ARTICLE

Cost-effectiveness Analysis of the National Perinatal Hepatitis B Prevention Program AUTHORS: Carolina Barbosa, PhD,a Emily A. Smith, MPH,b Thomas J. Hoerger, PhD,a Nancy Fenlon, RN, MS,c Sarah F. Schillie, MD, MPH, MBA,b Christina Bradley, BS,a and Trudy V. Murphy, MDb aRTI International, Chicago, Illinois; and bDivision of Viral Hepatitis, National Center for HIV/AIDS, Viral Hepatitis, STD and TB Prevention, and cImmunization Services Division, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia

KEY WORDS cost-effectiveness, QALY, cost, perinatal, infection, hepatitis B, immunization programs ABBREVIATIONS ACIP—Advisory Committee on Immunization Practices CDC—Centers for Disease Control and Prevention CEAC—cost-effectiveness acceptability curve CI—confidence interval HBIG—hepatitis B immune globulin HBsAg—hepatitis B surface antigen HBV—hepatitis B virus HepB—hepatitis B ICER—incremental cost-effectiveness ratio PEP—postexposure prophylaxis PHBPP—Perinatal Hepatitis B Prevention Program PVST—postvaccination serologic testing QALY—quality-adjusted life-year Dr Barbosa participated in the concept and design of the study; the collection, analysis, and interpretation of data; and drafted and revised the manuscript. Ms Smith, Ms Fenlon, Dr Schillie, and Dr Murphy participated in the concept and design of the study; the collection, analysis, and interpretation of data; and revised the manuscript. Dr Hoerger participated in the concept and design of the study, the analysis and interpretation of data, and drafted and revised the manuscript. Ms Bradley participated in the collection of data and revised the manuscript. All authors approved the manuscript as submitted. www.pediatrics.org/cgi/doi/10.1542/peds.2013-0718 doi:10.1542/peds.2013-0718 Accepted for publication Nov 1, 2013 Address correspondence to Carolina Barbosa, PhD, RTI International, 230 West Monroe St, Ste 2100, Chicago, IL 606064901. E-mail: [email protected] (Continued on last page)

WHAT’S KNOWN ON THIS SUBJECT: Infant postexposure prophylaxis prevents perinatal hepatitis B (HepB) virus transmission and mortality and morbidity caused by chronic HepB virus infection. The US Perinatal Hepatitis B Prevention Program (PHBPP) identifies and manages infants born to HepB surface antigen–positive women. WHAT THIS STUDY ADDS: It presents the first estimates of the long-term costs and outcomes of postexposure prophylaxis with the PHBPP. It analyzes the effects of the PHBPP, and alternative immunization scenarios, on health and economic outcomes for the 2009 US birth cohort.

abstract OBJECTIVE: To analyze the cost-effectiveness of the national Perinatal Hepatitis B Prevention Program (PHBPP) over the lifetime of the 2009 US birth cohort and compare the costs and outcomes of the program to a scenario without PHBPP support. PHBPP’s goals are to ensure all infants born to hepatitis B (HepB) surface antigen–positive women receive timely postexposure prophylaxis, complete HepB vaccine series, and obtain serologic testing after series completion. METHODS: A decision analytic tree and a long-term Markov model represented the risk of perinatal and childhood infections under different prevention alternatives, and the long-term health and economic consequences of HepB infection. Outcome measures were the number of perinatal infections and childhood infections from infants born to HepB surface antigen–positive women, quality-adjusted life-years (QALYs), lifetime costs, and incremental cost per QALY gained. The health outcomes and total costs of each strategy were compared incrementally. Costs were evaluated from the health care system perspective and expressed in US dollars at a 2010 price base. RESULTS: In all analyses, the PHBPP increased QALYs and led to higher reductions in the number of perinatal and childhood infections than no PHBPP, with a cost-effectiveness ratio of $2602 per QALY. In sensitivity analyses, the cost-effectiveness ratio was robust to variations in model inputs, and there were instances where the program was both more effective and cost saving. CONCLUSIONS: This study indicated that the current PHBPP represents a cost-effective use of resources, and ensuring the program reaches all pregnant women could present additional public health benefits. Pediatrics 2014;133:243–253

PEDIATRICS Volume 133, Number 2, February 2014

243

Each year, an estimated 25 000 infants are born to hepatitis B surface antigen (HBsAg)-positive women in the United States.1 With no intervention, these infants have a 40% to 90% risk of hepatitis B virus (HBV) infection.2,3 Approximately 90% of infected infants develop chronic HBV infection, which carries a 25% risk of premature death from progressive damage to the liver, leading to cirrhosis, or cancer of the liver.4 Postexposure prophylaxis (PEP), consisting of hepatitis B immune globulin (HBIG) and hepatitis B (HepB) vaccine administered at birth, followed by completion of a 3-dose HepB vaccine series is 85% to 95% effective in preventing perinatal HBV infection.4 The Advisory Committee on Immunization Practices (ACIP) of the Centers for Disease Control and Prevention (CDC) recommends that all pregnant women receive HBsAg screening to ensure that infants of HBsAg-positive women receive PEP.4 ACIP recommends that infants of HBsAg-positive women receive HBIG and HepB vaccine within 12 hours of birth, complete the HepB series, and receive serologic testing (HBsAg and antibody to HBsAg) 1 to 2 months after completing the HepB vaccine series to determine outcomes (HepB immunity by vaccination, nonresponse to vaccination, or chronic HBV infection) and future management. ACIP also recommends that the small percentage (5%) of infants who fail to respond to an initial HepB vaccine series and remain uninfected receive an additional HepB vaccine series and repeat postvaccination serologic testing (PVST). In 1990, the CDC began funding Perinatal Hepatitis B Prevention Programs (PHBPP) in all 64 immunization grantees through the 317 program. PHBPP’s goals are to ensure all infants born to HBsAg-positive women receive timely PEP, complete HepB vaccine series, obtain PVST after series completion, and continue medical care appropriate 244

BARBOSA et al

to their outcomes. These goals are accomplished through case management activities. Programs report outcomes annually to the CDC.1,4 In 2008, provisional estimates were that 48% of the 25 600 annual estimated births to HBsAg-positive women were identified and case-managed by PHBPP.1 Most HBsAg-positive women and infants case-managed by PHBPP receive medical care funded by private insurance or Medicaid, although data on the exact proportions of each have not been collected.1 More information on PHBPP activities is described in Appendix A of the Supplemental Information. Although economic analyses of the burden and prevention impact of HBV transmission have been conducted,2,5–10 no study has focused on the long-term costs and outcomes of PEP through the current national PHBPP. This study analyzes the effects of the PHBPP on health and economic outcomes attributable to perinatal and early childhood infection (postnatal infection, before age 5 years), determined over the lifetime of the 2009 US birth cohort. Costs and outcomes of the current PHBPP (“PHBPP” strategy) are compared incrementally to a strategy represented by the ACIP recommendations without PHBPP support (“No PHBPP” strategy). The study’s main hypotheses were as follows1: the number of perinatal and childhood infections will be lower in the PHBPP strategy,2 the higher qualityadjusted life-years (QALYs) gained with PHBPP will justify the higher program costs, and the program will be costeffective when compared with the No PHBPP strategy.

compared and can help with understanding the value for money of each strategy.11,12 Where 1 strategy does not dominate (ie, is not both more effective and less costly), costs and effects are combined in the form of incremental cost-effectiveness ratios (ICERs), defined as the difference in costs (C) divided by the difference in mean effectiveness (E), (Cj2Ci)/(Ej2Ei), where j is a more costly strategy than i. The ICER represents the additional cost required to achieve 1 additional unit of outcome. An optimal intervention is one with an ICER that is not more than the decision maker’s intrinsic valuation for an additional unit of the outcome.11,12 In a cost-effectiveness analysis, health benefits associated with the strategies under comparison can be measured and valued by using natural units of outcome (eg, life-years), number of infections, or they can reflect the value individuals place on their health, called utilities. QALYs are a widely used measure of health benefits that incorporate mortality and morbidity estimates in a single index. Cost per QALY estimates allow a comparison between different health care interventions that compete for the same pool of resources.

Overview

This cost-effectiveness analysis compared 2 strategies. In the base-case analysis, the No PHBPP strategy assumes that there is no program targeted at identification and management of infants born to HBsAgpositive mothers and that infants do not receive HBIG or PVST. In the PHBPP strategy, a proportion of the infants born to HBsAg-positive mothers are identified prospectively and managed through the national PHBPP. Those not captured by the PHBPP follow the No PHBPP strategy.

Economic evaluation provides a framework to allocate resources to effective strategies. Full economic evaluations are evaluations where the costs and consequences of at least 2 strategies are

The costs and outcomes of the strategies analyzed in this study were calculated by using a decision-analytic model. The model was constructed in 2 parts, a decision analytic tree and a

METHODS

ARTICLE

long-term Markov model, both following a cohort simulation approach.13,14

infection. The cost-effectiveness analysis focused on the last 2 outcomes and QALYs lost with each type of infection.

Decision Analytic Tree

The decision tree model base-case parameters and sources are summarized in Table 1. The 24 784 cohort considered in this analysis was the estimated number of infants born to HBsAg-positive mothers in 2009.1,15,16 More than 95% of women were screened for HBsAg in the United States in 2003–2004.4,15 Consistent with this estimate, and so that all infants born to HBsAg-positive women are modeled, both strategies assume a 100% screening coverage. Vaccine efficacy is assumed to confer immunity to chronic

The decision tree estimated the expected number of infants, born to HBsAgpositive women, who were perinatally and postnatally infected and was used in all cost-effectiveness analyses. The decision tree was built in TreeAge Pro 2011 (TreeAge Software, Inc, Williamstown, MA) (Fig 1). All infants born to HBsAgpositive women follow either the No PHBPP or the PHBPP strategy. Four mutually exclusive outcomes were modeled: seroprotection, susceptibility, perinatal infection, and childhood

and acute HBV infection acquired perinatally and during childhood before age 5 years. Efficacy varies by the time of first dose of HBIG and HepB vaccine administration, or HepB vaccine without HBIG, and by completion of vaccine series within 9 months or longer periods. The major determinant of the effectiveness of PEP for infants of HBsAg-positive mothers is on-time administration of the initial dose of HepB vaccine and HBIG.17 Because most available data use ,24 hours of life as the optimal timing for administering the first dose of HepB vaccine and HBIG, we also use ,24 hours rather than the ACIP-recommended #12 hours. Infants born to HBsAg-positive

FIGURE 1 Decision tree model. Decision node represented by squares, chance node by circles, and terminal nodes by triangles. Markov model entered in M circles. HBsAg+, hepatitis B surface antigen positive; Hep1Vacc, first dose of HepB vaccine; Vacc., vaccine; m, months; PVST, postvaccination serologic test.

PEDIATRICS Volume 133, Number 2, February 2014

245

TABLE 1 Model Parameters: Base-Case Values and Their Sources, Variations in Sensitivity Analysis and Assumptions, and Probabilistic SA Distributionsa

Parameter Name, Reference No.

Population and risk of infection Births per year16 P mother is HBsAg+1,15 N infants to HBsAg+ mother P HBsAg+ mother is HBeAg+38 P transmission HBeAg+3,39-42 P transmission HBeAg-3,39-42 P of transmission HBsAg+c

P of childhood infection among not seroprotected2 Vaccination coverage for No PHBPP P Hep1Vacc #24 h27,28

SAb

Base-Case

4 130 665 0.006 24 784 0.3 0.832 0.153 0.3567

0.3

Upper Value

Lower Value

— 0.008 — 0.36 1 0.314 0.5610

— 0.004 — 0.24 0.734 0.037 0.2043

0.36

0.24

.544

.544

0

P Hep1Vacc+HBIG #24h

0

0.544

0

P Hep1Vacc 24–72 h27,28

0.081

0.081

0.000

P Hep1Vacc+HBIG 24–72h

0

0.081

0

P Hep1Vacc 72 h–3 mo27,28 P no dose27,28 P completion in 9 mo29 P PVST

0.313 0.063 0.639 0

— — 0.626 0.560

— — 0.652 0

Probabilistic SA Assumption

— CI in ref 15 — 620% Highest and lowest value from 5 studies Highest and lowest value from 5 studies Calculated with upper and lower values of seroprevalence and transmission estimates 620% Variable set as: 0.544 2 P Hep1Vacc+HBIG #24 h Upper value assumes all infants receiving first dose within 24 h in NIS27,28 data also receive HBIG Variable set as: 0.081 2 P Hep1Vacc+HBIG 24–72 h Variable set as: P Hep1Vacc+HBIG #24 h*(0.081/0.544) — — CI in Smith 20121 Upper value assumes PVST rates equal to those under program support

Vaccination coverage for PHBPP P Hep1Vacc+HBIG #24h1 0.855 — — — P Hep1Vacc #24 h 0.032 — — — P Hep1Vacc+HBIG 24–72 h 0.104 — — — P Hep1Vacc 24–72 h 0.001 — — — P Hep1Vacc 72 h–3 mo 0.001 — — — P no dose 0.007 — — — P completion in 9 mo 0.723 0.868 0.578 620% P PVST 0.560 1 0.448 220% to 1 P HBs+ woman is identified 0.479 1 0.479 Assumed base value as lower bond and and captured by PHBPP 100% coverage for upper bond Efficacy of HBV vaccine and HBIG by time of administration and completion of vaccination series HBIG+Hep1Vacc #24 h and 0.920 0.970 0.830 Upper and lower values from CI in Lee remaining doses on time26 et al 200626 for HBIG+vaccine on time HBIG+Hep1Vacc #24 h and 0.830 0.970 0.830 Upper and lower values from CI in Lee remaining doses not on timed et al 200626 for HBIG+vaccine on time HBIG+Hep1Vacc 24–72 h and 0.616 0.828 0.460 Assuming only 10% cut on efficacy as remaining doses on timee upper value (0.92*0.9). Lower value is a conservative assumption of cutting efficacy in half (0.92/2) HBIG+Hep1Vacc 24–72 h and 0.556 0.747 0.415 Assuming only 10% cut on efficacy as remaining doses not on timef upper value (0.83*0.9). Lower value is a conservative assumption of cutting efficacy in half (0.83/2) Hep1Vacc #24 h and remaining 0.720 0.800 0.600 Upper and lower values from CI in Lee doses on time26 et al 200626 for vaccine on time 0.600 0.800 0.600 Upper and lower values from CI in Lee Hep1Vacc #24 h and remaining et al 200626 for vaccine on time doses not on timeg Hep1Vacc 24–72 h and remaining 0.482 0.648 0.360 Assuming only 10% cut on efficacy as doses on timeh upper value (0.72*0.9). Lower value is a conservative assumption of cutting efficacy in half (0.72/2)

246

BARBOSA et al

Distribution

SE

a

—

— 0.00392 — 0.1176 NA NA —

—

b b b b

—

— b

0.1176

dirichlet

NA

—

dirichlet —

—

NA —

b

4 154 33 —

— 385 — 10 31 183 —

4

10

18 556

15 555

2 —

—

2746

31 365

—

dirichlet dirichlet b —

NA NA NA —

10 677 2132 10 734 —

23 434 31 979 6064

dirichlet dirichlet dirichlet dirichlet dirichlet dirichlet b b b

NA NA NA NA NA NA NA NA NA

4339 162 528 5 5 36 3669 6697 12 033

736 4913 4547 5070 5070 5039 1406 5336 13 088

b

0.1372

3

0

b

0.1372

5

1

—

—

—

—

—

—

—

—

b

0.196

3

1

b

0.196

3

2

—

—

—

—

ARTICLE

TABLE 1 Continued Parameter Name, Reference No.

Base-Case

Hep1Vacc 24–72 h and remaining doses not on timei

0.402

SAb Upper Value

Lower Value

0.540

0.300

Hep1Vacc 72 h–3 mo and remaining 0.323 0.648 doses on timej Hep1Vacc 72 h–3 mo and remaining 0.269 0.540 doses not on timek Cost parameters Cost per 1 vaccine dose43 $22 — Administrative costs for 1 $14 — vaccine dose44 $121 $146 Cost of HBIG45 Cost of anti-HBsAg Test46 $15 — Blood drawn cost46 $3 — $15 — Cost of HBsAg screening test46 Cost of outpatient consultation47 $61 — Cost of identification in PHBPP $783 $929 per infantl Cost of management in PHBPP $769 $890 per infantl Markov model inputsm Cost due to HBV perinatal and childhood infection Cost of perinatal infection $24 117 $28 941 diagnosed at infection Cost of perinatal infection not $9831 $11 797 diagnosed at infection Cost of childhood infection $11 643 $13 972 diagnosed at infection Cost of childhood infection not $3713 $4455 diagnosed at infection QALYs loss due to HBV perinatal and childhood infection QALY loss due to perinatal 0.7656 0.9187 infection diagnosed at infection QALY loss due to perinatal 1.1851 1.4221 infection not diagnosed at infection QALY loss due to childhood 0.2560 0.3072 infection diagnosed at infection 0.5161 0.6193 QALY loss due to childhood infection not diagnosed at infection

0.241 0.201

— — $97 — — — — $472 $552

Probabilistic SA Assumption

Assuming only 10% cut on efficacy as upper value (0.6*0.9). Lower value is a conservative assumption of cutting efficacy in half (0.6/2) Assuming same SA values as when first dose is given 24–72 h Assuming same SA values as when first dose is given 24–72 h — — 620% — — — — For upper value the bootstrapped of CI used, for lower value the total cost for all grantees divided by the total number of infants was used

Distribution

SE

a

b

—

—

—

—

—

—

—

—

—

—

—

—

— —

— —

— —

— —

— — — — — g

— — — — — 75

— — — — — 110

— — — — —

g

62

154

5

7

$19 294

620%

g

9454

7

3706

$7865

620%

g

3854

7

1511

$9315

620%

g

4564

7

1789

$2970

620%

g

1455

7

570

0.6125

620%

g

0.3001

7

0.1

0.9481

620%

g

0.4646

7

0.2

0.2048

620%

b

0.1003

5

13

0.4129

620%

b

0.2023

3

2

HBeAg+, hepatitis B e antigen positive; HBsAg+, hepatitis B surface antigen positive; Hep1Vacc, first dose of HepB vaccine; NA, not applicable; NIS, National Immunization Survey; P, probability; SA, sensitivity analysis; —, no value was assumed in the sensitivity analysis or no distribution was assigned to the parameter. a Calculation of all inputs detailed in Appendix A of the Supplemental Information. b Upper and lower values for sensitivity analyses were informed by the CI of the parameter estimate or if CIs were not available, upper and lower values were set at 620% of the base value. For other variables, upper and lower values were based on previous studies and educated guesses. c Calculated as (P transmission HBeAg+*P HBsAg+mother HBe+) + (P transmission HBeAg-*(1 2 P HBsAg+mother HBe+)). d Assuming low end of Lee et al26 CI for HBIG+Hep1Vacc #24 h and remaining doses on time. e Calculated as 0.92*0.67, assuming a decrease in efficacy of 33%.17,18 f Calculated as 0.87*0.67, assuming a decrease in efficacy of 33%.17,18 g Assuming low end of Lee et al26 CI for Hep1Vacc #24 h and remaining doses on time. h Calculated as 0.72*0.67, assuming a decrease in efficacy of 33%.17,18 i Calculated as 0.6*0.67, assuming a decrease in efficacy of 33%.17,18 j Calculated as 0.4824*0.67, assuming a decrease in efficacy of 33%.17,18 k Calculated as 0.402*0.67, assuming a decrease in efficacy of 33%.17,18 l Authors’ calculations by using self-reported grantee data to CDC from PHBPP grantees in 49 US states (excluding Alaska), 5 cities (Chicago, Houston, New York City, Philadelphia, San Antonio), and the District of Columbia. CIs around individual cost estimates were calculated by using bootstrapping.48 m Inputs from Markov model described in Appendix A of the Supplemental Information; and Markov model structure described in Appendix B of the Supplemental Information and in Hoerger et al.49

PEDIATRICS Volume 133, Number 2, February 2014

247

women who do not complete the HepB vaccine series, or who have delays in completing the series, are significantly more likely to develop chronic infection than infants who receive timely PEP.18,19 The number of children receiving PHBPP services was calculated from PHBPP data reported to the CDC (infants born in 20081 and 2009, CDC unpublished data), which included the number of identified and case-managed infants born to HBsAg-positive women and the number receiving HBIG and HepB vaccine.1 Unit costs included HepB vaccine dose, vaccine administration, program costs, screening all pregnant women, HBIG, and PVST with outpatient consultation. Markov Model A Markov model of disease progression was used to estimate the HepB-related lifetime medical costs and QALY losses associated with perinatal and childhood HBV infections. These were then used as inputs to the decision tree. The Markov model was built in Microsoft Visual Studio 2008. We defined early diagnosis as diagnosis of infection soon after infection (100% probability of diagnosis for infants who are serologically tested) and late diagnosis as detection of infection at a 1% annual rate,20–22 or when the infection becomes symptomatic. We estimated hepatitis-related costs and QALY loss for early and late diagnosis and for perinatal and childhood infection. Table 1 shows that perinatal infections have higher costs and greater QALY losses than childhood infections, because chronic HBV infection is more common after perinatal infection (90% vs 30%).4 For both types of infections, early diagnosis leads to earlier and more monitoring, and more effective treatment than late diagnosis, resulting in higher costs but lower QALY loss. Appendices A and B in the Supplemental Information provide detailed explanations of the calculation of each parameter in Table 1 and

248

BARBOSA et al

a detailed description of the Markov model, respectively. Analysis Outcomes The combined decision tree and Markov models quantified the number of perinatal and early childhood infections from infants born to HBsAg-positive women. The other main outcome measures were QALYs, lifetime costs, and incremental cost per QALY gained. Costs were evaluated from the health care system perspective and expressed in 2010 US dollars. Future costs and QALYs were discounted at a 3% annual rate.11,12 For each strategy, expected health outcomes were the number of perinatal and childhood infections and QALYs lost with infection. Expected total costs were the sum of strategy costs and the present value of discounted future health care costs associated with monitoring and treating HBV infection and its complications. Cost-effectiveness Analysis ICERs were estimated.23 The ICER represents the additional cost to achieve 1 less QALY loss due to infection. The preferred strategy is determined by comparing the ICER to what decision makers are willing to pay for an additional QALY. No consensus has been reached on decision makers’ willingness to pay in the United States, although a $50 000/QALY benchmark has been used in several studies.24 Probabilistic sensitivity analysis was conducted to incorporate uncertainty in model parameters. Monte Carlo simulation was used to calculate the combined impact of the model’s various uncertainties.25 The distributions used are presented in Table 1. Uncertainty in the model was described using a costeffectiveness acceptability curve (CEAC)23 to graphically show the probability that 1 strategy is more cost-effective than the other based on decision makers’ willingness to pay for an additional QALY.11,23

Sensitivity Analyses One-way sensitivity analyses were conducted to estimate the impact of changing each key parameter individually. A series of multiway sensitivity analyses changed several variables simultaneously to create 6 hypothetical scenarios: (1) pessimistic and optimistic vaccine efficacy scenarios, using lower or upper bounds of the confidence interval (CI) from Lee et al26 for first dose of vaccine and HBIG and 50% or 10% cut in the remaining variables’ efficacy, respectively; (2) no maternal screening cost: the cost of screening all pregnant women in both strategies was excluded; (3) same diagnosis probability: assumed all perinatal and childhood infections were diagnosed at infection (early diagnosis); (4) perfect PHBPP: assumed that all infants born to HBsAg-positive women were identified and managed by a perfect program (100% coverage), all infants received HBIG and HepB vaccine within 24 hours and completed all doses within 9 months, and all infants were serologically tested postvaccination; (5) No PHBPP, where all first HepB doses #72 hours are given with HBIG, no PVST after completion; and (6) No PHBPP, where all first HepB doses #72 hours are given with HBIG, with PVST after completion. The 2 last scenarios relax the assumption of no administration of HBIG and PVST under No PHBPP. They estimate the impact on the ICER of compliance with recommended PEP in the absence of the PHBPP. In the absence of PHBPP, a proportion of infants born to HBsAgpositive women likely still receive HBIG and possibly PVST. Because there are no data on administration rates of HBIG and PVST rates after series completion among infants not followed by PHBPP, scenarios 5 and 6 assumed that all first doses of HepB given before 72 hours, based on National Immunization Survey data,27–29 were accompanied by

ARTICLE

HBIG. Scenario 6 also assumed that PVST would be done at the same rate as in the program strategy.

makers are willing to pay at least $2602 per QALY gained, the PHBPP can be considered cost-effective.

Because program costs relied on selfreported data (as detailed in Appendix A in the Supplemental Information), which might not have captured all the identification and management-related costs, a threshold analysis on program costs was conducted.

Figure 2 presents the CEAC. The probability that the PHBPP is cost-effective increases as health gains are more highly valued; if decision makers are willing to pay ∼$6000 for an additional QALY, the PHBPP is cost-effective. Sensitivity Analyses

RESULTS Base-Case Results Table 2 presents the cost-effectiveness results for the base-case analysis. The lifetime costs for the No PHBPP strategy were the costs of screening all pregnant women ($60 018 562), the cost of routine vaccination ($1 943 035), and the long-term medical costs associated with late diagnosis of perinatal and childhood infections ($52 364 741). The PHBPP had higher lifetime costs ($120 319 857) and better health outcomes relative to No PHBPP. The lifetime costs of the PHBPP included the costs of screening all pregnant women ($60 018 562), identifying infants born to HBsAg-positive women ($9 304 521), managing those infants ($8 977 101), the cost of PEP with HBIG and HepB vaccine, with additional HepB series and PVST for infants who failed to respond to the initial HepB vaccine series ($4 221 035), and long-term HepBrelated medical costs ($37 798 101). Compared with No PHBPP, the PHBPP was associated with 2351 fewer total infections (1485 perinatal and 866 childhood), 2304 less QALYs lost, and an ICER of $2602 per QALY. Hence, if decision

Figure 3 shows the effect of the 1-way sensitivity analyses on the ICERs. The probability of HBV transmission had the highest impact on the ICER. With higher transmission probabilities, the program was more cost-effective, and with lower transmission probabilities, the ICER was higher. Overall, changing each variable had a moderate to small

effect, and the ICER never surpassed $6500 per QALY. Table 3 presents the results of the multiway sensitivity analyses. Under a pessimistic vaccine efficacy scenario, the PHBPP was slightly more costly and less effective than in the base-case scenario, with a slightly higher ICER ($3060 per QALY). Under the optimistic vaccine efficacy scenario, the PHBPP had lower costs and higher QALYs gained than in the base-case scenario. Excluding the cost of screening all pregnant women did not have an impact on the ICER. With early diagnosis of all infections, the PHBPP was costsaving (less costly and more effective than vaccination only). This is caused by the higher cost and lower QALY loss related to early diagnosis of perinatal

FIGURE 2 Base-case results in the form of a CEAC. The CEAC shows the probability that 1 strategy is preferred to the other, for different maximum willingness to pay for an additional QALY. As decision makers are willing to pay more for an additional QALY, the more costly and effective strategy is preferred.

TABLE 2 Base-Case Estimates of Lifetime Costs, Infections, Perinatal Infections, Childhood Infections, and QALYs Lost for No PHBPP and PHBPP, Together With Incremental Analysis

No PHBPP PHBPP PHBPP Versus No PHBPP a

Total Costa

Total Infections

Perinatal Infections

Childhood Infections

Total QALYs Losta

ICER ($/ QALY)

$114 325 741 $120 319 857 $5 994 115

6816 4464 22351

4423 2938 21485

2393 1527 2866

6476 4173 22304

$2602

Costs and QALYs are discounted at a 3% annual rate. Costs in US dollars at a 2010 price base.

PEDIATRICS Volume 133, Number 2, February 2014

249

FIGURE 3 Sensitivity analysis. Shows the range of ICERs (horizontal axis; $/QALY) for lower and upper bound input values (vertical axis), when the PHBPP is compared with no program. For example, higher transmission rate leads to a more cost-effective program and vice versa.

TABLE 3 Multiway Sensitivity Analysis: What-If Scenarios Total Cost, $a

Total Infections

Total Perinatal Infections

(1a) Pessimistic vaccine efficacy scenario No PHBPP 124 070 020 8084 5246 PHBPP 131 789 282 5571 3682 (1b) Optimistic vaccine efficacy scenario No PHBPP 95 993 164 4430 2874 PHBPP 104 482 875 2669 1752 (2) No maternal screening cost No PHBPP 54 307 179 6816 4423 PHBPP 60 301 294 4464 2938 (3) Same diagnosis probability (all diagnosed at infection) No PHBPP 196 488 335 6816 4423 PHBPP 170 959 339 4414 2938 (4) Perfect PHBPP No PHBPP 114 325 741 6816 4423 PHBPP 133 862 135 1257 1109 (5) No PHBPP where all first HepB doses #72 h are given with HBIG, no PVST No PHBPP 103 458 773 5098 3308 PHBPP 114 658 166 3570 2357 (6) No PHBPP where all first HepB doses #72 h are given with HBIG, with PVST No PHBPP 119 464 037 4941 3308 PHBPP 122 996 909 3487 2357 a

Costs and QALYs are discounted at a 3% annual rate. Costs in US dollars at a 2010 price base.

250

BARBOSA et al

Total Childhood Infections

Total QALYs Losta

ICER, $ ($/QALY)

2838 1889

7681 5159

3060

1555 918

4209 2509

4994

2393 1527

6476 4173

2602

2393 1476

3998 2627

Cost saving

2393 148

6476 922

3517

1790 1212

4844 3322

7358

1632 1130

4297 3037

2804

ARTICLE

and childhood infections. Under the assumption of a perfect program, the higher costs of the program ($133 862 135 vs $120 319 857 in the base-case) led to many more infections avoided and hence, less QALYs lost. Under this assumption, there were only 1257 infections and the ICER was $3517 per QALY. Providing HBIG to all infants who received a first dose of HepB within 72 hours in the No PHBPP strategy resulted in an ICER of $7358 per QALY. This is related to the higher impact of this assumption in the No PHBPP strategy where the higher costs related to HBIG administration are offset by lower future complication costs and QALYs lost related to HepB infections. However, also providing PVST in the No PHBPP strategy reduced the ICER of scenario 5 to $2804 per QALY because PVST leads to earlier treatment but reduces only childhood infections and not perinatal infections, which are associated with higher future costs. The threshold analysis on the program costs showed that assuming a $50 000 per QALY maximum willingness to pay, the cost of PHBPP identification could be up to $54 000 per infant and the costs of PHBPP management could be up to $68 000 per infant. This is well above the base-case values of $783 and $769 for the costs of identification and management per infant.

DISCUSSION When the PHBPP is compared with a strategy that follows ACIP recommendations without the support of the program, the ICER is $2602 per QALY, and when this strategy also considers HBIG and HBIG with PVST, the ICER is between $7358 and $2804 per QALY, which by most standards suggests that the PHBPP is cost-effective in avoiding perinatal and childhood infections.30,31 The costs and outcomes of immunization strategies to prevent HBV transmission in the United States have been evaluated.2 Comparing prevention of

PEDIATRICS Volume 133, Number 2, February 2014

perinatal HBV infection with a situation of no immunization, in 1991 Margolis et al2 reported a total of 4703 perinatal and childhood infections prevented from 11 400 HBsAg-positive women. To understand our results within the context of those presented by Margolis et al,2 we compared the PHBPP to a hypothetical scenario of no vaccination in a separate analysis (results available on request). Scaled to our estimate of 24 784 HBsAg-positive mothers, the number of infections prevented in Margolis et al2 would be analogous to 10 224 infections avoided, close to our estimate of 9159 infections prevented when the PHBPP is compared with a situation of no immunization. The slightly lower number of infections prevented in our models is caused by the use of a lower transmission rate (updated to take into account current seroprevalence estimates), and a slightly lower efficacy of PEP that takes into account current compliance rates by timing of first and remaining doses of PEP doses. Margolis et al2 did not account for PEP compliance (they assumed the same value as routinely administered vaccines) and did not cost a specific perinatal prevention program. Therefore, program cost estimates are not directly comparable. Margolis et al 2 presented costeffectiveness results from a third-part payer’s perspective in which only direct medical costs were included. They also presented net benefit estimates from a broader perspective in which medical direct costs and morbidity- and mortality-related work-loss costs were included. When a broader perspective was adopted, Margolis et al2 reported that the perinatal prevention strategy generated cost savings. Following our health care system perspective on costs and health economic guidelines that recommend that mortality-related productivity losses should not be included in cost-effectiveness studies with QALYs, we excluded mortality-related

work-loss costs.11,12 Thus, we do not compare our ICER with that presented by Margolis et al,2 because the authors did not present ICERs in the form of cost per QALYs and included different cost categories. Other data limitations could have affected the results. The HepB vaccine efficacy with HBIG starting within 24 hours of birth was taken from a recent meta-analysis based on 3 trials, and HepB vaccine efficacy without HBIG was based on 4 trials.32 Despite the small number of studies, the estimates are within the range of results from studies not included because they were not randomized controlled trials.33,34 The efficacy of HepB vaccine and HBIG when the first dose is delayed and when all vaccine doses are not completed on time was determined by the authors’ calculations relying on 2 studies.17,18 These areas of uncertainty were considered in the probabilistic and multiway sensitivity analyses. Another important data limitation relates to the cost of the PHBPP. Program costs were based on aggregate data reported by PHBPP grantees. We excluded outliers and used bootstrap techniques to account for imperfections of the data. However, it is possible that cross-subsidies (eg, office space, utilities, shared equipment, hospital personnel time for communication, and liaising with PHBPP staff) were not captured in the reports. More accurate costs are needed for future analyses. In addition to the data limitations, more general issues remain. Some benefits of the PHBPP have been underestimated. Recent evidence suggests vaccination confers immunity into early adulthood.35–37 This study did not account for infections avoided at later stages in life (eg, during adolescence and adulthood). We focused on the impact of perinatal vaccination in the cohort of infants born to HBsAg-positive mothers in 2009. Our approach also did not take into account 251

herd immunity, which would require a dynamic transmission model approach in which the risk of infection changes over time. Our results showed that the program is a potentially cost-effective intervention, even without the additional benefits that would accrue from dynamic effects like herd immunity. A dynamic transmission model might further enhance this result. Finally, the benefits of PHBPP might have been further underestimated because we did not include maternal evaluation and management of household and sexual contacts.

CONCLUSIONS This study presents the first costeffectiveness analysis of the US PHBPP. The simulation of a perfect PHBPP scenario showed that an expansion of the current recommendations to all infants at risk would lead to a reduced number of perinatal and childhood infections, with an ICER of $3517 per QALY. Our model predictedthatthecurrentPHBPPrepresented a cost-effective use of resources, and that expansion of the program could represent an economically efficient method to prevent further morbidity and mortality.

ACKNOWLEDGMENTS We acknowledge the contributions by Perinatal Hepatitis B Prevention Coordinators Julie Lazaroff, New York City Department of Health and Mental Hygiene, and Patrick Fineiis, Michigan Department of Community Health, for providing useful comments on the model; and by Qian Li from the CDC’s National Center for Immunization and Respiratory Diseases who analyzed data from the 2010 National Immunization Program to determine HepB vaccine coverage at intervals relevant to this analysis.

8. Jacobs RJ, Saab S, Meyerhoff AS. The cost effectiveness of hepatitis immunization for US college students. J Am Coll Health. 2003; 51(6):227–236 9. Kim SY, Billah K, Lieu TA, Weinstein MC. Cost effectiveness of hepatitis B vaccination at HIV counseling and testing sites. Am J Prev Med. 2006;30(6):498–506 10. Krahn M, Guasparini R, Sherman M, Detsky AS. Costs and cost-effectiveness of a universal, school-based hepatitis B vaccination program. Am J Public Health. 1998;88(11):1638–1644 11. Drummond M, Sculpher M, Torrance G, O’Brien B, Stoddart G. Methods for the Economic Evaluation of Health Care Programmes. 3rd ed. Oxford, UK: Oxford University Press; 2005 12. Gold MR, Siegel JE, Russel LB, Weinstein MC. Cost Effectiveness in Health and Medicine. New York, NY: Oxford University Press; 1996 13. Sonnenberg FA, Beck JR. Markov models in medical decision making: a practical guide. Med Decis Making. 1993;13(4):322–338 14. Hunink M, Glaziou P, Siegel J, Weeks J, Pliskin J, Elstein A. Decision Making in Health and Medicine. Integrating Evidence and Values. Cambridge, UK: Cambridge University Press; 2001 15. Koumans EHA, Rosen J, van Dyke MK, et al; ABC and DHAP/RTI teams. Prevention of motherto-child transmission of infections during pregnancy: implementation of recommended interventions, United States, 2003-2004. Am J Obstet Gynecol. 2012;206(2):e1–, e11 16. Martin JA, Hamilton BE, Ventura SJ, et al. Births: final data for 2009. Natl Vital Stat Rep. 2011;60(1):1–70

17. Marion SA, Pastore MT, Pi DW, Mathias RG. Long-term follow-up of hepatitis B vaccine in infants of carrier mothers. Am J Epidemiol. 1994;140(8):734–746 18. Kohn MA, Farley TA, Scott C. The need for more aggressive follow-up of children born to hepatitis B surface antigen-positive mothers: lessons from the Louisiana Perinatal Hepatitis B Immunization Program. Pediatr Infect Dis J. 1996;15(6):535–540 19. Tharmaphornpilas P, Rasdjarmrearnsook AO, Plianpanich S, Sa-nguanmoo P, Poovorawan Y. Increased risk of developing chronic HBV infection in infants born to chronically HBV infected mothers as a result of delayed second dose of hepatitis B vaccination. Vaccine. 2009;27(44):6110–6115 20. Asian Liver Center. Hepatitis B statistics for Asians and Pacific Islanders. Stanford School of Medicine. Available at: http://liver. stanford.edu/Education/faq.html. Accessed September 9, 2012 21. Lin SY, Chang ET, So SK. Why we should routinely screen Asian American adults for hepatitis B: a cross-sectional study of Asians in California. Hepatology. 2007;46(4):1034–1040 22. Shuler CM, Fiore AE, Neeman R, et al. Reduction in hepatitis B virus seroprevalence among U.S.-born children of foreign-born Asian parents—benefit of universal infant hepatitis B vaccination. Vaccine. 2009;27 (43):5942–5947 23. Briggs A, Sculpher M, Klaxton C. Decision Modelling for Health Economic Evaluation. New York, NY: Oxford University Press; 2006 24. Grosse SD. Assessing cost-effectiveness in healthcare: history of the $50,000 per QALY

REFERENCES 1. Smith EA, Jacques-Carroll L, Walker TY, Sirotkin B, Murphy TV. The national Perinatal Hepatitis B Prevention Program, 1994-2008. Pediatrics. 2012;129(4):609–616 2. Margolis HS, Coleman PJ, Brown RE, Mast EE, Sheingold SH, Arevalo JA. Prevention of hepatitis B virus transmission by immunization. An economic analysis of current recommendations. JAMA. 1995;274(15):1201– 1208 3. Beasley RP, Trepo C, Stevens CE, Szmuness W. The e antigen and vertical transmission of hepatitis B surface antigen Am J Epidemiol. 1977;105(2):94–98 4. Mast EE, Margolis HS, Fiore AE, et al; Advisory Committee on Immunization Practices (ACIP). A comprehensive immunization strategy to eliminate transmission of hepatitis B virus infection in the United States: recommendations of the Advisory Committee on Immunization Practices (ACIP) part 1: immunization of infants, children, and adolescents. MMWR Recomm Rep. 2005;54(RR-16):1–31 5. Goldstein ST, Zhou FJ, Hadler SC, Bell BP, Mast EE, Margolis HS. A mathematical model to estimate global hepatitis B disease burden and vaccination impact. Int J Epidemiol. 2005;34(6):1329–1339 6. Hutton DW, Tan D, So SK, Brandeau ML. Cost-effectiveness of screening and vaccinating Asian and Pacific Islander adults for hepatitis B. Ann Intern Med. 2007;147(7): 460–469 7. Pisu M, Meltzer MI, Lyerla R. Costeffectiveness of hepatitis B vaccination of prison inmates. Vaccine. 2002;21(3-4):312– 321

252

BARBOSA et al

ARTICLE

25.

26.

27.

28.

29.

30.

31.

32.

threshold. Expert Rev Pharmacoecon Outcomes Res. 2008;8(2):165–178 Doubilet P, Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabilistic sensitivity analysis using Monte Carlo simulation. A practical approach. Med Decis Making. 1985;5(2):157–177 Lee C, Gong Y, Brok J, Boxall EH, Gluud C. Hepatitis B immunisation for newborn infants of hepatitis B surface antigen-positive mothers. Cochrane Database Syst Rev. 2006;2006 (2):CD004790 Centers for Disease Control and Prevention (CDC). Estimated vaccination coverage for hepatitis B vaccine among children from birth to 3 days of age 2010. Available at: www2a. cdc.gov/nip/coverage/nis/nis_iap3.asp? fmt=v&rpt=tab03_antigen_state&qtr=Q1/ 2010-Q4/2010. Accessed September 4, 2012 Centers for Disease Control and Prevention (CDC). Estimated vaccination coverage for hepatitis B vaccine among children from birth to 3 days of age 2009. www2a.cdc.gov/ nip/coverage/nis/nis_iap2.asp?fmt=v&rpt= tab03_antigen_state&qtr=Q1/2009-Q4/2009. Accessed September 4, 2012 Centers for Disease Control and Prevention (CDC). Estimated vaccination coverage with individual vaccines by 3 months of age 2010. Available at: www2a.cdc.gov/nip/coverage/ nis/nis_iap3.asp?fmt=v&rpt=tab04_3mo_ iap&qtr=Q1/2010-Q4/2010. Accessed September 4, 2012 Braithwaite RS, Meltzer DO, King JT Jr, Leslie D, Roberts MS. What does the value of modern medicine say about the $50,000 per quality-adjusted life-year decision rule? Med Care. 2008;46(4):349–356 Neumann PJ, Weinstein MC. Legislating against use of cost-effectiveness information. N Engl J Med. 2010;363(16):1495–1497 Lee CF, Gong Y, Brok J, Boxall EH, Gluud C. Effect of hepatitis B immunisation in newborn infants of mothers positive for hepatitis B surface antigen: systematic review and meta-analysis. BMJ. 2006;332(7537):328–336

33. Beasley RP, Hwang LY. Postnatal infectivity of hepatitis B surface antigen-carrier mothers. J Infect Dis. 1983;147(2):185–190 34. Schillie SF, Murphy TV. Seroprotection after recombinant hepatitis B vaccination among newborn infants: a review. Vaccine. 2013;31 (21):2506–2516 35. McMahon BJ, Dentinger CM, Bruden D, et al. Antibody levels and protection after hepatitis B vaccine: results of a 22-year follow-up study and response to a booster dose. J Infect Dis. 2009;200(9):1390–1396 36. Poovorawan Y, Chongsrisawat V, Theamboonlers A, et al. Evidence of protection against clinical and chronic hepatitis B infection 20 years after infant vaccination in a high endemicity region. J Viral Hepat. 2011;18(5): 369–375 37. Zhu C-L, Liu P, Chen T, et al. Presence of immune memory and immunity to hepatitis B virus in adults after neonatal hepatitis B vaccination. Vaccine. 2011;29(44): 7835–7841 38. Ott JJ, Stevens GA, Wiersma ST. The risk of perinatal hepatitis B virus transmission: hepatitis B e antigen (HBeAg) prevalence estimates for all world regions. BMC Infect Dis. 2012;12(1):131 39. Hwang LY, Roggendorf M, Beasley RP, Deinhardt F. Perinatal transmission of hepatitis-B virus- role of maternal HBeAg and anti-HBc IgM. J Med Virol. 1985;15(3): 265–269 40. Okada K, Kamiyama I, Inomata M, Imai M, Miyakawa Y, Mayumi M. e antigen and antie in the serum of asymptomatic carrier mothers as indicators of positive and negative transmission of hepatitis B virus to their infants. N Engl J Med. 1976;294(14): 746–749 41. Shiraki K, Yoshihara N, Sakurai M, Eto T, Kawana T. Acute hepatitis-B in infants born to carrier mothers with the antibody to hepatitis-B antigen. J Pediatr. 1980;97(5): 768–770

42. Stevens CE, Neurath RA, Beasley RP, Szmuness W. HBeAg and anti-HBe detection by radioimmunoassay—correlation with vertical transmission of hepatitis-B virus in Taiwan. J Med Virol. 1979;3(3):237–241 43. Centers for Disease Control and Prevention (CDC). CDC vaccine price list. 2012. Available at: www.cdc.gov/vaccines/programs/ vfc/awardees/vaccine-management/price-list/ index.html. Accessed September 4, 2012 44. Miriti MK, Billah K, Weinbaum C, et al. Economic benefits of hepatitis B vaccination at sexually transmitted disease clinics in the US. Public Health Rep. 2008;123(4): 504–513 45. Centers for Medicare & Medicaid Services. Fee schedule reimbursement rate. 2012. Available at: https://www.cms.gov/transmittals/ downloads/R1209CP.pdf. Accessed September 4, 2012 46. Centers for Medicare & Medicaid Services. Clinical diagnostic laboratory fee (CLAB). 2012. Available at: www.cms.gov/Medicare/MedicareFee-for-Service-Payment/ClinicalLabFeeSched/ index.html?redirect=/ClinicalLabFeeSched/ 02_clinlab.asp. Accessed September 4, 2012 47. Centers for Medicare & Medicaid Services. Physician fee schedule. 2012. Available at: https://www.cms.gov/apps/physician-feeschedule/search/search-criteria.aspx. Accessed September 4, 2012 48. Briggs AH, Wonderling DE, Mooney CZ. Pulling cost-effectiveness analysis up by its bootstraps: a non-parametric approach to confidence interval estimation. Health Econ. 1997;6(4):327–340 49. Hoerger TJ, Wittenborn J, Uzunangelov V, Simpson S. Costs and Benefits Associated With Different Policy Approaches to Preventing and Controlling Viral Hepatitis: Final Report Prepared for the Department of Health and Human Services Office of the Assistant Secretary for Policy Evaluation. Research Triangle Park, NC: RTI International; 2012

(Continued from first page) PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). Copyright © 2014 by the American Academy of Pediatrics FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose. FUNDING: RTI International’s work on the project was supported by Centers for Disease Control and Prevention contract 200-2009-30991 TO3. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the Agency for Toxic Substances and Disease Registry. POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

PEDIATRICS Volume 133, Number 2, February 2014

253