Comprehensive Cost-Utility Analysis of Newborn Screening Strategies

Aaron E. Carroll and Stephen M. Downs

Pediatrics 2006;117;S287-S295
DOI: 10.1542/peds.2005-2633H

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://www.pediatrics.org/cgi/content/full/117/5/S1/S287

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright © 2006 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from www.pediatrics.org by on September 20, 2010

SUPPLEMENT ARTICLE

Comprehensive Cost-Utility Analysis of Newborn

Screening Strategies
Aaron E. Carroll, MD, MSa,b, Stephen M. Downs, MD, MSa,b

a Children’s Health Services Research, Indiana University School of Medicine, Indianapolis, Indiana; bRegenstrief Institute for Health Care, Indianapolis, Indiana

The authors have indicated they have no financial relationships relevant to this article to disclose.

ABSTRACT
BACKGROUND. Inborn errors of metabolism are a significant cause of morbidity and
death among children. Inconsistencies in how individual states arrive at screening
www.pediatrics.org/cgi/doi/10.1542/
strategies, however, lead to marked variations in testing between states. peds.2005-2633H
OBJECTIVE. To determine the cost-effectiveness of each component test of a multitest doi:10.1542/peds.2005-2633H
newborn screening program, including screening for phenylketonuria, congenital The views expressed in this article are
those of the authors and do not necessarily
adrenal hyperplasia, congenital hypothyroidism, biotinidase deficiency, maple represent the views of the Indiana
syrup urine disease, galactosemia, homocystinuria, and medium-chain acyl-CoA University School of Medicine.
dehydrogenase deficiency. Key Words
newborn, screening, decision analysis,
METHODS. A decision model was used, with cohort studies, government reports, cost-effectiveness, cost utility
secondary analyses, and other sources. Discounted costs, quality-adjusted life- Abbreviations
PKU—phenylketonuria
years (QALYs), and incremental cost-effectiveness ratios were measured. MS/MS—tandem mass spectrometry
CAH— congenital adrenal hyperplasia
RESULTS. All except 2 screening tests dominated the “no-test” strategy. The 2 excep- CH— congenital hypothyroidism
tions were screening for congenital adrenal hyperplasia, which cost slightly more MSUD—maple syrup urine disease
than $20 000 per QALY gained, and screening for galactosemia, which cost QALY— quality-adjusted life-year
CP— cerebral palsy
$94 000 per QALY gained. The screening test with the lowest expected cost was MCAD—medium-chain acyl-CoA
tandem mass spectrometry. The results found in our base-case analysis were stable dehydrogenase

across variations in nearly all variables. In instances in which changes in risks, Accepted for publication Dec 27, 2005
Address correspondence to Aaron E. Carroll,
sequelae, costs, or utilities did affect our results, the variation from base-case MD, MS, Riley Research 330, 699 West Dr,
estimates was quite large. Indianapolis, IN 46074. E-mail: aaecarro@iupui.
edu
CONCLUSIONS. Newborn screening seems to be one of the rare health care interven- PEDIATRICS (ISSN Numbers: Print, 0031-4005;
tions that is beneficial to patients and, in many cases, cost saving. Over the long Online, 1098-4275); published in the public
domain by the American Academy of
term, funding comprehensive newborn screening programs is likely to save money Pediatrics
for society.

PEDIATRICS Volume 117, Number 5, May 2006 S287

Downloaded from www.pediatrics.org by on September 20, 2010
I NBORN ERRORS OF metabolism are a significant cause
of morbidity and death among children.1–4 Technology
has advanced remarkably since screening began for phe-
tive of this study was to determine the cost-effectiveness
of a comprehensive newborn screening program, includ-
ing screening for PKU, congenital adrenal hyperplasia
nylketonuria (PKU) ⬎40 years ago.5 Although we have (CAH), congenital hypothyroidism (CH), biotinidase de-
technology today to screen for a host of diseases, which ficiency, maple syrup urine disease (MSUD), galac-
diseases are screened for differs according to state. Every tosemia, homocystinuria, and medium-chain acyl-CoA
year ⬃4 million dried blood spots are analyzed for met- dehydrogenase (MCAD) deficiency, alone or in combi-
abolic and hematologic disorders and endocrinopathies.6 nation with diseases detectable with MS/MS methods.
These diseases account for ⬃3000 cases of potentially
disabling or fatal outcomes each year.6 The diseases for METHODS
which we screen and the manner in which we screen
vary significantly according to state. These decisions Study Design
should be based on disease prevalence, costs of testing, We developed a decision model with decision analysis
treatment, and complications, and effectiveness of software (DATA 4.0; TreeAge Software, Williamstown,
timely intervention. Inconsistencies regarding how MA), to compare various strategies for screening for
states arrive at these decisions, however, lead to marked newborn diseases. We adopted the recommendations of
variations in testing among states. the Panel on Cost-Effectiveness in Health and Medicine
Tandem mass spectrometry (MS/MS) offers the abil- for conducting and reporting a base-case analysis.17 We
ity to increase drastically the number of diseases for assumed a societal perspective to produce results, per-
which we screen.6–8 More than 20 disorders of fatty acid mitting comparisons across different interventions, and
oxidation and organic acidemia are already screened for expressed results in terms of discounted costs, quality-
routinely with MS/MS methods.9 The number of dis- adjusted life-years (QALYs), and incremental cost-effec-
eases screened for and the number of states using the tiveness ratios.
technology are increasing rapidly.10
The significant start-up costs of MS/MS screening, Decision Model Structure and Assumptions
however, have prevented some states from adopting it as To frame the model for analysis, we made a number of
part of their standard screening process.11 Many have simplifying assumptions. We assumed that any child
pointed out that most of the diseases for which MS/MS could have, at most, 1 of the conditions for which the
methods screen have a relatively low prevalence.9,10 tests screen. We used a multiplicative utility model, that
MS/MS screening can lead to false-positive or abnormal is, for outcomes with ⬎1 complication, we multiplied
results for well children, yielding increased costs and the quality adjustments (utilities) for all relevant com-
unnecessary parental angst.12,13 In addition, some dis- plications together. We assumed the existence of a com-
eases are so severe that, even with an ideal screening prehensive screening program that ensures timely test-
program, many children would not be treated in time to ing, result notification, and appropriate treatment of all
prevent morbidity and death. positive screens. Finally, we assumed that MS/MS meth-
Other data argue for the merits of broader screening ods would be used to screen for multiple conditions.
even in light of these issues. Although costs for wider
screening can be significant, the costs saved through The Decision Model
preventing hospitalizations for diagnosis or long-term Figure 1 shows the decision model structure. The deci-
disability are also significant.11 Studies show that even sion tree consisted of a decision node whose branches
diseases without a consistently effective treatment can represent 8 screening tests, that is, PKU, CAH, CH, bio-
benefit from early diagnosis.7 Evidence indicates that, tinidase deficiency, MSUD, galactosemia, homocystin-
overall, screening reduces stress while improving out- uria, and MS/MS tests. We assumed that MS/MS would
comes.12 be used to detect PKU, biotinidase deficiency, MSUD,
A more objective method for determining the useful- galactosemia, and homocystinuria, as well as MCAD
ness of any new technology is cost-effectiveness analysis. deficiency. In a second analysis, we compared MS/MS
A number of these analyses were conducted previously directly with a screening panel that included PKU, bio-
for other aspects of newborn screening14,15 or the use of tinidase deficiency, MSUD, galactosemia, and homo-
MS/MS screening for a small number of disease process- cystinuria detection with conventional methods.
es.1,9,10 Currently, programs do not screen for biotinidase Following each test branch is a chance node whose
deficiency and galactosemia with MS/MS methods, but branches represent the 8 conditions that the child may
there is clear evidence that galactosemia could be de- have, ie, PKU, CAH, CH, biotinidase deficiency, MSUD,
tected with this technology.16 galactosemia, homocystinuria, or MCAD deficiency, or
To date, no cost-effectiveness study has examined the “no disease.” Each disease branch is followed by a sub-
prospect of incorporating broad MS/MS testing into a tree that represents potential sequelae of the disease. In
comprehensive newborn screening program. The objec- Fig 1, final outcomes are shown only for homocystin-

S288 CARROLL, DOWNS

Downloaded from www.pediatrics.org by on September 20, 2010
FIGURE 1
Decision model. Branches from the square decision node represent the various screening tests. Following each of these is a subtree representing each of the conditions that might be
present. Only one of these subtrees is shown. Following each condition is a subtree depicting the possible outcomes of the conditions. Final outcomes are shown for homocystinuria
(HCY) only. # indicates the probability of the lower branch is 1 minus the probabilities of the branches above. BIOT indicates biotinidase deficiency; GAL, galactosemia; Dz, disease.

uria. PKU was associated with a risk of mild, moderate, ical and administrative data sources. We searched Med-
or severe developmental delay and/or seizure disorder. line (from January 1980 to November 2002) for studies
CAH could lead to early childhood death (average life with the terms “newborn screening” and any of the
expectancy: 1 year). CH could cause mild, moderate, or disease entities we were investigating. We also searched
severe delay. Biotinidase deficiency could lead to any online Internet databases for the same terms. We sup-
combination of seizure disorder, severe developmental plemented this strategy by hand-searching citation lists
delay, hearing loss, or blindness. MSUD could lead to from articles identified in the electronic search and re-
any combination of severe or mild cerebral palsy (CP) or cent reviews and cost-effectiveness analyses of other
mild, moderate, or severe developmental delay. Galac- newborn screening strategies.
tosemia could lead to death in infancy or moderate or The prevalence of each disease was estimated, when
severe delay. Homocystinuria had a risk of mild, mod- possible, from the National Newborn Screening Report.18
erate, or severe delay and/or vision loss. MCAD defi- Supplemental information was taken from the American
ciency could lead to early death (modeled as death at 1 Academy of Pediatrics newborn screening fact sheets
year, on average) or CP and/or mild-to-severe develop- and secondary analyses.8,19 We determined the sensitiv-
mental delay. Children with no disease face a risk of a ities and specificities of the various screening tests
false-positive newborn screening result and its attendant through national reports and other secondary analy-
expenses. ses.10,13,18,20 Probabilities of various sequelae and the ef-
The probability of each of the outcomes with screen- fectiveness of early intervention in preventing these se-
ing, P(outcome兩screen), was represented by the formula quelae were also estimated from the newborn screening
P(outcome兩screen) ⫽P(outcome) ⫻ [1 ⫺ (efficacy ⫻ sen- fact sheets, secondary data analyses, and cohort studies
sitivity)], where P(outcome) is the probability of the of patients with the diseases in question.19,21–28 Base-case
outcome (eg, CP or delay), without screening or treat- probabilities for major complications are summarized in
ment, derived from the literature (see below), efficacy is Table 1.
the efficacy of early detection and treatment in prevent-
ing the outcome, and sensitivity is the sensitivity of the Estimation of Life Expectancy
screening test for the disease. We estimated the average life expectancy of normal
patients from the National Vital Statistics Reports published
Data and Assumptions by the Centers for Disease Control and Prevention.29 We
Probabilities of sequelae, quality-adjusted survival rates, estimated the effect of disability on life expectancy
and costs for treatment were derived from various clin- through a cohort study on the subject.30

SUPPLEMENT S289
Downloaded from www.pediatrics.org by on September 20, 2010
 TABLE 1 Data Used In Decision Tree
Base Case Reference Threshold
Test characteristics
Sensitivity of newborn screening tests (for all conditions) 1 13 Table 4
Specificity of BIOT screening test 0.9998 13 0.958
Specificity of CAH screening test 0.992 13
Specificity of CH screening test 0.985 13 0.970
Specificity of GAL screening test 0.997 13
Specificity of HCY screening test 0.9993 20 0.994
Specificity of MS/MS testing 0.9975 10 0.854
Specificity of MSUD screening test 0.9987 10 0.997
Specificity of PKU screening test 0.998 13 0.881
Disease prevalence, cases per 100 000 infants
BIOT 1.1 18 0.2
CAH 4.9 18 3500
CH 39.6 18 28.0
GAL 1.5 18
HCY 0.4 18 0.16
MCAD deficiency 9.0 19
MSUD 0.6 18 0.48
PKU 6.6 18 0.8
Probability of disease causing sequelae
BIOT, hearing loss 0.76 25
BIOT, severe developmental delay 0.50 24
BIOT, seizure disorder 0.70 24
BIOT, vision loss 0.50 24
Risk of death from CAH in first 5 y 0.10 Expert opinion
CH, mild developmental delay 0.27 19 0.018
CH, moderate developmental delay 0.021 19
CH, severe developmental delay 0.019 19 0.0082
GAL, neonatal death 0.14 27 ⬎0.27
GAL, moderate developmental delay 0.90 26
GAL, severe developmental delay 0.10 26
HCY, blindness 1 22 0.4
HCY, mild developmental delay 0.11 21
HCY, moderate developmental delay 0.33 21
HCY, severe developmental delay 0.05 21
MCAD deficiency, moderate CP 0.10 28
MCAD deficiency, neonatal death 0.204 21
MCAD deficiency, mild developmental delay 0.05 9
MCAD deficiency, severe developmental delay 0.05 9
MSUD, moderate CP 0.50 23
MSUD, severe CP 0.25 23 0.14
MSUD, severe developmental delay 0.50 23 0.34
PKU, mild developmental delay 0.05 24
PKU, moderate developmental delay 0.475 24
PKU, severe developmental delay 0.475 24
PKU, seizure disorder 0.25 24
Effectiveness of early screening in preventing sequelae
BIOT, developmental delay 1 19
BIOT, hearing loss 1 19
BIOT, seizure disorder 1 19
BIOT, vision loss 1 19
CAH, preventing death from crisis 0.8 Expert opinion
CH, developmental delay 1 19 0.67
GAL, neonatal death 0.66 21
GAL, developmental delay 0 21
HCY, blindness 1 22 0.31
HCY, developmental delay 1 22
MCAD deficiency, neonatal death 0.98 21
MSUD, CP 0.93 23 0.59
MSUD, developmental delay 0.53 19, 24 0.36
PKU, developmental delay 1 24 0.01
PKU, seizure disorder 1 24

S290 CARROLL, DOWNS

Downloaded from www.pediatrics.org by on September 20, 2010
 TABLE 1 Continued
Base Case Reference Threshold
Cost of screening tests
BIOT $1.83 33 $14
CAH $3.63 33
CH $4.59 33 $8.70
GAL $3.79 33
HCY $0.84 33 $2.45
MS/MS $16.02 9 $58
MSUD $2.49 33 $2.90
PKU $3.43 33 $38.20
Follow-up evaluation of false-positive result (for any test) $300 Expert opinion Table 4
Cost of caring for disease
BIOT $6592 21
CAH $10 000 Expert opinion
CH $9439 34
GAL $9439 34
HCY $122 515 34
MCAD deficiency $10 173 21
MSUD $122 515 34
PKU $122 515 34
Cost of sequelae
Blindness $581 688 $180 000 (HCY)
Hospitalization before death $27 809 36
MCAD deficiency crisis $308 315 21
Deafness $445 255 14
Mild developmental delay $44 192 35, expert opinion $3000
Moderate developmental delay $77 079 35, expert opinion
Severe developmental delay $1 042 110 35 Table 4
Moderate CP $77 079 35, expert opinion
Severe CP $946 528 35 $540 000 (MSUD)
Seizure disorder $216 848 21
Life expectancy, y
Normal 77.2 29
Mild developmental delay 74 30
Moderate developmental delay 67.6 30
Severe developmental delay 58.6 30
Utilities of having sequelae
Blindness 0.514 32
Deafness 0.8611 31
Mild developmental delay 0.7396 31
Moderate developmental delay 0.5625 31
Severe developmental delay 0.3909 31
Moderate CP 0.5625 31
Severe CP 0.3909 31
Base-case estimates are presented with the reference from which those estimates were drawn. Also presented are the threshold values obtained from sensitivity analyses (see text). BIOT indicates
biotinidase deficiency; GAL, galactosemia; HCY, homocystinuria.

Quality-of-Life Adjustments urban emergency department. These were the most rel-
We adjusted life expectancy for quality of life by using evant utilities available for these outcomes. Utilities for
health state utilities.17 We calculated QALYs by multiply- blindness were taken from a study that developed an
ing remaining years of life by the utility value for each equation for estimating utility on the basis of visual
health state. Utilities for most disabilities were drawn acuity.32 We did not adjust utility if a disease state did not
from a study assessing parents’ utilities by using a stan- result in a disability that was thought to affect quality of
dard gamble, adjusting for different severities when nec- life. Base-case estimates for utilities are summarized in
essary.31 The study assessed utilities for outcomes of Table 1.
occult bacteremia and meningitis, which resemble
closely outcomes of the diseases for which newborns are Costs
screened. Specifically, parent utilities were assessed for The costs for the specific screening tests were derived
deafness and various degrees of developmental delay from 2 sources, ie, a PriceWaterhouseCoopers analysis of
and CP. Respondents in the study were parents of chil- newborn screening costs and a prior cost-effectiveness
dren between 2 and 24 months of age who visited an study of MS/MS screening.9,33 These costs incorporated

SUPPLEMENT S291
Downloaded from www.pediatrics.org by on September 20, 2010
amortized start-up costs and operating costs. The costs of to $1 million. To test the impact of changing multiple
treating selected diseases over the course of a lifetime assumptions, we developed a “pessimistic” case analysis
were estimated with data from the Office of Technology that biased the model against newborn screening. In the
Assessment analysis of strategies for newborn screening, pessimistic case, we eliminated indirect costs from the
data from a Washington State report on the cost-effec- estimated costs of developmental delay and CP.35 We
tiveness of some screening strategies, and, in 1 case, eliminated the costs of developmental delay from any
expert opinion.21,34 Costs of treating sequelae of different outcomes that included CP. We increased the cost of
diseases were estimated from sources including a Mor- MS/MS to $20, the high end of published estimates. We
bidity and Mortality Weekly Report on economic costs as- increased the cost of evaluating false-positive screening
sociated with mental retardation, CP, hearing loss, and test results to $1000. We used low-end estimates for the
visual impairment; an epidemiologic study of end-of-life prevalence of MCAD deficiency (1 case per 18 000 in-
care costs; and prior secondary analyses.14,21,35,36 All costs fants) and PKU (1 case per 20 000 infants). We also used
were adjusted to 2004 US dollars with an inflation cal- a low-end estimate of the risk of death resulting from
culator on the Web site of the Department of Labor, CAH (3%).
Bureau of Labor Statistics.37 For this study, costs were a
complex mixture of medical and nonmedical costs, and
RESULTS
we elected to adjust costs with the General Consumer
Price Index, rather than the Medical Care Component Base-Case Analysis
Consumer Price Index. Base-case estimates for costs are The results of the base-case analysis are shown in Table
summarized in Table 1. 2. The first column of Table 2 shows the testing strategies
in ascending order of average cost. The second column
Time Preference shows the average cost of each strategy. This includes
To reflect patient preferences for health outcomes and the cost of the test and the average costs (per person
having material goods sooner rather than later, we dis- screened) associated with all of the treatments and out-
counted all costs and health effects at an annual rate of comes for the diseases represented in the model. The
3%. third column shows the difference in average costs be-
tween the testing strategy and not testing. A negative
Calculation of Incremental Cost-Effectiveness Ratios number means the strategy saves money, on average,
For each treatment strategy, we calculated the expected over not screening. The fourth column shows the aver-
total costs by multiplying the probability of each unique age effectiveness of each strategy in QALYs. The fifth
outcome with its associated costs and then adding these column shows the difference in effectiveness between
values for all possible outcomes. Total QALYs for each each strategy and not testing. A positive number means
treatment were calculated in a similar manner. We then there is a gain in QALYs. The sixth column is the average
calculated the incremental cost-effectiveness of each cost-effectiveness ratio for each strategy, and the sev-
screening strategy by dividing the difference in costs by enth column is the incremental cost-effectiveness ratio,
the difference in QALYs. the incremental cost per QALY gained. Tests marked as
“Dominates” save money and improve outcomes rela-
Sensitivity Analysis tive to not testing.
We used 1-way sensitivity analysis to identify important All except 2 screening tests dominated the no-testing
model uncertainties. All probabilities were tested across strategy. That is, these screening tests improved out-
the full range from 0 to 1. Costs were analyzed from $0 comes and reduced costs relative to not screening. The 2

TABLE 2 Results of the Base-Case Analysis

Strategy Cost Incremental Effectiveness Incremental C/E Incremental C/E
Cost Effectiveness
MS/MS testing $55 ⫺$43 77.19358 0.004 0.72 Dominates
PKU test $63 ⫺$35 77.19264 0.00306 0.81 Dominates
BIOT test $85 ⫺$13 77.19008 0.0005 1.1 Dominates
CH test $93 ⫺$5 77.19272 0.00314 1.21 Dominates
HCY test $96 ⫺$2 77.18977 0.00019 1.24 Dominates
MSUD test $97 ⫺$1 77.18978 0.0002 1.26 Dominates
No test $98 77.18958 1.26
GAL test $102 $5 77.18963 0.00005 1.32 $94 000
CAH test $103 $6 77.18986 0.00028 1.34 $20 357
Presented are overall cost and effectiveness, incremental cost and incremental effectiveness over no testing, and both cost-effectiveness and incremental cost-effectiveness ratios. All strategies were
compared with not testing. C/E indicates cost-effectiveness; BIOT, biotinidase deficiency; HCY, homocystinuria; GAL, galactosemia.

S292 CARROLL, DOWNS

Downloaded from www.pediatrics.org by on September 20, 2010
 TABLE 3 Comparison of Costs and Effectiveness TABLE 5 Cost-Effectiveness Analysis With the Pessimistic Case
Strategy Cost Incremental Effectiveness Incremental C/E Incremental Analysis Assumptions
Cost Effectiveness C/E Strategy Cost Incremental Effectiveness Incremental C/E Incremental
MS/MS $49.00 77.1957 0.64 Dominates Cost Effectiveness C/E
Panel $52.80 $3.70 77.19358 ⫺0.00212 0.68 No test $39.5 77.192 0.51
MS/MS testing for PKU, biotinidase deficiency, MSUD, galactosemia, and homocystinuria was Panel $40.5 $1.00 77.196 0.00322 0.52 310.56
compared with a panel of conventional testing methods for the same diseases. MS/MS also MS/MS $43.5 $3.00 77.196 0.00062 0.56 4838.71
tests for medium-chain acyl-CoA deficiency, for which there is no other screening test available.
Cost-effectiveness (C/E) of screening for PKU, biotinidase deficiency, MSUD, galactosemia, and
C/E indicates cost-effectiveness.
homocystinuria with a panel of conventional testing methods was compared with that of
MS/MS testing for the same conditions plus medium-chain acyl-CoA dehydrogenase defi-
ciency and no testing.
exceptions are screening for CAH, which costs slightly
more than $20 000 per QALY gained, and screening for
galactosemia, which costs approximately $94 000 per as specificity decreased. When the specificity of a screen-
QALY. The screening test with the lowest expected cost ing test was lowered, a larger proportion of unaffected
is MS/MS. children had false-positive test results, which resulted in
In addition to comparing each test with a no-testing increased costs to rule out disease. When the specificity
strategy, we compared a panel of conventional tests for of a test was below a threshold value, some tests were no
PKU, biotinidase deficiency, MSUD, galactosemia, and longer cost saving (Table 1); however, threshold values
homocystinuria with MS/MS methods to screen for the were below those reported in the literature.
same conditions and MCAD deficiency. The cost and The prevalence of a disease also affected its incremen-
effectiveness comparison is shown in Table 3. When tal cost. Biotinidase deficiency, CH, homocystinuria,
used in this multiplex manner, MS/MS was less expen- MSUD, and PKU all became costly when the prevalence
sive and more effective than the comparable panel of was decreased to values between 3 and 10 times lower
conventional screening tests. than baseline values. When the prevalence of CAH was
very high (3.5%), screening for CAH became cost saving.
Sensitivity Analysis The cost of screening for galactosemia was not sensitive
Because most screening tests were dominant over not to prevalence, and the cost of MS/MS screening was not
screening, we conducted sensitivity analysis on all of the sensitive to the prevalence of MCAD deficiency.
variables in the model, to determine at what levels the The results were relatively insensitive to the rates at
tests were no longer cost saving. The fourth column of which individual sequelae occurred. Screening for CH
Table 1 shows the threshold value for each variable, at was not cost saving if the rate of mild developmental
which the corresponding screening strategy becomes delay without treatment was less than one tenth of
more costly than not screening. baseline values or if the risk of severe delay was de-
In the base-case analysis, we assumed that the sensi- creased by approximately one half. If the risk of blind-
tivity of all of the screening tests was 100%. When we ness resulting from biotinidase deficiency was ⬍40%,
varied test sensitivity from 0% to 100%, the dominant then screening for biotinidase deficiency alone was no
strategies remained dominant to sensitivities well below longer cost saving. Screening for MSUD alone was not
those in the literature, as follows: PKU, 10%; CH, 67%; cost saving if, without treatment, the risk of severe CP
biotinidase deficiency, 13%; MSUD, 83%; homocystin- was ⬍14% or that of severe developmental delay was
uria, 37%; MS/MS, 27% (Table 4). ⬍34%. Because MS/MS screens for many conditions, it
Because of the costs of evaluating false-positive re- remained cost saving despite changes in any one of these
sults, the dominant testing strategies became more costly parameters. Testing for galactosemia had a positive net
cost in all sensitivity analyses; however, when the risk of
neonatal death was ⬎27%, the cost of screening fell
TABLE 4 Thresholds at Which Tests Are No Longer Cost Saving
below $50 000 per QALY saved.
Strategy Sensitivity Cost of Cost of Cost of
The cost savings with the screening tests depended in
False-Positive Severe Delay Severe CP
Result some cases on the effectiveness of treatment in prevent-
ing specific sequelae. Screening for CH was not cost
MS/MS testing 0.27 ⬎$10 000 a a

PKU test 0.1 a a a saving if it was ⬍67% effective in preventing severe

BIOT test 0.13 $600 a a delay. Screening for homocystinuria was not cost saving
CH test 0.67 ⬎$10 000 $450 000 a
if treatment was ⬍31% effective in preventing blind-
HCY test 0.37 $2600 a a
ness. MSUD screening was not cost saving if treatment
MSUD test 0.83 ⬎$10 000 $685 000 $540 000
b b b b
was ⬍59% effective in preventing CP or ⬍36% effective
GAL test
CAH test b b b b in preventing delay.
BIOT indicates biotinidase deficiency; HCY, homocystinuria; GAL, galactosemia. Our results were not sensitive to the costs of screening
a Screening strategy is cost saving at any value of the variable. tests. Screening test costs that exceeded savings from
b Screening strategy is not cost saving at any value of the variable.

SUPPLEMENT S293
Downloaded from www.pediatrics.org by on September 20, 2010
disease costs were between 2 and 10 times reported costs sults must be minimized to realize the cost savings of
of screening tests. Conversely, screening for galactosemia these screening strategies.
or CAH was not cost saving even if tests could be per- Our pessimistic case analysis was intended to elimi-
formed for free. nate biases favoring newborn screening. With these as-
Lifetime costs of diseases and their sequelae are diffi- sumptions, newborn screening with conventional test-
cult to estimate. However, we found that the cost saving ing or MS/MS methods was not cost saving. However,
for most of these screening tests was either insensitive or the incremental costs per QALY were very low, com-
sensitive at values significantly below reported rates. pared with conventional health economic bench-
Our pessimistic case analysis biased the model against marks.38,39
screening. Under this set of assumptions, newborn As with all cost-effectiveness analyses, our study is
screening with MS/MS methods was not cost saving limited by the validity of our assumptions about risks,
relative to a conventional screening panel, and the costs, and quality of life. For example, we assumed that
screening panel was not cost saving relative to no data from small cohort studies would predict accurately
screening (Table 5). The conventional screening panel the occurrence of sequelae in the population at large.
cost $310.56 per QALY saved over not screening. Com- We were forced to make some broad assumptions about
pared with this panel, MS/MS screening cost $4838.71 lifetime costs that could not take into account changes in
per QALY saved. technology or care. It is also inherently difficult to esti-
mate quality of life across different disease states, espe-
cially with respect to children. However, we found that
DISCUSSION incremental cost-effectiveness was relatively unaffected
Newborn screening for PKU, CH, biotinidase deficiency, by even large changes in nearly all of our variables.
MSUD, and homocystinuria individually not only was Another serious weakness of this study is the cost
cost-effective but actually was cost saving in our base- estimates for most of the conventional screening tests.
case analysis. Use of MS/MS to screen for PKU, biotini- We used estimates from PriceWaterhouseCoopers.33 This
dase deficiency, MSUD, and homocystinuria, as well as study merely compiled global estimates by various new-
MCAD deficiency, had even greater cost savings because born screening programs. There was no effort to divide
of the multiplicity of conditions detected with a single fixed and marginal costs, and programs were not con-
test. This was true even when MS/MS was compared sistent in which costs were included in their estimates.
directly with a panel of available conventional tests for Nevertheless, these are the best estimates currently
the same conditions. Screening for CAH had a net cost available. A more-careful microeconomic analysis of
per QALY gained; however the cost was less than the screening program costs would be needed to improve these
$50 000 per QALY used conventionally as a benchmark estimates; however, considering our sensitivity analyses, it
for cost-effectiveness.38,39 seems unlikely that more carefully derived costs would
At more than $90 000 per QALY, screening for galac- show that newborn screening was not cost-effective.
tosemia generally would not be considered cost-effec- Newborn screening seems to be one of the rare health
tive. This is because identifying galactosemia early may care interventions that is beneficial to patients and, with
prevent a neonatal death but doing so induces high some assumptions, cost saving. Over the long term,
lifetime costs because of sequelae. In our sensitivity anal- funding comprehensive newborn screening programs
yses, we found that under no circumstances would may save money for society. Even with pessimistic as-
screening for galactosemia be cost saving. However, if sumptions, newborn screening seems to be highly cost-
the risk of neonatal death resulting from galactosemia is effective. Although additional refinements of cost esti-
as high as 27%, then screening for galactosemia is cost- mates may be justified, they are unlikely to change these
effective according to conventional criteria. In addition, conclusions materially. Cost-effectiveness analysis will
the marginal cost of screening for galactosemia with continue to be a useful way to consider which screening
MS/MS or in a panel of newborn screens seems to be tests to include.
cost-effective.
Some results are sensitive to changes in the effective- ACKNOWLEDGMENTS
ness of screening and therapy, which emphasizes the This work was conducted under a subcontract from the
importance of an effective timely system for notification American College of Medical Genetics (agreement 240-
and follow-up evaluation of positive screening tests. Loss 01-0038), with funding from the Maternal and Child
of effectiveness in many areas may nullify the cost sav- Health Bureau, Health Resources and Services Adminis-
ings and health benefits of the screen. Similarly, the tration. Dr Carroll is supported by a grant from the
quality of the screening tests is important, specifically in National Institutes of Health (K23 DK67879-01).
minimizing false-positive results. The sensitivity of our We thank Alex Kemper, MD, MPH, MS, Tracy Lieu,
results to the specificity of tests and the cost of evaluating MD, MPH, and Scott Grosse, PhD, for their thoughtful
false-positive results emphasizes that false-positive re- comments on a draft of this manuscript.

S294 CARROLL, DOWNS

Downloaded from www.pediatrics.org by on September 20, 2010
REFERENCES 19. American Academy of Pediatrics, Committee on Genetics.
1. Venditti LN, Venditti CP, Berry GT, et al. Newborn screening by Newborn screening fact sheets. Pediatrics. 1996;98:473–501
tandem mass spectrometry for medium-chain acyl-CoA dehy- 20. National Newborn Screening and Genetics Resource Center. Ho-
drogenase deficiency: a cost-effectiveness analysis. Pediatrics. mocystinuria. In: National Newborn Screening Report—2000. Austin,
2003;112:1005–1015 TX: National Newborn Screening and Genetics Resource Center;
2. American Academy of Pediatrics, Committee on Genetics. 2003:1–11. Available at: http://genes-r-us.uthscsa.edu/resources/
Newborn screening fact sheets. Pediatrics. 1989;83:449 – 464 newborn/00/ch7㛭complete.pdf. Accessed June 23, 2004
21. Washington State Department of Health. Least Burden and
3. Levy HL, Albers S. Genetic screening of newborns. Annu Rev
Cost Benefit Analysis, Newborn Screening for Metabolic Dis-
Genomics Hum Genet. 2000;1:139 –177
orders, WAC 246-650; August 12, 2003. Available at: www.
4. Waisbren SE, Levy HL. Expanded screening of newborns for
sboh.wa.gov/Meetings/Meetings_2003/2003-10_15/documents/
genetic disorders. JAMA. 2004;291:820 – 821
Tab09-NBS_analysis.pdf. Accessed June 23, 2004
5. Guthrie R, Susi A. A simple phenylalanine method for detect-
22. Cruysberg JR, Boers GH, Trijbels JM, Deutman AF. Delay in
ing phenylketonuria in large populations of newborn infants.
diagnosis of homocystinuria: retrospective study of consecutive
Pediatrics. 1963;32:338 –343
patients. BMJ. 1996;313:1037–1040
6. Centers for Disease Control and Prevention. Using tandem 23. Naughten ER, Jenkins J, Francis DE, Leonard JV. Outcome of
mass spectrometry for metabolic disease screening among new- maple syrup urine disease. Arch Dis Child. 1982;57:918 –921
borns. MMWR Morb Mortal Wkly Rep. 2001;50:1–22. Available 24. Scriver CR. The Metabolic and Molecular Bases of Inherited Disease.
at: www.cdc.gov/mmwr/preview/mmwrhtml/rr5003a1.htm. 8th ed. New York, NY: McGraw-Hill; 2001
Accessed May 13, 2004 25. Wolf B, Spencer R, Gleason T. Hearing loss is a common
7. American College of Medical Genetics/American Society of feature of symptomatic children with profound biotinidase
Human Genetics Test and Technology Transfer Committee deficiency. J Pediatr. 2002;140:242–246
Working Group. Tandem mass spectrometry in newborn 26. Shield JP, Wadsworth EJ, MacDonald A, et al. The relationship
screening. Genet Med. 2000;2:267–269 of genotype to cognitive outcome in galactosaemia. Arch Dis
8. Naylor EW, Chace DH. Automated tandem mass spectrometry Child. 2000;83:248 –250
for mass newborn screening for disorders in fatty acid, organic 27. Badawi N, Cahalane SF, McDonald M, et al. Galactosaemia: a
acid, and amino acid metabolism. J Child Neurol. 1999;14(suppl controversial disorder: screening and outcome: Ireland
1):S4 –S8 1972–1992. Ir Med J. 1996;89:16 –17
9. Insinga RP, Laessig RH, Hoffman GL. Newborn screening with 28. Pollitt RJ, Leonard JV. Prospective surveillance study of me-
tandem mass spectrometry: examining its cost-effectiveness in dium chain acyl-CoA dehydrogenase deficiency in the UK.
the Wisconsin Newborn Screening Panel. J Pediatr. 2002;141: Arch Dis Child. 1998;79:116 –119
524 –531 29. Arias E, Smith BL. Deaths: preliminary data for 2001. Natl Vital
10. Schoen EJ, Baker JC, Colby CJ, To TT. Cost-benefit analysis of Stat Rep. 2003;51:1– 44
universal tandem mass spectrometry for newborn screening. 30. Bittles AH, Petterson BA, Sullivan SG, Hussain R, Glasson EJ,
Pediatrics. 2002;110:781–786 Montgomery PD. The influence of intellectual disability on life
11. Wood JC, Magera MJ, Rinaldo P, Seashore MR, Strauss AW, expectancy. J Gerontol A Biol Sci Med Sci. 2002;57:M470 –M472
Friedman A. Diagnosis of very long chain acyl-dehydrogenase 31. Bennett JE, Sumner W II, Downs SM, Jaffe DM. Parents’
deficiency from an infant’s newborn screening card. Pediatrics. utilities for outcomes of occult bacteremia. Arch Pediatr Adolesc
2001;108(1). Available at: www.pediatrics.org/cgi/content/ Med. 2000;154:43– 48
full/108/1/e19 32. Sharma S, Brown GC, Brown MM, et al. Converting visual
12. Waisbren SE, Albers S, Amato S, et al. Effect of expanded acuity to utilities. Can J Ophthalmol. 2000;35:267–272
newborn screening for biochemical genetic disorders on child 33. PriceWaterhouseCoopers. Newborn Screening Programs: An
outcomes and parental stress. JAMA. 2003;290:2564 –2572 Overview of Cost and Financing. New York, NY: Price
WaterhouseCoopers; 2002. Available at: www.marchofdimes.
13. Kwon C, Farrell PM. The magnitude and challenge of false-
com/files/Final㛭PWC㛭NBS㛭Report2.pdf. Accessed June 3, 2004
positive newborn screening test results. Arch Pediatr Adolesc
34. US Congress, Office of Technology Assessment. Data and
Med. 2000;154:714 –718
methods used in OTA’s cost-effectiveness analysis of strategies
14. Keren R, Helfand M, Homer C, McPhillips H, Lieu TA. Pro-
for newborn screening. In: Healthy Children: Investing in the
jected cost-effectiveness of statewide universal newborn hear-
Future. Washington, DC: US Government Printing Office; 1988:
ing screening. Pediatrics. 2002;110:855– 864
236 –241. Publication OTA-H-345. Available at: www.
15. Panepinto JA, Magid D, Rewers MJ, Lane PA. Universal versus
wws.princeton.edu/cgi-bin/byteserv.prl/⬃ota/disk2/1988/8819/
targeted screening of infants for sickle cell disease: a cost- 881919.PDF. Accessed June 3, 2004
effectiveness analysis. J Pediatr. 2000;136:201–208 35. Centers for Disease Control and Prevention. Economic costs
16. Jensen UG, Brandt NJ, Christensen E, Skovby F, Norgaard- associated with mental retardation, cerebral palsy, hearing loss,
Pedersen B, Simonsen H. Neonatal screening for galactosemia and vision impairment: United States, 2003. MMWR Morb Mor-
by quantitative analysis of hexose monophosphates using tan- tal Wkly Rep. 2004;53:57–59
dem mass spectrometry: a retrospective study. Clin Chem. 2001; 36. Angus DC, Barnato AE, Linde-Zwirble WT, et al. Use of inten-
47:1364 –1372 sive care at the end of life in the United States: an epidemio-
17. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. logic study. Crit Care Med. 2004;32:638 – 643
Recommendations of the Panel on Cost-Effectiveness in Health 37. US Department of Labor, Bureau of Labor Statistics. Inflation
and Medicine. JAMA. 1996;276:1253–1258 calculator. Available at: www.bls.gov/cpi. Accessed July 14, 2004
18. National Newborn Screening and Genetics Resource Center. Na- 38. Goldman L, Gordon DJ, Rifkind BM, et al. Cost and health im-
tional Newborn Screening Report—2000. Austin, TX: National New- plications of cholesterol lowering. Circulation. 1992;85:1960 –1968
born Screening and Genetics Resource Center; 2003. Available at: 39. Hirth RA, Chernew ME, Miller E, Fendrick AM, Weissert WG.
http://genes-r-us.uthscsa.edu/resources/newborn/00chapters.html. Willingness to pay for a quality-adjusted life year: in search of
Accessed June 3, 2004 a standard. Med Decis Making. 2000;20:332–342

SUPPLEMENT S295
Downloaded from www.pediatrics.org by on September 20, 2010
 Comprehensive Cost-Utility Analysis of Newborn Screening Strategies
Aaron E. Carroll and Stephen M. Downs
Pediatrics 2006;117;S287-S295
DOI: 10.1542/peds.2005-2633H
Updated Information including high-resolution figures, can be found at:
& Services http://www.pediatrics.org/cgi/content/full/117/5/S1/S287
Citations This article has been cited by 5 HighWire-hosted articles:
http://www.pediatrics.org/cgi/content/full/117/5/S1/S287#othera
rticles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Premature & Newborn
http://www.pediatrics.org/cgi/collection/premature_and_newbor
n
Permissions & Licensing Information about reproducing this article in parts (figures,
tables) or in its entirety can be found online at:
http://www.pediatrics.org/misc/Permissions.shtml
Reprints Information about ordering reprints can be found online:
http://www.pediatrics.org/misc/reprints.shtml

Downloaded from www.pediatrics.org by on September 20, 2010