DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY SYSTEMATIC REVIEW

Economic evaluation and cost of interventions for cerebral palsy:

a systematic review
SOPHY T F SHIH 1 | UTSANA TONMUKAYAKUL 1 | CHRISTINE IMMS 2 | DINAH REDDIHOUGH 3,4 |
H KERR GRAHAM 5 | LIZ COX 6 | ROB CARTER 1

1 Deakin Health Economics, Centre for Population Health Research, Faculty of Health, Deakin University, Geelong, Victoria; 2 Centre for Disability and Development
Research, Australian Catholic University, Fitzroy, Victoria; 3 Murdoch Children’s Research Institute, Parkville, Victoria; 4 The University of Melbourne, Parkville, Victoria;
5 Royal Children’s Hospital, Parkville, Victoria; 6 Victorian Managed Insurance Authority, Melbourne, Victoria, Australia.
Correspondence to Sophy T F Shih at Deakin Health Economics, Centre for Population Health Research, Faculty of Health, Deakin University, 221 Burwood Highway, Burwood, Vic. 3125,
Australia. E-mail: [email protected]

This systematic review is commented on by Marceniuk on page 532 of this issue.

PUBLICATION DATA AIM Economic appraisal can help guide policy-making for purchasing decisions, and
Accepted for publication 20th November treatment and management algorithms for health interventions. We conducted a systematic
2017. review of economic studies in cerebral palsy (CP) to inform future research.
Published online 10th January 2018. METHOD Economic studies published since 1970 were identified from seven databases. Two
reviewers independently screened abstracts and extracted data following the Preferred
ABBREVIATIONS Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Any
CEA Cost-effectiveness analysis discrepancies were resolved by discussion.
CUA Cost-utility analysis RESULTS Of 980 identified references, 115 were included for full-text assessment. Thirteen
ICER Incremental cost-effectiveness articles met standard criteria for a full economic evaluation, two as partial economic
ratio evaluations, and 18 as cost studies. Six were full economic evaluations alongside clinical
ITB Intrathecal baclofen studies or randomized controlled trials, whereas seven involved modelling simulations. The
QALY Quality-adjusted life-year economic case for administration of magnesium sulfate for imminent preterm birth is
RCT Randomized controlled trial compelling, achieving both health gain and cost savings. Current literature suggests
intrathecal baclofen therapy and botulinum toxin injection are cost-effective, but stronger
evidence for long-term effects is needed. Lifestyle and web-based interventions are
inexpensive, but broader measurement of outcomes is required.
INTERPRETATION Prevention of CP would avoid significant economic burden. Some
treatments and interventions have been shown to be cost-effective, although stronger
evidence of clinical effectiveness is needed.

Cerebral palsy (CP) describes a group of developmental The precise aetiology of CP is still unclear. Risk factors
disorders of movement and posture, causing activity for CP include preterm birth, multiple pregnancy, intra-
restriction or disability attributed to disturbances occurring amniotic infection, perinatal inflammation, low maternal
in the fetal or infant brain. The motor impairment may be thyroid hormone levels, perinatal asphyxia, placenta abnor-
accompanied by a seizure disorder and by impairment of malities, fetal growth retardation, and neonatal hyperbiliru-
sensation, cognition, communication and/or behaviour, and binaemia. While CP involves damage to the central
by secondary musculoskeletal problems.1 CP has a preva- nervous system, clinical symptoms of CP are predomi-
lence of approximately 1 in 500 neonates, with 17 million nantly observed in the musculoskeletal system. In addition,
people affected worldwide.2 The overall prevalence of CP 31% of children with CP born between 1993 and 2006
in high-income countries is 2.11 per 1000 live births,3 and had epilepsy, 5% were blind, 2% were deaf, and 44% had
2.0 to 2.8 per 1000 live births in low- and middle-income intellectual impairment.7
countries.4 In Australia, after a long period of stable preva- CP is a lifelong condition, with profound impacts on
lence at 2 to 2.5 per 1000 live births, the rate of CP the individuals, as well as their family. Mothers of chil-
declined to 1.4 to 2.1 per 1000 live births between 2007 dren with CP have poorer mental and physical health
and 2009.5 The downward trend is particularly evident in outcomes than mothers of children with typical develop-
infants born extremely preterm. The severity and complex- ment.8,9 From a broader perspective, CP also significantly
ity of CP has also declined. The CP registers in some Aus- affects education and welfare systems. Given the burden
tralian states have estimated the current CP prevalence to and impact of CP, improvements in our knowledge base
be less than 1.5 per 1000 live births.6 to guide treatment and prevention is crucial. While

© 2018 Mac Keith Press DOI: 10.1111/dmcn.13653 543

clinical research has progressed our understanding of risk What this paper adds
factors and efficacy, our knowledge of cost-effectiveness • Cost-effectiveness evidence shows prevention is the most significant strategy.
and how best to utilize available budgets is poorly docu- • Some treatments are cost-effective, but stronger evidence for long-term
mented. effectiveness is required.
We conducted a systematic review to provide informa- • Comparison of treatment costs is challenging owing to variations in method-
ologies and varying clinical indications.
tion on the economic aspects of CP that have been
researched and to identify the strengths/weaknesses and carried out to identify additional relevant studies. The
gaps in the current economic research. We aim to address search of published studies was performed in December
the cost and cost-effectiveness of interventions focusing on 2015 and updated in April 2017.
CP for individuals from the pre-/perinatal period through
to adolescence in studies reporting clinical investigations Inclusion and exclusion criteria
and/or modelling simulations. This study is, to our knowl- Two reviewers (STFS and UT) independently screened all
edge, the first systematic review on economic evidence for titles and abstracts identified. Publications meeting the
CP-specific interventions. inclusion criteria were included for full-text assessment.
Discrepancies between two screening results were discussed
METHOD (STFS, UT, RC) and resolved. Inclusion for full-text
A systematic review was conducted, following the assessment were those articles reporting costs, cost esti-
Cochrane Systematic Review and the Preferred Reporting mates, cost-effectiveness, cost–benefit, or economic evalua-
Items for Systematic Reviews and Meta-Analyses tion in CP. Studies that only reported CP as one of the
(PRISMA) guidelines, to identify economic studies on outcomes measured, rather than investigating interventions
interventions for children with CP.10,11 for CP prevention or treatment, were excluded. For exam-
ple, a wide range of perinatal interventions to prevent
Search strategy adverse birth outcomes have been studied. The studies that
Seven electronic databases (Ovid MEDLINE, Embase, included CP as one of the outcomes with less than 10% of
CINHAL, PsycINFO, Econlit, Health Economic Evalua- cases reported were excluded from the current review.
tion Database, and NHS Economic Evaluation Database Economic evaluation studies were appraised against the
in the Cochrane Library) were searched with the following checklist criteria for assessing economic evaluation set by
search terms: (1) ‘cerebral palsy’; (2) ‘health care costs’ and Drummond et al.12 The checklist of Drummond et al.
‘cost analysis’; and (3) ‘economic evaluation’ (including consists of 10 important areas: (1) well-defined questions;
‘cost-effectiveness analysis’ and ‘cost-benefit analysis’). The (2) comprehensive description of the competing alterna-
search strategies for various databases are documented in tives; (3) established effectiveness; (4) identification of rele-
Table I. The literature search was restricted to articles vant costs and consequences; (5) appropriate measurement
written in English and published from 1970 to the present. of costs and consequences; (6) credible valuation of costs
A manual search of references in selected articles was also and consequences; (7) costs and consequences adjusted for
timing; (8) incremental analysis; (9) uncertainty analysis;
(10) discussion of issues of concern to users.
Table I: Search strategies for various databases

Database Search terms Data extraction

A data extraction template was developed and utilized for
Ovid Medline
1 “cerebral palsy”.sh
each of the studies selected for full-text analysis. Extracted
2 cost*.ab. or “costs and cost analysis”.sh. or information included author names, year of publication,
“cost benefit analys*”.ab. or “cost-benefit country of study, aim of study, study type, population
analysis”.sh. or “health care costs”.xs
3 1 and 2
included, intervention studied, comparison group, time
Embase horizon of study, currency and reference year, main results,
#1 ‘cerebral palsy’:ab,ti and comments for discussion. Data were extracted inde-
#2 ‘cost’:ab,ti
#3 ‘cost effectiveness analysis’:ab,ti
pendently by two reviewers (STFS and UT). Discrepancies
#4 ‘cost benefit analysis’:ab,ti on extracted data were cross-checked and discussed, with
#5 ‘health care costs’:ab,ti final decisions agreed between STFS, UT, and RC.
#6 #2 OR #3 OR #4 OR #5
#7 #1 AND #6
CINAHL, PsycINFO, HEED, EcoLit databases Outcome measures
S1 MJ cerebral palsy or TI cerebral palsy or AB cerebral Two outcome measures were targeted in the review: ‘cost’
palsy
S2 TX (cost effective analysis or cost effective* or cost
(defined according to the method used in the study, e.g.
benefit analysis or economic evaluation) total cost, per case, per episode, lifetime, etc.) and the
S3 TX (randomized controlled trial or rct) ‘cost-effectiveness ratio’ (e.g. incremental cost-effectiveness
S4 S1 AND S2 AND S3
ratio [ICER]). Costs were normally reported as costs of
Sh, subheading; ab, abstract; xs, exploded subheading; ti, title; MJ, specific treatment for CP. Health economics studies
word in major subject healding; S, search.

544 Developmental Medicine & Child Neurology 2018, 60: 543–558

presenting the burden of disease or cost of illness related RESULTS
to CP were excluded from the present paper and reported There were 980 references identified from the electronic
elsewhere.13 The focus of the present paper was to provide databases. After removal of duplicates, 706 titles and
a value judgement of efficiency for interventions targeting abstracts were screened. Based on our inclusion criteria,
CP, whereas the burden of disease or cost of illness studies 115 articles were included for full-text assessment. Of these,
presented a description of the magnitude of the condition 15 articles met the criteria of economic evaluation and 18
or impact of the problem without enough information to publications reported cost of treatment. The PRISMA
make value judgements. Although studies reporting cost of flowchart, demonstrating the identification, screening, eligi-
specific treatment or therapy do not assist in making a bility, and inclusion of studies, is shown in Figure 1.
value judgement of efficiency, they were included in the The 15 included articles reported economic evaluation
present paper as they provided information about budget results of ICER for the following interventions: prevention
implications of interventions for CP. For the purposes of of CP (n=4, including one conference abstract);14–17 botuli-
this review, arguable assumptions made in included studies, num toxin A (BoNT-A) injection (n=4, including one con-
if any, were also recorded to flag variations between ference abstract);18–21 intrathecal baclofen (ITB) therapy
studies. (n=3);22–24 functional electrical stimulation (n=1);25

Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. aBoD studies were reported elsewhere. NHS,
National Health Service; EED, Economic Evaluation Database; HEED, Health Economic Evaluations Database; BoD, burden of disease.

Review 545
proximal femoral hardware retention (n=1);26 web-based system’ perspective, the model predicted an ICER of CAN
home program (n=1);27 and lifestyle intervention (n=1).28 $2083 per QALY gained and CAN$28 755 per CP
Seven of these economic evaluations were conducted by a averted. With the decision threshold for cost-effectiveness
modelling approach, whereas six were economic evaluations of ‘$50 000 per QALY’ applied in countries like Australia,
alongside clinical studies or randomized controlled trials the UK, and Canada, a result of CAN$2000 per QALY
(RCTs). The results of the data extractions from the eco- would suggest strong cost-effectiveness.
nomic evaluation studies are listed in Table II. Two studies The study of Cahill et al. also suggested that, compared
were classified as partial economic evaluations (i.e. func- with no treatment, magnesium sulfate was the dominant
tional electrical stimulation and proximal femoral hardware strategy (saving costs and gaining more health benefits)
retention) as either no comparison group data was presented when given to females with preterm, premature rupture of
or no comparison of outcome and cost were carried out. membranes, as well as females at risk of delivery before
There were 18 publications that reported on the costs of 28 weeks’ gestation.16 However, the intervention was not
treatment and therapy in 15 studies, either as part of a cost-effective when the risk reduction in moderate-to-
clinical study assessing effectiveness, or as a dedicated eco- severe CP by the magnesium treatment was less than 14%.
nomic study. Among these, six studies were cost analyses
or budget impact analyses of BoNT-A; four involved ITB Other preventative interventions for CP
therapy; four were cost studies of various surgical proce- One study reported a trial-based economic evaluation of
dures; two were assessments of gait analysis; and two stud- two fetal surveillance strategies focused specifically on the
ies focused on community-based care and a specialized prevention of CP.17 The trial results indicated a reduction
seating delivery model. Details of included economic eval- of CP cases through the surveillance strategies of car-
uation and cost analysis studies are discussed blow. diotocography plus ST-segment waveform analysis of the
fetal electrocardiogram versus cardiotocography alone.
Economic evaluation studies on prevention of CP Without a common metric like QALYs, this result is lim-
Magnesium sulfate ited in informing the ‘value-for-money’ of surveillance per
Administration of magnesium sulfate in females at risk of se, but is useful for decisions around technical efficiency
preterm birth to prevent CP is the most promising inter- (i.e. how to offer surveillance to whom with what modality,
vention of those studied to date. Two decision-analytic if a decision is to be made to offer fetal surveillance). We
modelling studies were carried out by Cahill et al. and are unable to conclude whether or not this intervention is
Bickford et al. to assess the cost-effectiveness of this inter- cost-effective by the threshold of $50 000 per QALY.
vention.14,16 Both studies modelled the cost and outcome However, fetal surveillance is significantly less cost-effec-
for a lifetime horizon and were undertaken from both the tive by the ICER of €167 854 per one CP case prevented
‘health system’ perspective to account for resources compared with the ICER of CAN$28 755 per CP averted
incurred in the health care sector, including out-of-pocket by magnesium intervention.
costs to the individuals, and the ‘societal’ perspectives to
capture a broader range of impacts. Economic evaluation studies on treatment for CP
In the study by Cahill et al.,16 cost was estimated in US BoNT-A injection
dollars and outcomes were reported as quality-adjusted life- Three studies have assessed the cost-effectiveness of
year (QALY). The strategies were also compared in terms BoNT-A injection in the management of spasticity, but
of ‘cases prevented’, ‘neonatal or infant deaths’, and cost- they demonstrate inconsistent results.19–21 In a modelled
effectiveness ratios. In the study by Bickford et al.,14 CP evaluation in the Australian setting, conducted as part of
was the only neonatal outcome included in the models and an application to the Pharmaceutical Benefits Advisory
all costs were presented in 2011 Canadian dollars (CAN$). Committee for drug listing, the case is based on equivalent
Data from four RCTs were pooled to obtain the probabili- efficacy of BoNT-A injection and serial casting. The
ties of CP of various severities for the Canadian study.29–32 effects of BoNT-A lasted longer and therefore it was
For imminent preterm birth, use of magnesium sulfate regarded as the preferred treatment.21 The efficacy of the
was a dominant strategy that is both less costly and more treatment was modelled from RCTs including Australian
effective than the alternative of no treatment from both patients.33,34 A cost consequences analysis was undertaken
the ‘health system perspective’ and the broader ‘societal in which the costs and consequences of treatment were
perspective’.14 The broader perspective included productiv- presented separately and only the direct medical costs were
ity costs, such as all costs associated with lost labour mar- included. The additional cost for BoNT-A injection, dis-
ket productivity for adults with CP, and social costs counted at 5% annually, was 793 Australian dollars (AUD
consisting of all costs associated with specialized education, $) for patients with hemiplegia and AUD$867 for diplegia,
specialized housing, and lost labour market productivity over a treatment interval of 3.7 years. The study concluded
for primary care providers of children with CP. that the additional costs associated with BoNT-A over and
For threatened preterm birth, the intervention was dom- above serial casting treatment were modest and could be
inant from the ‘societal’ perspective and cost-effective from offset by indirect costs (e.g. travel costs) if analysed from a
a ‘health system’ perspective. From the narrower ‘health societal perspective.

546 Developmental Medicine & Child Neurology 2018, 60: 543–558

 Table II: Data extraction from studies reporting an economic evaluation of interventions for cerebral palsy (CP)

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

Magnesium sulfate
Bickford et al.14 Canada To assess the cost- Decision-tree models Life expectancies at Magnesium sulfate No treatment Patients in whom PTB From a health system and All costs are presented in
effectiveness of and probabilistic birth for fetal at <32+0wks’ a societal perspective 2011 Canadian dollars
administering sensitivity analyses neuroprotection gestation is respectively, savings of ($1CAN=$1US=£0.62)
magnesium sulfate using TreeAge imminent or CAN$2242 and CAN Both costs and QALYs
to patients in whom threatened $112,602 are obtained for were discounted at 3%
PTB at <32wks’ each QALY gained annually for the base
gestation Savings of CAN$30 942 case analyses
and CAN$1 554 198 is
obtained for each case of
CP averted when
magnesium sulfate is
administered to patients
in whom PTB is imminent
From a health system
perspective and a societal
perspective respectively,
a cost of CAN$2083 is
incurred and savings of
CAN$108 277 is obtained
for each QALY gained
and a cost of CAN$28 755
is incurred and a saving
of CAN$1 494 500 is
obtained for each case of
CP averted when
magnesium sulfate is
administered to patients
in whom PTB is
threatened
Cahill et al.16 USA To estimate the cost- Decision analytic and Life expectancy of 75y Magnesium sulfate No treatment Females carrying Magnesium sulfate for
effectiveness of cost-effectiveness therapy for the singleton neuroprophylaxis led to
magnesium sulfate model prevention of CP pregnancies between lower costs (USD$1739 vs
neuroprophylaxis for in infants born 24wks 0/7d and USD$1917) and better
all females at risk for preterm 31wks 6/7d gestation outcomes (56.684 vs
PTB before 32wks at high risk for 56.678 QALYs)
from a societal spontaneous PTB For every 10,000 females at
perspective 32wks’ gestation due risk for preterm birth
to either PTL or treated with magnesium,
PPROM USD$1.8 million were
saved and 52 QALYs
were gained

Review
547
 Table II: Continued

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

15
Cahill et al. USA To estimate the cost- Decision analytic and Magnesium sulfate No treatment Magnesium for
effectiveness of cost-effectiveness therapy for the neuroprophylaxis was the
magnesium model prevention of CP dominant strategy (i.e.
neuroprophylaxis for in infants born less expensive and more
all females at risk for preterm effective)
PTB before 32wks Considering an annual
from a societal birth rate of 4 million
perspective births, a strategy for
magnesium sulfate for

548 Developmental Medicine & Child Neurology 2018, 60: 543–558

neuroprophylaxis would
prevent 904 cases of
moderate-to-severe CP
each year
Vijgen et al.17 the Netherlands To establish the cost- A trial-based Fetal surveillance Cardiotocography The ICER was €167 854 per Resources used were
effectiveness of two economic evaluation during labour with only prevented case of CP documented by specific
strategies of fetal from a health care cardiotocography when using ST analysis items in the CRFs
surveillance during provider perspective plus ST analysis vs cardiotocography only, Unit costs for ST analysis
labour to prevent CP and varies between were based on
€78 719 and €393 594 depreciation costs of the
when using probabilities medical devices purchase
between 6% and 30% prices, as well as training
In the trial the number of costs for obstetricians
cases with metabolic The probability of
acidosis was 20 in the ST developing CP after
analysis group (n=2827) metabolic acidosis was
and 30 in the 14% as derived from a
cardiotocography group literature review
(n=2840)
Addition of ST analysis
reduces the risk of CP at
a considerable cost per
case
 Table II: Continued

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

BoNT-A injection
Tapias et al.18 Spain To conduct a CMA of CMA alongside a Average time of Abo-BoNT-A Ona-BoNT-A 936 spastic children No significant difference in The analysis was
Abo-BoNT-A vs Ona- observational follow-up: 18.6mo for aged 2–18y treated in the treatment conducted from the
BoNT-A in treating longitudinal study Abo-BoNT-A and the paediatric effectiveness between the perspective of the NHS,
patients with 19.06mo for Ona- neurology unit of a two BoNT-A. Therefore, a considering direct costs
paediatric spasticity BoNT-A Spanish hospital CMA was conducted only: pharmacological
The mean pharmacological costs (manufacturer’s
cost per patient and year selling price) and costs of
was €480 for Ona-BoNT-A visits in 2016 Euros if the
and €287 for Abo-BoNT-A, ‘typical patient’ profile is
which represents annual considered in sensitivity
savings in favour of Abo- analysis, the potential
BoNT-A of €193 in pharmacological savings
pharmacological cost with Abo-BoNT-A would
The total annual direct cost be €295 per patient per
obtained was €839 for year
Ona-BoNT-A and €631 for
Abo-BoNT-A,
representing a difference
of €208 per year in favour
of treatment with Abo-
BoNT-A
Catsman-Berrevoets the Netherlands To compare RCT BoNT-A followed by Intensive PT only 65 children with CP No statistically significant
et al.19 effectiveness and iPT (n=41) (n=24) aged 4–12y evidence was found for
cost-effectiveness of added value of BoNT-A
BoNT-A+iPT vs only injections in BoTN-A+iPT
intensive PT treatment for the primary
treatment in a RCT outcomes
Average treatment costs
(including plaster and
Ankle-Foot-Orthosis) were
significantly higher in
BoNT-A+iPT than in iPT
(€8963 vs €6182; p=0.001)

Review
549
 Table II: Continued

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

Yagudina et al.20 Russia To conduct a CEA of Decision tree model Abo-BoNT- Standard therapy The Abo-BoNT-A+standard The data on drugs efficacy
Abo-BoNT- A+standard includes centrally therapy helps to avoid (measured as proportion
A+standard therapy, therapy vs Ona- acting muscle surgical intervention in of patients with spastic
Ona-BoNT- BoNT-A+standard relaxant, 93% of patients vs 90% of forms of CP, avoided
A+standard therapy therapy vs physiotherapy, patients with Ona-BoNT- orthopaedic surgery at
and standard therapy standard therapy casting, and A+standard therapy and second year of therapy)
solely in patients solely orthosis for 48% of patients with was obtained from
with spastic CP standard therapy. available clinical trials.
The Abo-BoNT- The following costs were
A+standard therapy has taken into account: the
the lowest cost- costs of
effectiveness ratio pharmacotherapy,
(RUB$11 509 /USD$215) in inpatient and outpatient

550 Developmental Medicine & Child Neurology 2018, 60: 543–558

comparison with drug care, costs of adverse
therapy Ona-BoNT- events, sanatorium–resort
A+standard therapy (RUB medical care, casting
$12 872/USD$238) and orthopaedic surgeons,
standard therapy with a disability pensions,
centrally acting muscle disabled child care
relaxant without BoNT-A benefit
(RUB$27 715/USD$467)
by the end of 2y of
treatment
Houltram et al.21 Australia To compare two A cost consequence BoNT-A in the Serial casting Children with CP The RCTs demonstrated The economic evaluation
methods of analysis with a management of equivalent efficacy of was undertaken as part of
conservative simple economic equinus in children BoNT-A and serial an application to the
management of calf modelling approach with CP casting; however, with PBAC to obtain
spasticity and to establish resource BoNT the effect lasted reimbursement for
equinus gait  utilization by longer and was clearly BOTOX
intramuscular treatment arm the preferred treatment Only direct medical costs
injection of BoNT-A For patients with were considered (BoNT-
and serial casting hemiplegia the total A, medical personnel
additional cost, time and medical
discounted at 5% consumables)
annually, for BoNT-A is The costs associated with
AUD$793 the treatment of adverse
For patients with diplegia events have not been
the total additional cost considered because the
for BoNT-A is AUD$867 adverse effect is
reversible
 Table II: Continued

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

ITB therapy
Bensmail et al.22 France To assess the cost- Modelling study Various treatment ITB therapy Compared with Patients with disabling ITB therapy model Direct medical costs were
effectiveness of ITB sequences over 2y conventional spasticity and revealed a lower cost measured in Euros (2006)
therapy medical treatments functional (€59 391 vs €88 272; and based on a French
for disabling dependence caused p<0.001) and an overall retrospective cost survey
spasticity by any neurological more favourable cost- at Raymond Poincar

e
disease effectiveness ratio Hospital by Bensmail
22
(€75 204/success vs et al.
€148 822/success;
p<0.001), compared with
conventional medical
management without ITB
Hoving et al.23 the Netherlands To compare the costs Combined CEA/CUA CITB in children Standard treatment Gaining one QALY cost, on Health effects were
and health effects of alongside the Dutch with intractable only average, €32 737 (€28 273, expressed in terms of a
CITB with those of national study on the spastic CP using the UK EQ-5D VAS for individual
standard treatment efficacy and safety of index) problems and QALYs,
only, from the health CITB for intractable Additional mean annual derived from the Dutch
care perspective spasticity in children costs of CITB €3732 and EQ-5D index
with CP the mean intervention- Included intervention costs
related health care costs and other health care
at €4226 per year costs. For the latter, data
were collected by means
of a questionnaire and a
cost diary
Costs were estimated for
the year 2003 in Euros.
We discounted costs that
were not available for the
year 2003 with 4% per
year according to the
Dutch guidelines
de Lissovoy USA To assess the cost- CUA using ITB therapy for the Alternative medical Children with severe ITB therapy increased the HUI-2 to rate health states
et al.24 effectiveness of ITB mathematical treatment of and surgical spasticity of cerebral 5y cost of treatment by associated with the
among children with modelling and severe spasticity therapy origin who have not $49 000 relative to course of treatment by a
severe spasticity of computer simulation associated with CP responded to less alternative treatment panel of clinicians
cerebral origin invasive treatments accompanied by an Data on treatment costs
such as oral average gain of 1.2 QALY representative of these
medications; over a The net result was an children were derived
5y episode of incremental cost- from a health insurance
treatment effectiveness ratio of USD claims database that
$42 000 per QALY included both commercial
All costs were adjusted to and Medicaid data
base year 2003. Both
costs and QALYs are
discounted at an annual
rate of 3%

Review
551
 Table II: Continued

References Country Aim Study type Time horizon Population Comparison Population Study results Comments

Other treatment/intervention
Comans et al.27 Australia To estimate the cost- CEA alongside a 20wks The ‘Move it to Usual care 102 children aged 8– The intervention group
effectiveness of a randomized improve it’ (Mitii) 18y with unilateral had significantly higher
training system for controlled trial therapy is a web- spastic CP (GMFCS proportions of
improvements in based system level I or II) responders in two
upper-limb function designed to subscales of the AMPS
for children with facilitate intensive and the two subscales of
unilateral CP motor planning, the COPM. Power to
upper limb, gross detect a significant
motor, and difference was at least
cognitive 88% for all four outcomes
rehabilitation (AMPS-M, AMPS-P,
COPM-P, COPM-S

552 Developmental Medicine & Child Neurology 2018, 60: 543–558

responder)
ICER ranged from AU$3078
to $4191 per one unit of
effectiveness measured
by AMPS or COPM.
Bootstrapped credible
intervals ranged from
AUD$1553 for AMPS-P to
AUD$9574 for COPM-P.
Slaman et al.28 the Netherlands To evaluate the cost CUA alongside a 6mo lifestyle Usual care, no 57 adolescents and No significant differences QALYs were derived from
utility of a lifestyle single-blind, RCT intervention intervention to young adults with between groups were the SF-36 questionnaire
intervention among consisting of improve physical spastic CP found for direct medical using the SF-6D
adolescents and physical fitness behaviour and costs or productivity
young adults with CP training combined fitness costs
with counselling A cost-utility ratio of
sessions focusing –€23 664 per QALY was
on physical found for the lifestyle
behaviour and intervention vs no
sports treatment
participation

QALY, quality-adjusted life-year; PTB, preterm birth; PTL, preterm labour; PPROM, preterm premature rupture of membranes; ICER, incremental cost-effectiveness ratio; CRF, case report
form; BoNT-A, botulinum toxin A; CMA, cost-minimization analysis; NHS, National Health Service; PT, physiotherapy; RCT, randomized controlled trial; iPT, intensive physiotherapy; CEA,
cost-effectiveness analysis; PBAC, Pharmaceutical Benefits Advisory Committee; ITB, intrathecal baclofen; CUA, cost-utility analysis; CITB, continuous ITB; EQ-5D, EuroQol-5D; VAS, visual
analogue scale; HUI, Health Utilities Index; GMFCS, Gross Motor Function Classification System; AMPS, Assessment of Motor and Process Skills; COPM, Canadian Occupational Perfor-
mance Measure; SF-36; Short Form-36; SF-6D, Short Form-6D; ST, segment waveform analysis of the fetal electrocardiogram.
 Further, a more recent RCT (2015) showed no statisti- In a further study, an ICER of USD$42 000 per QALY
cally significant evidence for the added value of BoNT-A in a CUA was reported.24 The likelihood that ITB has a
treatment followed by intensive physiotherapy compared cost per QALY of less than or equal to USD$50 000 was
with intensive physiotherapy alone.19 The trial found greater than 70%. Similar to the modelling study discussed
trends towards the intervention effect in favour of only above,22 the CUA compared ITB with a conventional
intensive physiotherapy for improving gross motor func- medical approach among children who had not responded
tion (p=0.095), decreasing sedentary behaviour during to less-invasive treatments such as oral medications. A
everyday physical activity (p=0.087), and improving quality mathematical model simulated the experience of two
of life (p=0.066). The addition of BoNT-A to physiother- groups of children, an ITB group and an alternative treat-
apy improved everyday physical activity over 24 weeks’ ment group, followed over a 5-year treatment period.
follow-up (p=0.064), with a significantly higher treatment Based on results of 15 studies selected by the authors to
cost (€8963 vs €6182; p=0.001). The higher treatment cost identify the typical symptom profile of a child with severe
did not seem to be warranted for the improved daily physi- spasticity, five health states were established to describe a
cal activity in the short-term, but long-term impact is typical child receiving ITB. Utility weights for each health
uncertain. state were rated by a panel of nine clinicians using the
A third modelling study constructed a decision-tree model Health Utilities Index.35 By drawing random samples with
to simulate the effects of Abo-BoNT-A, Ona-BoNT-A, and bootstrapping techniques from the appropriate sets of util-
standard therapy.20 Treatment efficacy, measured as the pro- ity and cost values, the model created a set of data points
portion of patients with spastic CP who avoided orthopaedic for members of each of the two cohorts. ICERs were
surgery at the end of 2 years of therapy, was obtained from derived from the bootstrapping data sets of cost and utility
available clinical trials. Abo-BoNT-A plus standard therapy that generated QALY over a 5-year period.
is the most cost-effective treatment choice with the lowest In contrast to these desk-top modelling studies, a CEA/
ICER. Similar results were found in another, more recent CUA was carried out alongside a Dutch national study on
economic study of BoNT-A treatment in which two types of the efficacy and safety of ITB therapy for children with
BoNT-A (i.e. Abo-BoNT-A and Ona-BoNT-A) were com- CP.23 Data were collected from a sample of 15 young people
pared in 895 paediatric patients aged 2 to 18 years with spas- aged between 7 years and 17 years at the time of pump
ticity.18 This was a cost-minimization analysis, where the implantation with Gross Motor Function Classification Sys-
treatment effectiveness was equivalent between the two tem levels III to V. The economic evaluations compared the
drugs, conducted alongside a longitudinal observational costs and health effects of ITB with standard treatment only,
study. The total direct cost (pharmacological and medical for a 1-year period. Standard treatment included physical
visits) difference was €208 per year per child in favour of therapy, occupational therapy, and/or rehabilitation. Both
treatment with Abo-BoNT-A. the CEA and the CUA was undertaken from the health care
perspective, taking into account all relevant resources con-
ITB sumed. Additional health effects of ITB were assessed by
Despite the limited efficacy base, there were three eco- using the visual analogue scale for individual problems in
nomic evaluation studies on ITB, two modelling studies the CEA and the EuroQol-5D in the CUA.36 Owing to the
and one cost-effectiveness analysis (CEA)/cost-utility analy- small sample size, bootstrapping methods were undertaken
sis (CUA) alongside a national study on the efficacy of to verify the reliability of the results with 1000 replications.
ITB.22–24 The modelling studies aimed to simulate real-life The results showed ITB therapy was cost-effective by
scenarios. improving health outcome at a reasonable cost with an
One modelling study compared the cost-effectiveness of ICER of €32 737 per QALY (using the Dutch EuroQol-5D
ITB used as the first-line treatment with all other conven- index) and €28 273 per QALY (using the UK EuroQol-5D
tional treatment options offered to patients with spastic- index).
ity.22 The comparator was a package of specific current
treatments based on the most established French treatment Other therapies and interventions for treating CP
patterns, namely physical therapy only, oral antispasticity Recently, a CEA was conducted alongside an RCT to esti-
agents, focal spasticity treatments, neurosurgical interven- mate the cost and benefits of providing a multimodal web-
tions, nursing care, and ITB (plus ITB potential adjust- based program delivered at home to facilitate intensive
ment dose plus potential pump removal). Two decision motor planning, upper limb, gross motor, and cognitive
trees were constructed by simulation models using com- rehabilitation.27 The participants of the RCT were 102
puter programming languages that aimed to replicate real- children with spastic unilateral CP aged 8 to 18 years with
life clinical practices. Using ITB as the first-line strategy Gross Motor Function Classification System level I or II
in severely impaired individuals with disabling spasticity and Manual Ability Classification Scale levels I and III.
had a significantly higher success rate than conventional The ICER results reported as ‘cost per proportion of
medical management (78.7% vs 59.3%; p<0.001). ITB was responders’, defined as the minimum clinically important
considered to be the dominant strategy providing greater difference by either 0.3 logits on the Assessment of Motor
effectiveness at a lower cost. or Process Skills or 2 points on the Canadian Occupational

Review 553
Performance Measure, ranged from AUD$3078 to AUD therapy (ITB-free).41 Cost projections were developed over
$4191 compared with usual care. With modest costs and a a 30-year time horizon. Costs in the month of implant and
significant difference in proportion of responders for the in the following year were USD$26 375 more than con-
intervention group, the authors concluded that the inter- ventional management. However, financial break-even
vention offered a cost-effective program adjunct to direct occurred between the second and third years postITB
rehabilitation for limited costs and greater gains in health implant. The lifetime analysis indicated that ITB was cost
outcomes. saving, with USD$8009 saved per patient per year com-
A CUA of a lifestyle intervention was evaluated along- pared with conventional therapy (3% discount rate; 2007
side an RCT for 57 adolescents and young adults with CP reference year). Most of the savings were derived from
aged 16 to 24 years.28 The analysis examined a 6-month reductions in inpatient admissions, physician office visits,
lifestyle intervention consisting of physical fitness training and outpatient physiotherapy. However, another study that
combined with counselling sessions (focused on physical compared 9 months before and after implantation of ITB
behaviour and sports participation), compared with the indicated no significant difference in total costs.42
control group who continued with usual care. Intervention
costs, direct medical costs, and productivity costs were Surgical procedures
assessed, with 2009 Dutch reference unit prices. Quality of A micro-costing study detailing cost components associated
life was measured using the Short Form-36 and converted with an intervention, was conducted to determine health
into Short-Form-6D utility scores.37 The preliminary care costs of upper-extremity surgical correction in 39 chil-
results showed the intervention group gained 0.0131 dren with spastic CP at a Dutch hospital.43 The average
QALYs with a lower annual total cost of €310, compared hospital cost was €6813 per child (reference year 2014),
with the control group. However, none of the comparisons consisting of medical costs from the first contact until
in cost or outcome between the two groups were statisti- 9 months after surgery. Rehabilitation costs were estimated
cally significant. Bootstrapping results showed 86% of at €3599 per child with an average of 3.5 months duration
ICERs were less €20 000 per QALY. The study suggests a of the rehabilitation program.
lifestyle intervention is cost-saving or cost-effective com- A costing exercise to produce patient level costing data
pared with offering no intervention to improve movement for all instrumented scoliosis corrections was performed to
behaviour and fitness among young people with CP. inform a national tariff for paediatric spinal surgery in the
UK.44 A total cost of £20 340 was estimated from 23
Costing studies of treatment for CP patients with non-idiopathic scoliosis with neuromuscular,
BoNT-A injection CP, congenital, and syndromic scoliosis. Another retro-
Two budget impact analyses, undertaken in the UK and the spective review of 74 surgical patients with neuromuscular
Russian Federation, showed that Abo-BoNT-A was a less scoliosis (28% with CP) indicated a total (SD) surgical cost
costly treatment than other BoNT-A injections, for example of USD$50 096 (USD$23 998).45 Major contributors to
Ona-BoNT-A or Inco-BoNT-A.38,39 In the UK study, the cost of scoliosis surgery were implants, inpatient and
treatment with BoNT-A for patients with upper-limb spas- intensive care unit costs, and bone grafts.
ticity was less costly than ‘best supportive care’ per patient
per year. Note that the meaning of ‘best supportive care’ Gait analysis
varies from country to country. In this study an increased One study concluded that computerized gait analysis was a
uptake of Abo-BoNT-A resulted in a 5-year saving of £6 potentially useful technology in the management of children
283 829 from the UK payer’s perspective. with walking disabilities, but its efficacy had not been estab-
A retrospective clinical notes review in Germany on chil- lished.46 Later, a retrospective study in 462 ambulatory
dren treated with BoNT-A showed an 85% reduction in the patients with CP was conducted to compare the number of
percentage of children requiring surgery and 60% shorter procedures and total costs between groups of patients under-
average length of stay than the control group who would be going gait analysis versus no gait analysis.47 Adjusting for age,
eligible but did not receive the treatment.40 The total cost of CP type, ambulatory status, Gross Motor Function Classifica-
managing a patient receiving BoNT-A during their first year tion System level, and follow-up time, patients in the gait
of treatment was found to be €16 700. The comparable cost analysis group had more procedures (gait analysis: 5.8; no gait
of managing a control group patient was €33 800. The analysis: 4.2; p<0.001) and higher costs (gait analysis: CAN
researchers concluded that the use of BoNT-A released $43 006; no gait analysis: CAN$35 215; p<0.001) during
resources for alternative use during the first year after treat- index surgery but less subsequent surgery after the index sur-
ment, without any loss of clinical improvement. gery. Patients in the NGA group were twice as likely to have
undergone additional surgery than patients in the gait analysis
ITB therapy group (adjusted hazard ratio 2.1; p=0.002).
A retrospective database analysis using actuarial methods
was carried out to investigate the cost associated with ITB DISCUSSION
therapy for adjunct spasticity control versus continued con- When interpreting the findings of this review, it is impor-
ventional medical management in the absence of ITB tant to note that the term ‘economic evaluation’ has a very

554 Developmental Medicine & Child Neurology 2018, 60: 543–558

specific meaning and flows from the overarching economic outcome measures, such as cost per QALY in CUA, is
principle of ‘opportunity cost’ – that in choosing one preferable to using different outcomes in cost-effectiveness
action we give up the benefits that might flow from measured by different studies. A number of studies in this
alternative actions. This underlies the concept of efficiency review reported incomparable ICERs, so that a value
– that of maximizing net benefit (i.e. benefit gained vs judgement could not be formed. For example, cost per
benefit forgone) with resource use, which is the mechanism proportion of responders by the specific measurements in
that enables measurement of benefits gained or forgone. In the multimodal web-based program RCT are of limited
applying this principle, economic evaluation must involve a use as they cannot be compared to other interventions
comparison of alternatives (often ‘current practice’ vs an without this specific outcome measurement.27 Caution is
‘option for change’) and must involve an analysis of both always required when comparing one study to another as
costs and benefits. A comparison of only outcomes may assumptions are sometimes made, particularly in modelling
establish efficacy/effectiveness, a comparison of only costs studies in which assumptions made for the model parame-
may establish the cheapest option; however, both are ters could impact on the conclusions.
required for cost-effectiveness. Similarly, a comparison of The studies reported provide evidence that prevention of
cost and outcomes for a single intervention is a cost-out- CP using magnesium sulfate in preterm births is cost-
come description; it is not an economic evaluation. Eco- effective.14–16 Substantial lifetime costs attributable to CP
nomic evaluation can be conducted either alongside have been reported in the USA and Europe, with the latest
clinical studies, such as RCTs, or by a modelling approach estimate up to USD$921 000 per person.49–51 In Australia,
that constructs mathematical relationships to simulate rele- the annual cost per person with CP was estimated at AUD
vant consequences, such as disease progression, using the $43 431 and at AUD$115 000 if the disability and prema-
best available information. The modelling study is helpful ture death was included.52 In Australia, the number of pre-
in the absence of data certainty or when real-life studies and perinatally acquired CP cases over the 1993 to 2006
are too difficult to conduct. period was estimated to be 4817.6 If effective prevention of
The generic concept of efficiency is further refined to tar- pre- and perinatally acquired CP can be achieved, the
get one or both of two key questions: (1) should an interven- reduction in CP cases in Australia could be translated into
tion be undertaken or ceased (i.e. value-for-money or direct and indirect financial costs of AUD$67million per
‘allocative efficiency’); and (2) if an intervention is to be year averted by applying the Australian annual cost per
undertaken, how should it be designed (i.e. ‘technical effi- person with CP.52 Put another way, the reduction in CP
ciency’). There are different types of economic evaluations to cases equates to a lifetime cost of USD$4.4billion averted,
answer these questions, mainly differentiated on the basis of if the US estimate of lifetime costs is applied.50 Current
how they measure benefit. Evaluations that focus on the mea- economic modelling studies on the use of magnesium sul-
surement of clinical and/or physical outcomes are called fate suggest the intervention is a ‘dominant strategy’ for
CEAs and are often closely linked to the primary and sec- imminent preterm birth.16 With this ‘back of the envelope’
ondary outcomes of trials. Evaluations that focus on measure- calculation, it can readily be seen that effective prevention
ment of quality of life are called CUAs and require strategies like magnesium sulfate could bring substantial
preference-based utility instruments. Evaluations that focus relief to the health care system, to the welfare system, to
on measuring benefit in dollar terms are called cost–benefit individuals and their families, and to government.
analyses. The different types of economic evaluation have dif- Administration of magnesium sulfate in females at risk
ferent credentials in terms of whether they can address alloca- of preterm birth is the most promising intervention to pre-
tive and/or technical efficiency. The most preferred vent CP in terms of cost-effectiveness to date among the
intervention is, in economic terms, a dominant intervention, 15 included economic evaluation publications. The inter-
which both saves cost and improves health outcomes. Follow- vention of magnesium sulfate for fetal neuroprotection is
ing the notion of ‘dominant’, a cost-effective intervention is either dominant or very cost-effective, depending on the
one that improves health at additional cost with a lower ICER risk level of preterm birth and the perspective taken by the
value. However, there is no explicit ICER threshold estab- analysis. Strong evidence of the benefit of magnesium sul-
lished; instead, a general rule of thumb, for example less than fate in preventing CP among infants born very preterm has
$50 000 per QALY is commonly used in countries like Aus- led to the development of guidelines for its use which have
tralia, the UK, and Canada.48 been endorsed by the Australasian College of Obstetricians
It is challenging to compare cost and cost-effectiveness and Gynaecologists and the National Health and Medical
of different treatment and interventions in a direct way, as Research Council.53 However, the uptake of these guideli-
the decision contexts and the health care systems are nes has been limited and magnesium sulfate is only used as
diverse. More importantly, the treatment and interventions a strategy to prevent CP in a small number of females.
for CP all have different indications so that clinicians and Nevertheless, 57% of CP cases occur in infants born at
policy-makers should not make decisions merely based on term where magnesium sulfate has no role in preventing
the economic considerations. Clinical judgement and eco- CP.5
nomic value judgement should be critically examined side In contrast, the fetal surveillance during labour interven-
by side. Judging the economic credential using common tion was not cost-effective, with ICERs that varied between

Review 555
€78 719 and €393 594 per CP case prevented. In the procedures. There is a debate, however, about its overall
majority of CP cases, the underlying injury occurred before clinical benefits. Two RCTs investigating BoNT-A injec-
labour onset. There is evidence of peripartum asphyxia in tion frequency concluded that yearly injection versus every
only about 10% of term births resulting in CP. Accord- 4 months achieved the same treatment outcomes for
ingly, no matter how effective intrapartum fetal surveil- lower-limb spasticity in children with CP.62,63 Substantial
lance is in detecting fetal hypoxia/asphyxia, it will only savings, in medical costs and cost to the families in time
afford opportunities for mitigation in a small number of and travel, would be made if the BoNT-A regimen was
cases. The ICER results are limited in assessing the techni- reduced from every 4 months to yearly injections. Based
cal efficiency to determine what is the best way to under- on the current economic evaluation assessment, use of
take fetal surveillance, with what modality, and in which BoNT-A injection may be cost-effective, but further
target population. New research areas in the prevention of research is needed on its overall cost-effectiveness for dif-
CP in high-risk populations include use of antioxidant ferent applications in CP management.
therapies (e.g. melatonin) in the perinatal period to protect Many budget impact analyses and costing studies
the fetus, particularly the developing brain, against oxida- assessed the cost of BoNT-A versus control groups. The
tive stress in pregnancy and at birth.54 Further evidence of selection of comparator ranged from ‘best supportive care’
efficacy and effectiveness is required. to ‘usual care’. The meaning of ‘best supportive care’ or
In addition to prevention of CP, the economic analysis ‘usual care’ will vary from country to country and also
of CP included studies of BoNT-A and ITB. Modelling within countries. Therefore, it is hard to draw conclusions
simulation studies suggest that ITB could be cost-effective as to whether cost savings in using BoNT-A to manage
or cost-saving, although its efficacy is still to be verified by CP in one country are applicable to another country.
stronger evidence.22–24 However, inconsistent results of the The current economic evaluation literature regarding
cost-effectiveness for BoNT-A injection have been ITB therapy suggests that the intervention is cost-effective
reported from four economic evaluation studies, mainly in the short term and could be a dominant strategy in the
owing to the additional effectiveness of the treatment com- long term. However, the efficacy and effectiveness of the
pared with the chosen comparator, for example serial cast- therapy has not been well established and stronger evi-
ing, intensive physiotherapy, or standard treatment.19–21 dence is required. According to a ‘systematic review of sys-
Furthermore, some results are indicative and do not consti- tematic reviews on best available intervention evidence for
tute strong evidence of value-for-money without evaluating children with CP’, ITB therapy was graded as a ‘yellow’
long-term effect or reporting ICER by QALYs. BoNT-A intervention, that is, where predominantly low-quality sup-
injection is better for purely dynamic equinus but often it porting evidence is available and the size of the gains var-
is used for mixed equinus when there is some contracture. ied between studies.61 It was graded as a ‘probably do it’
Serial casting might by slightly superior to BoNT-A injec- intervention, but quality and well-designed clinical trials
tion when there is a little more contracture.33,55 This raises are necessary to verify efficacy. ITB may benefit children
the importance of critical considerations in the clinical with severe spasticity of cerebral origin who have not
context. In contrast, for economic considerations, without responded to less invasive treatments such as oral
a unified comparator it is difficult to draw conclusions medications.
about the economic value of the treatment. Nevertheless, it Cost-effectiveness ratios were presented by some studies
is clear from studies comparing different types of BoNT-A that were not economic evaluations, that is, by cost out-
that Abo-BoNT-A is the most economical choice com- come descriptions. This can create confusion for readers
pared with other types with equivalent treatment effect. and such studies need to come with a warning that the
There is evidence in the literature that BoNT-A injec- information provided is descriptive and does not report
tion is effective for CP management (e.g. spasticity, motor efficiency. Economic evaluations should report incremental
function), but the results are inconsistent. BoNT-A injec- costs in relation to incremental outcomes. However,
tion for the upper limbs has been used with good efficacy ICERs were not often reported. There are various guideli-
if combined with therapy to reduce spasticity and improve nes for critical appraisal/reporting of economic evaluation
hand function;56,57 other studies have shown that BoNT-A studies; but no universally accepted criterion standard is
injection is ineffective for hip displacement.58,59 One ani- used.12,64,65
mal study suggests that repeat BoNT-A injections may
have potential harm in dramatically reducing muscle CONCLUSION
torque and producing fibrosis.60 BoNT-A injection is con- It is clear from the present systematic review that the eco-
sidered effective for the following goals: reduction of nomics of CP is under-researched and more economic
upper- and lower-limb spasticity; improved walking abili- studies in this topic, as well as long-term clinical studies,
ties; improved hand function and performance of func- are needed to provide robust evidence to inform value
tional hand activities in combination with occupational judgements. At this time, successful prevention in CP
therapy; and reduction in drooling.61 BoNT-A has been would clearly avoid significant costs. The administration of
used in young children to reduce spasticity in multiple magnesium sulfate for imminent preterm birth is a domi-
muscles before children are old enough to undergo surgical nant strategy resulting in less cost and more benefit

556 Developmental Medicine & Child Neurology 2018, 60: 543–558

compared with no treatment. However, guidelines for the economic outcomes of interventions aiming to prevent CP
use of magnesium sulfate in preterm labour are not well or to improve outcomes for individuals with CP is critical
adhered to and it will not benefit the large proportion of for policy-making and service delivery.
infants born at term.
Economic evaluation of ITB therapy suggests that the A CK N O W L E D G E M E N T S
intervention is cost-effective, but stronger evidence for We are grateful for the comments and feedback on our work
long-term effects is needed. Implantation of the ITB drug from the members of the project consultation group: Professor
delivery system is expensive and makes no difference in the Dinah Reddihough, Murdoch Children’s Research Institute, Mel-
short term; however, ITB may be a cost-saving interven- bourne; Professor H Kerr Graham, Royal Children’s Hospital,
tion in the long term. BoNT-A injection is more costly Melbourne; Professor Christine Imms, Australian Catholic
than other conventional therapies (e.g. serial casting and University; Professor Nadia Badawi, Cerebral Palsy Alliance; Dr
physiotherapy), but the evidence for additional treatment Elise Davis, University of Melbourne; Professor Eve Blair,
benefits is inconclusive and the long-term effects are University of Western Australia; Professor Iona Novak, Cerebral
uncertain. Palsy Alliance; Dr Sarah McIntyre, Cerebral Palsy Alliance; Dr
A web-based home therapy program is considered very Helen Bourke-Taylor, Monash University; Professor Euan Wal-
cost-effective with improvements in motor skills and occu- lace, Monash University; Professor Michael Fahey, Monash
pational performance. A lifestyle intervention could be University. We thank the Victorian Managed Insurance Authority
cost-saving or cost-effective compared with offering no and Centre of Research Excellence in Cerebral Palsy for funding
intervention to improve movement behaviour and fitness. support for this systematic review. Any opinions, findings and
Gait analysis is a potentially useful technology. conclusions, or recommendations expressed in this paper are those
There are large gaps in the evidence that is available of the authors and do not necessarily reflect the views of the Vic-
about the effectiveness of interventions, as well as a lack of torian Managed Insurance Authority. The authors have stated that
high-quality economic appraisal in clinical studies. Contin- they had no interests that might be perceived as posing a conflict
uing to build evidence about both the clinical and or bias.

REFERENCES
1. Rosenbaum P, Paneth N, Leviton A, et al. A report: the 2011. Available from: http://handbook.cochrane.org (ac- in the treatment of spasticity in children with cerebral
definition and classification of cerebral palsy April 2006. cessed 15 October 2015). palsy with botulinum toxin type A: an observational,
Dev Med Child Neurol Suppl 2007; 109: 8–14. 11. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA longitudinal, retrospective study. Farm Hosp 2016; 40:
2. Graham HK, Rosenbaum P, Paneth N, et al. Cerebral Group. Preferred reporting items for systematic reviews 412–26.
palsy. Nat Rev Dis Primers 2016; 2: 15083. and meta-analyses: the PRISMA statement. PLoS Med 19. Catsman-Berrevoets CE, Bussmann JBJ, Pangalila RF,
3. Oskoui M, Coutinho F, Dykeman J, Jett
e N, Pring- 2009; 6: e10000097. et al. Treatment with botulinum toxin of children with
sheim T. An update on the prevalance of cerebral palsy: 12. Drummond M, Sculpher M, Claxton K, Stoddart G, cerebral palsy has no added therapeutical value or cost-
a systematic review and meta-analysis. Dev Med Child Torrance G. Methods for the Economic Evaluation of effectiveness for gross motor function, everyday physical
Neurol 2013; 55: 509–19. Health Care Programmes. Oxford, UK: Oxford Univer- activity levels or quality of life when combined with
4. Gladstone M. A review of the incidence and prevalance, sity Press UL, 2015. intensive functional physiotherapy. Eur J Paediatr Neurol
types and aetiology of childhood cerebral palsy in resource- 13. Tonmukayakul U, Shih S, Carter R. Literature Review 2015; 19: S11.
poor settings. Ann Trop Paediatr 2010; 30: 181–96. & Synthesis for VMIA – Cerebral Palsy: Interventions 20. Yagudina R, Kulikov A, Ugrekhelidze D. Cost-effective-
5. Australian Cerebral Palsy Register Group. Australian & Health Economics. Geelong, Vic.: Deakin University, ness analysis of botulinum toxin type A treatment for
Cerebral Palsy Register Report 2016. Sydney, NSW: 2016. cerebral palsy. Value Health 2015; 18: A734.
Australia, 2016. 14. Bickford CD, Magee LA, Mitton C, et al. Magnesium 21. Houltram J, Noble I, Boyd RN, Corry I, Flett P, Gra-
6. Australian Cerebral Palsy Register Group. Australia and sulphate for fetal neuroprotection: a cost-effectiveness ham HK. Botulinum toxin type A in the management of
the Australian Cerebral Palsy Register for the birth analysis. BMC Health Serv Res 2013; 13: 527. equinus in children with cerebral palsy: an evidence-
cohort 1993 to 2006. Dev Med Child Neurol 2016; 58 15. Cahill A, Odibo A, Stout M, Grobman W, Macones G, based economic evaluation. Eur J Neurol 2001; 8(Suppl.
(Suppl. 2): 3–4. Caughey A. Is magnesium sulfate therapy for the pre- 5): 194–202.
7. Reid SM, Meehan E, McIntyre S, Goldsmith S, Badawi vention of cerebral palsy in preterm infants cost-effec- 22. Bensmail D, Ward AB, Wissel J, et al. Cost-effectivenss
N, Reddihough DS. Temporal trends in cerebral palsy tive? Am J Obstet Gynecol 2009; 201: S83. modeling of intrathecal baclofen therapy versus other
by impairment severity and birth gestation. Dev Med 16. Cahill AG, Odibo AO, Stout MJ, Grobman WA, interventions for disabling spasticity. Neurorehabil Neural
Child Neurol 2016; 58(Suppl. 2): 25–35. Macones GA, Caughey AB. Magnesium sulfate therapy Repair 2009; 23: 546–52.
8. Bourke-Taylor H, Howie L, Law M, Pallant JF. Self- for the prevention of cerebral palsy in preterm infants: a 23. Hoving MA, Evers SM, Ament AJ, van RE, Vles JS,
reported mental health of mothers with a school-aged decision-analytic and economic analysis. Am J Obstet Dutch Study Group on Child Spasticity. Intrathecal
child with a disability in Victoria: a mixed method Gynecol 2011; 205: 542.e1–7. baclofen therapy in children with intractable spastic
study. J Paediatr Child Health 2012; 48: 153–9. 17. Vijgen SM, Westerhuis ME, Opmeer BC, et al. Cost- cerebral palsy: a cost-effectiveness analysis. Dev Med
9. Eker L, T€
uz€
un EH. An evaluation of quality of life of effectiveness of cardiotocography plus ST analysis of the Child Neurol 2008; 50: 450–5.
mothers of children with cerebral palsy. Disabil Rehabil fetal electrocardiogram compared with cardiotocography 24. de Lissovoy G, Matza LS, Green H, Werner M, Edgar
2004; 26: 1354–9. only. Acta Obstet Gynecol Scand 2011; 90: 772–8. T. Cost-effectiveness of intrathecal baclofen therapy for
10. Cochrane Handbook for Systematic Reviews of Inter- 18. Tapias G, Garc
ıa-Romero M, Crespo C, Cuesta M, the treatment of severe spasticity associated with cere-
ventions Version 5.1.0 [updated March 2011]. [Internet] Forn
e C, Pascual-Pascual SI. Cost-minimization analysis bral palsy. J Child Neurol 2007; 22: 49–59.

Review 557
25. Taylor P, Humphreys L, Swain I. The long-term cost- 38. Abogunrin S, Hortobagyi L, Remak E, Dinet J, Gabriel 53. Badawi N. Cerebral palsy – at last it’s time to talk about
effectiveness of the use of Functional Electrical Stimula- S, Bakheit AM. Budget impace analysis of botulinum prevention and maybe even cure. Dev Med Child Neurol
tion for the correction of dropped foot due to upper toxin A therapy for upper limb spasticity in the United 2012; 54(Suppl. 5): 5.
motor neuron lesion. J Rehabil Med 2013; 45: 154–60. Kingdom. Clinicoecon Outcomes Res 2015; 7: 185–93. 54. Miller SL, Wallace EM, Walker DW. Antioxidant ther-
26. Truong WH, Novacheck T, Shin E, et al. Should rou- 39. Yagudina R, Kulikov A, Ugrekhelidze D. Budget impace apies: a potential role in perinatal medicine. Neuroen-
tine proximal femoral hardware removal be performed analsysis of botulinum toxin type A treatment for cere- docrinology 2012; 96: 13–23.
in children with cerebral palsy? Dev Med Child Neurol bral palsy in the Russian Federation. Value Health 2015; 55. Hayek S, Gershon A, Wientroub S, Yizhar Z. The effect
2013; 55(Suppl. 3): 13. 18: A752. of injections of botulinum toxin type A combined with
27. Comans T, Mihala G, Sakzewski L, Boyd RN, Scuffham 40. Ruiz FJ, Guest JF, Lehmann A, et al. Costs and conse- casting on the equinus gait of children with cerebral
P. The cost-effectiveness of a web-based multimodal quences of botulinum toxin type A use. Management of palsy. J Bone Joint Surg 2010; 92: 1152–9.
therapy for unilateral cerebral palsy: the Mitii random- children with cerebral palsy in Germany. Eur J Health 56. Olesch CA, Greaves S, Imms C, Reid SM, Graham HK.
ized controlled trial. Dev Med Child Neurol 2017; 59: Econ 2004; 5: 227–35. Repeat botulinum toxin-A injections in the upper limb
756–61. 41. Saulino M, Guillemette S, Leier J, Hinnenthal J. Medi- of children with hemiplegia: a randomized controlled
28. Slaman J, van den Berg-Emons R, Tan SS, et al. Cost- cal cost impact of intrathecal baclofen therapy for severe trial. Dev Med Child Neurol 2010; 52: 79–86.
utility of a lifestyle intervention in adolescents and plasticity. Neuromodulation 2015; 18: 141–9. 57. Hoare BJ, Wallen MA, Imms C, Villanueva E, Rawicki
young adults with spastic cerebral palsy. J Rehabil Med 42. Morton RE, Gray N, Vloeberghs M. A controlled study HB, Carey L. Botulinum toxin A as an adjunct to treat-
2015; 47: 338–45. of intrathecal baclofen in non-ambulant children with ment in the management of the upper limb in children
29. Marret S, Marpeau L, Zupan-Simunek V, et al. Magne- cerebral palsy. Dev Med Child Neurol 2010; 52(Suppl. 4): with spastic cerebral palsy (UPDATE). Cochrane Data-
sium sulphate given before very-preterm birth to protect 56–7. base Syst Rev 2010; CD003469.
infant brain: the ranomised controlled PREMAG trial. 43. Alewijnse JV, van Rooijen EM, Kreulen M, Smeulders 58. Graham HK, Boyd R, Carlin JB, et al. Does botulinum
BJOG 2007; 114: 310–18. MJC, Tan SS. A microcosting study of the surgical cor- toxin a combined with bracing prevent hip displacement
30. Mittendorf R, Dambrosia J, Pryde PG, et al. Association rection of upper extremity deformity in children with in children with cerebral palsy and “hips at risk”? A ran-
between the use of antenatal magnesium sulfate in pre- spastic cerebral palsy. Dev Neurorehabil 2017; 20: 173–8. domized, controlled trial. J Bone Joint Surg Am 2008;
term labor and adverse health outcomes in infants. Am J 44. Cole AA, Blythe ES, Breakwell LM, Michael ALR. 90: 23–33.
Obstet Gynecol 2002; 186: 1111–18. Patient level costing for instrumented paediatric scoliosis 59. Willoughby K, Ang SG, Thomason P, Graham HK.
31. Crowther CA, Hiller JE, Doyle LW, Haslam RR, Aus- correction. Eur Spine J 2014; 23: S133–4. The impact of botulinum toxin A and abduction bracing
tralasian Collaborative Trial of Magnesium Sulphate 45. Diefenbach C, Ialenti MN, Lonner BS, Kamerlink JR, on long-term hip development in children with cerebral
(ACTOMg SO4) Collaborative Group. Effect of magne- Verma K, Errico TJ. Hospital cost analysis of neuro- palsy. Dev Med Child Neurol 2012; 54: 743–7.
sium sulphate given for neuroprotection before birth: a muscular scoliosis surgery. Bull Hosp Jt Dis (2013) 2013; 60. Minamoto VB, Suzuki KP, Bremner SN, Lieber RL,
randomized controlled trial. JAMA 2003; 290: 2669–76. 71: 272–7. Ward SR. Dramatic changes in muscle contractile and
32. Rouse DJ, Hirtz DG, Thom E, et al. A randomized, 46. Hailey D, Tomie J. An assessment of gait analysis in the structural properties after 2 botulinum toxin injections.
controlled trial of magnesium sulfate for the prevention rehabilitation of children with walking difficulties. Dis- Muscle Nerve 2015; 52: 649–57.
of cerebral palsy. N Engl J Med 2008; 359: 895–905. abil Rehabil 2000; 22: 275–80. 61. Novak I, Mcintyre S, Morgan C, et al. A systematic
33. Corry IS, Cosgrove AP, Duffy CM, McNeill S, Taylor 47. Wren TA, Kalisvaart MM, Ghatan CE, et al. Effects of review of interventions for children with cerebral palsy:
TC, Graham HK. Botulinum toxin A compared with preoperative gait analysis on costs and amount of sur- state of the evidence. Dev Med Child Neurol 2013; 55:
stretching casts in the treatment of spastic equinus: a gery. J Pediatr Orthop 2009; 29: 558–63. 885–910.
randomised prospective trial. J Pediatr Orthop 1998; 18: 48. Henry DA, Hill SR, Harris A. Drug prices and value for 62. Hastings-Ison T, Blackburn C, Rawicki B, et al. Injec-
304–11. money: the Australian Pharmaceutical Benefits Scheme. tion frequency of botulinum toxin A for spastic equinus:
34. Flett PJ, Stern LM, Waddy H, Connell TM, Seeger JD, JAMA 2005; 294: 2630–2. a randomized clinical trial. Dev Med Child Neurol 2016;
Gibson SK. Botulinum toxin A versus fixed cast stretch- 49. Centers for Disease Control and Prevention (CDC). 58: 750–7.
ing for dynamic calf tightness in cerebral palsy. J Paedi- Economic costs of birth defects and cerebral palsy – 63. Kanovsk

y P, Bares M, Severa S, Richardson A, Dysport
atr Child Health 1999; 35: 71–7. United States, 1992. MMWR Morb Mortal Wkly Rep Paediatric Limb Spasticity Study Group. Long-term effi-
35. Horsman J, Furlong W, Feeny D, Torrance G. The 1995; 44: 694–9. cacy and tolerability of 4-monthly versus yearly botuli-
Health Utilities Index (HUI): concepts, measurement 50. Centers for Disease Control and Prevention (CDC). Eco- num toxin type A treatment for lower-limb spasticity in
properties and applications. Health Qual Life Outcomes nomic costs associated with mental retardation, cerebral children with cerebral palsy. Dev Med Child Neurol 2009;
2003; 1: 54. palsy, hearing loss, and vision impairment – United States, 51: 436–45.
36. Lamers LM, Stalmeier PFM, McDonnell J, Krabbe 2003. MMWR Morb Mortal Wkly Rep 2004; 53: 57–9. 64. Drummond MF, Jefferson TO. Guidelines for authors
PFM, van Busschbach JJ. Measuring the quality of life 51. Kruse M, Michelsen SI, Flachs EM, Brønnum-Hansen and peer reviewers of economic submissions to the BMJ.
in economic evaluations: the Dutch EQ-5D tariff. Ned H, Madsen M, Uldall P. Lifetime costs of cerebral palsy. The BMJ Economic Evaluation Working Party. BMJ
Tijdschr Geneeskd 2005; 149: 1574–8. Dev Med Child Neurol 2009; 51: 622–8. 1996; 313: 275–83.
37. Brazier J, Roberts J, Deverill M. The estimation of a 52. Access Economics. The Economic Impact of Cerebral 65. Gold M, Siegel J, Russell L, Weinstein M. Cost-Effec-
preference-based measure of health from the SF-36. J Palsy in Australia in 2007. Sydney, NSW: Access Eco- tiveness in Health and Medicine. Oxford, UK: Oxford
Health Econ 2002; 21: 271–92. nomics Pty Limited for Cerebral Palsy Australia, 2008. University Press, 1996.

558 Developmental Medicine & Child Neurology 2018, 60: 543–558

 DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY SYSTEMATIC REVIEW

RESUMEN

ECONOMICA
EVALUACION


Y DE COSTOS DE LAS INTERVENCIONES PARA LA PARALISIS
SISTEMATICA
CEREBRAL: UNA REVISION

OBJETIVO La evaluacio
n econo
mica puede ayudar a orientar la formulacio
n de pol
ıticas para las decisiones de compra y los
algoritmos de tratamiento y gestio
n para las intervenciones de salud. Realizamos una revisio
n sistema
tica de los estudios
econo
micos en para

lisis cerebral (PC) para informar la investigacio
n futura.
M
ETODO Los estudios econo
micos publicados desde 1970 se identificaron a partir de siete bases de datos. Dos revisores
analizaron de forma independiente los resu
menes y extrajeron los datos siguiendo los lineamientos de los
ıtems de informes
preferidos para revisiones sistema
ticas y metaana
lisis (PRISMA). Cualquier discrepancia se resolvio
mediante discusio
n.
RESULTADOS De 980 referencias identificadas, 115 se incluyeron para la evaluacio
n de texto completo. Trece art
ıculos cumplieron
con los criterios esta
ndar para una evaluacio
n econo
mica completa, dos como evaluaciones econo
micas parciales y 18 como
estudios de costos. Seis fueron evaluaciones econo
micas completas junto con estudios cl
ınicos o ensayos controlados aleatorios,
mientras que siete incluyeron simulaciones de modelado. El argumento econo
mico para la administracio
n de sulfato de magnesio
para un parto prematuro inminente es convincente, logrando tanto un aumento de la salud como un ahorro en los costos. La
literatura actual sugiere que el tratamiento con baclofeno intratecal y la inyeccio
n de toxina botul
ınica son rentables, pero se
necesitan pruebas ma
s so
lidas de los efectos a largo plazo. Las intervenciones mediadas por pa
ginas webs (web-based
interventions) e intervenciones focalizadas en el estilo de vida son econo
micas, pero se requiere una medicio
n ma
s amplia de los
resultados.
INTERPRETACION
La prevencio
n de ls PC evitar
ıa una carga econo
mica significativa. Se ha demostrado que algunos tratamientos
e intervenciones son rentables, aunque se necesitan pruebas ma
s so
lidas de la eficacia cl
ınica.

RESUMO
AVALIACß AO ^
~ ECONOMICA ~ PARA PARALISIA CEREBRAL: UMA REVISAO
ß OES
E CUSTO DAS INTERVENC ~ SISTEMATICA

OBJETIVO A apreciacßa~o econo^ mica pode ajudar na elaboracßa~o de pol
ıticas para deciso~ es sobre aquisicßa~o de servicßos, e nos
algoritmos para manejo e tratamento relacionados a intervencßo ~ es em sau
de. No
s realizamos uma revisa
~o sistema
tica dos estudos
econo^ micos em paralisia cerebral (PC) para informar futuras pesquisas.
M
ETODO Estudos econo^ micos publicados desde 1970 foram identificados em sete bases de dados. Dois revisores avaliaram
independentemente os resumos e extra
ıram dados seguindo as diretrizes dos Itens preferidos para reportar em reviso ~ es
sistema
ticas e metana
lises (PRISMA). Quaisquer discrepa ^ncias foram resolvidas por discussa ~o.
RESULTADOS Das 980 refere^ncias identificadas, 115 foram inclu
ıdas para avaliacßa~o do texto completo. Treze artigos atenderam
aos crite
rios para avaliacßa~ o econo^ mica completa, dois como avaliacßa ~ o econo ^ mica parcial, e 18 como estudos de custo. Seis foram
~ es econo
avaliacßo ^ micas completas acompanhando estudos cl
ınicos, enquanto sete envolveram simulacßo ~ es de modelos. O caso
econo^ mico relativo a  administracßa ~o de sulfato de magne
sio para o nascimento prematuro iminente e
atraente, atingindo tanto
ganho em sau
de quanto economia de custos. A literatura atual sugere que as terapias com baclofeno intratecal e injecßa ~o de
toxina botul
ınica sa ~o custo-efetivas, mas evide ^ncias mais fortes quanto aos efeitos de longo prazo sa ~ o necessa ~ es

rias. Intervencßo
sobre o estilo de vida e baseadas na internet sa ~ o baratas, mas medidas de resultados mais amplas sa ~o necessa
rias.
INTERPRETAC ~ A prevencßa~o da PC evitaria custos econo^ micos significativos. Alguns tratamentos e intervencßo~ es te^m se
ß AO
mostrado custo-efetivos, embora evide ^ncias mais fortes de efetividade cl
ınica sejam necessa
rias.