Rh37101250

British Journal of Rheumatology 1998;37:1102–1109
COST-EFFECTIVENESS AND COST-UTILITY OF COMBINATION THERAPY IN
EARLY RHEUMATOID ARTHRITIS: RANDOMIZED COMPARISON OF COMBINED
STEP-DOWN PREDNISOLONE, METHOTREXATE AND SULPHASALAZINE WITH
SULPHASALAZINE ALONE
A. C. VERHOEVEN, J. C. BIBO, M. BOERS,* G. L. ENGEL and
Sj VAN DER LINDEN for the COBRA TRIAL GROUP
Department of Internal Medicine and Rheumatology, Maastricht University Hospital, Maastricht and *Department of Clinical Epidemiology, University Hospital ‘de Vrije Universiteit’, Amsterdam, The Netherlands Objective. Assessment of the cost-effectiveness and cost-utility of early intervention in rheumatoid arthritis (RA) patients, with combined step-down prednisolone, methotrexate and sulphasalazine, compared to sulphasalazine alone.
Methods. Multicentre 56 week randomized double-blind trial with full economic analysis of direct costs and utility analysis with rating scale and standard gamble measurement techniques.
Results. The combined-treatment group included 76 patients and the sulphasalazine group 78 patients. The mean total costs per patient in the first 56 weeks of follow-up were $5519 for combined treatment and $6511 for treatment with sulphasalazinealone (P = 0.37). Out-patient care, in-patient care and non-health care each contributed about one-third to the total costs. Thecombined-treatment group appeared to generate savings in the length of hospital stay for RA, non-protocol drugs and costsof home help, but comparisons were not statistically significant. Protocol drugs and monitoring were slightly more expensivein the combined-treatment group. Clinical, radiographic and functional outcomes significantly favoured combined treatmentat week 28 (radiography also at week 56). Utility scores also favoured combined treatment.
Conclusion. Combined treatment is cost-effective due to enhanced efficacy at lower or equal direct costs.
K : Early rheumatoid arthritis, Cost-effectiveness, Cost–utility analysis, Combined treatment, DMARDs,Glucocorticoids, Prednisolone, Methotrexate, Sulphasalazine, Randomized double-blind trial.
R arthritis (RA) is a systemic disease with symmetrical inflammation of joints in the upper and lower extremities as the most important feature.
Patients suffer from pain, stiffness, impaired function reported extensively [5]. Briefly, in a 1 yr double- in daily life and at work, increased dependence on blind randomized clinical trial, 156 RA patients (ACR family and friends, and decreased participation in criteria [6 ]), aged 18–70 yr, were randomly assigned leisure activities. RA is associated with morbidity, to two treatment groups. All patients had early, active worsening of quality of life and mortality [1, 2].
disease. No prior treatment with second-line anti- The monetary cost of RA is high due to increased rheumatic medication, apart from antimalarials, was use of out-patient medical services, increased hospital- allowed. One group was treated with a combination ization rates and major work disability in the course of sulphasalazine, methotrexate and prednisolone; the of the disease [3]. Intervention studies that include an other group was treated with sulphasalazine and economic evaluation in RA are rare [4] and, to our double placebo. Prednisolone and methotrexate (or knowledge, non-existent in early RA.
the placebos) were tapered and stopped after week 28 A combination therapy regimen of step-down and week 40, respectively, while sulphasalazine was prednisolone, methotrexate and sulphasalazine, tested continued. All patients had calcium supplementation against sulphasalazine alone in early RA patients, (1 g/day) prescribed for as long as they used prednis- demonstrated excellent clinical response, low toxicity olone, and folic acid (1 mg/day) for as long as they and slowing of radiographic progression [5] (COBRA used methotrexate. The primary clinical outcome was trial; COmbinatietherapie Bij Reumatoı¨de Artritis).
a pooled index [7]; a composite measure that reflects The current full economic analysis addresses the ques- each patient’s clinical improvement. This measure is tion of whether this combined treatment is also cost- suited for the comparison of clinical benefits between groups. In this report, the benefits are also expressedin terms of improved physical function and utility.
All utility assessments were performed by trainedindependent assessors who contacted the patients onlyat baseline and four times thereafter. This way, the Submitted 15 December 1997; revised version accepted 2 June assessors stayed blind for effects of high-dose prednis- olone during the first 6 weeks of the protocol. The Epidemiology VE9-78, VU University Hospital, PO Box 7057, 1007 study protocol was approved by research and medical ethics committees in all participating hospitals (nine VERHOEVEN ET AL.: COST-EFFECTIVENESS OF COMBINATION THERAPY clinical centres in The Netherlands and one in were included. In addition, bone density measurement Belgium). All patients gave informed consent in and thorax radiography were considered mandatory in writing before they entered the study protocol between patients before starting combined treatment to assess pre- existing osteopenia and to exclude pulmonary tubercu-losis; these costs were attributed to combined-treatment only. Remaining costs strictly related to the execution of Costs were elicited from a societal perspective. To the trial were not included, e.g. costs of outcome assess- evaluate the economic consequences of combined treat- ment, only direct costs were considered. Direct costs Costs of non-protocol medication (prescription and are costs that are directly related to the intervention.
‘over-the-counter’ drugs) were assessed by records from These are detailed below in five parts: (1) costs of the the diaries. Four different classes of medication were intervention (protocol drugs and monitoring); (2) costs distinguished: non-steroidal anti-inflammatory drugs of non-protocol drugs; (3) other costs of out-patient (NSAIDs) and analgetics, gastroprotective agents, care; (4) costs of in-patient care; (5) direct non-medical disease-modifying anti-rheumatic drugs (DMARDs) costs. The period of follow-up comprised 56 weeks in (apart from protocol medication) and miscellaneous.
Apart from medication, health care utilization Costs were primarily expressed in 1995 Dutch included visits to the general practitioner, specialists, guilders and subsequently converted into US dollars physiotherapists, and any diagnostic and all thera- ($) at the 1994 Purchasing Power Parities rate of peutic procedures ordered as a consequence of these 2.143:1 [8]. Where possible, specific cost prices were visits. The costs of consultation of a general practi- derived from cost research performed in the University tioner ($15) or physiotherapist ($16) were based on charges. The costs of out-patient department specialist internal report, February 1996). These cost prices were consultations (ranging from $22 to $55 per consulta- generalized to all participating clinical centres. Data tion) were based on records of 26 (17%) trial patients on health care utilization were obtained from patient in the University Hospital Maastricht and generalized diaries specifically developed for this study. Patients per treatment group to patients from other participat- were asked to complete forms for 56 weeks; these ing clinical centres. Although registered, costs for any weekly forms were collected at every control visit.
aids (prostheses and ortheses) and adjustments to the Additional data were collected from hospital records patients’ homes are not reported here (the volume of and biannual structured interviews. In the case of different aids and adjustments is relatively small, and doubt, the hospital record was decisive (e.g. to deter- the reimbursement systems both in The Netherlands mine the exact duration of hospitalization). Patients and Belgium vary substantially by municipality).
were asked to report every health care utilization, Costs of hospitalization were priced at $400/day in regardless of whether it was related to their disease, a university or non-university hospital (based on cost because the symptoms and signs as well as the side- research) and $110/day in a rehabilitation centre effects of treatment in RA are very heterogeneous, and (charge derived from the literature [10]).
The direct non-medical costs comprise costs for Total intervention costs comprise costs of the thera- professional or non-professional help at home, costs peutic intervention itself and costs related to the necessary related to loss of leisure time of the patient and an monitoring of adverse effects. The costs of protocol accompanying person where necessary, transportation medication were calculated per patient by multiplying costs, and out-of-pocket expenses for disease-related prices (1995 Dutch pharmacy standards, handling costs activities and purchases. Costs of paid and unpaid help included) by the volume of protocol medication in the were derived from the time estimates registered in the first 56 weeks. The costs of calcium and folic acid supple- weekly patient diary. The price for qualified housekeep- ments were only attributed to the combined-treatment ing was set at $11. Unpaid help was valued at 80% of group, although patients in both treatment groups had the price for professional help. Loss of leisure time these prescribed. The costs of monitoring were based on due to visiting health care-providing institutes was a post hoc consensus among the trial rheumatologists valued at $2.12/hour, a price derived from a literature who were knowledgeable on the prevalence of toxicity survey [10]. Travel costs were calculated from the that had occurred during the trial. They set the frequency number of visits (excluding those for trial purposes of out-patient rheumatology consultation at seven in the only) and costs per visit. Kilometre distances were first year for either treatment, the frequency of laboratory gathered from a route-planning computer package that monitoring for sulphasalazine treatment at seven and uses postal area codes. The kilometre price was set at that for combined treatment at eight. With respect to the $0.27, an amount also in use for reimbursement pur- extensiveness of laboratory monitoring, we referred to poses. Out-of-pocket expenses include such costs as the programme that was used during the trial. This is swimming in extra heated pools and costs of alternative open to discussion, and will be dealt with in the sensitivity health care (calculated from patients’ reports in the analysis. Annual radiography of hands and feet to assess joint damage was considered to be part of normal clinical As the time frame of follow-up in this report is only practice in patients with early RA and the integral mon- a little over 1 yr, no discounting of future costs (or itoring schedule of both treatment groups. These costs benefits) back to current value was carried through.
BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 10 Arthritis-related indirect costs, such as sick leave or of the baseline and biannual assessments of utility occupational disability, are not reported here, but will yielded an approximation of QALY gained in each be elucidated in a separate article with a follow-up treatment group. We restricted our time window to the first 56 weeks of the protocol in every patient, i.e. wedid not extrapolate by multiplying the result with a Efficacy measures: clinical effects life expectancy approximation. Direct costs in both The primarily clinical outcome was a pooled index.
treatment groups were related to the gained QALYs This composite measure comprised each patient’s standardized improvement in erythrocyte sedimenta-tion rate ( ESR), grip strength, tender joint count, observer’s global assessment and improvement in func- Sensitivity analysis tested the robustness of the cost tional ability (scored from the MacMaster Toronto estimates obtained. Volumes, as well as prices, especi- Arthritis patient preference interview [11]). A second- ally those solely based on assumptions, were varied to ary endpoint was radiographic damage of joints in evaluate the resulting change in the costs. The focus hands and feet, expressed in a quantitative score was on the influence of set prices of hospital admissions for erosions and joint space narrowing [12]. Many and help at home, as in this study other volumes were clinicians feel that this type of radiological damage based on empirical data. Each of these two prices were score represents cumulative disease activity [13].
varied by adding and subtracting 25%. The relative Radiographs were read by two trained independent price for help by non-professionals was also reduced observers unaware of treatment allocation. Physical from the original 80 to 0% of professional charges.
function was evaluated by a validated Dutch version Monitoring costs were set at a fixed level in the primary of the Health Assessment Questionnaire [14, 15] filled analysis. The monitoring schedule is the consensus of out by the patients at baseline, and weeks 16, 28, 40 the trial rheumatologists achieved in completion of the study. The economic consequences of different mon- Interpolation of each patient’s scores at five time itoring schedules (i.e. a 25% increase) were considered, points yielded a time-integrated score reflecting average as not only the frequency, but also the extensiveness, clinical benefit and disability during the 56 week follow- of the monitoring schedule for combined treatment is up period. Direct costs in both treatment groups were open to discussion. The monitoring costs for treatment related to the clinical benefits; between-group differ- with sulphasalazine alone were derived from a pub- ences in time-integrated units of the pooled index, lished ACR recommendation [18]. Here, we did not change the frequency of laboratory checks as our integrated functional disability scores may be used in To analyse the difference in costs between the In addition to traditional measures of efficacy, utilit- combined-treatment and sulphasalazine group, costs ies by standard gamble and rating scale methods were per patient-year (of 56 weeks) were calculated and measured at baseline, and weeks 28 and 56. A utility expressed as means per patient per group. As costs are is a single comprehensive outcome measure that reflects cumulative, the mean is a useful statistic because of its the value or preference that respondents assign to a direct relationship to the sum. Although the distribu- particular health state. This value is expressed on a tion of costs was skewed, the number of patients in scale ranging from 1 (perfect health) to 0 (death), and our trial permitted parametric (Student’s t) testing for takes into account both the positive treatment effects between-group differences in mean aggregate costs and the negative side-effects. We elicited utilities from (central limit theorem). In the secondary analyses on patients participating in the trial by means of the the differences in volumes, non-parametric Mann- Maastricht Utility Measurement Questionnaire, a reli- Whitney tests were preferred. All differences in volume able and validated adaptation in Dutch of the were tested per semester because a priori maximal MacMaster Utility Measurement Questionnaire [16, contrast was expected in the first semester considering 17]. It is administered as an interview. Utility measure- the withdrawal of combined treatment after week 28.
ment using this questionnaire comprises the rating No adjustments were made for multiple testing. All scale and the standard gamble methods. The rating analyses were performed on an intention-to-treat basis.
scale measures utilities directly by asking the patients to place health states on a thermometer scale (i.e.
vertical visual analogue scale). The standard gamble In one patient (randomized to combined treatment), method derives utilities from the patients’ responses to protocol medication was stopped within 1 week decision situations under risk. A chance board with a because his disease was in remission at baseline.
probability wheel was used as a visual aid to facilitate Another patient (randomized to sulphasalazine treat- ment) dropped out in week 2 due to adverse effects The quality adjusted life year (QALY ) is a measure (skin rash) and was lost to follow-up for most cost that expresses effects in terms of both quality of life parameters. Data from these two patients were and survival. A calculation of the area under the curve VERHOEVEN ET AL.: COST-EFFECTIVENESS OF COMBINATION THERAPY combined-treatment group included 76 patients and for combined treatment compared to sulphasalazine the sulphasalazine group 78 patients. Data on costs (P = 0.27). This was mainly due to a significantly were complete for all these patients. Both groups were lower number of in-hospital days in admitted patients balanced in most important demographic and prog- during the first semester: 515 vs 190 days (P = 0.05).
In the first semester, the number of hospital admissions The mean total costs per patient in the first 56 weeks was 18 in the combined-treatment group vs 24 in the of follow-up were $5519 for combined treatment sulphasalazine group; in the second semester, there and $6511 for treatment with sulphasalazine alone were seven admissions in both groups. Admissions (P = 0.37; Table II ). Out-patient care, in-patient care were shorter in the combined-treatment group; mean and non-health care each contributed about one-third to the total costs. In the first semester, total costs were Total direct non-medical costs were a mean $1840 almost twice those in the second semester. By defini- for patients in the combined-treatment group and tion, combined treatment (protocol drug cost) was $2133 in the sulphasalazine group (P = 0.54). Home more expensive than sulphasalazine: $326 vs $181 per help accounted for 94%. Not all of these costs were patient. Likewise, the monitoring schedules cost $839 really spent, as not only professional but also voluntary per year in the combined-treatment group vs $559 in help by spouse, family or friends was appraised. The the sulphasalazine group ( Table III ). The costs of demand decreased slightly in the second semester.
drugs outside the protocol were significantly lower Patients in the combined-treatment group reported less in the combined-treatment group ($237 vs $329; help than patients in the sulphasalazine group, but this P < 0.001). This was mainly due to lower use of NSAIDS, analgetics and gastroprotective drugs in the The clinical effects have been fully reported elsewhere combined-treatment group, especially during the first [5]. Briefly, at week 28, the mean improvement in terms of the pooled index was 1.4 for the combined- Apart from the intervention under study and non- treatment group and 0.8 for the sulphasalazine group protocol drugs, the costs of ambulant care were slightly (P < 0.0001). At week 56, these values were 1.1 vs 0.9.
lower in the combined-treatment group. Patients in the Comparison of time-integrated indices (with four combined-treatment group consulted their general follow-up measurements and baseline by definition practitioners more often, but paid less visits to physio- equal to zero) shows mean improvement of 1.1 vs 0.7 and ergotherapists (Table IV ). Taken together, the (P = 0.0001). Radiographic damage scores of both effect of protocol treatment predominated to make groups were comparable at baseline. At this time, 23% combined treatment $219 more expensive than sulpha- of the patients in the combined-treatment group and salazine in out-patient costs (P = 0.007).
19% in the sulphasalazine group showed no damage.
Costs of in-patient care were a mean $917 lower After 56 weeks, the median progression in the radio-graphic damage score was two points in the combined-treatment group vs six in the sulphasalazine group (P = 0.004). Baseline and follow-up data on disability Baseline characteristics of the study patients, according to treatment of 154 patients were available (76 combined-treatment and 78 sulphasalazine). Baseline scores of both groups were comparable. At week 28, the improvement in the disability score was 1.1 for the combined-treatment (P < 0.0001). At week 56, improvement compared to baseline was 0.8 vs 0.6 (P = 0.06). Comparison of 56 weeks time-integrated scores showed mean improve- ment scores of 0.83 vs 0.50 (P = 0.0003).
Utility assessments including baseline and two follow-up assessments were available for 67 patients in the combined-treatment group and 75 in the sulphasal- azine group. Baseline utility assessment resulted in similar scores for both treatment groups; scores with the rating scale method were 0.55 (.. 0.18) and 0.55 (0.20) for the combined-treatment and sulphasalazine group, respectively; scores with standard gamble were 0.78 (0.16) and 0.76 (0.19). At week 28, rating scale utility scores increased by 0.24 (... 0.02) in the combined-treatment and 0.15 (0.02) in the sulphasal- azine group (P = 0.006). Standard gamble utility scoresincreased by 0.10 (0.02) in the combined-treatment *Mean ± .., count and percentages, or median (minimum– and 0.06 (0.02) in the sulphasalazine group (P = 0.05).
At week 56, most of the between-group contrast seen †Patients with available baseline radiographs; combined treatment group n = 74, sulphasalazine group n = 75.
after the first semester was lost: mean improvement BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 10 VERHOEVEN ET AL.: COST-EFFECTIVENESS OF COMBINATION THERAPY Costs of possible monitoring schedules during the first 56 weeks of treatment For combined treatment (post hoc consensus rheumatologists, see Patients and methods):8 × : laboratory tests: ESR, haemoglobin, haematocrit, white blood cell count ( WBC ) including differentiation, platelet count and mean corpuscular volume, ALAT, ASAT, bilirubin, alkaline phosphatase, glucose, creatinine, sodium, potassium,phosphate, calcium, albumin and total protein, urinalysis (albumin and glucose) 1 × : bone densitometry lumbar spine and one hip For sulphasalazine treatment (post hoc consensus rheumatologists):7 × : laboratory tests (see above) For sulphasalazine treatment according to ACR recommendations [18]:7 × : laboratory tests: haemoglobin, WBC including differentiation and platelet count Volumes of out-patient care. Number of visits to health care professionals* *COBRA, combined treatment of step-down prednisolone, methotrexate and sulphasalazine. SSZ, sulphasalazine. Intention-to-treat analysis; COBRA: n = 76; SSZ: n = 78.
with rating scale 0.18 (0.03) in the combined-treatment and not significantly different in both groups; in the group vs 0.16 (0.02) in the sulphasalazine group, and combined-treatment group, bone density in the lumbar with standard gamble 0.07 (0.02) vs 0.07 (0.02). The spine decreased by a mean of 1.3% in 56 weeks.
area under the curve calculation of the utility scores In the sensitivity analysis, when the price of hospital- demonstrated a significantly better gain of 0.06 QALY ization was reduced by 25%, the mean total cost assessed by rating scale (P = 0.01; 95% CI: 0.02; 0.10).
remained $849 lower in the combined-treatment group.
Assessed by standard gamble, the difference is 0.02 The same adjustment in the charge for help at home gained QALY (P = 0.33; −0.02; 0.06).
resulted in a smaller between-group contrast of $900.
The efficacy of the combined treatment in clinical When the price for help at home by non-professionals outcomes is superior to sulphasalazine alone. The total was set at zero, mean total costs of combined-treatment costs are slightly lower, although not significantly so remained $443 lower than costs of treatment with in the combined group. Accordingly, relative cost- sulphasalazine. Finally, the consequences of adjust- efficacy favours combined treatment; it will be cost- effective when implemented in patients comparable to ( Table III ). The adjusted monitoring costs were $1049 those included in this COBRA trial. Likewise, signific- (+25%) and $357 (−36%) for combined treatment and antly better utility scores and equal costs result in sulphasalazine, respectively; this still resulted in $580 better cost–utility ratios in the combined-treatment lower total costs for combined treatment.
No extra toxicity due to the extra medication in the combined-treatment group occurred. On the contrary, In the setting of a randomized trial, this full eco- significantly fewer patients in the combined-treatment nomic analysis revealed combined treatment with step- group stopped protocol medication due to adverse down prednisolone, methotrexate and sulphasalazine events or lack of therapy effect; 6 vs 23 in the sulpha- to be more effective than sulphasalazine alone at equal salazine group (P = 0.0008). The drop-outs in the or lower costs. The combined-treatment group had combined-treatment group also occurred later. Loss of lower expenses for non-protocol medication and bone mineral density in spine and hips was modest, in-patient care, and lower costs outside the health care BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 10 system that offset the higher costs for protocol medica- had active disease, and restrictions to the number or duration of hospitalizations would most likely have The efficacy of the combined treatment in clinical as worsened the outcomes in the more frequently well as radiological outcomes is superior to sulphasal- hospitalized sulphasalazine group [21]. Moreover, the azine alone. As the total direct costs in the combined- costs of health care found in this study do not seem to be ‘out of range’ in comparison with figures from cost-efficacy favours combined treatment. Utility an American health care setting [9, 20].
scores, as a generic measure of therapy benefits, also In comparison with sulphasalazine, combined treat- favoured combined treatment. Concordant with other ment is the dominant therapeutic option, due to equal reported studies, the scores derived with the standard or lower costs and enhanced efficacy. This full eco- gamble method were at an absolute higher level and nomic analysis confirms that combined treatment with least responsive to change. This has been attributed to step-down prednisolone, methotrexate and sulphasal- the risk-aversive attitude of patients with a putatively azine may be a rational choice in early and active RA.
non-fatal or chronic disease like RA [19].
The time frame of this study is long compared to A most other studies, but still relatively short for a The COBRA trial was funded by grant 92-045 chronic disease like RA, and definitely too short to evaluate the incidence of late effects. Continuing Netherlands, of which ACV was a research fellow. We follow-up of the included patients will provide import- thank Pharmacia and Upjohn, Sweden, for providing ant answers on costs and outcomes in the long run.
sulphasalazine free of charge, and are grateful to Mrs Prednisolone might have induced modest (non- Carla Bakker for her contribution in the development significant, mean <1%) and partially reversible bone of the utility interview, to Mrs Marjolein Braeken for loss, but this did not result in symptomatic fractures trial management and to Mrs Lily Heusschen-Houben or complaints in any of the patients.
for data entry. The COBRA trial group comprised the Like most clinical trials, this trial was primarily following centres and rheumatologists: Danie¨l den designed to study clinical benefits and thus probably Hoed Kliniek, Rotterdam, H. M. Markusse; Medisch underpowered to prove cost benefits. A post hoc power Spectrum Twente and Twenteborg Hospital, Enschede calculation reveals minimum group sizes of 374 neces- and Almelo, M. A. F. J. van de Laar; University sary for the between-group difference found to reach Hospital, Maastricht, M. Boers; Pellenberg Hospital, the level of significance (P < 0.05, two-sided). The Catholic University, Leuven, Belgium, R. Westhovens; only significant difference in cost found favoured the Jan van Breemen Instituut, Amsterdam, J. C. van sulphasalazine group, i.e. $219 lower costs for out- Denderen; Bronovo Hospital, The Hague, D. van Zeben; patient care. However, this can be largely attributed University Hospital, Leiden, B. A. C. Dijkmans; Reinier to the per protocol (fixed ) difference in intervention de Graaf Gasthuis, Delft, A. J. Peeters; Sint Laurentius Hospital, Roermond, P. Jacobs; Medisch Centrum, Sensitivity analysis showed the robustness of the conclusions in the economic analysis. The direct costsof combined treatment are not significantly lower (and must thus be presumed to be similar to the costs of 1. Gordon DA, Hastings DE. Rheumatoid arthritis.
treatment with sulphasalazine alone). On the other Clinical features: early, progressive and late disease. In: hand, calculations with considerable alternations in Klippel JH, Dieppe PA, eds. Rheumatology. St Louis: the assumptions of charges and extensiveness of mon- 2. Wolfe F, Hawley DJ, Cathey MA. Clinical and health itoring schedules consistently showed lower costs for status measures over time: prognosis and outcome combined treatment. As the primary analysis of the trial showed significantly less drop-out due to therapy failure and adverse events in the combined-treatment 3. McIntosh E. The cost of rheumatoid arthritis. Br group, a less frequent and extensive monitoring sched- ule than that performed during the trial might be 4. Bosi Ferraz M, Maetzel A, Bombardier C. Critical appropriate in normal clinical practice. Also, the neces- appraisal of economic evaluation published in the field sity of initial bone densitometry—in this analysis still of rheumatology and related disciplines. Arthritis Rheum regarded as mandatory for combined treatment—is questionable with regard to the observed modest effects 5. Boers M, Verhoeven AC, Markusse HM et al. Randomised comparison of combined step-down pred- nisolone, methotrexate and sulphasalazine with sulpha- The generalizability of our findings is principally salazine alone in early rheumatoid arthritis. Lancet restricted to the health care systems in The Netherlands and Belgium. The health care system in these countries 6. Arnett FC, Edworthy SM, Bloch DA et al. The American is characterized by universal access and equal facilities Rheumatism Association revised criteria for the classi- for all inhabitants, and small distances between fication of rheumatoid arthritis. Arthritis Rheum patients’ homes and clinical centres. From a North American perspective, patients were frequently hospit- 7. Smythe HA, Helewa A, Goldsmith CH. ‘Independent alized [20]. Notably, at baseline, all included patients assessor’ and ‘pooled index’ as techniques for measuring VERHOEVEN ET AL.: COST-EFFECTIVENESS OF COMBINATION THERAPY treatment effects in rheumatoid arthritis. J Rheumatol 15. Siegert CEH, Vleming LJ, Vanderbroucke JP, Cats A.
Measurement of disability in Dutch rheumatoid arthritis 8. Centraal Economisch Plan. The Hague: SDU, 1996.
patients. Clin Rheumatol 1984;3:305–9.
9. Gabriel SE, Crowson CS, Campion ME, O’Fallon WM.
16. Bennett K, Torrance GR, Tugwell P. Methodological Direct medical costs unique to people with arthritis.
challenges in the development of utility measure of health-related quality of life in rheumatoid arthritis.
Control Clin Trials 1991;12 (suppl.):118–28.
10. Rutten FFH, van Ineveld BM, van Ommen R, van Hout 17. Bakker CH, Rutten-van Mo¨lken MPMH, van Doorslaer BA, Huijsman R. Directe kosten binnen de gezond- EKA, Bennet K, van der Linden S. Health related utility heidszorg. In: Kostenberekening bij gezondheidszorg- assessment by rating scale and standard gamble in onderzoek; richtlijnen voor de praktijk. Utrecht: van patients with ankylosing spondylitis or fibromyalgia.
Patient Educ Counsel 1993;20:145–52.
11. Tugwell P, Bombardier C, Buchanan WW, Goldsmith 18. American College of Rheumatology ad hoc committee CH, Grace E, Hanna B. The MACTAR patient prefer- on clinical guidelines. Guidelines for monitoring drug ence disability questionnaire: an individualized func- therapy in rheumatoid arthritis. Arthritis Rheum tional priority approach for assessing improvement in physical disability in clinical trials in rheumatoid arth- 19. Goossens MEJB, Rutten-van Mo¨lken MPMH, Leidl ritis. J Rheumatol 1987;14:446–51.
RMJ, Bos SGPM, Vlaeyen JWS, Teeken-Gruben NJG.
12. van der Heijde DMFM, van Riel PL, Nuver-Zwart IH, Gribnau FW, van de Putte LBA. Effects of hydroxy- Randomized clinical trial. II Economic evaluation.
chloroquine and sulphasalazine on progression of joint damage in rheumatoid arthritis. Lancet 1989;i:1036–8.
20. Lanes SF, Lanza LL, Radensky PW et al. Resource 13. van Leeuwen MA, van der Heijde DMFM, van Rijs utilization and cost of care for rheumatoid arthritis and wijk MH et al. Interrelationship of outcome measures osteoarthritis in a managed care setting. The importance and process variables in early rheumatoid arthritis. A of drug and surgery costs. Arthritis Rheum 1997;40: comparison of radiologic damage, physical disability, joint counts, and acute phase reactants. J Rheumatol 21. Vliet Vlieland TPM, Zwinderman AH, Vandenbroucke JP, Breedveld FC, Hazes JMW. A randomised clinical 14. Fries JF, Spitz PW, Young DY. The dimensions of trial of in-patient multidisciplinary treatment versus health outcome: the Health Assessment Questionnaire, routine outpatient care in active rheumatoid arthritis.
disability and pain scales. J Rheumatol 1982;9:789–93.

Source: http://cobratherapie.nl/pdf/Verhoeven%20et%20al,%201998.pdf