Epilepsy Research (2013) 104, 151—157

journal homepage: www.elsevier.com/locate/epilepsyres

Cost-effectiveness analysis for the use of serum

antiepileptic drug level monitoring in children
diagnosed with structural-metabolic epilepsy
Muhannad R.M. Salih a,∗, Mohd. Baidi Bahari b, Asrul Akmal Shafie c,
Mohamed Azmi Ahmad Hassali c, Omer Qutaiba B. Al-lela d,
Arwa Y. Abd a, Vigneswari M. Ganesan e

a
College of Pharmacy, Al-Rasheed University, Baghdad, Iraq
b
Unit of Clinical Pharmacy and Pharmacy Practice, Faculty of Pharmacy, AIMST University, Kedah, Malaysia
c
Discipline of Social and Administrative Pharmacy, School of Pharmaceutical Sciences, Universiti Sains Malaysia, Penang, Malaysia
d
Department of Pharmacy Practice, Kulliyyah of Pharmacy, International Islamic University Malaysia, Pahang, Malaysia
e
Department of Paediatrics, Hospital Pulau Pinang, Penang, Malaysia

Received 30 May 2012; received in revised form 21 September 2012; accepted 30 September 2012
Available online 25 October 2012

KEYWORDS Summary Treatment with antiepileptic drugs is commonly guided by serum level mon-
Cost-effectiveness itoring. Such monitoring requires expensive laboratory equipment and products. However,
analysis; well-conducted studies on the cost-effectiveness of therapeutic drug monitoring for antiepilep-
Structural-metabolic tic drugs are lacking particularly in patients with structural-metabolic epilepsy. The study
epilepsy; aims to assess the cost-effectiveness of serum level monitoring services in the management
Paediatrics; of children with structural-metabolic epilepsy during the first year of diagnosis.
Malaysia A retrospective cost-effectiveness analysis was conducted from the provider perspective. It
included patients attended a paediatric neurology clinic. The effectiveness measures used in
this analysis were the number of patients that achieved ≥50% reduction in seizure frequency,
and the number of patients with 3-month seizure free. Medical records of the patients were
reviewed for the required information. Medical chart/billing data obtained from the hospital
were collected to estimate the resources used (One Malaysian Ringgit MYR is equivalent to 0.31
USD). The recruited children were followed for one year following their first visit.
The average cost effectiveness ratio for the monitored patients (MYR 2735 per patient that
achieved a ≥50% reduction in seizure frequency) was lower than that for non-monitored patients
(MYR 2921 per patients that achieved a ≥50% reduction in seizure frequency), with incremental
cost-effectiveness ratio of MYR 2357 per one additional patient that achieved a ≥50% reduction
in seizure frequency. The average cost effectiveness ratios for monitored and non-monitored

∗ Corresponding author at: College of Pharmacy, Al-Rasheed University, Baghdad, Iraq. Tel.: +964 7713226872.
E-mail address: muhanad rmk@yahoo.com (M.R.M. Salih).

0920-1211/$ — see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.eplepsyres.2012.09.012
152 M.R.M. Salih et al.

group were MYR 8279 and MYR 6433 per patient with a 3-month seizure-free period, respectively,
with incremental cost-effectiveness ratio of MYR 29,666 per one additional patient with a 3-month
seizure-free period.
In terms of the effectiveness measures used, serum level monitoring of antiepileptic drugs was
found to be cost-effective. However, the incremental cost-effectiveness ratio was found to be
sensitive to the cost of management.
© 2012 Elsevier B.V. All rights reserved.

Introduction Methods

The present budget constraint measures in health have Setting

called for economic appraisal of health care services
and pharmaceuticals for patient care. However, decision
based on price alone could be misleading (Shakespeare
and Simeon, 1998). A high initial cost is usually associ-
ated with developments in the management of epilepsy.
Alternatively, the developments may perhaps, provide long-
term savings if a considerable number of patients receiving
the new modalities become seizure-free. Hence, it has
become essential to assess the economic burden of health
care options in addition to their therapeutic effective-
ness. It is no longer sufficient to merely exhibit a good
level of efficacy for a specific treatment or service when
a less expensive remedy with satisfactorily positive out-
comes is more preferable (Beran, 1999). Therefore, for
situations in which a number of competing drugs or
amenities are present, economic evaluation that com-
pares the expenditure and outcomes of treatment with
services or drugs are required (Shakespeare and Simeon,
1998).
During the last three decades, treatment with older-
generation antiepileptic drugs (AEDs) (phenytoin, primi-
done, phenobarbital, valproic acid, and carbamazepine)
has been guided by serum level monitoring. Consequently,
a huge expansion in laboratory equipment and thera-
peutic drug monitoring (TDM) service has been observed
in many health care providers. However, well-conducted,
patient-oriented studies on the cost-effectiveness of TDM
for AEDs are lacking. Moreover, there are contradictory
thoughts on the importance of using serum-level moni-
toring to optimise AED therapy (Touw et al., 2007). In
view of that, some medical researchers were questioning
the growth of TDM and whether its importance has been
overstated (Sjoqvist, 1985; Spector et al., 1988; Mcinnes,
1989).
The cost for epilepsy management shows a high level
of discrepancy in relation to the frequency, severity, and
type of seizure (Begley et al., 1999; Pachlatko et al.,
2008). In this fashion, more consistent figures can be
obtained when patients are selected based on the under-
lying type of cause (i.e., aetiology). The term symptomatic,
which was replaced by structural-metabolic, is often sub-
stituted for the concept of a poor prognosis (Engel,
2001; Berg et al., 2010). Moreover, the application of
TDM in paediatric populations may be very different than
that in adult populations (Patsalos et al., 2008). Accord-
ingly, the study aims to assess the cost-effectiveness
of TDM services in the management of children with
structural-metabolic epilepsy during the first year of diag-
nosis.
The study was conducted at the Paediatric Neurology Clinic,
Hospital Pulau Pinang. This hospital is the main public hos-
pital in Penang Island. It is a tertiary referral centre for the
Northern Territory, Peninsular Malaysia. At present, Hospital
Pulau Pinang has 3293 employees, with 115 specialists and
412 medical officers. It has 1090 beds, covering a catch-
ment area of approximately 750,000 people (The Ministry of
Health Malaysia, February, 2011).
Study design
A retrospective cost-effectiveness analysis (CEA) was imple-
mented. It is an annual prevalence-based study on patients
that attended the Paediatric Neurology Clinic, Hospital
Pulau Pinang. The total costs of epilepsy management
were estimated from the provider (i.e., hospital) perspec-
tive, using a bottom-up, microcosting analysis. The list of
cost items included outpatient clinic visits, hospitalisation,
electroencephalography (EEG), neuroimaging, TDM, labo-
ratory tests, AEDs, and non-AEDs. Medical records of the
patients were reviewed for the required information. Med-
ical chart/billing data (i.e., case reports) obtained from
the hospital (i.e., provider) were collected to estimate the
resources used. In most cases, prices or cost data were stan-
dardised for the year 2010. (One Malaysian Ringgit MYR is
equivalent to 0.31 USD). All patients who met the study
criteria were included in the study, and no sampling tech-
nique was considered. The earliest to the latest chart entry
used as data were 1992—2010. A unified set of inclusion and
exclusion criteria were espoused to arrive at one cohort of
patients. Data on each patient were collected for the one
year period following the first visit.
Sample size calculation
The Power and Sample Size Program (PS) software version
3.0.12, January 2009, portrayed a good option to calculate
the sample size for the current investigation (Dupont and
Plummer, 1990). This is a free programme developed by the
Department of Preventive Medicine, Vanderbilt University
School of Medicine, Nashville, Tennessee, USA. The percent-
ages of achieving complete seizure control (i.e., seizure
free) for both groups (TDM and non-TDM) were taken from
a previous study (Rane et al., 2001). The non-TDM group
showed an 8% complete seizure control compared with the
TDM group of 44%. The level of statistical significance (˛)
and power of the study (1 − ˇ) used were 0.05 and 0.9,
Cost-effectiveness analysis for the use of serum antiepileptic drug level monitoring 153

respectively. These clinical data (complete seizure control)

Table 1 Demographic characteristics.
were entered into the software to perform the sample size
calculation for this study. By running the programme, the Characteristics Patient n (%)e Pf
calculated sample size was 29 patients for each group.
Group A Group B

Inclusion criteria Age (year)b 0.972

<4 7.0 (21.90) 17.0 (19.30)
All patients who satisfied the study criteria were included. A 4—<8 13.0 (40.60) 34.0 (38.60)
total of 132 paediatric patients were enrolled in the study. 8—<12 8.0 (25.0) 24.0 (27.30)
Inclusion criteria were as follows: (a) age ≥2 years; (b) a ≥12 4.0 (12.5) 13.0 (14.80)
newly diagnosis of structural-metabolic epilepsy based on Gender 0.417
clinical, electrophysiological, and imaging investigations; Male 17.0 (53.10) 54.0 (61.40)
(c) therapeutic management with AEDs; and (d) three or Female 15.0 (46.90) 34.0 (38.60)
more visits during the first year from the referral time. Race 0.892
Malay 10.0 (31.30) 29.0 (33.0)
Chinese 13.0 (40.60) 38.0 (43.10)
Study approval
Indian 9.0 (28.10) 21.0 (23.90)
Development 0.762
This study was approved by the National Institutes of Health
GDD/IDc 17.0 (53.10) 44.0 (50.0)
(NIH) and by the Medical Research Ethics Committee (MREC),
Normal 15.0 (46.90) 44.0 (50.0)
Ministry of Health Malaysia (Registration ID: NMRR-09-931-
Seizure type 0.036a
4714).
Focal 20.0 (62.5) 36.0 (40.90)
Generalised 12.0 (37.5) 52.0 (59.10)
Definition of groups Type of AEDsd 0.436
Group I 19.0 (59.40) 59.0 (67.0)
After completing the data collection process, patients were Group II 13.0 (40.60) 29.0 (33.0)
divided into two groups: TDM group, all patients who a P < 0.05 is considered statistically significant.
received TDM services at least once during the investigative b Age categorisation depended on the baseline assessment.
year; and non-TDM group, all patients who did not receive c GDD/ID: global developmental delay/intellectual disability.
TDM services. d Group I: all patients who only received old generation of

AEDs. Group II: all patients who received new generation of AEDs
Cost-effectiveness analysis as add on therapy.
e Group A: all patients who received TDM services at least once

To achieve the study goals and to answer the research
questions, the cost-effectiveness ratio (CER) and the incre-
mental cost-effectiveness ratio (ICER) were calculated. This
calculation was performed to assess the use of TDM in
the management of paediatric patients diagnosed with
structural-metabolic epilepsy. This CEA was done according
to the following equations:
C minimum and maximum annual costs were used as the plau-
CER = (1)
E
C1 − C2
ICER = (2)
E1 − E2
CER is the cost-effectiveness ratio; C is the costs; E is the
effectiveness; ICER is the incremental cost-effectiveness
ratio.
Eq. (1) gives information concerning the cost per patient
attaining the target, whereas Eq. (2). is the ratio of the
change in cost of an intervention as compared with the con-
trol to the change in outcomes between the two groups
(Wagner and Goldstein, 2004).
Two effectiveness measures were used in this economic
analysis, including the number of patients that achieved
≥50% reduction in seizure frequency, and the number of
patients with 3-month seizure free. These measures were
assessed based on the differences from the time of the base-
line evaluation (i.e., first visit) to the end of the follow-up
period (i.e., 1-year following the first visit). Because there
was uncertainty in a number of variables used to measure
during the follow-up period. Group B: all patients who did not
receive TDM services.
f Pearson Chi-square.
costs, one-way sensitivity analysis was conducted to check
the robustness of CEA results. The annual cost of epilepsy
management was chosen as an uncertain parameter. The
sible costs for the sensitivity analysis.
Results
Description
One hundred thirty two patients were recruited in this obser-
vational study and 12 of them were excluded. Missing or
ambiguous data and doubt associated with the aetiology
of epilepsy were the main reasons for the exclusion. The
mean age of the studied patients was 7.23 ± 3.55 years.
The patients’ age ranged from 2 to 15 years. Weight of the
patients showed a mean of 23.48 ± 12.71 kg with minimum
and maximum values of 8.3, 74 kg respectively (Table 1).
The ethnic origin of the patients was follows: Chinese, 41%;
Malays, 31%; and Indians, 28%. Number of male (54%) was
to some extent higher than females (46%). There was no
association between the utilisation of TDM services and
demographic characteristics (age, gender, and race).
154 M.R.M. Salih et al.

Table 2 Annual cost of epilepsy management by the use of TDM.a

Characteristic Group Ab Group Bc

Mean ± SD Percentiles (25—50—75) Mean ± SD Percentiles (25—50—75)

Outpatient clinic visitsd 74.80 ± 21.65 56.65—70.82—84.98 63.41 ± 19.79 56.65—56.65—70.82

Hospitalisation 464.75 ± 1251.39 0.00—0.00—286.00 107.25 ± 445.00 0.00—0.00—0.00
EEG 28.52 ± 20.97 0.00—43.29—43.54 22.31 ± 22.91 0.0—21.61—43.49
Neuroimaging 36.09 ± 114.01 0.00—0.00—0.00 49.12 ± 128.52 0.00—0.00—0.00
Laboratory tests 75.38 ± 164.79 0.00—0.00—105.51 63.66 ± 127.00 0.00—0.00—50.57
AEDsd 1318.77 ± 822.08 0.00—0.00—105.51 1129.42 ± 1146.42 0.00—0.00—50.57
Non-AEDs 64.64 ± 152.40 0.00—0.00—13.24 92.75 ± 309.01 0.00—0.00—6.56
TDM 19.50 ± 38.92 0.00—0.00—27.7 — —
Total costd 2136.12 ± 1727.97 1192.24—1531.23—2529.85 1527.95 ± 1321.14 593.45—1106.85—1968.80
a Cost is expressed in Malaysian Ringgit (MYR).
b Group A: all patients who received TDM services at least once during the investigative year.
c Group B: all patients who did not receive TDM services.
d Mann—Whitney test, P < 0.05 is considered statistically significant.

Fifty percent of the patients had global developmental
delay/intellectual disability (GDD/ID). Likewise, gener-
alised epilepsy was the diagnosed seizure type in about
half of the patients. Child development did not appear
to be an important factor to affect the utilisation pat-
tern of TDM. However, patients with focal seizure found
to be more associative with the use of TDM than gener-
alised seizure patients (P = 0.036, Chi-square test). Majority
of the patients were suffering from recurrent seizure attacks
and only a few proportion of them were seizure-free at
the baseline assessment. During the follow-up period, two-
thirds of the patients received only old generation AEDs
(valproic acid, carbamazepine, clonazepam, phenobarbital,
and phenytoin) and only one-third of the patients received
the newer agents (lamotrigine, topiramate, vigabatrin, and
gabapentin) as adjuvant therapy.
Thirty-two patients received TDM services throughout the
follow-up period (i.e., one year). Fifty assays were done
for those TDM patients; eight were done during admissions
and 42 tests were performed during clinic visits. Valproic
acid was the most requested AED for serum monitoring in
the investigated population. It constituted 42% of the total
assays done during the follow-up period. Subsequently, car-
bamazepine made up 38% of the assays. While phenytoin and
phenobarbital located at the bottom of the ranking table;
they formed 12%, 8% of the tests respectively.
frequency compared to the baseline was higher in TDM
(78.10%) than in non-TDM group (52.30%). The CER for TDM
patients (MYR 2735 per patient that achieved a ≥50% reduc-
tion in seizure frequency) was lower than that for non-TDM
patients (MYR 2921 per patients that achieved a ≥50% reduc-
tion in seizure frequency). Consequently, the ICER for TDM
versus non-TDM group was MYR 2357 per one additional
patient that achieved a ≥50% reduction in seizure frequency
(Table 3).
There was a 25.8% increase in the proportion of patients
with a 3-month seizure-free period from baseline to the end
of the follow-up period for TDM patients and 23.75% for non-
TDM patients. The CER for TDM and non-TDM group were
MYR 8279 and MYR 6433 per patient with a 3-month seizure-
free period, respectively. Ultimately, the ICER was MYR
29,666 per one additional patient with a 3-month seizure-
free period.
Sensitivity analysis
The ICER gaining one additional patient that achieved a
≥50% reduction in seizure frequency ranged from MYR 1366
for the minimum annual cost and increased to MYR 13,249
when the maximum annual cost was used (Table 4). The ICER
gaining one additional patient with a 3-month seizure-free
period ranged from MYR 17,194 to MYR 166,746.

Cost-effectiveness analysis Discussion

The total annual cost of management for 120 patients with
structural-metabolic epilepsy was MYR 202,816 (i.e., MYR
1690.13 ± 1458.37 per patient per year). Most of the cost
items included in this study showed no significant differ-
ence between the monitored and non-monitored patients
(Table 2). However, the costs of outpatient clinic visits, and
AEDs were significantly higher in the TDM than the non-TDM
patients (P = 0.004 and P = 0.025). Patients who received TDM
services had higher management costs than those who had
not received TDM services (P = 0.01). However, the propor-
tion of patients that achieved a ≥50% reduction in seizure
The individualisation of drug therapy via the measurement
and interpretation of AED levels is not a recent approach.
Therapeutic drug monitoring services have been used since
the 1960s (Touw et al., 2007). However, there are still con-
tradictory views concerning the impact of TDM in optimising
AED therapy. In the context of limited and constrained bud-
getary resources, the cost of each intervention should be
justified. Although the utilisation of TDM services is justi-
fied for targeted or specialty populations, such as patients
with refractory epilepsy (Schumacher and Barr, 2001), a full
economic evaluation should be used as an assessment tool
Cost-effectiveness analysis for the use of serum antiepileptic drug level monitoring 155

to determine the true value of intervention. Hence, looking

ICERc (MYR)
merely at the clinical outcomes may not justify the expen-
diture.

2357
29,666
On searching the literature, a paucity of cost analysis
studies concerning the impact of TDM for AEDs was found.
In the conclusion of a paper published on behalf of the Cost-
Effectiveness of Therapeutic Drug Monitoring Committee of

CERb (MYR)
the International Association for Therapeutic Drug Monitor-
ing and Clinical Toxicology, TDM of old AEDs was noted to
2921
6433
be cost-effective (Touw et al., 2005). This conclusion was
based on only one retrospective pharmacoeconomic anal-
ysis in India (Rane et al., 2001). In that evaluation study,
Group A: all patients who received TDM services at least once during the investigative year. Group B: all patients who did not receive TDM services.
Effectiveness (%)f

cost was estimated from two perspectives (i.e., hospital

and patient). From the hospital perspective, the CER for
patients who received TDM was 13.03 rupees per one seizure
saved per year. The CER from the patient perceptive was
2.67 rupees per one seizure saved per year. Moreover, the
52.30
23.75

cost of TDM to the provider per seizure prevented was 22.35

rupees, while the cost to the patient was 4.5 rupees.
Two years later, the same authors of the above-cited
1-Year (b) (%)

review article published an update concerning the cost-

effectiveness of TDM (Touw et al., 2007). The same
conclusion was determined, which was that ‘‘TDM of anti-
33.75

epileptic drugs can be cost-effective’’. However, in this

—

review, an additional retrospective observational study was

used to support the stated conclusion (Bond and Raehl,
RSF: number of patients that achieved a ≥50% reduction in seizure frequency compared to the baseline.
Baseline (a) (%)

2006). In that study, hospitals without pharmacists who man-

aged AED therapy showed a higher rate of death, higher
length of hospital stay, higher total Medicare charges, higher
laboratory charges, and a higher rate of aspiration pneumo-
nia than hospitals with pharmacist-managed AED therapy. As
10
—
B

pharmacist management is highly dependent on the use of

TDM for AEDs, it can be speculated that TDM is helpful in
CER (MYR)

optimising therapeutic management for epileptic patients.

In comparison, these findings cannot be easily compared
to the present results. These cited studies used differ-
b

2735
8279

ent populations, settings, and practices as compared with

the current study. In this study, both of the calculated
ICERs were lower than the threshold value recommended
Cost-effectiveness analysis for the use of TDM services.

f
Effectiveness (%)

by the World Health Organisation through the programme

of CHOosing Interventions that are Cost Effective (WHO-
TSF: number of patients with a 3-month seizure-free period.

CHOICE). Along these lines, an ICER less than three times

of the Gross Domestic Product per capita (i.e., interna-
78.10

tional dollar 20,845) is considered cost-effective for a

25.8

public policy intervention in the World Health Organisa-

tion Western Pacific Region B (World Health Organization,
ICER: incremental cost-effectiveness ratio.

2005). This was estimated to be MYR 36,145 based on the

1-Year (b) (%)

2005 implied Purchasing-Power-Parity conversion rate for

Malaysian currency of 1.734 (International Monetary Fund,
32.25

2009). However, results of a sensitivity analysis showed that

the ICER for the number of patients with 3-month seizure
—

CER: cost-effectiveness ratio.

free was sensitive to the cost of management.

In terms of the effectiveness measures used in the cur-
Baseline (a) (%)

rent study, serum level monitoring of AEDs was found to

Effectiveness = b − a.

be cost-effective in the management of children diagnosed

with structural-metabolic epilepsy. However, the ICER for
Groupa

one of the effectiveness measure (number of patients with

6.45

3-month seizure free) was found to be sensitive to the cost of

—
A

management. These findings may not directly apply to other

Table 3

countries. A lot of aspects should be considered, including

RSFd
TSFe

the incidence and severity of the disease, the availability of

a
b
c
d
e
f

health care resources, clinical practice patterns and relative

156
Table 4 One-way cost sensitivity analysis of TDM by using the minimum and maximum annual cost.a
Groupb
A B
CERc (MYR)
RSFe
Minimum annual cost 536
Maximum annual cost 11,413
TSFf
Minimum annual cost 1623
Maximum annual cost 34,548
a Minimum cost: MYR 418.80 for Group A and MYR 66.31 for Group B. Maximum cost: MYR 8913.59 for Group A and MYR 5495.29 for
Group B.
b Group A: all patients who received TDM services at least once during the investigative year. Group B: all patients who did not receive
TDM services.
c CER: cost-effectiveness ratio.
d ICER: incremental cost-effectiveness ratio.
e RSF: number of patients that achieved a ≥50% reduction in seizure frequency compared to the baseline.
f TSF: number of patients with a 3-month seizure-free period.
costs (Sculpher et al., 2004). For this reason, it is rational
for national guidelines to request that analyses should be
related to the local setting. Although the study objective
was addressed, it might be of significant value if future stud-
ies assess the cost-effectiveness of serum AED monitoring for
longer follow-up period (i.e., 2—3 years after diagnosis).
Acknowledgments
The authors would like to thank the Head of Paediatrics
Institute, Hospital Kuala Lumpur, Dr. Hussain Imam bin
Hj Muhammad Ismail for his help and supervision. The
authors would like to acknowledge the Institute of Postgrad-
uate Studies, Universiti Sains Malaysia for their support in
carrying out this work through providing the USM Fellow-
ship and the USM-RU-Postgraduate Research Grant Scheme
(1001/PFARMASI/842026).
References
Begley, C.E., Annegers, J.F., Lairson, D.R., Reynolds, T.F., 1999.
Methodological issues in estimating the cost of epilepsy. Epilepsy
Res. 33, 39—55.
Beran, R.G., 1999. Economic analysis of epilepsy care. Epilepsia 40,
19—24.
Berg, A.T., Berkovic, S.F., Brodie, M.J., Buchhalter, J., Cross, J.H.,
Van Emde Boas, W., Engel, J., French, J., Glauser, T.A., Mathern,
G.W., 2010. Revised terminology and concepts for organization
of seizures and epilepsies: report of the ILAE Commission on Clas-
sification and Terminology, 2005—2009. Epilepsia 51, 676—685.
Bond, C.A., Raehl, C.L., 2006. Clinical and economic outcomes of
pharmacist-managed antiepileptic drug therapy. Pharmacothe-
rapy 26, 1369—1378.
Dupont, W.D., Plummer Jr., W.D., 1990. Power and sample size cal-
culations: a review and computer program. Control. Clin. Trials
11, 116—128.
Engel Jr., 2001. A proposed diagnostic scheme for people with
epileptic seizures and with epilepsy: report of the ILAE Task
Force on Classification and Terminology. Epilepsia 42, 796—803.
M.R.M. Salih et al.
ICERd (MYR)
CERc (MYR)
126 1366
10,507 13,249
279 17,194
23,138 166,746
International Monetary Fund, 2009. World Economic and Financial
Surveys: World Economic Outlook Database. International Mon-
etary Fund, Washington, DC, Available: http://www.imf.org/
external/pubs/ft/weo/2009/01/weodata/index.aspx (accessed
10.07.11) (online).
Mcinnes, G.T., 1989. The value of therapeutic drug monitoring to
the practising physician — an hypothesis in need of testing. Br.
J. Clin. Pharmacol. 27, 281—284.
Pachlatko, C., Chisholm, D., Meinardi, H., Sander, J.W., 2008.
The economic impact of epilepsy. In: Engel, J., Pedley, T.A.,
Aicardi, J. (Eds.), Epilepsy: A Comprehensive Textbook. Lippin-
cott Williams & Wilkins, Philadelphia.
Patsalos, P.N., Berry, D.J., Bourgeois, B.F., Cloyd, J.C., Glauser,
T.A., Johannessen, S.I., Leppik, I.E., Tomson, T., Perucca, E.,
2008. Antiepileptic drugs — best practice guidelines for thera-
peutic drug monitoring: a position paper by the subcommission
on therapeutic drug monitoring, ILAE Commission on Therapeu-
tic Strategies. Epilepsia 49, 1239—1276.
Rane, C.T., Dalvi, S.S., Gogtay, N.J., Shah, P.U., Kshirsagar, N.A.,
2001. A pharmacoeconomic analysis of the impact of therapeutic
drug monitoring in adult patients with generalized tonic—clonic
epilepsy. Br. J. Clin. Pharmacol. 52, 193—195.
Schumacher, G.E., Barr, J.T., 2001. Therapeutic drug monitoring: do
the improved outcomes justify the costs? Clin. Pharmacokinet.
40, 405—409.
Sculpher, M.J., Pang, F.S., Manca, A., Drummond, M.F., Golder, S.,
Urdahl, H., Davies, L.M., Eastwood, A., 2004. Generalisability in
economic evaluation studies in health care: a review and case
studies. Health Technol. Assess. 8, 1—192.
Shakespeare, A., Simeon, G., 1998. Economic analysis of
epilepsy treatment: a cost minimization analysis compar-
ing carbamazepine and lamotrigine in the UK. Seizure 7,
119—125.
Sjoqvist, F., 1985. Interindividual differences in drug response: an
overview. In: Rowland, M., Sheiner, L.B., Steimer, J. (Eds.), Vari-
ability in Drug Therapy. Raven Press, New York.
Spector, R., Park, G.D., Johnson, G.F., Vesell, E.S., 1988. Therapeu-
tic drug monitoring. Clin. Pharmacol. Ther. 43, 345—353.
The Ministry of Health Malaysia. February, 2011. Available:
http://hpp.moh.gov.my/v2/ (accessed).
Touw, D.J., Neef, C., Thomson, A.H., Vinks, A.A., 2005. Cost-
effectiveness of therapeutic drug monitoring: a systematic
review. Ther. Drug Monit. 27, 10—17.
Cost-effectiveness analysis for the use of serum antiepileptic drug level monitoring 157

Touw, D.J., Neef, C., Thomson, A.H., Vinks, A.A., 2007. Cost-
effectiveness of therapeutic drug monitoring: an update. EJHP
Sci. 13, 83—91.
Wagner, T.H., Goldstein, M.K., 2004. Behavioral interventions and
cost-effectiveness analysis. Prev. Med. 39, 1208—1214.
World Health Organization, 2005. CHOosing Interventions
that are Cost Effective (WHO-CHOICE): Cost-Effectiveness
Thresholds. World Health Organization, Geneva, Available:
http://www.who.int/choice/costs/CER levels/en/index.html
(accessed 10.07.11) (online).