGA with 2006 AAP risk factors, and 32–35 wGA with ≤1 risk factor based on the rates among late–pre-term infants in the Forbes et al.58 study (33–36 wGA). This is lower than the 30–48.4% rates of ICU admission among premature infants with RSV lower respiratory tract infection (LRI) previously described by Horn and Smout59. Using probabilities from Forbes et al.58 instead of Horn and Smout59 is more conservative because it biases the model towards no prophylaxis.

Case fatality in intensive care unit

We modeled a case fatality rate in the ICU of 6.1% for infants of 32–34 wGA with 2009 AAP risk factors, 32–35 wGA with 2006 AAP risk factors, and 32–35 wGA with ≤1 risk factor and 4.3% for <32 wGA as per Gunville et al.60 This study examined 271 patients who were admitted to the pediatric ICU with respiratory distress, the majority with RSV LRI. These data are also consistent with an analysis by Thorburn et al.61 who observed an RSV pediatric ICU case fatality rate of 4.1% among premature infants hospitalized in the UK. We chose the point estimates from Gunville et al.60 because they are more recent, germane to pre-term populations of interest, and are US data. We only modeled mortality for those requiring an ICU stay. By choosing to incorporate mortality only in an ICU setting and not incorporating mortality for all other patients, we are using a conservative approach.

Palivizumab effectiveness

The relative risk reduction (RRR) for RSV hospitalization with palivizumab used in the model was from the post-hoc analysis of infants of ≤35 wGA without CLDP from the IMpact-RSV Clinical Study3,20. Specifically, we used a 75.1% RRR in RSV hospitalization rates for infants ≤32 wGA (1.78% palivizumab; 7.14% placebo). We used an 82.2% RRR (1.78% palivizumab; 10% placebo)3,20 for infants of 32–34 wGA with 2009 AAP risk factors, 32–35 wGA with 2006 AAP risk factors, and 32–35 wGA with ≤1 risk factor.

Relative risk reduction on intensive care unit admission among hospitalized infants

We incorporated a 63% RRR (11.1% palivizumab [2/13] vs 30% placebo [12/31]) in the incidence of ICU admissions among those who were hospitalized with RSV based on post-hoc analysis of infants ≤35 wGA without CLDP from the IMpact-RSV Clinical Study3,20. The 63.0% RRR in ICU admissions was used across all cohorts in the model.

Costs

Costs were categorized as palivizumab prophylaxis costs and RSV-related direct costs and indirect costs (Table 2). Palivizumab prophylaxis costs included prophylaxis acquisition and administration costs, direct non-medical costs (i.e., transportation costs), and indirect costs (i.e., caregiver time lost from work and usual activities attributable to prophylaxis). RSV direct costs were specific to a Medicaid population and based on an analysis of total healthcare costs within 12 months of an initial RSV episode using a national Medicaid claims database24. RSV indirect costs consisted of caregiver time lost from work and usual activities attributable to severe RSV. Costs have been adjusted to US 2010 dollars, when necessary, using the medical care services component of the consumer price index62.

Table 2.  Cost of drug administration and RSV-related outcomes adjusted to 2010.

Variable	Infant population	Base case inputs	Source for base case inputs	Sensitivity analysis range	Source for sensitivity analysis range
Direct medical costs 12 mo after RSV infection compared with placebo	<32 wGA, ≤6 mo CA	$39,399	Shi et al24	$31,519–$47,279	Assumption (±20%)
	32–34 wGA, ≤3 mo CA with 2009 AAP RF	$20,160	Shi et al24	$16,128–$24,192	Assumption (±20%)
	32–35 wGA, ≤6 mo CA with 2006 AAP RF
	32–35 wGA, ≤6 mo CA with ≤1 RF
Cost of palivizumab	All groups	100 mg, $1190.97 50 mg, $631.10	Estimated average public spend for palivizuma20,27	100 mg, $952.77–$1429.16 50 mg, $504.88–$757.32	Assumption (±20%)
				For maximum value: WAC scenario, 100 mg, $2027.80 50 mg, $1073.88	Thomson Healthcare, Inc.77
Administration, prophylaxis	All groups	$38	CPT 96372	$30–$43	Range for CPT code
Average hourly cost for the time lost due to RSV infection (indirect cost)	All groups	$19	US Department of Labor62	$15–$23	Assumption (±20%)
Transportation costs	All groups	$6	Luce et al65	$5–$7	Assumption (±20%)

AAP = American Academy of Pediatrics; CA = chronologic age; CPT = current procedural terminology; RSV = respiratory syncytial virus; WAC = wholesale acquisition cost; wGA = weeks gestational age.

Note: owing to weight-based dosing of palivizumab, the model calculations are based on a cost-per-mg value that is derived from the average cost per mg for 50-mg and 100-mg vials of palivizumab, respectively.

The recommended dosage of palivizumab is 15 mg/kg delivered intramuscularly monthly during the RSV season. We modeled dosing regimens that started in November lasting through the end of March. To determine the cost of prophylaxis, the model determined the number of doses an infant received based on their month of birth and predicted their corresponding chronologic and gestation-adjusted age at each monthly administration of prophylaxis for the general population of infants. Dosing stopped at the end of March.

The amount of each dose (15 mg/kg) depends on the infant’s weight at prophylaxis administration, which is predicted by the chronologic and gestation-adjusted age (i.e., the infant’s CA minus the number of weeks of prematurity) at prophylaxis administration using gestational age-adjusted infant growth curves63. More specifically, the 50th percentile growth curve for very low birth weight infants (defined as birth weight ≤1500 g) was used to predict patient weight at time of administration for infants <32 wGA63. The 50th percentile growth curve for low-birth weight infants (defined as birth weight 1501–2500 g) was used to predict patient weight for moderately premature infants63. To adjust CA for prematurity, the mean gestational age for each cohort is required. Using the 2006 US Natality data64, our estimate for the mean gestational age for infants <32 wGA was 28.6 weeks. The mean gestational age was 33.3 wGA for infants 32–34 wGA with 2009 AAP risk factors. The mean gestational age was 34.2 wGA for infants 32–35 wGA with 2006 AAP risk factors and 32–35 wGA with ≤1 risk factor. Because of the weight-based dosing of palivizumab, the model calculations are based on a cost-per-milligram value that is derived from the average cost per milligram for 100-mg and 50-mg vials of palivizumab, respectively.

The model includes an estimated average public payer spend, which was calculated by taking the average reimbursement methodology of the top 10 palivizumab-using states in 2010 under Medicaid programs (ingredient cost and dispense fee) minus the unit rebate amount from the fourth quarter of 201020,27. Providers who dispense outpatient drugs to Medicaid patients are reimbursed by the state Medicaid based on a formula determined by the individual state program. The reimbursement formula varies from state to state and often contains two components that are denoted as ingredient cost and dispensing fee, which are combined to produce a single total reimbursement to the provider. In addition, federal law mandates that state Medicaid programs receive a significant rebate directly from the drug manufacturer for each unit of product dispensed.

RSV-related costs

Direct costs associated with RSV were based on a recent analysis of utilization and cost data from a national Medicaid database, including healthcare data from patients across 12 states24,45. A retrospective cohort study was conducted among a population of Medicaid-insured premature infants (33–36 wGA) with RSV LRI (n = 9180) and without RSV LRI (n = 19,110) during two RSV seasons (2003–2005). The primary outcome was average total healthcare cost within the 1-year period following the first RSV LRI diagnosis date. In this population of Medicaid-insured infants who were hospitalized with RSV, the total direct medical costs within 1 year after an RSV LRI hospitalization episode was $22,288 and significantly higher compared with the 1-year costs among a comparison group of infants without RSV LRI ($4823), representing an incremental cost difference of $17,465 (p < 0.001)24,45.

Our cost-effectiveness model of palivizumab was from the societal perspective and, therefore, included indirect costs. Indirect costs associated with RSV included caregiver time lost from work or usual activities (8 h per hospitalization day) resulting from RSV hospitalization. The cost of caregiver time off work was based on the hourly compensation from the US Department of Labor62. Transportation costs were taken from Luce et al.65

Quality-adjusted life-years and utilities

Health utilities were applied in our economic analysis. Similar to economic models for vaccines, our model used a baseline utility of 1.0 for infants29,48–52. The utility decrement for RSV hospitalization and post-hospitalization recovery period was estimated in a prospective controlled study of children and parents with and without RSV hospitalization66 (Table 3). To estimate total QALYs lost, we used the hospital length of stay for estimating the QALYs lost from an RSV hospitalization and a 60-day period for estimating QALYs lost from the post-hospitalization recovery period66.

Table 3.  Health utilities incorporated in the cost-effectiveness model.

Variable	Infant population	Base case inputs	Source for base case inputs	Sensitivity analysis range	Source for Sensitivity analysis range
Baseline utility	All groups	1	Assumption	–	–
Hospitalization disutility	All groups	− 0.3972	Leidy et al.66	−0.3534 to −0.4520	Leidy et al.66
Post-hospitalization recovery period disutility	All groups	−0.0911	Leidy et al.66	−0.0728 to −0.1093	Assumption (±20%)
Duration, post-hospitalization recovery period, d	All groups	60	Leidy et al.66	48–72	Assumption (±20%)
Age at time of RSV hospitalization, mo	All groups	2.4	Palmer et al.45	1.9–2.9	Assumption (±20%)
Average life expectancy of general US population, y	All groups	77.45	Centers for Disease Control and Prevention46	58.1–96.8	Assumption (±25%)

RSV = respiratory syncytial virus.

Analyses

The primary outcome was the ICER, calculated as the difference between the average costs for each group divided by the difference between the average QALYs for each group. An ICER within the range of other pediatric preventative therapies such as currently recommended vaccines ($0–$157,000 QALY)49,67–74 was considered cost-effective.

Sensitivity analyses

Extensive one-way and probabilistic sensitivity analyses were conducted to ascertain which variables were most influential in the model. Variables that had a substantial impact on the model results included background risk of RSV hospitalization, societal cost of palivizumab in the public sector, and palivizumab efficacy. The parameter ranges (minimum and maximum values) used were identified from the literature or through post-hoc analysis of palivizumab clinical trial studies wherever possible.

The background risk for RSV hospitalization was an influential variable in previous economic models. There are no Medicaid studies that have examined the risk of RSV hospitalization among the exact cohorts in our model; therefore, the minimum and maximum values for the background rates of RSV hospitalization used in sensitivity analysis were based on extrapolation on the available evidence in the literature. The minimum background risk of RSV hospitalization for infants <32 wGA, ≤6 months CA was taken from the Boyce et al.15 study, which examined rates of RSV hospitalization among infants and children in Tennessee Medicaid. Boyce et al. found that the rate of RSV hospitalization among infants of ≤28 wGA (0 to <6 months) was 9.4%. We used a maximum value (34.2%) from Boyce et al.15, that reflects the average rates of RSV hospitalization between CLDP and CHD infants (0–6 months). We assumed that the upper-bound parameter estimate of RSV hospitalization for infants <32 wGA could be similar to rates among infants with comorbidities given that extremely premature infants have immature lung development and function. In addition, a majority of these infants are exposed to mechanical ventilation which introduces additional trauma to the lungs. The minimum background rate of RSV hospitalization for infants of 32–34 wGA with 2009 AAP risk factors and 32–35 wGA with 2006 AAP risk factors (8%) was taken from Boyce et al.15 and represented the RSV hospitalization rate of infants 33 to <36 wGA (0–6 months). We also used the RSV hospitalization rate of 34.2% from Boyce et al.15 to represent the upper-bound parameter estimate of RSV hospitalization. For 32–35 wGA with ≤1 risk factor, we used 1.7% from Sampalis et al.37 to represent the lower-bound parameter estimate and 6.6% from Wegner et al.23 to represent the upper-bound parameter estimate. The 6.6% RSV hospitalization rate was among infants 32–35 wGA who did not receive RSV prophylaxis under North Carolina Medicaid. The minimum and maximum values for RRR of RSV hospitalization were taken from the 95% CI of the post-hoc analysis of the IMpact-RSV Clinical Study3,20.

Data regarding case fatality rates among RSV infants in the ICU setting are few. For a lower-bound parameter estimate, we examined a 3.0% mortality rate among RSV hospitalized infants (CHD infants <2 years of age) based on Yount et al.75 For an upper-bound parameter estimate, we used an 8.11% mortality rate after RSV hospitalization among premature infants born at 32–35 wGA from Sampalis et al.76

The estimated societal public spend for palivizumab was an influential variable in the model. We calculated ±20% of the estimated societal public spend for palivizumab. Furthermore, we examined the 2010 wholesale acquisition cost (WAC) for palivizumab to represent an upper-bound parameter estimate in sensitivity analysis (100-mg vial, $2027.80; 50-mg vial, $1073.88)77. The values for other variables in one-way sensitivity analyses are presented in Tables 1–3.

Probabilistic sensitivity analysis

A probabilistic sensitivity analysis (PSA)78 was also conducted across all model parameters simultaneously. Monte Carlo techniques were used according to pre-specified model distribution assumptions (beta, gamma, normal, and lognormal) based on the characteristics of each parameter78. The background risk of RSV hospitalization, the background risk of ICU admission, case fatality rate in the ICU, drug wastage, and discount rates were based on beta distributions. The RRR of RSV hospitalization and RRR of ICU admission were based on log-normal distributions. Cost data were based on gamma distributions because these take into account the likely skew and variability with these parameters. The estimation of the distribution ranges was based on the published data or, if not reported, a confidence interval was assumed to equal 20% of the deterministic value and the standard errors were calculated. Results of the probabilistic sensitivity analysis were based on 1000 Monte Carlo simulations.

Results

Palivizumab prophylaxis was cost-effective compared with no RSV prophylaxis among high-risk infants. In the base-case analysis, infants <32 wGA and ≤6 months CA who received prophylaxis with palivizumab had lower average costs of care compared with those without prophylaxis, resulting in a cost savings of $2339 (Table 4). The QALY gained for infants <32 wGA was 0.046. Therefore, among infants <32 wGA and ≤6 months CA, palivizumab was a dominant strategy compared with no prophylaxis (i.e., saved costs and improved QALY). Infants of 32–34 wGA and ≤3 months CA with 2009 AAP risk factors who received RSV prophylaxis had higher average costs of care compared with those not receiving prophylaxis for a net cost increase of $864 per infant. Infants with palivizumab prophylaxis had, on average, fewer hospitalization costs than infants without palivizumab prophylaxis ($519 vs $2921, respectively). Prophylaxis with palivizumab resulted in an additional 0.0539 QALY gained per infant. The incremental cost per (additional) QALY gained was $16,037. For infants 32–35 wGA, ≤6 months CA with 2006 AAP risk factors, there was a net increase cost of $2060 and a net QALY gained of 0.0539, resulting in an ICER of $38,244. In contrast, for infants 32–35 wGA, ≤6 months with ≤1 risk factor, there was a net increase cost of $3853 and a net QALY gained of 0.014, resulting in an ICER of $281,892. Results for these populations are presented in the context of ICERs for various vaccines in the US49,70–74 (Figure 2).

Figure 2.  Benchmark of palivizumab cost-effectiveness results with those of vaccines. AAP = American Academy of Pediatrics; DTaP = diptheris, tetanus, and acellular pertussis vaccine; HPV = human papillomavirus vaccine; RF = risk factor; TIV = trivalent inactivated influenza vaccine; wGA = weeks gestational age.

Table 4.  Net costs and QALYS between palivizumab prophylaxis and no prophylaxis among the 4 cohorts.

	<32 wGA, ≤6 mo CA*			32–34 wGA, ≤3 mo CA* with AAP 2009 RF^†			32–35 wGA, ≤6 mo CA* with AAP 2006 RF^‡			32–35 wGA, ≤6 mo CA* with ≤1 RF^$
	Palivizumab	No Prophylaxis	Difference	Palivizumab	No Prophylaxis	Difference	Palivizumab	No Prophylaxis	Difference	Palivizumab	No Prophylaxis	Difference
Prophylaxis costs	$3287	0	$3287	$3266	0	$3266	$4462	0	$4462	$4462	$0	$4462
RSV-related costs	$1867	$7493	−$5627	$519	$2921	−$2402	$519	$2921	−$2402	$132	$741	−$609
Total costs	$5154	$7493	−$2339	$3785	$2921	$864	$4981	$2921	$2060	$4594	$741	$3853
QALY	30.907	30.861	0.046	30.908	30.854	0.0539	30.908	30.854	0.0539	30.911	30.897	0.014
ICER	–	–	Dominant	–	–	$16,037	–	–	$38,244	–	–	$281,892

AAP = American Academy of Pediatrics; CA = chronologic age; ICER = incremental cost-effectiveness ratio calculated as incremental costs/incremental QALYs; QALY = quality-adjusted life-year; RF = risk factor; RSV = respiratory syncytial virus; wGA = weeks gestational age.

*Chronologic age at the start of the RSV season (November 1).

^†Having siblings <5 y of age and/or attending day care.

^‡Having 2 of the following: environmental air pollutants, congenital abnormalities of the airways, severe neuromuscular disease, school-aged siblings, or day care attendance.

^$Risk factors described in Sampalis et al37.

One-way and probabilistic sensitivity analyses

One-way sensitivity analysis determines the impact of each variable on baseline cost-effectiveness results (Tables 5–8 in the Appendix). Overall, the baseline results for palivizumab remained robust in one-way sensitivity analyses across all high-risk cohorts. Greater variability in sensitivity analysis results were observed among infants of 32–35 wGA with <1 risk factor. In infants <32 wGA, ≤6 months CA, palivizumab prophylaxis was either dominant or within accepted CEA thresholds in one-way sensitivity analyses. Palivizumab was cost-effective in infants 32–34 wGA with 2009 AAP risk factors and infants 32–35 wGA with 2006 AAP risk factors when the background RSV hospitalization rate was decreased from 13.4% to 8% (i.e., the minimum parameter estimate). Baseline cost-effectiveness results were also within CEA thresholds when using 2010 WAC palivizumab costs (i.e., the maximum parameter estimate without incorporating manufacturer rebates). Overall, the background risk of RSV hospitalization, the cost of palivizumab, and the efficacy of palivizumab were some of the most influential variables on the base case results across cohorts.

Table 5.  One-way sensitivity analysis for infants <32 wGA, ≤6 months CA.

	Baseline results		Dominant
	Input		ICER
	Minimum	Maximum	Input minimum	Input maximum
RRR of RSV hospitalization (75.1%)	27.2%	91.5%	$35,146	Dominant
Background risk of RSV hospitalization (18.5%)	9.4%	34.2%	$18,366	Dominant
Discount rate for outcomes (3%)	0%	5%	Dominant	Dominant
Average cost of palivizumab per mg* ($12.27)	$9.81^†	$14.72^† $20.88^‡	Dominant	Dominant Dominant
Increased direct medical costs 12 mo after RSV infection compared with placebo ($39,399)	$31,519	$47,279	Dominant	Dominant
Probability of case fatality in ICU (4.3%)	3.0%	8.11%	Dominant	Dominant
Background risk of ICU admission (19.5%)	17.8%	48.4%	Dominant	Dominant
RRR of ICU admission (63%)	0%	94.8%	Dominant	Dominant

CA = chronologic age; ICER = incremental cost-effectiveness ratio; ICU = intensive care unit; RRR = relative risk reduction; RSV = respiratory syncytial virus; wGA = weeks gestational age.

Dominant means that RSV prophylaxis administration is cost-saving and improves health outcomes compared to a no RSV prophylaxis strategy.

*Owing to weight-based dosing of palivizumab, the model calculations are based on a cost-per-mg value that is derived from the average cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^†Derived from ±20% of the average base case cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^‡Derived from the average 2010 wholesale acquisition costs for the 100-mg vial ($2027.80) and 50-mg vial ($1073.88), respectively.

Table 6.  One-way for infants 32–34 wGA, ≤3 months CA with 2009 AAP RF.

	Baseline results		$16,037
	Input		ICER
	Minimum	Maximum	Input minimum	Input maximum
Background risk for RSV hospitalization (13.4%)	8%	34.2%	$56,964	Dominant
Cost of palivizumab per mg* ($12.27)	$9.81^†	$14.72^† $20.88^‡	$5077	$26,997 $54,521
RRR of RSV hospitalization (82.2%)	45.4%	94.2%	$42,981	$9062
Discount rate for outcomes (3%)	0%	5%	$6539	$23,588
Increased direct medical costs 12 mo after RSV infection compared with placebo ($20,160)	$16,128	$24,192	$24,284	$7789
Background risk of ICU admission (21.9%)	17.8%	48.4%	$19,467	$7476
Probability of ICU death (6.1%)	3%	8.11%	$30,713	$12,149
RRR ICU admission (63%)	0%	94.8%	$18,088	$15,167

AAP = American Academy of Pediatrics; CA = chronologic age; ICER = incremental cost-effectiveness ratio; ICU = intensive care unit; RF = risk factor; RRR = relative risk reduction; RSV = respiratory syncytial virus; wGA = weeks gestational age.

Dominant means that RSV prophylaxis administration is cost-saving and improves health outcomes compared to a no RSV prophylaxis strategy.

*Owing to weight-based dosing of palivizumab, the model calculations are based on a cost-per-mg value that is derived from the average cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^†Derived from ±20% of the average base case cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^‡Derived from the average 2010 wholesale acquisition costs for the 100-mg vial ($2027.80) and 50-mg vial ($1073.88), respectively.

Table 7.  One-way for infants 32–35 wGA, ≤6 months CA with 2006 AAP RF.

	Baseline results		$38,244
	Input		ICER
	Minimum	Maximum	Input minimum	Input maximum
Background risk for RSV hospitalization (13.4%)	8%	34.2%	$94,162	Dominant
Cost of palivizumab per mg* ($12.27)	$9.81^†	$14.72^† $20.88^‡	$23,034	$53,455 $91,655
Probability of case fatality in ICU (6.1%)	3.0%	8.11%	$73,244	$28,974
Discount rate for outcomes (3%)	0%	5%	$15,594	$56,252
RRR of RSV hospitalization (82.2%)	45.4%	94.2%	$69,489	$30,157
Background risk of ICU admission (21.9%)	17.8%	48.4%	$46,425	$17,830
Increased direct medical costs 12 mo after RSV infection compared with placebo ($20,160)	$16,128	$24,192	$46,492	$29,997
RRR of ICU admission (63%)	0%	94.8%	$43,135	$36,170

AAP = American Academy of Pediatrics; CA = chronologic age; ICER = incremental cost-effectiveness ratio; ICU = intensive care unit; RRR = relative risk reduction; RSV = respiratory syncytial virus; wGA = weeks gestational age.

*Owing to weight-based dosing of palivizumab, the model calculations are based on a cost-per-mg value that is derived from the average cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^†Derived from ±20% of the average base case cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^‡Derived from the average 2010 wholesale acquisition costs for the 100-mg vial ($2027.80) and 50-mg vial ($1073.88), respectively.

Table 8.  One-way sensitivity analysis for infants 32–35 wGA, ≤6 months CA with ≤1 RF.

	Baseline results		$281,892
	Input		ICER
	Minimum	Maximum	Input minimum	Input maximum
Background risk of RSV hospitalization (3.4%)	1.7%	6.6%	$608,380	$123,595
Case fatality in the ICU (6.1%)	3%	8.11%	$539,866	$213,560
Discount rate for outcomes (3%)	0%	5%	$114,944	$414,623
Cost of palivizumab per mg* ($12.27)	$9.81^†	$14.72^† $20.88^‡	$221,945	$341,839 $492,392
Background risk of ICU admission (21.9%)	17.8%	48.4%	$342,188	$131,419
RRR of RSV hospitalization (82%)	45.4%	94.2%	$360,314	$261,592
Increased direct medical costs 12 mo after RSV infection compared with placebo ($20,160)	$16,128	$24,192	$290,140	$273,645
RRR of ICU admission (63%)	0%	94.8%	$317,941	$266,602

CA = chronologic age; ICER = incremental cost-effectiveness ratio; ICU = intensive care unit; RF = risk factor; RRR = relative risk reduction; RSV = respiratory syncytial virus; wGA = weeks gestational age.

*Owing to weight-based dosing of palivizumab, the model calculations are based on a cost-per-mg value that is derived from the average cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^†Derived from ± 20% of the average base case cost per mg for 100-mg and 50-mg vials of palivizumab, respectively.

^‡Derived from the average 2010 wholesale acquisition costs for the 100-mg vial ($2027.80) and 50-mg vial ($1073.88), respectively.

Probabilistic sensitivity analysis was conducted to allow the joint effect of parameter uncertainty to be simultaneously assessed (Figures 3–10 in the Appendix). The scatterplot reflects the incremental cost and incremental QALYs generated for palivizumab compared with no prophylaxis for each run of the probabilistic sensitivity analysis. The cost-effectiveness acceptability curves present the probability of cost-effectiveness at different willingness-to-pay thresholds for an additional QALY. For <32 wGA infants, 6 months CA, palivizumab was found to be cost-effective at all willingness-to-pay thresholds for an additional QALY ($0–$300,000 QALY). For infants 32–34 wGA with 2009 AAP risk factors and infants 32–35 wGA with 2006 AAP risk factors, palivizumab was found to be cost-effective the majority of the time at relatively low willingness-to-pay thresholds for an additional QALY. For infants 32–35 wGA with ≤1 risk factor, palivizumab was generally not cost-effective.

Figures 3 and 4.  Probabilistic analyses <32 wGA, ≤6 months CA. CA = chronologic age; QALY = quality-adjusted life-year; wGA = weeks gestational age.

Figures 5 and 6.  Probabilistic analyses for 32–34 wGA, ≤3 months CA with AAP 2009 RF. AAP = American Academy of Pediatrics; CA = chronologic age; QALY = quality-adjusted life-year; RF = risk factor; wGA = weeks gestational age.

Figures 7 and 8.  Probabilistic analyses 32–35 wGA, ≤6 months CA with AAP 2006 RF. AAP = American Academy of Pediatrics; CA = chronologic age; QALY = quality-adjusted life-year; RF = risk factor; wGA = weeks gestational age.

Figures 9 and 10.  Probabilistic analyses 32–35 wGA, ≤6 months CA with ≤1 RF. CA = chronologic age; QALY = quality-adjusted life-year; RF = risk factor; wGA = weeks gestational age.

Discussion

To our knowledge, this is the first cost-utility analysis of palivizumab in a Medicaid population. It has been conducted according to the guidelines set forth by the US Panel on Cost-Effectiveness in Health and Medicine29. We found that palivizumab was cost-effective among various high-risk premature infants without CLDP or CHD in a Medicaid population. Palivizumab was cost-effective compared with no RSV prophylaxis among infants <32 wGA, ≤6 months CA; infants 32–34 wGA, ≤3 months CA with 2009 AAP risk factors; and infants 32–35 wGA, ≤6 months CA with 2006 AAP risk factors. Palivizumab was not cost-effective among infants 32–35 wGA, ≤6 months CA with ≤1 risk factor. Results were marginally favorable among infants with 32–34 wGA, ≤3 months CA with 2009 AAP risk factors compared with 32–35 wGA, ≤6 months CA with 2006 AAP risk factors because the former cohort of infants weigh less than their older counterparts, resulting in lower palivizumab costs. Our cost-effectiveness analysis of palivizumab is the first to examine 2006 and 2009 AAP eligibility guidelines for premature infants within the 32–35 wGA group. Our findings indicate that palivizumab is cost-effective in both a broader (2006 AAP) and narrower (2009 AAP) group of 32–35 wGA infants covered by Medicaid who have multiple risk factors. Given the favorable cost-effectiveness findings for both groups, future policy should focus on identifying an algorithm that captures most high-risk infants in this GA group.

The methodological approach used in previous economic analyses of palivizumab in Medicaid populations focused on measuring cost per RSV hospitalization avoided and did not estimate QALYs. These studies concluded that RSV prophylaxis was not cost-effective in their respective Medicaid populations21–23. These studies, however, did not examine the impact of ICU care and other direct medical costs associated with severe RSV disease. A recent systematic review of cost-effectiveness analyses of palivizumab conducted by Smart and colleagues79 found that palivizumab was cost-effective when health utilities were incorporated in the model, whereas studies that used observational data instead of clinical trial data and did not incorporate health utilities tended to have lower measured study quality using criteria based on the Quality of Health Economic Studies79. A cost-effectiveness result represented by QALY is the recommended methodology by health economists that allows comparisons of the value of a drug or treatment across multiple disease areas28,29.

Previous economic models of palivizumab estimated the burden of serious RSV disease using claims data and examined International Classification of Diseases, Ninth Edition, RSV-specific codes only21. This methodology underestimates the rate of RSV hospitalization because most of these hospitalizations are billed in claims data as unspecified bronchiolitis or unspecified pneumonia. Flaherman et al80 recently published RSV testing rates from Kaiser Permanente in Northern California and estimated that 65% to 85% of severe RSV disease is underdiagnosed. Low rates of RSV testing are not surprising given that the American Academy of Pediatrics does not routinely recommend RSV testing for clinically diagnosed bronchiolitis because the knowledge gained from such testing rarely alters management decisions or outcomes18.

Furthermore, many of the previous economic studies did not incorporate federally mandated manufacturer rebates under the Medicaid program. A task force of health economists from the International Society of Pharmacoeconomics suggests that with public health programs, “the drug cost must represent the actual acquisition costs, incorporating any rebates or discounts”26.

The background rate of RSV hospitalization was an influential variable in our economic analysis of palivizumab. Evidence suggests that the background risk of RSV hospitalization is higher among ethnic minorities or Medicaid infants1,4–10. Some studies have found increased prematurity rates and low-birth weight among minorities, including African Americans81, and prematurity status increases the risk of severe RSV disease. Ethnic minorities also face increased exposure to potential risk factors for RSV, such as tobacco exposure, as reported in the 1998 Surgeon General’s Report on Tobacco Use Among U.S. Racial/Ethnic Minority Groups82. African Americans (37.6%) and American Indians and Alaska natives (36.95%) were more likely than other groups to report exposure to environmental tobacco smoke at home. The burden of RSV hospitalization among Medicaid infants may also be due to issues related to low compliance rates with RSV prophylaxis83.

The cost-effectiveness results of RSV prophylaxis in our base-case analysis for AAP guideline-eligible, high-risk infants were comparable with those of recent pediatric vaccine intervention studies. Our results for <32 wGA, <6 month CA are similar to DTaP vaccine, ie, both are cost-saving and increase health outcomes compared with no intervention. Ray et al84 examined the benefits of pneumococcal conjugate vaccination among children ≤23 months of age and found cost-effectiveness ratios of $112,000 per life-year saved without herd immunity and $7500 with herd immunity. Similarly, Goldie et al85 found that the most cost-effective strategy for human papillomavirus 16/18 vaccine was starting vaccination at 12 years of age combined with triennial conventional cytologic screening beginning at 25 years of age, which had a cost-effectiveness ratio of $60,000 per QALY. Finally, Shepard et al86 examined the strategy of universal conjugate meningococcal vaccination of all US adolescents starting at 11 years of age and found the cost-effectiveness ratio to be $121,000 per life-year saved, whereas the cost-effectiveness of meningococcal vaccine (2 doses, all 11- and 16-year-olds) was found to be $157,000 cost per QALY gained73. Our palivizumab cost-effectiveness results for infants 32–34 wGA, ≤3 months CA with 2009 AAP risk factors and infants 32–35 wGA, ≤6 months CA with 2006 AAP risk factors are similar to the lower ICER results of a number of recommended vaccines by both the Advisory Committee on Immunization Practices and the AAP.

Our analysis makes several important conservative assumptions. First, we did not attribute a reduction in hospital length of stay resulting from palivizumab prophylaxis87. Second, the model did not include RSV disease presenting in outpatient settings. The incidence of outpatient RSV LRI visits and associated costs can be substantial on the healthcare system45,88,89. Third, we conservatively did not assume a mortality risk reduction due to palivizumab55 and modeled a background case fatality rate among RSV ICU patients only. Fourth, we did not value parents’ quality-of-life decrements when their child was hospitalized with RSV66. Some have advocated that pediatric cost-effectiveness analyses should value the effect on both the parents and child because these outcomes are real for both parties.90 Fifth, per the Panel on Cost-Effectiveness in Health and Medicine, we did not value the lifetime earnings lost (indirect costs) of infants because we include infant quality of life. Sixth, we did not assume a reduction in recurrent wheezing91. Should any of these aforementioned conservative assumptions be included in the analysis, the cost-effectiveness profile of RSV prophylaxis with palivizumab would likely improve.

There are several limitations to this economic analysis. More evidence on RSV-specific mortality with and without palivizumab prophylaxis is needed. Non-compliance with palivizumab is a common problem in Medicaid populations, and future economic analyses can determine the cost-effectiveness of noncompliance versus compliance with palivizumab. Our economic model did not incorporate other strategies to prevent RSV infection, such as the effect of parent education, but they are likely to be consistent between prophylaxis and no-prophylaxis groups. Therefore, inclusion of these strategies should not have a major impact on the results. Our results are not generalizable to commercially insured infant populations or infants outside of the US. Despite these limitations, the strengths of our model include using Medicaid-specific healthcare data and incorporating federally mandated rebates in estimating the societal public spend for palivizumab. We were able to ascertain the important drivers of cost-effectiveness for RSV prophylaxis for each subpopulation of interest in our sensitivity analysis. Finally, we specifically utilized the efficacy in non-CLDP premature infants (75.1%–82.2% RRR)3,20.

In sum, the morbidity and associated healthcare needs of Medicaid-insured infants and children remain a public health priority in the US. This vulnerable population includes a disproportionate number of premature infants, those born with low birth weight, racial/ethnic minorities, and infants with increased exposure to environmental risk factors for RSV disease, such as tobacco exposure. Additional data are needed regarding the burden of RSV disease among the Medicaid populations in various geographic regions of the US, access to care with palivizumab, and long-term health outcomes.

CONCLUSIONS

This is the first cost-utility analysis of palivizumab in a Medicaid population. Using conservative assumptions, we conclude that RSV prophylaxis was highly cost-effective for a population of Medicaid infants <32 wGA and ≤6 months CA (palivizumab was a dominant strategy compared with no prophylaxis); 32–34 wGA, ≤3 months CA with 2009 AAP risk factors; and 32–35 wGA, ≤6 months CA with 2006 AAP risk factors. RSV prophylaxis was not cost-effective for infants 32–35 wGA, ≤6 months CA with ≤1 risk factor. Incorporating risk factor data favorably changes the cost-effectiveness profile of RSV prophylaxis because the risk of RSV hospitalization increases among infants who have >1 risk factor. Palivizumab is cost-effective in a population of Medicaid infants owing to the combination of lower cost structure for healthcare items and services, manufacturer rebates, and higher rates of RSV disease. Cost-effectiveness results for palivizumab among high-risk infants in a Medicaid population were found to be in the range of those for vaccines recommended by Advisory Committee on Immunization Practices and AAP. The de facto assumption that RSV prophylaxis is not cost-effective should be reexamined. Palivizumab administration with full-season dosing among high-risk infants in the Medicaid population was cost-effective.

Transparency

Declaration of funding:

The study was funded by MedImmune, LLC

Declaration of financial/other relationships:

PM and AM have disclosed that they are employees of MedImmune LLC and own shares in the company. MP and LW have disclosed that they have served as consultants for MedImmune for this economic analysis. MP is on MedImmune’s speakers bureau for palivizumab. CMRO Peer Reviewers on this manuscript have disclosed that they have no relevant financial relationships.

Acknowledgments:

Editorial assistance for this manuscript was provided by Susan Myers, MSc, and Gerard P. Johnson, PhD, of Complete Healthcare Communications, Inc., Chadds Ford, PA, USA, and was funded by MedImmune.