Doi:10.1016/j.cct.2005.08.003

Contemporary Clinical Trials 27 (2006) 1 – 12 The use of unequal randomisation ratios in clinical trials: A review J.C. Dumville a,*, S. Hahn b, J.N.V. Miles a, D.J. Torgerson a a Area 4, York Trials Unit, Department of Health Sciences, University of York, York, YO10 5DD, United Kingdom b Medical Research Collaborating Center, Seoul National University College of Medicine, Seoul National University Hospital, Seoul 110-744, Republic of Korea Received 28 September 2004; accepted 11 August 2005 Objective: To examine reasons given for the use of unequal randomisation in randomised controlled trials (RCTs).
Main Measures: Setting of the trial; intervention being tested; randomisation ratio; sample size calculation; reason given forrandomisation.
Methods: Review of trials using unequal randomisation.
Databases and sources: Cochrane library, Medline, Pub Med and Science Citation Index.
Results: A total of 65 trials were identified; 56 were two-armed trials and nine trials had more than two arms. Of the two-arm trials,50 trials recruited patients in favour of the experimental group. Various reasons for the use of unequal randomisation were given.
Six studies stated that they used unequal randomisation to reduce the cost of the trial, with one screening trial limited by theavailability of the intervention. Other reasons for using unequal allocation were: avoiding loss of power from drop-out or cross-over, ethics and the gaining of additional information on the treatment. Thirty seven trials papers (57%) did not state why they hadused unequal randomisation and only 14 trials (22%) appeared to have taken the unequal randomisation into account in theirsample size calculation.
Conclusion: Although unequal randomisation offers a number of advantages to trials the method is rarely used and is especiallyunder-utilised to reduce trial costs. Unequal randomisation should be considered more in trial design especially where there arelarge differences between treatment costs.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Unequal randomisation; Randomised controlled trials Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Learning curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Other . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
* Corresponding author. Tel.: +441904321325; fax: +441904 321383.
E-mail address: [email protected] (J.C. Dumville).
1551-7144/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.cct.2005.08.003 J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reasons for unequal randomisations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sample size calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Maximise available participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Differential loss to follow-up and cross over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Treatment experience. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Unequal allocation in practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reasons for limited use of unequal allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 In randomised controlled trials (RCTs) the allocation of participants to experimental and control groups is usually balanced leading to approximately equal group sizes. However, whilst balanced group sizes will maximise a studyTsstatistical power, the use of unequal randomisation ratios will only significantly reduce the power of a study if theratio is 3 : 1 or more There are a number of situations, discussed below, where unequal group sizes in a trial maybe beneficial. Indeed, it has been argued that the use of unequal randomisation could and should be substantiallyincreased In most RCTs there is probably an unequal distribution of either research or treatment costs between the trial arms.
When this is the case, economic efficiency can potentially be improved using unequal randomisation. For a fixed trialbudget, using unequal randomisation to favour the least expensive trial arm allows more participants to be recruitedthus increasing the power of the trial Where there is a fixed total sample size available for recruitment, unequalrandomisation can confer large financial savings to a trial with limited impact on statistical power Furthermore,when there is a fixed quantity of treatments available for one trial arm, but more than twice the number of trialparticipants are available, trial efficiency can be improved using unequal allocation. Allocating more participants tothe treatment with no restrictions on availability will increase the statistical power of the trial by utilising all theavailable participants.
There are situations where unequal randomisation that favours the experimental group is useful. In evaluations of a new technology that has a dlearningT element, for example a new surgical procedure, more participants might beallocated to the intervention group to reduce the impact of the dlearning curveT on the results of the trial Allocatingmore participants to the experimental treatment allows the clinician to climb the learning curve more rapidly andallows for an analysis that compares both treatments when at the top of this curve.
Some have advocated unequal allocation in order to expose fewer participants to the hazards of a treatment in a trial This view pre-supposes that there is a high degree of certainty regarding the risks and benefits accruing to thetreatment under evaluation. Yet, as many RCTs are conducted when there is equipoise, this view cannot generally beused as justification for unequal allocation in trials. However, there may be cases, for example in clinical drug trials,where the beneficial nature of a treatment, known to have unpleasant side-effects, is still being investigated. Hereunequal allocation could, potentially, be justified on ethical grounds.
J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 Unequal allocation can be useful in studies that compare a group with an individual treatment (e.g., group versus individual cognitive behavioural therapy) where it helps to obtain enough participants for the group therapy. Usingunequal allocation to increase the number of groups will increase the study’s power for a fixed total sample size.
Additionally, a larger number of participants may be allocated to one group when a higher drop-out rate is anticipated, as this will enhance statistical power for a dper protocolT analysis.
Whilst the literature cites a number of reasons why unequal allocation may be beneficial in RCTs, to date there has been no review undertaken that assesses why this method has been used in practice. We undertook a review of trialsthat employed unequal randomisation to investigate why it had been used and to discuss these applications further.
The Cochrane Library, Medline, Pub Med and the Science Citation Index were searched for all paper abstracts that contained words and phrases related to unequal randomisation. These were: bunequal or unbalanced, randomis(z)ation, allocation or ratioQ.
Papers from this literature search were also supplemented by the authorsT detailed personal knowledge of trials that had unequal group sizes and from the bibliographies of papers discussing unequal randomisation. Trials that used acluster design were excluded from this review.
Double data extraction was carried out by JD and SH. Data was extracted on: the setting of the trial (primary care, secondary care or university); the intervention being tested; the randomisation ratio; the sample size calculation usedand the reason given for unequal randomisation. A sample size calculation was recorded as being outlined when somedetails of the calculation were provided. We recorded that unequal randomisation had been taken into account in thesample size calculation if the authors explicitly stated they had accounted for unequal randomisation or we were able torepeat and confirm the sample size calculation for the proposed ratio. Conversely, unequal randomisation was recordedas having not been taken into account if the authors’ suggested sample size calculation was for equal group sizes; if wecould not follow the sample size calculation from the details supplied we recorded the calculation as being unclear.
If the reason for unequal randomisation was not given in the paper, where possible, authors were e-mailed to gain further information about why they used unequal randomisation.
We identified 65 trials that used unequal randomisation . One further paper was identified but was excluded from the study as, although it used unequal randomisation (2 : 1 : 2 : 2 : 1 : 1), of the six arms, only three arms were compared inthe study and these were balanced. Most included trials (~80%) were identified from personal knowledge andbibliographies rather than from databases. Details of the 65 included trials, in terms of settings, interventions, allocationratios, sample size calculation, and reasons for use of unequal allocation, are shown in Appendix 1. summarisesthe trial characteristics. Many trials papers (n = 37; 57%) did not state why they used unequal randomisation in theirdesign; after e-mailing authors 15 replies clarifying the reason for use of unequal randomisation were received.
Of the trials, 56 (86%) were two armed and nine trials had more than two arms. Unequal randomisation favouring the experimental group was used most frequently with the allocation favouring the control group(s) used only seventimes ).
The reasons for unequal randomisation given in each study are detailed in the Appendix. Of the seven studies where allocation favoured the control group, four trials used unequal allocation to reduce financial costs as the J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 a In 2 of these trials unequal randomisation was performed in only some of the centres.
experimental treatment or intervention was more expensive than the control . One further trial was alarge screening programme, where the authors state there were a limited number of screenings available: thusthey wanted to maximise the power of the study for this finite number of examinations A second screeningtrial did not give a clear reason for the use of unequal randomisation that favoured the control group but it islikely to be for similar reasons as the first . The remaining trial that allocated in favour of the control groupdid so for reasons related to statistical considerations, however, this reasoning did not make the use of unequalrandomisation clear Finally, one of the two trials that compared analogous treatments also used unequalrandomisation to reduce cost by minimising the number of participants in the most expensive treatment group The remaining studies mostly had more participants in their treatment group(s) than their control group. Five trials used unequal allocation as they were concerned about drop-out or transfer from the treatment group Six trials used unequal randomisation to increase patient acceptability of the trial and therefore recruitment ratesand three trials stated ethical reasons for maximising participants exposure to the treatment. Individual trials also gave other specific reasons for using unequal randomisation to maximise allocationto the experimental group. These were: the late starting of one arm of the trial limited variability of the controlgroup more participants required in the treatment group for the next phase of the study ; to give increasedpower for a secondary analysis increased availability of the intervention service compared to the controlservice and to ensure maximum use of the available counselling intervention Only one trial clearly attributed the used of an unbalanced allocation to overcome a learning curve and to gain more experience of a treatment However, gaining more experience of a treatment, including increasedsafety and toxicity data, were given as reasons for the use of unequal randomisation in thirteen further trials J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 Of the 65 trials, 40 (62%) were recorded as having outlined a sample size calculation. Fourteen of these 40 studies appeared to have taken the unequal randomisation into account. The remaining 26 studies that outlined a sample sizecalculation either did not take unequal randomisation into account (n = 5) or did not supply enough detail to establishwhether unequal randomisation had been accounted for (n = 21).
We found few RCTs that employed unequal randomisation. Because unequal randomisation was rarely described in paper abstracts, most papers for this review were identified from extensive combined personalknowledge of clinical trials. This non-systematic approach could be viewed as a possible limitation of thisstudy. Of those trials identified as having unbalanced groups, many gave no apparent reason for using thismethod. The use of unequal randomisation is an important methodological feature of a trial and studies shouldgive a reason for its use. Trials that did discuss their use of unequal randomisation describe doing so for avariety of reasons.
Although the use of unequal allocation can lead to considerable cost savings, economic efficiency was one of the least frequently cited for using unequal allocation. Those studies that did use unequal randomisation to confer savingsappeared to do so successfully. For example, in a screening trial where screening cost o70 per person, 195,000participants were randomised with unbalanced allocation leading to savings of more than o2 million with very littleloss in power Similarly, unequal allocation in a trial of hip protectors reduced the overall cost of the trial by10% (Torgerson, unpublished observation).
Three studies, in which the availability of one treatment was not restricted used unequal allocation to improve statistical power. For example, had Turnbull et al. used equal allocation rather than unequalallocation they would have recruited over 100 less patients into their intervention group and thus into theirtrial.
Unequal randomisation can improve power in three armed trials when a dose comparison between two drug arms is being carried out as the differences between drug dosages would be expected to be more muted compared with theplacebo comparison. In this review no trial gave this as a reason for using unequal randomisation although one three-armed trial used this design 4.3. Differential loss to follow-up and cross over Unequal allocation was used because of anticipated differences in drop-out or treatment cross-over rates between groups in five studies . The advantage of unequal randomisation in this situation is that it willpreserve statistical power for an don treatmentT or dper protocolT analysis. However, it is important to note that whenlarge drop-out or cross-over rates are anticipated, unequal randomisation will not reduce the risk of bias throughdrop-out nor will it prevent ddilutionT effects of participants crossing over into the other treatment arm. Intention totreat analysis, that is analysing all participants in their originally allocated groups, must still be carried out as thisgives a conservative but unbiased estimate of treatment effects. The five studies identified here did undertake anintention to treat analysis.
Thirteen studies exposed increased numbers of participants to the experimental treatment or procedures to gain more experience in using the treatment. This included the collection of safety and toxicity data including adverse J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 events. Additionally, further studies that did not state a reason for the use of unequal randomisation were phase I–III-type trials that tested drugs, radiological or surgical treatments and often reported safety and adverse event data. Wepostulate that these studies used unequal randomisation to increase the amount of information on the new treatment.
This application of unequal randomisation has been only briefly discussed in the literature and has been acknowl-edged as being potentially controversial Only one study used unequal allocation specifically because of learningcurve effects . However, this study did not appear to examine whether there was any interaction betweentreatment effects and patient recruitment. If learning curve effects are a reason for unequal allocation, it would belogical to look for any effect of learning and, if observed, present an analysis of participants recruited don the topT ofthe learning curve.
Practically, carrying out unequal allocation adds little complexity when compared to using balanced allocation.
The process of randomisation must simply ensure that, in the case of a 2 : 1 ratio, double the number of peoplerandomly enter one group compared to the other.
It is important when using unequal allocation to ensure that a correct sample size calculation is performed, and ideally a statistician should be consulted. In a majority of the studies we examined it did not appear that unequalrandomisation had been taken into account for the sample size calculation. This is particularly serious wherestudies have used randomisation ratios of 3 : 1 or more where the power of the study will have been significantlyreduced.
Input from a statistician is also useful in the analysis of data from trials using unequal randomisation. The statistical tests commonly used to analyse data from trials, such as the t-test, ANOVA, ANCOVA and regression analysis, makean assumption about the homogeneity of the variance of the residuals across the levels of the independent variables. Ifthe variances are not equal, but groups are equal this does not seriously affect the type 1 error rate. However, wherethe groups are not equally allocated this does not always hold. Thus the decision to use unequal randomisation musttake into account the potential for unequal variance.
4.6. Reasons for limited use of unequal allocation There may be several reasons why trials rarely employ unbalanced allocation. Some might, incorrectly, consider unequal groups sizes as somehow being dunscientificT. Indeed, when advocating unequal randomisation, one of theauthors (DT) has had such feedback on two different occasions from grant-giving bodies. Also, the need to maximisethe power of the study might be the over-riding concern among some trial statisticians and for a given total samplesize, balanced allocation, usually, gives the most statistical power. However, for a fixed sample size, unequalrandomisation of 2 : 1 can be used without a serious loss of power . One reason why unequal allocation is notused in factorial trials may be the view that this makes the analysis more complicated. When computations of trialswere done dby handT completely balanced factorial designs were easier to calculate; however, with the use of moderncomputers this justification no longer holds.
Resistance to unequal randomisation may stem from concern that its use may require the recruitment of an increased number of participants, since recruitment is often a limiting factor in trials. Indeed, where there are a smallnumber patients that can be recruited unequal randomisation may not be advantageous. However, as previously noted,where there is a fixed budget, savings introduced by use of unequal randomisation may allow more sites to berecruited to the trial — thus allowing increased recruitment, which can increase the power of a study Furthermore,for a fixed sample size potential cost savings from using unequal allocation should be balanced against the loss ofpower that unequal randomisation will cause.
Unequal randomisation is not commonly undertaken. Although by no means suitable for every trial, there are a number of situations where unequal randomisation can be useful, especially by conferring financialsavings on trials — yet unequal randomisation is rarely used for this reason and could be implemented moreoften.
Effect of direct payment on questionnaire Drop out (due to transfer) expected intreatment group Drop out (due to transfer) expected intreatment group Prevention of otitis media with inactivated Lamivudine in children with chronic hepatitis Effectiveness of influenza vaccine in healthy, Cognitive–behavioural therapy in the treatment Comparison of zolmitriptan and sumatriptan Discontinuing drug therapy in HIV-infected Peginterferon-a2a and ribavirin combination exposure of difficult to treat patientsto the experimental treatment — not Adefovir dipivoxil for the treatment of chronic Prophylactic implantation of a defibrillator Effects of antenatal day care for medical Evidence of real world effectiveness of a Laparoscopic versus abdominal hysterectomy Dermatan sulphate for the prevention of deep Pirenzepine ophthalmic gel in myopic children therapy with losartan and hydrochlorothiazide Insulin vs. human insulin in type 1 diabetes Effects of bicarbonate/lactate-based dialysis Endoscopic versus open carpel tunnel release: Uterine artery embolization vs. abdominal Pantoprazole in the prevention of peptic ulcers Large-scale prostrate cancer screening trials Zidovudine and Zalcitabine in HIV patients Neutron radiation therapy for head and neck Effects of Alendronate and risedronate on bone Effect of sibutramine in weight maintenance Nicardipine therapy in ambulatory elderly Sustained-release formulation of ibuprofen Trandolapril a new angiotensin converting Hypocaloric high protein diet spares lean body Radiotherapy in the treatment of lung cancer Cost effectiveness of transplantation types for Cost-effectiveness analysis of improved blood pressure control in hypertensive patients a Primary care (PC), Secondary care (SC), University (U).
b Reason for unequal randomisation not stated in paper but given on e-mail.
J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 [1] Pocock SJ. Clinical trials: a practical approach. Chichester7 John Wiley and Sons; 1995.
[2] Torgerson D, Campbell M. Unequal randomisation can improve the economic efficiency of clinical trials. J Health Serv Res Policy [3] Avins AL. Can unequal be more fair? Ethics, subject allocation, and randomised clinical trials. J Med Ethics 1998;24:401 – 8.
[4] Birks YF, Porthouse J, Addie C, et al. Randomised controlled trial of hip protectors among women living in the community. Osteoporos Int [5] Gail M, Williams R, Byar D, Brown C. How many controls? J Chronic Dis 1976;29:723 – 31.
[6] Rosenberger WF, Lachin JM. Randomization in Clinical Trials. John Wiley and Sons, New York.
[7] Abraham E, Park YC, Covington P, Conrad SA, Schwartz M. Liposomal prostaglandin E1 in acute respiratory distress syndrome: a placebo- controlled, randomized, double-blind, multicenter clinical trial. Crit Care Med 1996;24:10 – 51.
[8] Adinolfi LE, Utili R, Tonziello A, Ruggiero G. Effects of alpha interferon induction plus ribavirin with or without amantadine in the treatment of interferon nonresponsive chronic hepatitis C: a randomised trial. Gut 2003;52:701 – 5.
[9] Agnelli G, Cosmi B, Di FP, et al. A randomised, double-blind, placebo-controlled trial of dermatan sulphate for prevention of deep vein thrombosis in hip fracture. Thromb Haemost 1992;67:203 – 8.
[10] ALLHAT Collaborative Research Group. Major cardiovascular events in hypertensive patients randomized to doxazosin vs. chlorthalidone: the antihypertensive and lipid-lowering treatment to prevent heart attack trial. JAMA 2000;283:1967 – 75.
[11] Atkin WS, Edwards R, Wardle J, et al. Design of a multicentre randomised trial to evaluate flexible sigmoidoscopy in colorectal cancer screening. J Med Screen 2001;8:137 – 44.
[12] Bae CY, Cho CY, Cho K, et al. A double-blind, placebo-controlled, multicenter study of cerebrolysin for Alzheimer’s disease. J Am Geriatr [13] Bartlett JD, Niemann K, Allred, Houde B, Edmondson MJ, Crocett RS. A tolerability study of pirenzepine ophthalmic gel in myopic children.
[14] Bianchi Porro G, Lazzaroni M, Imbesi V, Montrone F, Santagada T. Efficacy of pantoprazole in the prevention of peptic ulcers, induced by non-steroidal anti-inflammatory drugs: a prospective, placebo-controlled, double-blind, parallel-group study. Dig Liver Dis 2000;32:201 – 8.
[15] Canal N, Imbimbo BP. Relationship between pharmacodynamic activity and cognitive effects of eptastigmine in patients with Alzheimer’s disease. Clin Trials Ther Effects, vol. 60. 1996. p. 218 – 28.
[16] Casley-Smith JR, Wang CT, Casley-Smith JR, Zi-hai C. Treatment of filarial lymphoedema and elephantiasis with 5,6-benzo-alpha-pyrone (coumarin). BMJ 1993;307:1037 – 41.
[17] Cocconi G, Bella M, Zironi S, et al. Fluorouracil, doxorubicin, and mitomycin combination versus PELF chemotherapy in advanced gastric cancer: a prospective randomized trial of the Italian Oncology Group for Clinical Research. J Clin Oncol 1994;12:2687 – 93.
[18] Critchley JAJH, Gilchrist N, Ikeda L, et al. A randomized, double masked comparison of the antihypertensive efficacy and safety of combination therapy with losartan and hydrochlorothiazide verses captopril and hydrochlorothiazide in elderly and younger patients. CurrTher Res Clin E 1996;57:392 – 407.
[19] Cunningham D, Pyrhonen S, James RD, et al. Randomised trial of irinotecan plus supportive care versus supportive care alone after fluorouracil failure for patients with metastatic colorectal cancer. Lancet 1998;352:1413 – 8.
[20] Deeks SG, Wrin T, Liegler T, et al. Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV- infected patients with detectable viremia. N Engl J Med 2001;344:472 – 80.
[21] de Koning HJ, Auvinen A, Berenguer Sanchez A, et al. European Randomized Screening for Prostate Cancer (ERSPC) Trial. International Prostate Cancer Screening Trials Evaluation Group. Large-scale randomized prostate cancer screening trials: program performances in theEuropean Randomized Screening for Prostate Cancer Trial and the Prostate, Lung, Colorectal and Ovary Cancer Trial. Int J Cancer2002;97:237 – 44.
[22] Dowrick C, Dunn G, Luis Ayuso-Mateos J, et al. Problem solving treatment and group psychoeducation for depression: multicentre randomized controlled trial. BMJ 2000;321:1450 – 4.
[23] Epstein O, Lee R.g, Boss AM, Juan S, Cook DG, Scheuer PJ. d-Penicillamine treatment improves survival in primary biliary cirrhosis. Lancet [24] Errington RD, Ashby D, Gore SM, et al. High energy neutron treatment for pelvic cancers: study stopped because of increased mortality.
[25] Fischl MA, Stanley K, Collier AC, et al. Combination and monotherapy with zidovudine and zalcitabine in patients with advanced HIV disease. The NIAID AIDS Clinical Trials Group. Ann Intern Med 1995;122:24 – 32.
[26] Garry R, Fountain J, Mason S, et al. The evaluate study: two parallel randomised trials, one comparing laparoscopic with abdominal hysterectomy, the other comparing laparoscopic with vaginal hysterectomy. BMJ 2004;328:129 – 37.
[27] Geraud G, Olesen J, Pfaffenrath V, et al. Comparison of the efficacy of zolmitriptan and sumatriptan: issues in migraine trial design.
[28] Goodfield MJD, Andrew L, Evans EGV. Short term treatment of dermatophyte onychomycosis with terbinafine. BMJ 1992;304:1151 – 3.
[29] Goodwin PJ, Leszcz M, Ennis M, et al. The effect of group psychosocial support on survival in metastatic breast cancer. N Engl J Med [30] Griffin TW, Davis R, Hendrickson FR, Maor MH, Laramore GE. Fast neutron radiation therapy for unresectable squamous cell carcinomas of the head and neck: the results of a randomized RTOG study. Int J Radiat Oncol 1994;10:2217 – 21.
J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 [31] Hadziyannis SJ, Sette H, Morgan TR, et al. Peginterferon-a2a and ribavirin combination therapy in chronic hepatitis C. Ann Intern Med [32] Hadziyannis SJ, Tassopoulos NC, Heathcote EJ, et al. Adefovir dipivoxil for the treatment of hepatitis B e antigen-negative chronic hepatitis [33] Hoberman A, Greenberg DP, Paradise JL, et al. Effectiveness of inactivated influenza vaccine in preventing acute otitis media in young children: a randomized controlled trial. JAMA 2003;290:1608 – 16.
[34] Home PD, Lindholm A, Riist A, for the European Insulin Aspart Study Group. Insulin aspart vs. human insulin in the management of long- term blood glucose control in type 1 diabetes mellitus: a randomized controlled trial. Diabet Med 2000;17:762 – 71.
[35] Hosking D, Adami S, Felsenberg D, et al. Comparison of change in bone resorption and bone mineral density with once weekly alendronate and daily risedronate: a randomized placebo-controlled study. Curr Med Res Opin 2003;19:383 – 94.
[36] Hundley VA, Cruickshank FM, Lang GD, et al. Midwife managed delivery unit: a randomised controlled comparison with consultant led care.
[37] Infant Health and Development Program. Enhancing the outcome of low-weight premature infants. JAMA 1990;263:3035 – 42.
[38] International Collaborative Ovarian Neoplasm Group. Paclitaxel plus carboplatin versus standard chemotherapy with either single-agent carboplatin or cyclophosphamide, doxorubicin, and cisplatin in women with ovarian cancer: the ICON3 randomised trial. Lancet2002;360:505 – 15.
[39] James WP, Astrup A, Finer N, et al. Effect of sibutramine on weight maintenance after weight loss: a randomised trial. STORM Study Group.
Sibutramine trial of obesity reduction and maintenance. Lancet 2000;356:2119 – 25.
[40] Johnson D, Waddell RD, Pelton SI, et al. Randomised trial of intradermal Mycobacterium vaccae or intradermal hepatitis B immunisation in children with HIV infection. Vaccine 1999;17:2583 – 7.
[41] Jonas MM, Kelley DA, Mizerski J, et al. Clinical trial of lamivudine in children with chronic hepatitis. N Engl J Med 2002;346:1706 – 13.
[42] Jones S, Holmes CJ, Krediet RT, et al., Bicarbonate/Lactate Study Group. Bicarbonate/lactate-based peritoneal dialysis solution increases cancer antigen 125 and decreases hyaluronic acid levels. Kidney Int 2001;59:1529 – 38.
[43] Knapp MJ, Knopman DS, Solomon PR, et al. A 30-week randomized controlled trial of high-dose tacrine in patients with Alzheimer’s disease.
The Tacrine Study Group. JAMA 1994;271:985 – 91.
[44] Krakoff LR. Nicardipine monotherapy in ambulatory elderly patients with hypertension. Am Heart J 1989;117:250 – 4.
[45] MacDermid JC, Richards RS, Roth JH, Ross DC, King GJW. Endocopic versus open carpel tunnel release: a randomised trial. J Hand Surg [46] MacVicar J, Dobbie G, Owen-Johnstone L, Jagger C, Hopkins M, Kennedy J. Simulated home delivery in hospital: a randomised controlled trial. Br J Obstet Gynaecol 1993;100:316 – 23.
[47] Moss AJ, Zareba W, Hall WJ, et al. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med 2002;346:877 – 83.
[48] Mundinger MO, Kane RL, Lenz ER, et al. Primary care outcomes in patients treated by nurse practitioners or physicians: a randomized trial.
[49] Nichol KL, Mendelman PM, Mallon KP, et al. Effectiveness of live, attenuated intranasal influenza virus vaccine in healthy, working adults: a randomized controlled trial. JAMA 1999;282:137 – 44.
[50] O’Connor TP, Anderson AM, Lennox B, Muldoon C. A novel sustained-release formulation of ibuprofen provides effective once-daily therapy in the treatment of rheumatoid arthritis and osteoarthritis. Br J Clin Pract 1993;4:10 – 3.
[51] Ondo WG, Levy JK, Dat Vuong K, Hunter C, Jankovic J. Olanzapine treatment for dopaminergic-induced hallucinations. Mov Disord [52] Ott E, Nussmeier NA, Duke PC, et al. Efficacy and safety of the cyclooxygenase 2 inhibitors parecoxib and valdecoxib in patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg 2003;125:1481 – 92.
[53] Patat A, Surjus A, Le Go A, Granier J. Safety and tolerance of single oral dose of trandolapril (RU 44.570), a new angiotensin converting enzyme inhibitor. Eur J Clin Pharmacol 1989;36:17 – 23.
[54] Piatti PM, Monti F, Fermo I, et al. Hypocaloric high-protein diet improves glucose oxidation and spares lean body mass: comparison to hypocaloric high-carbohydrate diet. Metabolism 1994;43:1481 – 7.
[55] Pinto I, Chimeno P, Romo A, et al. Uterine fibroids: uterine artery embolization versus abdominal hysterectomy for treatment — a prospective randomized and controlled clinical trial. Radiology 2003;226:425 – 31.
[56] Rizzoli R, Greenspan SL, Bone G 3rd, et al. Two-year results of once-weekly administration of alendronate 70 mg for the treatment of postmenopausal osteoporosis. J Bone Miner Res 2002;17:1988 – 96.
[57] Roberts PJ, Roberts C, Sibbald B, Torgerson DJ. The effect of direct payment or a lottery on questionnaire response rates: a randomised controlled trial. J Epidemiol Community Health 2000;54:71 – 2.
[58] Rodgers H, Atkinson C, Bond S, Sudde M, Dobson R, Curless R. Randomized controlled trial of a comprehensive stroke education program for patients and care givers. Stroke 1999;30:2585 – 91.
[59] Rudolf Jon behaf of HES in Acute Stroke Study Group. Hydroxyethyl starch for hypervolemic hemodilution in patients with acute ischemic stroke: a randomized, placebo-controlled phase II safety study. Cerebrovasc Dis 2002;14:33 – 41.
[60] Saag MS, Powderly WG, Cloud GA, et al. Comparison of amphotericin B with fluconazole in the treatment of acute AIDS-associated cryptococcal meningitis. The NIAID Mycoses Study Group and the AIDS Clinical Trials Group. N Engl J Med 1992;326:83 – 9.
[61] Sarosdy MF, Schellhammer PF, Sharifi R, et al. Comparison of goserelin and leuprolide in combined androgen blockade therapy. Urology [62] Saunders M, Dische S, Barrett A, et al., on behalf of the CHART Steering Committee. Continuous hyperfractionated accelerated radiotherapy (CHART) versus conventional radiotherapy in non-small-cell lung cancer: a randomised multi-centred trial. Lancet 1997;350:161 – 5.
J.C. Dumville et al. / Contemporary Clinical Trials 27 (2006) 1–12 [63] Soto J, Fuya P, Herrera R, Berman J. Topical paromomycin/methylbenzethonium chloride plus parenteral meglumine antimonate as treatment for American cutaneous leishmaniasis: controlled study. Comment. Clin Infect Dis 1998;26:56 – 8.
[64] Straus WL, Qazi SA, Kundi Z, Nomani NK, Schwartz B. Antimicrobial resistance and clinical effectiveness of co-trimoxazole versus amoxycillin for pneumonia among children in Pakistan: randomised controlled trial. Pakistan Co-trimoxazole Study Group. Lancet 1998;352:270 – 4.
[65] Suckfu¨ll M., for the Hearing Loss Study Group. Firinogen and LDL apherresis in treatment of sudden hearing loss: a randomised multi-centred [66] Sundar S, Jha TK, Thakur CP, et al. Oral miltefosine for Indian visceral leishmaniasis. N Engl J Med 2002;347:1739 – 46.
[67] Turkington D, Kingdon D, Turner T. Effectiveness of a brief-cognitive-behavioural therapy intervention in the treatment of schizophrenia. Br J [68] Turnbull DA, Wilkinson C, Gerard K, et al. Clinical, psychological, and economic effects of antenatal care for three medical complications of pregnancy: a randomised controlled trial of 395 women. Lancet 2004;363:1104 – 8.
[69] UK Prospective Diabetes Study Group. Cost effectiveness analysis of improved blood pressure control in hypertensive patients with type 2 diabetes: UKPDS 40. BMJ 1998;317:720 – 6.
[70] van Agthoven M, Vellenga E, Fibbe WE, Kingma T, Uyl-de Groot CA. Cost analysis and quality of life assessment comparing patients undergoing autologous peripheral blood stem cell transplantation or autologous bone marrow transplantation for refractory or relapsed non-Hodgkin’s lymphoma or Hodgkin’s disease: a prospective randomised trial. Eur J Cancer 2001;37:1781 – 9.
[71] Zhu SH, Anderson CM, Tedeschi GJ, et al. Evidence of real-world effectiveness of a telephone quitline for smokers. N Engl J Med 2002;

Source: http://www.jeremymiles.co.uk/mestuff/publications/p39.pdf