‘Order from chaos’, an unsound foundation for decisions in health care?: Non-adherence to
disease-specific guidelines for pharmacoeconomic studies in rheumatoid arthritis
Amr Makady1*, Gerardus W.J. Frederix2, Anke M. Hövels2
1. Graduate School of Life Sciences, Utrecht University, Utrecht, the Netherlands;
2. Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht University, Utrecht, the
Netherlands.
*Corresponding Author:
E-mail: makady.amr@gmail.com
(Word Count: 3314 words)
Abstract
Objectives: Biologic drugs used in treatment of rheumatoid arthritis (RA) increase economic burden
of RA on health care systems. Thus, pharmacoeconomic (PE) studies of treatment options in RA are
valuable for decision-makers in the health care sector. Previous literature has demonstrated the
presence of wide variations in PE study outcomes and methodology. This study aimed to firstly
examine whether major variations still exist in PE study outcomes and methodology for RA
treatment, and secondly, assess whether PE studies in RA treatment options adhered to RA diseasespecific guidelines presented by OMERACT-ILAC.
Methods: A literature review was conducted for PE studies evaluating TNF-α inhibitors use in RA.
Four different databases (e.g. Embase, NIHR-EED) were searched for PE studies that focussed on
Adalimumab between October 2003 and May 2013. Methodological quality of included studies was
checked using the CHEC-checklist. Data extraction forms were used to retrieve information such as
study outcomes (QALY’s, costs, ICER’s) and modelling methods and parameters (e.g. time horizon,
sources for costs and effectiveness data). Finally, information retrieved from all studies was
compared to recommendations proposed by the OMERACT-ILAR guideline of 2002.
Results: Ten studies were identified that met all inclusion criteria and were consequently included in
our analysis. All studies met at least 12 of the 19 items of the CHEC checklist for quality and 4 studies
met all 19 items. Study outcomes varied considerably in QALY’s calculated, costs and ICER’s. Patient
subtypes investigated, modelling methods, sources for cost and effectiveness data also varied
significantly. Only 1 of the 12 recommendations presented in the OMERACT-ILAR guideline was
unanimously implemented in all 10 studies of our review. Only 1 study was found to contain all
elements of guideline recommendations, albeit with some limitations
Conclusions: Considerable variations still exist in outcomes and methodologies for PE studies of
Adalimumab in the treatment of RA. Adherence to disease-specific guidelines for the conduct of PE
studies in RA is very low. Development of strict disease-specific guidelines in RA and subsequent
adoption by re-imbursement agencies is vital to ensure comparability and validity of PE study
outcomes during decision-making in the health care setting.
(Word Count: 348 words)
Introduction
Rheumatoid arthritis (RA) is a chronic, progressive, systemic autoimmune disorder affecting synovial
joints. Worldwide estimates state that RA affects between 0.5% and 1% of the global population
(Emery, 2006; Sommer et al., 2005), often within patients’ years of peak productivity (between 30-50
years of age)(Adam Rindfleisch & Muller, 2005). As a result, RA places a heavy burden on society,
both through the high direct and indirect medical costs to the healthcare system, as well as through
economic loss due to loss in productivity subsequent to increased disability of patients (Emery, 2006;
Fautrel et al., 2007). According to Jacobbson et al., costs of living with RA in Sweden were €12,020
per patient per year, of which 41% were attributable solely to direct costs of therapeutic agents
(Jacobsson et al., 2007). Meanwhile, 55% of such costs were attributable to indirect costs, mainly due
to early retirement and long-term sick leave.
The introduction of Infliximab, the first TNF-α inhibitor, in 1998 for the treatment of RA provided a
positive therapeutic addition to traditional disease-modifying anti-rheumatic drugs (tDMARD’s).
Numerous clinical trials have demonstrated the greater efficacy of treatment of RA with TNF-α
inhibitors in comparison to tDMARD’s (Maini et al., 1999; Moreland & Schiff, 1999; Weinblatt et al.,
2003). However, the therapeutic advantage of such agents is offset by their high costs (Jacobsson et
al., 2007; Michaud, Messer, Choi, & Wolfe, 2003). Therefore, proof of cost-effectiveness of TNF-α
inhibitors is essential in order to justify reimbursement of such expensive products in light of
increased prescribing patterns of biologicals for RA treatment and limited health care
budgets(Grijalva, Chung, Stein, Mitchel, & Griffin, 2008; McBRIDE et al., 2011).
Previous literature has demonstrated that notable variations exist between outcomes of
pharmacoeconomic (PE) studies, as well as between study methodologies (e.g. chosen modelling
methods and sources for data on costs and clinical effectiveness) (Maetzel, Ferraz, & Bombardier,
1998). Consequently, comparability and validity of PE study outcomes become problematic issues to
reimbursement agencies, affecting their decision-making process (Rawlins, Barnett, & Stevens, 2010).
In an attempt to standardise PE study methodology, thereby increasing the comparability, validity
and credibility of PE study outcomes, guidelines for study conduct have been established by national
institutions. For example, the Dutch College for Health Insurance (CVZ), National Institute for Clinical
Excellence (NICE) in the United Kingdom, and Academy of Managed Care Pharamcy (AMCP) in the
United States each published guidelines detailing study characteristics such as modelling
methodology, parameterisation and data sources (AMCP, 2009; CVZ, 2006; NICE, 2013). Despite the
presence of these guidelines, significant differences in modelling methods and study outcomes still
exist within PE studies in rheumatology (Barbieri et al., 2007).
The presence of such differences in PE study characteristics in rheumatology prompted the creation
of disease-specific guidance for PE studies in rheumatology by the OMERACT Health Economics
Group; an international, independent initiative of health experts working under the aegis of the
International League for Rheumatology (ILAR) to develop and validate clinical and radiographic
outcome measures in rheumatoid arthritis, osteoarthritis, psoriatic arthritis, fibromyalgia, and other
rheumatic diseases (OMERACT, 2013). The OMERACT-ILAR guideline (2002) outlines a reference case
for RA, where a reference case represents a proposed set of minimum criteria that all PE studies
should include, allowing comparability across studies (Gabriel, Tugwell, & Drummond, 2002). To our
knowledge, despite the high quality of the OMERACT-ILAC guideline, formal adoption of its
recommendations has not yet occurred by national reimbursement agencies.
Therefore, the purposes of this study are two-fold. Firstly, this study aims to establish whether
significant differences in fact still exist in PE study outcomes, methodologies and parameters, for
recent PE studies evaluating Adalimumab, as one example of biological TNF-α inhibitors. Secondly,
this study aims to evaluate whether recent PE studies for Adalimumab adhere to disease-specific
recommendations for PE studies in RA established by the OMERACT-ILAC guideline of 2002.
Methods
Search Strategy
A systematic literature review was conducted to identify PE studies assessing Adalimumab between
October 2003 and May 2013 (figure 1). The use of October 2003 as a starting date corresponds to the
date of first licensing of Adalimumab by the European Medicines Agency (EMA) (EMA, 2013). In
addition, adherence to OMERACT-ILAR guideline recommendations was assumed, since all PE studies
included in our study were published after 2003, i.e. post-publication of the OMERCAT-ILAC
guideline. Both clinical and economic evaluation databases were employed in this review, namely:
Embase, MedLine (PubMed), The Cochrane Library, and the National Institute for Health - Research
Centre for Reviews and Dissemination (NIHR-CRD). Search terms encompassed different forms of PE
studies (e.g. economic evaluation, cost-effectiveness, cost-utility), different licensed TNF-α
antagonists (specifically: infliximab, etanercept and adalimumab), and rheumatoid arthritis (RA). A
full list of the search terms used are provided in the table below (table 1).
Abstracts of the studies generated by the search were screened against inclusion and exclusion
criteria (table 2). Any studies that did not meet all inclusion criteria, or met one of the three exclusion
criteria were excluded from analysis.
Methodological Quality
The CHEC checklist was used to assess the methodological quality of studied included (Evers,
Goossens, De Vet, Van Tulder, & Ament, 2005). This list contains 19 items, selected through a Delphi
process by 23 experts in the field of health economics to assess the quality of PE evaluations. All
publications were assessed by two reviewers using the following characteristics: presented in text,
presented but with limitations and not presented in text.
Data Extraction
Based upon author consensus, two Excel data extraction forms were produced in order to
systematically obtain relevant data form studies selected for analysis.
The first form contained information related to study outcomes, i.e. calculated quality-adjusted lifeyears (QALY’s), incremental costs, incremental cost-effectiveness ratios (ICER’s), and patient subtype
data. All costs from different studies were converted into the 2012 value using inflation data based
on consumer price index (CPI) (Global Rates, 2013). Due to the use of different currencies in different
studies, costs were then converted to US Dollars using the purchase price parity (PPP) index (OECD,
2013). Finally, all values were converted to Euros using exchange rates provided by the European
Central Bank (ECB, 2013). ICER’s were calculated by dividing the adjusted costs in Euros by the
reported QALY’s.
The data extracted from individual studies for the second form included: author, publication year,
comparator used, TNF-α inhibitors studied, type of PE analysis conducted, country within which
study was conducted, model structure used, perspective, discount rate, time-horizon, sources for
clinical effectiveness data, costs and utility measurement data, adverse event modelling, and study
sponsorship by drug manufacturer.
Comparison to OMERACT-ILAC Guidline
Characteristics of PE studies included in our literature survey were compared to the reference case
described by Gabriel et al. in the OMERACT-ILAC guideline for PE studies in rheumatology (table 3), in
order to assess whether studies in literature were indeed comparable across various variables. One
variable from the reference case, “classification and reporting of adverse events”, was excluded from
analysis, owing to the fact that it is currently under development by OMERACT.
Results
Study selection
The Embase and Medline databases search yielded 423 and 184 hits respectively. Meanwhile, the
Cochrane Library and NIHR-CRD databases resulted in 68 and 67 hits respectively. In total, 720 hits
were generated using our search strategy (figure 1).
Titles and abstracts of all search hits from each separate database were individually compared to
inclusion and exclusion criteria. 306 hits did not meet all inclusion criteria, thus were excluded from
analysis. 9 hits studied the use of TNF-α antagonists for conditions other than rheumatoid arthritis
thus were also excluded. Meanwhile, 380 hits comprised documents other than published economic
evaluations and were excluded from our literature review. The remaining 47 hits, which had met all
inclusion criteria, were then checked for duplicate entries. Any hits consisting of the abstract only,
without access to the original article, were excluded from our final analysis.
Eventually, 10 articles were identified that met all inclusion criteria and to which access to the full
manuscript was available. All 10 articles were included in our review.
Methodological Quality
Of the 19 items of the CHEC check-list for methodological quality, 12 were present in all 10 included
studies (table 4). These items included: a clearly described research question and competing
treatment alternatives, appropriate PE study design and time horizon selection, clear identification
and appropriate measurement of relevant costs and outcomes (including cost discounting), validity
of conclusions according to data provided, and discussion regarding the generalisability of study
results.
Of all included studies, 4 contained all relevant information for all 19 CHEC items: Nguyen et al.[5],
Spalding et al.[7], Wu et al[9], and Malottki et al[10].The remaining 6 studies also represented relevant
data for at least 17 of the 19 CHEC criteria.
Moreover, the absence of relevant data was relatively low in all 10 included studies. The most
notable absence of data was in item number 18 of the CHEC list, concerning the indication of no
conflict of interest of study researcher(s) and funder(s).
Data Extraction
Outcomes of PE studies (QALY’s, Costs, ICER’s and Patient subtypes)
Six medical interventions involving Adalimumab and a comparator drug were identified within
articles included in our review (table 5). The quality of life-adjusted years (QALY) values calculated by
different studies, within the same intervention group, varied greatly. For example, QALY’s gained
ranged from 0.08[8] to 2.41[9] when Adalimumab monotherapy was compared to Methotrexate (MTX)
and traditional disease-modifying anti-rheumatic drugs (tDMARD’s). Similarly, incremental costs
calculated within each study differed significantly. For example, incremental costs values calculated
for Adalimumab monotherapy versus MTX/tDMARDs were between €7,859[7] and €104,112[9].
Consequently, incremental cost-effectiveness ratios (ICER’s) values calculated were different for
studies within the same medical intervention group. For example, ICER’s calculated for Adalimumab
monotherapy versus MTX/tDMARD’s ranged from 30,090[1] Euros/QALY to 183,743[2] Euros/QALY. An
even broader range could be seen in the group comparing Adalimumab + MTX to MTX/tDMARD’s,
with ICER’s values calculated between 19,788[6] Euros/QALY and 226,670[2] Euros/QALY.
It is worth noting that the category of patient subtypes was inconsistent among different PE studies
(table 5). In some studies, cost-utility of TNF-α antagonists was evaluated for patients with moderate
to severe RA[1,2,5,6,9], while in others their cost-utility was evaluated for patients with early RA[2,3].
Additionally, some studies analysed TNF-α antagonists’ cost-utility as first-line agents[2,7], while
others analysed their use as third-line agents[2]. Finally, one study[10] analysed cost-effectiveness for
patients who have failed to respond to their first TNF-α antagonist.
Modelling methods and parameters
Modelling methods used in the selected PE studies varied considerably (table 6); 5 of the 10 studies
used patient-level, discrete-event simulation models[2,3,6,8], while 3 of the 10 studies used Markov
modelling strategies[5,7,9]. Moreover, different methods were used within the category of Markov
modelling (e.g. standard Markov modelling, probabilistic Markov models and Markov cohort
models).
Data regarding safety and toxicity were either not reported[6,10], were not available[4,5], estimated
from literature[6], or extracted from RCT’s[7] or observational studies[1,2] (table 6). Similarly, sources
for direct medical costs, more specifically for data on prices not related to medication cost, differed
for different studies. Some studies derived such costs from published guidelines[2,6], national
databases[1,3,4,5,6,8,9], published studies[6,5,8] or expert opinion[1,2]. Meanwhile, sources and methods for
utility measures also differed between studies. For example, some studies correlated health
assessment questionnaire (HAQ) data from randomised clinical trials (RCT’s) directly to utility values
using linear regression models, while others converted EuroQoL-5D data to HAQ measures first
before calculating utility.
On the other hand, data for clinical effectiveness of TNF-α antagonists was extracted from RCT’s in 8
of the 10 included studies (table 6). One study extracted effectiveness data from a national, diseasespecific database[4], while the other used Bayesian meta-analysis of data from published RTC’s[5].
Full study funding by the manufacturer of Adalimumab was only observed for 3 of the 10 included
studies. One study was partially funded by the manufacturer.
Comparison to OMERACT-ILAC Guidline
Only 1 of the 12 recommendations made by Gabriel et al. in the OMERACT-ILAR guideline was
unanimously implemented in all 10 studies of our review; namely, that relating to duration of
treatment (table 7). Meanwhile 3 were implemented in all studies but with some limitations:
recommendations regarding valuation of health (i.e. QALY’s), resource use and discontinuation of
treatment.
The article by Bansback et al[1] was the only study found to contain all elements of guideline
recommendations, albeit with some limitations regarding two guideline requirements (table 7). Nine
of the 10 included studies contradicted 1 or more guideline requirements, the maximum number of
contradictions coinciding with the article by Malotkki et al[10].
A detailed version of the table has been included in the appendix, displaying the examples of
limitations on data provided to meet certain guideline requirements (see appendix, table 1).
Discussion
Despite the presence of PE study guidelines, outcome measures presented by different studies in the
forms of QALY’s, direct medical costs and ICER’s varied considerably within our review. For example,
the range of QALY values presented within the group of studies comparing Adalimumab
monotherapy to MTX/ tDMARD’s were between 0.14 and 2.41. Similarly, calculated costs ranged
between €7,859[7] and €104,112[9], while values of ICER’s were between 30,090[1] Euros/QALY and
183,743[2] (table 5). On the other hand, analysis for the presence of CHEC checklist items in all 10
included studies of our review led to the observation that authors adhered to most
recommendations for good methodological quality of PE studies. This indicates that disparities in
reported outcomes of PE studies are not a result of non-adherence to general guidelines for the
quality of PE study methodology.
One of the main reasons behind the widely diverging outcomes was the choice of patient subtypes.
In the study by Chen et al. (2006), it was demonstrated that ICER values varied for different patient
subtypes. For example, early RA patients using TNF-α inhibitors as third-line agents had an ICER of
45,905[2] Euros/QALY while late RA patients using TNF-α inhibitors as third-line agents had an ICER of
183,743[2] Euros/QALY.
However, on comparing calculated QALY’s and costs between studies by Chen et al. and Spalding et
al. for use of TNF-α inhibitors as first-line agents against RA (i.e. within the same patient subtype
group), we found that QALY measurements by Spalding et al. were much lower in comparison to
Chen et al. (0.14 and 0.6508, respectively). Meanwhile, costs calculated by Spalding et al. were
€7,859[7] compared to €45,483[2] calculated by Chen et al. There were similar discrepancies in QALY
and costs values between both studies when Adalimumab + MTX was compared to tDMARD’s;
QALY’s calculated were 0.05[7] compared to 1.0619[2], while costs were €9,733[7] compared to
€42,538[2] respectively (table 5).
Analysis of modelling methods, model parameters, resources for data used in outcome calculation
and sources for study funding was performed to uncover further potential reasons for differences in
reported outcomes. First and foremost, modelling methods were split between discrete-event
modelling and Markov modelling, with study-specific differences within each sub-group of modelling
methods. Secondly, sources for toxicity and safety data associated with drug use were either lacking
or derived from widely varying sources (e.g. observational studies, expert opinion, or estimates from
previous literature). Thirdly, data for direct medical costs not associated with drug prices were
derived from different sources in included studies (e.g. published guidelines, expert opinion, or
previous literature). Finally, the sources and methods used for derivation and calculation of utility
weights differed between included articles.
The presence of such notable discrepancies in patient subtypes, modelling methods and data sources
has an evident effect on study outcomes, which subsequently has a negative impact on the
comparability, validity and accuracy of results of PE studies (OMERACT, 2013; Rawlins et al., 2010).
This is a concern for regulatory authorities and re-imbursement agencies that rely on PE studies for
decision-making (Rawlins et al., 2010).Without consistent modelling methodologies and parameter
definitions, decision-makers would not be able to compare or verify values of PE studies performed
at a national level with others performed in different settings. This, in turn, might negatively impact
re-imbursement decisions, ultimately affecting the lives of many patients in need.
One approach to help increase comparability and validity of PE study outcomes is through the
establishment and adoption of disease-specific guidelines for study methodology. Such a guideline
has already been introduced by OMERACT-ILAC for economic evaluations in rheumatology (Gabriel et
al., 2002). However, upon comparing parameters and methods for PE study conduct of all 10
included articles in our review to recommendations made by Gabriel et al., we found that only 1 of
the 12 recommendations were met unanimously by all studies. Moreover, only one[1] of the included
studies met all recommendations made in the OMERACT-ILAR guideline, albeit with some limitations.
Moreover, despite the high quality of the OMERACT-ILAC guideline, its requirements provide some
leeway in study design without the call for relevant justification by the study authors. For example,
no sensitivity analysis is required if authors use alternative sources for resource use costs, valuation
of health methods or mortality rates. This can be demonstrated by the limitations of data provided
regarding sources for mortality rates presented in table 1 of the appendix.
These results demonstrate that we are currently facing a problem, as disease-specific
recommendations made to increase the validity and comparability of economic evaluations are not
taken into regard. Recently, other disease-specific guidelines for the methodology of PE evaluation of
therapeutic treatments in early endocrine breast cancer were introduced by Frederix et al in 2011
(Frederix, Severens, Hövels, Raaijmakers, & Schellens, 2012). Such recommendations encompassed
a reference case for the modelling methods and parameterisation of PE models used. Similar
recommendations were made by Mittman et al. (2013) for the updating of national Canadian
guidelines for economic evaluations to suit oncology-specific PE evaluations (Mittmann et al., 2012).
Owing to the fact that such guidelines are not currently formally adopted by reimbursement
agencies, we can also assume that adherence to their recommendations will not occur. This begs the
question: is it worthwhile to devise disease-specific guidelines for the above-mentioned purposes,
bearing in mind that formal adoption and implementation of their recommendations does not occur?
Our systematic review has two main limitations. Differences in outcomes between different PE
studies were only compared for Adalimumab. Bearing in mind that there are a number of other TNFα inhibitors available on the market (e.g. Infliximab, Etanercept), as well as recently-introduced
generic products, which have varying costs, clinical efficacy and safety profiles (Aaltonen et al., 2012;
Nam & Emery, 2010), it would be valuable to establish if such variance in outcomes also occurs in PE
studies for such agents.
Although evidence generated by this review suggests a wide variance exists in QALY, costs and ICER
values between different PE studies, the paucity in numbers of included studies prevents the
establishment of a statistical correlation between the reasons behind such a variance. The small
population of studies also prevents the quantification of the effect of variance in different modelling
parameters and methods on PE study outcome.
Conclusion
In conclusion, significant variations in RA patient subtypes, modelling methods, modelling
parameters and data sources were still observed between studies included in this review for PE
studies on Adalimumab. Furthermore, published disease-specific guidelines for PE studies in RA were
not adhered to in PE studies of Adalimumab for the treatment of RA.
Disease-specific guidelines serve a critical role in increasing comparability, validity and credibility of
PE studies. Although recommendations by OMERACT-ILAC are of high quality, they provide significant
leeway in PE study design without requesting relevant justification for such deviation. We propose
that efforts must be invested by relevant authorities to develop stricter disease-specific guidelines
with well-defined requirements that demand justification in case of deviation from reference cases.
More importantly though, formal adoption and implementation of the developed guidelines must
follow, otherwise, non-adherence to recommendations will ensue.
In doing so, decisions regarding re-imbursement of pharmaceutical products would be based upon a
more robust scientific foundation. This is of particular importance for the provision of healthcare to
RA patients, whose treatment often requires the chronic use of expensive biologic drugs.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Aaltonen, K. J., Virkki, L. M., Malmivaara, A., Konttinen, Y. T., Nordström, D. C., & Blom, M. (2012).
Systematic Review and Meta-Analysis of the Efficacy and Safety of Existing TNF Blocking Agents in
Treatment of Rheumatoid Arthritis. PLoS ONE, 7(1), e30275. Retrieved from
http://dx.doi.org/10.1371%2Fjournal.pone.0030275
Adam Rindfleisch, J., & Muller, D. (2005). Diagnosis and Management of Rheumatoid Arthritis.
American Family Physician, 72(6)
AMCP. (2009). The AMCP Format for Formulary Submissions Version 3.0 Academy of Managed Care
Pharmacy.
Bansback, N. J., Brennan, A., & Ghatnekar, O. (2005). Cost effectiveness of adalimumab in the
treatment of patients with moderate to severe rheumatoid arthritis in Sweden. Annals of the
Rheumatic Diseases, 64(7), 995-1002. doi:10.1136/ard.2004.027565
Barbieri, M., Drummond, M. F., Puig Junoy, J., Casado Gomez, M. A., Ballina Garcia, J. F., Blasco Segura,
P., & Poveda Andres, J. L. (2007). Critical appraisal of pharmacoeconomic studies comparing TNF[alpha] antagonists for rheumatoid arthritis treatment. Expert Review of Pharmacoeconomics &
Outcomes Research, 7(6), 613-626. Retrieved from
http://sfx.library.uu.nl/utrecht?sid=OVID:ovftdb&id=pmid:&id=doi:10.1586%2F14737167.7.6.613&iss
n=1473-7167&isbn=&volume=7&issue=6&spage=613&pages=613626&date=2007&title=Expert+Review+of+Pharmacoeconomics+%26+Outcomes+Research&atitle=Criti
cal+appraisal+of+pharmacoeconomic+studies+comparing+TNF%5Balpha%5D+antagonists+for+rheumatoid+arthritis+treatment.&aulast=Barbieri&pid=%3Cauthor%3
EBarbieri%2C+Marco%3BDrummond%2C+Michael%3BPuig+Junoy%2C+Jaume%3BCasado+Gomez%2C
+Miguel%3BBallina+Garcia%2C+Javier%3BBlasco+Segura%2C+Pilar%3BPoveda+Andres%2C+Jose%3C%
2Fauthor%3E%3CAN%3E00136282-20071200000015%3C%2FAN%3E%3CDT%3EReview%3C%2FDT%3E
Chen, Y. F., Jobanputra, P., Barton , P., & Jowett, S. (2006). A systematic review of the effectiveness of
adalimumab, etancercept and infliximab for the treatment of rheumatoid arthritis in adults and an
economic evaluation of their cost-effectiveness. Health Technology Assessment, 10(42)
CVZ. (2006). Guidelines for pharmacoeconomic research, updated version. Diemen, the Netherlands:
College voor Zorgversekeringen.
DAVIES, A., CIFALDI, M. A., SEGURADO, O. G., & WEISMAN, M. H. (2009). Cost-Effectiveness of
Sequential Therapy with Tumor Necrosis Factor Antagonists in Early Rheumatoid Arthritis. The Journal
of Rheumatology, 36(1), 16-26. doi:10.3899/jrheum.080257
ECB. (2013). European Central Bank: Euro foreign exchange rates. Retrieved 06/18, 2013, from
https://www.ecb.int/stats/exchange/eurofxref/html/index.en.html
EMA. (2013). Humira: EPAR summary for the public. (EPAR). United Kingdom: European Medicines
Agnecy.
Emery, P. (2006). Treatment of Rheumatoid Arthritis. BMJ, 332, 152-155.
Evers, S., Goossens, M., De Vet, H., Van Tulder, M., & Ament, A. (2005). Criteria list for assessment of
methodological quality of economic evaluations: Consensus on Health Economic Criteria. International
Journal of Technology Assessment in Health Care, 21(2), 240-5. Retrieved from
http://search.proquest.com/docview/210352722?accountid=14772
Fautrel, B., Clarke, A. E., Guillemin, F., Adam, V., St-Pierre, Y., Panaritis, T., . . . Penrod, J. R. (2007).
Costs of Rheumatoid Arthritis: New Estimates from the Human Capital Method and Comparison to the
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Willingness-to-Pay Method. Medical Decision Making, 27(2), 138-150.
doi:10.1177/0272989X06297389
Frederix, G. W. J., Severens, J. L., Hövels, A. M., Raaijmakers, J. A. M., & Schellens, J. H. M. (2012).
Reviewing the Cost-Effectiveness of Endocrine Early Breast Cancer Therapies: Influence of Differences
in Modeling Methods on Outcomes [Abstract]. Value in Health : The Journal of the International
Society for Pharmacoeconomics and Outcomes Research, 15(1) 94-105. Retrieved from
http://linkinghub.elsevier.com/retrieve/pii/S1098301511015646?showall=true
Gabriel, S. E., Tugwell, P., & Drummond, M. (2002). Progress towards an OMERACT-ILAR guideline for
economic evaluations in rheumatology. Annals of the Rheumatic Diseases, 61(4), 370-373.
doi:10.1136/ard.61.4.370
Global Rates. (2013). CPI Inflation - current international consumer price index inflation. Retrieved
06/14, 2013, from http://www.global-rates.com/economic-indicators/inflation/consumerprices/cpi/cpi.aspx
Grijalva, C. G., Chung, C. P., Stein, C. M., Mitchel, E. F., & Griffin, M. R. (2008). Changing patterns of
medication use in patients with rheumatoid arthritis in a Medicaid population. Rheumatology, 47(7),
1061-1064. doi:10.1093/rheumatology/ken193
Jacobsson, L., Lindroth, Y., Marsal, L., Juran, E., Bergstrom, U., & Kobelt, G. (2007). Rheumatoid
arthritis: what does it cost and what factors are driving those costs? Results of a survey in a
community‐derived population in Malmö, Sweden. Scand J Rheumatol, 36(3), 179-183.
doi:10.1080/03009740601089580
Kievit, W., Adang, E. M., Fransen, J., Kuper, H. H., van de Laar,M A F J., Jansen, T. L., . . . Van Riel,P C L
M. (2008). The effectiveness and medication costs of three anti-tumour necrosis factor α agents in the
treatment of rheumatoid arthritis from prospective clinical practice data doi:10.1136/ard.2007.083675
Maetzel, A., Ferraz, M. B., & Bombardier, C. (1998). A review of cost-effectiveness analyses in
rheumatology and other disciplines. Current Opinion in Rheumatology, 10(2), 136-140.
Maini, R., St Clair, E. W., Breedveld, F., Furst, D., Kalden, J., Weisman, M., . . . Lipsky, P. (1999).
Infliximab (chimeric anti-tumour necrosis factor α monoclonal antibody) versus placebo in rheumatoid
arthritis patients receiving concomitant methotrexate: a randomised phase III trial. The Lancet,
354(9194), 1932-1939. doi:http://dx.doi.org/10.1016/S0140-6736(99)05246-0
Malottki, K., Barton, P., Tsourapas, A., Uthman, A. O., Liu, Z., Routh, K., . . . Chen, Y. F. (2011).
Adalimumab, Etanercept, Infliximab, Rituximab and Abatacept for the Treatment of Rheumatoid
Arthritis After the Failure of a Tumour Necrosis Factor Inhibitor: A Systematic Review and Economic
Evaluation. Health Technology Assessment, 15(14)
McBRIDE, S., SARSOUR, K., WHITE, L. A., NELSON, D. R., CHAWLA, A. J., & JOHNSTON, J. A. (2011).
Biologic Disease-modifying Drug Treatment Patterns and Associated Costs for Patients with
Rheumatoid Arthritis. The Journal of Rheumatology, 38(10), 2141-2149. doi:10.3899/jrheum.101195
Michaud, K., Messer, J., Choi, H. K., & Wolfe, F. (2003). Direct medical costs and their predictors in
patients with rheumatoid arthritis. Arthritis & Rheumatism, 48(10), 2750-2762. doi:10.1002/art.11439
Mittmann, N., Evans, W. K., Rocchi, A., Longo, C. J., Au, H., Husereau, D., . . . Oh, P. I. (2012). Guidelines
for Health Technologies: Specific Guidance for Oncology Products in Canada [Abstract]. Value in Health
: The Journal of the International Society for Pharmacoeconomics and Outcomes Research, 15(3) 580585. Retrieved from http://linkinghub.elsevier.com/retrieve/pii/S1098301511036679?showall=true
Moreland, L. W., & Schiff, M. H. (1999). Etanercept therapy in rheumatoid arthritis. Annals of Internal
Medicine, 130(6), 478. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=afh&AN=1717041&site=ehost-live
Nam, J., & Emery, P. (2010). Aspects of TNF inhibitor therapy in rheumatoid arthritis. Modern
Rheumatology, 20(4), 325-330. doi:10.1007/s10165-010-0277-7
Nguyen, C., Bounthavong, M., Mendes, M. S., Christopher, M. D., Tran, J., Kazerooni, R., & Morreale, A.
(2012). Cost Utility of Tumour Necrosis Factor-α Inhibitors for Rheumatoid Arthritis.
PharmacoEconomics, 30(7), 575-593. doi:10.2165/11594990-000000000-00000
NICE. (2013). Guide to methods of technology appraisal 2013. United Kingdom: National Institute for
Health and Care Excellence.
OECD. (2013). Purchasing Power Parities for GDP and related indicators. Retrieved 06/18, 2013, from
http://stats.oecd.org/Index.aspx?DataSetCode=PPPGDP#
OMERACT. (2013). Outcome Measures in Rheumatology. Retrieved 06/28, 2013, from
http://www.omeract.org/
32. Rawlins, M., Barnett, D., & Stevens, A. (2010). Pharmacoeconomics: NICE's approach to decisionmaking. British Journal of Clinical Pharmacology, 70(3), 346-349. doi:10.1111/j.13652125.2009.03589.x
33. Soini, E. J., Hallinen, T. A., Puolakka, K., Vihervaara, V., & Kauppi, M. J. (2012). Cost-effectiveness of
adalimumab, etanercept, and tocilizumab as first-line treatments for moderate-to-severe rheumatoid
arthritis. Journal of Medical Economics, 15(2), 340-351. doi:10.3111/13696998.2011.649327
34. Sommer, O. J., Kladosek, A., Weiler, V., Czembirek, H., Boeck, M., & Stiskal, M. (2005). Rheumatoid
Arthritis: A Practical Guide to State-of-the-Art Imaging, Image Interpretation, and Clinical
Implications1. Radiographics, 25(2), 381-398. doi:10.1148/rg.252045111
35. Spalding, J., & Hay, J. (2006). Cost Effectiveness of Tumour Necrosis Factor-α Inhibitors as First-Line
Agents in Rheumatoid Arthritis. PharmacoEconomics, 24(12), 1221-1232. doi:10.2165/00019053200624120-00006
36. Wailoo, A. J., Bansback, N., Brennan, A., Michaud, K., Nixon, R. M., & Wolfe, F. (2008). Biologic drugs
for rheumatoid arthritis in the medicare program: A cost-effectiveness analysis. Arthritis &
Rheumatism, 58(4), 939-946. doi:10.1002/art.23374
37. Weinblatt, M. E., Keystone, E. C., Furst, D. E., Moreland, L. W., Weisman, M. H., Birbara, C. A., . . .
Chartash, E. K. (2003). Adalimumab, a fully human anti-tumor necrosis factor alpha monoclonal
antibody, for the treatment of rheumatoid arthritis in patients taking concomitant methotrexate: The
ARMADA trial. Arthritis & Rheumatism, 48(1), 35-45. doi:10.1002/art.10697
38. Wu, B., Wilson, A., Wang, F., Wang, S., Wallace, D. J., Weisman, M. H., & Lu, L. (2012). Cost
Effectiveness of Different Treatment Strategies in the Treatment of Patients with Moderate to Severe
Rheumatoid Arthritis in China. PLoS ONE, 7(10), e47373. Retrieved from
http://dx.doi.org/10.1371%2Fjournal.pone.0047373
Figure Captions
Figure 1: Selection of studies included in review
Figure 1 – Selection of studies included in review
Studies identified via Embase and screened manually (n= 423)
Studies identified via MedLine and screened manually (n=184)
Studies identified via The Cochrane Library and screened manually (n=68)
Studies identified via NIHR-CRD (n=67)
Study does not meet all inclusion criteria
(n=306)
436 studies of interest
Study assesses use of Adalimumab for conditions other
than RA (n=9)
427 studies of interest
Documents other than published economic
evaluations (e.g. reviews, expert opinion,
editorials, posters, etc.) (n=380)
47 studies of interest
Duplicates removed; Hits with access to abstract
only excluded from final study. (n=37)
10 studies included in
literature review
Table Captions
Table 1: Search strategies employed for literature search
Table 2: Inclusion and exclusion criteria for study inclusion
Table 3: OMERACT-ILAC reference case for RA
Table 4: CHEC checklist items presence in the 9 included studies
Table 5: Data extraction form 1: outcomes of PE studies (QALY’s, Costs, ICER’s and patient subtype
data)
Table 6: Data extraction form 2
Table 7: Comparison of included studies to recommendations by OMERACT-ILAR guideline
Table 1 - Search strategies employed for literature search
Strategy
Search String
Number
1.
((cost effectiveness utility benefit minimisation) AND tumour necrosis factor alpha
inhibitor) AND rheumatoid arthritis
2.
((cost effectiveness utility benefit minimisation) AND anti-TNF alpha) AND rheumatoid
arthritis
3.
((cost effectiveness utility benefit minimisation) AND adalimumab OR infliximab OR
etanercept) AND rheumatoid arthritis
4.
((economic evaluation) AND tumour necrosis factor alpha inhibitor) AND rheumatoid
arthritis
5.
((economic evaluation) AND anti-TNF alpha) AND rheumatoid arthritis
6.
((economic evaluation) AND adalimumab OR infliximab OR etancercept) AND rheumatoid
arthritis
Table 2 - Inclusion and exclusion criteria for study inclusion
Inclusion Criteria
1.
The article was published between October 2003 and May 2013.
2.
The article was published in English.
3.
The study population constituted of patients being administered TNF-α inhibitors, whether as first-,
second-, or third-line agents.
4.
The study focussed on the TNF-α inhibitor Adalimumab.
5.
The focus of the study was the determination of cost-effectiveness, cost-utility, cost-minimisation,
cost-benefit or cost-consequence of drug treatment.
6.
It is possible to extract information regarding the methodology, costs calculation and effect
determination from the study.
1.
2.
3.
Exclusion Criteria
Study does not meet all inclusion criteria.
Study assesses the use of TNF-α inhibitors for conditions other than rheumatoid arthritis.
Documents other than published economic evaluations (e.g. expert opinion, editorials, posters,
reviews, presentations, etc.).
Table 3 - OMERACT-ILAC reference case for RA
Methodological Issue
Recommendation
1.
Model horizon
One year
2.
Duration of treatment
Continuous
3.
Extrapolation beyond trial
Estimates of benefit based on trial data; estimates of withdrawal and
long term outcome based on observational data.
4.
Modelling beyond disease
No benefit or harm if treatment is stopped.
duration
5.
Synthesis of comparisons
Not recommended owing to uncertain validity of transitive
where clinical trials do not
comparisons.
exist
6.
Outcome measures
7.
8.
9.
Mortality
Valuation of health
Resource use
10.
11.
Discontinuation of
treatment
Therapeutic strategies
12.
Population risk stratification
 ACR 20 sustained for 6 months
 EULAR improvement criteria
 Clinical adverse events
Hazard for mortality from observational studies.
Values from general population using direct measurement.
Include all associated direct medical costs in the analysis, but report
indirect and non-medical costs separately.
Use discontinuation rates from observational studies.
Include modelling of most commonly used therapeutic strategy with
sensitivity analysis to consider other strategies.
Include clear definition of underlying population including low and
high risk groups.
Table 4 - CHEC checklist items presence in the 9 included studies
CHEC-List Question
1. Is the study population clearly described?
2. Are competing alternatives clearly described?
3. Is a well-defined research question posed in answerable form?
4. Is the economic study design appropriate to the stated
objective?
5. Is the chosen time horizon appropriate in order to include
relevant costs and consequences?
6. Is the actual persepective chosen appropriate?
7. Are all important and relevant costs for each alternative
identified?
8. Are all costs measured appropriately in physical units?
9. Are all costs valued appropriately?
10. Are all important and relevant outcomes for each alternative
identified?
11. Are all outcomes measured appropriately?
12. Are all outcomes valued appropriately?
13. Is an incremental analysis of costs and outcomes of
alternatives performed?
14. Are all future costs and outcomes discounted appropriately?
15. Are all important variables, whose values are uncertain,
appropriately subjected to sensitivity analysis?
16. Do the conclusions follow from the data reported?
17. Does the study discuss the generalizablity of the results to
other settings and patient/ client groups?
18. Does the article indicate that there is no potential conflict of
interest of study researcher(s) and funder(s)?
19. Are ethical and distributional issues discussed appropriately?
Legend: (Y=Yes; N=No; Lim.= Presented, but with limitations)
Ref. 1
Bansback
Y
Y
Ref. 2
Chen
Y
Y
Ref. 3
Davies
Y
Y
Ref. 4
Kievit
Y
Y
Ref. 5
Nguyen
Y
Y
Ref. 6
Soini
Y
Y
Ref. 7
Spalding
Y
Y
Ref. 8
Wailoo
Lim.
Y
Ref. 9
Wu
Y
Y
Ref. 10
Malottki
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Lim.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Lim.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
N
Lim.
N
Y
Y
Lim.
Lim.
Y
Y
Lim.
Y
Lim.
Lim.
Lim.
Lim.
Table 5 – Data extraction form 1: outcomes of PE studies (QALY’s, Costs, ICER’s and patient subtype data)
Interventions
Adalimumab monotherapy vs. DMARDs
(MTX)/ (tDMARDs)
Reference
Country
Incremental costs
(Euros)
QALYs gained
Cost per QALYs gained
(Euros)
Patient subtypes
1
Sweden
0,4733
14242
2
U.K.
0,2196
40350
183743 Late RA; TNF as 3rd Line
2
U.K.
0,9188
42177
45905 Early RA; TNF as 3rd Line
2
U.K.
0,6508
45483
69888 TNF inhibitors as first line
7
U.S.A.
0,14
7859
54446 TNF inhibitors as first line
9
China
2,41
104112
Adalimumab monotherapy vs. Anakinra
8
U.S.A.
0,20
28927
Adalimumab + MTX vs. DMARDs (MTX)
1
Sweden
1,1296
27943
24737 M/S RA
2
U.K.
0,4884
41757
85498 Late RA; TNF as 3rd Line
2
U.K.
1,0619
42538
40058 Early RA; TNF as 3rd Line
2
3
6
U.K.
U.S.A.
Finland
0,201
1,238
3,682
45561
47116
72859
226670 TNF inhibitors as first line
38058 Early RA
19788 M/S RA
7
U.S.A.
0,05
9733
194589 TNF inhibitors as first line
10
U.K.
0,75
28343
9
China
3,29
125171
Adalimumab + Rituximab vs. DMARDs
(tDMARDs)
30090 M/S RA
43200 HAQ score of 1.6 (M/S RA?)
144635 --
Inadequate response to initial
37791 TNF-alpha inhibitor
38046 HAQ score of 1.6 (M/S RA?)
Adalimumab + MTX vs. ETN + MTX
5
U.S.A.
Adalimumab + DMARDs/Prednisone
4
the Netherlands
-0,032
N/A
2293
N/A
-71656 M/S RA
N/A
TNF naive patiens
Table 6 - Data extraction form 2
Study
Authors
Publicatio
n Year
Comparator
TNF
Inhibitors
Studied
Type of
PE
Analysi
s
CUA
Country
Model
structure
Perspectiv
e
Discoun
t Rate
Bansbac
k et al.
2013
tDMARD
ADA, ETA,
INF
Chen et
al.
2006
(+/-) antiTNF
Davies et 2009
al.
Kievit et
al.
Nguyen
et al.
3% per
annum
TimeAdverse
horizon events
modelin
g
Lifetim Yes
e
Study
Sources for
sponsorship by safety data
manfufacturer
?
Yes
Observationa
l study
Sweden
Patient-based,
transition state
model
Policy
maker
ADA, ETA,
INF
CUA
U.K.
Discrete-event
simulation
Competing
sequential
therapies
ADA, ETA,
INF
CUA
U.S.A.
2013
None
ADA, ETA,
INF
CEA
2012
MTX
monotherap
y
ADA, ETA,
INF,
Certolizuma
b
CUA
NHS
6% per
annum
Lifetim
e
No
No
Ob study for Published
ETA & IN;
guidelines and
Assumtption expert opinion
s for ADA; UK
database for
tDMARDs
RCTs and
metaanalysis for
biologicals;
RCTs or
assumptions
for DMARDs
Individual
patient
simulation
(discrete-event
simulation)
Payer
3% per
annum
Lifetim
e
Yes
Yes
Ob study
Sweden;
Gebroek
RCT's for
each drug
Regression
formulas using
HAQ
the
Netherland
s
N/A
Not
mentioned
!
N/A
12
No
months
Partial funding
N/A
DREAM
database for
DAS-28
scores.
EuroQoL Eq-5D
and
physical/menta
l scales of SF36
Propensity
score for
differences
in
prognostic
factors for
ttt effects
U.S.A.
Markov model
Healthcare
payer
3% per
annum
5 years
No
N/A
Bayesian
analysis of
published
RCT's (metaanalysis)
Published
study
correlating ACR
response to
VAS
One-way
sensitvity
analysis for
variances in
model
parameters.
Scenario
analysis
(ACR). PSA
or second
order
Monte Carlo
for
uncertaintie
s in model
parameters..
Yes
Sources for
resource use
(i.e. costs)
Sources for
effectivenes
s data
Sources for
utility weights
Analysis of
uncertainty
National
databases,
arbistrary
calculations
and expert
opinion
2 RCTs for
ADA, Review
of published
articles for
ETA, IN,
tDMARDs
Linear
transformation
from HAQ; HUI
to HAQ
calculation
from published
study
Conversion
from HAQ to
QoL values via
equation based
on data set
Univariate
and
multivariate
SA
Analysource for
drug cost data;
Monitoring and
AE costs from
Medicare fee
shcedule
Drug costs
from Tariff list
by Dutch
Health Care
Insurance
Board; IN inpatient clinic
admin costs;
Drug costs
from AWP from
2009 Red Book;
Physician &
infusion costs
from CMS
Univariate
SA and
scnario
analyses.
Quasi-Cis
around
mean costs
and
benefits.
Probabilistic
SA
Soini et
al.
2012
MTX
monotherap
y as further
comparator
ADA, ETA,
CUA
INF,
tocilozumab,
rituximab
Finland
Probabilistic
microsimulatio
n (discreteevent model)
Public
healthcare
payer
3% per
annum
Lifetim
e
No
Yes
Not
mentioned.
Spalding
et al.
2006
MTX
monotherap
y
ADA, ETA,
INF
CUA
U.S.A.
Probabilistic
Markov model
Payer
3% per
annum
Lifetim
e
Yes
No
Published
literature
estimates;
RCT's one for
each drug
Wailoo
et al.
2008
Infliximab
ADA, ETA,
Anakinra
CUA
U.S.A.
Decision model
(patient-level
simulation)
Medicare
3% per
annum
Lifetim
e
No
No
4 RCT's for
ETA; 4 RCT's
for ADA; two
RCT's for INF
Wu et al. 2012
Competing
therapies
ADA, ETA,
INF,
Rituximab
CUA
China
Markov cohort
model
Chinese
healthcare
system
3% per
annum
Lifetim
e
Yes
No
Malotkki
et al.
tDMARD
ADA, ETA,
INF,
Rituximab,
Abatacept
CEA
U.K.
Individual
patient
simulation
(discrete-event
simulation)
NHS
3.5%
per
annum
Lifetim
e
No
No
OPV from
Chinese
Healthcare
Sytem, INP
from
Swedish
study; no
data on
toxicity
source
N/A
2011
Drug costs
from Finnish
Medicine Tariff
2011; Swedish
study for
hospitalization
costs; Routine
monitoring
protocol from
Statstics
Finland
Drug costs
from AWP from
2005; Routine
monitoring
costs from
Physician's Fee
and Coding
Guide; Lit for
adverse event
costs;
Monitoring
costs
calculated or
derived from
lit;
Hospoitalisatio
n fees from
study
US registry;
costs
associated to
HAQ scores
Phase III TOC
RCT
Estimated by
published non
linear mixed
model. Links
EQ-5D to HAQ
scores.
Multiple
sensitivity
analysis
scenarios
(one-way
sensitivity
analysis?).
RCT's for
data on
biologicals
(one for each
drug)
Percentage
reduction of
HAQ score
from RCT's
(raw measure)
Univariate
sensitivity
analysis.
4 RCT's for
ETA; 4 RCT's
for ADA; two
RCT's for INF
Chinese
healthcare
system
Review of
published
RCT's (2 for
ADA, 1 for
ETA and 1
for INF);
Review of
published
studies
EQ-5D index
and SF-36;
utilities
associated to
HAQ's
Review of
published
RCT's (2 for
ADA, 1 for ETA
and 1 for INF);
Review of
published
studies;
One-way
sensitivity
analysis and
Probabilistic
SA.
One-way
sensitivity
analysis and
Probabilistic
SA.
Not
mentioned.
Review of
Linear
published
transformation
studies:
from HAQ
RCT's,
manufacture
r dossiers
submitted to
NICE,
observationa
l studies
Probabilistic
SA
Table 7 - Comparison of included studies to recommendations by OMERACT-ILAR guideline
Methodological Issue
Rheumatoid Arthritis
Ref. 1
Ref. 2
Bansback
Chen
1 Model Horizon
Lifetime
Y
Y
2 Duration of treatment
Continuous
Y
Y
3 Extrapolation beyond
Estimates of benefit based on trial
trial duration
data; estimates of withdrawal and
Y
Y
long term outcomes based on
observational data
4 Modelling beyond trial
No benefit or harm if treatment is
Lim.
N
duration
stopped
5 Synthesis of
Not recommended
comparisons where
Y
Y
clinical trials do not exist
6
Outcome measures
7
Mortality
8
Valuation of health (i.e.
QALY)
Resource use
9
10 Discontinuation of
treatment
11 Therapeutic strategies
12 Population risk
stratification
i) ACR 20 sustained for 6 months**
ii) EULAR improvement criteria**
iii) Clinical adverse events
Hazard for mortality from
observational studies
Values from general population using
direct measurement***
Include all associated direct medical
costs in the analysis, but report
indirect and non-medical costs
seperately
Use discontinuation rates from
observational studies
Include modelling of most commonly
used therapeutic strategy with
sensitivity analysis to consider other
strategies
Include clear definition of underlying
population including low and high
risk groups
Legend: (Y=Yes; N=No; Lim.= Presented, but with limitations)
Ref. 3
Davies
Y
Y
Ref. 4
Kievit
N
Y
Ref. 5
Nguyen
N
Y
Ref. 6
Soini
Y
Y
Ref. 7
Spalding
Y
Y
Ref. 8
Wailoo
Y
Y
Ref. 9
Wu
Y
Y
Ref. 10
Malottki
Y
Y
Y
N
Lim.
Y
Lim.
Lim.
Lim.
Y
N
Y
Y
Lim.
Y
Lim.
Y
N
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
N
N
Lim.
Lim.
Y
Y
Lim.
Y
Y
Y
Y
N
Y
Y
N
N
N
Lim.
Y
Y
N
Y
Y
N
N
N
N
Lim.
Lim.
Lim.
N
N
Lim.
N
Y
Lim.
Lim.
Y
Y
Y
Y
Y
Lim.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Lim.
Y
Y
Y
Y
Y
Y
Y
Lim.
Lim.
Lim.
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
N
Y
Y
Y
Lim.
Lim.
Y
Y
Y
Y
N
Y
Y
Appendix 1 – Detailed table concerning OMERACT-ILAC recommendations
Table 1 - Comparison of included studies to recommendations by OMERACT-ILAR guideline (Detailed)
Methodological Issue Rheumatoid Arthritis
Ref. 1
Ref. 2
Ref. 3
Ref. 4
Bansback
Chen
Davies
Kievit
1 Model Horizon
Lifetime
Y
Y
Y
N
2 Duration of
Continuous
Y
Y
Y
Y
treatment
3 Extrapolation beyond Estimates of benefit based on
trial duration
trial data; estimates of
withdrawal and long term
outcomes based on
Y
Y
Y
N
observational data
4 Modelling beyond
trial duration
No benefit or harm if
treatment is stopped
5 Synthesis of
comparisons where
clinical trials do not
exist
6 Outcome measures
Not recommended
7 Mortality
i) ACR 20 sustained for 6
months**
ii) EULAR improvement
criteria**
iii) Clinical adverse events
Hazard for mortality from
observational studies
No. Reason
substantiated.
Ref. 5
Nguyen
N
Ref. 6
Soini
Y
Ref. 7
Spalding
Y
Ref. 8
Wailoo
Y
Ref. 9
Wu
Y
Ref. 10
Malottki
Y
Y
Y
Y
Y
Y
Y
Lim.
N
N
Y
Y
Y
Y
Y
Y
Y
Y
N
Lim.
Y
N
Y
Y
Adjusted data
from standard
life tables for
Swedish
population
Lim.
Adjusted
data from
standard
mortality
tables
Y
Adjusted
data from
US life
tables
Y
No. Reason
substantiated.
Withdrawal
Withdrawal Withdrawal
estimates not estimates
estimates
substantiated From NDB From
not OS
Danish
DANBIO
reg. not OS
Y
Y
Lim.
Y
N
Y
Y
Y
Y
N
Y
Y
N
Y
Y
Y
Y
Y
N
Y
Y
Y
N
N
N
DAS-28
instead
Costs only
N
Adjusted data
from Finnish
national
statistics
N
N
Y
N
Adjusted
data from
WHO
member
state life
tables
HAQ
instead
HAQ
instead
N
Adjusted
data
from
standard
mortality
tables
8 Valuation of health
(i.e. QALY)
9 Resource use
10 Discontinuation of
treatment
11 Therapeutic
strategies
12 Population risk
stratification
Values from general
population using direct
measurement***
Include all associated direct
medical costs in the analysis,
but report indirect and nonmedical costs seperately
Y
Y
Y
Y
Y
Y
Use discontinuation rates from
observational studies
Include modelling of most
commonly used therapeutic
strategy with sensitivity
analysis to consider other
strategies
Include clear definition of
underlying population
including low and high risk
groups
Legend: (Y=Yes; N=No; Lim.= Presented, but with limitations)
Y
Y
Y
Y
Estimated as
EQ5D values
based on HAQ
from trials
Travelling
costs included
in analysis
and not
reported
seperately
Time
Constant Not explicitly
horizon
rate per
explained
too short cycle due
to serious
infection
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
N
Y
Y
Y
Lim.
Lim.
Y
Y
Y
Y
N
Y
Y