Randomised controlled trials for policy interventions: a review of reviews and meta-regression.

OBJECTIVES: To determine whether randomised controlled trials (RCTs) lead to the same effect size and variance as non-randomised studies (NRSs) of similar policy interventions, and whether these findings can be explained by other factors associated with the interventions or their evaluation. DATA SOURCES: Two RCTs were resampled to compare randomised and non-randomised arms. Comparable field trials were identified from a series of health promotion systematic reviews and a systematic review of transition for youths with disabilities. Previous methodological studies were sought from 14 electronic bibliographic databases (Applied Social Sciences Index and Abstracts, Australian Education Index, British Education Index, CareData, Dissertation Abstracts, EconLIT, Educational Resources Information Centre, International Bibliography of the Sociological Sciences, ISI Proceedings: Social Sciences and Humanities, PAIS International, PsycINFO, SIGLE, Social Science Citation Index, Sociological Abstracts) in June and July 2004. These were supplemented by citation searching for key authors, contacting review authors and searching key internet sites. REVIEW METHODS: Analyses of previous resampling studies, replication studies, comparable field studies and meta-epidemiology investigated the relationship between randomisation and effect size of policy interventions. New resampling studies and new analyses of comparable field studies and meta-epidemiology were strengthened by testing pre-specified associations supported by carefully argued hypotheses. RESULTS: Resampling studies offer no evidence that the absence of randomisation directly influences the effect size of policy interventions in a systematic way. Prior methodological reviews and meta-analyses of existing reviews comparing effects from RCTs and non-randomised controlled trials (nRCTs) suggested that effect sizes from RCTs and nRCTs may indeed differ in some circumstances and that these differences may well be associated with factors confounded with design. No consistent explanations were found for randomisation being associated with changes in effect sizes of policy interventions in field trials. CONCLUSIONS: From the resampling studies we have no evidence that the absence of randomisation directly influences the effect size of policy interventions in a systematic way. At the level of individual studies, non-randomised trials may lead to different effect sizes, but this is unpredictable. Many of the examples reviewed and the new analyses in the current study reveal that randomisation is indeed associated with changes in effect sizes of policy interventions in field trials. Despite extensive analysis, we have identified no consistent explanations for these differences. Researchers mounting new evaluations need to avoid, wherever possible, allocation bias. New policy evaluations should adopt randomised designs wherever possible.

Spinal fixation surgery for acute traumatic spinal cord injury.

BACKGROUND: If the spine is unstable following traumatic spinal cord injury (SCI), surgical fusion and bracing may be necessary to obtain vertical stability and prevent re-injury of the spinal cord from repeated movement of the unstable bony elements. It has been suggested that this spinal fixation surgery may promote early rehabilitation and mobilisation. OBJECTIVES: To answer the question: is there a difference in functional outcome and other commonly measured outcomes between people who have a spinal cord injury and have had spinal fixation surgery and those who have not? SEARCH STRATEGY: The following databases were searched: AMED, CCTR, CINAHL, DARE, EMBASE, HEED, HMIC, MEDLINE, NRR, NHS EED. Searches were updated in May 2003 and MEDLINE was searched again in May 2007. The reference lists of retrieved articles were checked. SELECTION CRITERIA: Randomised controlled trials and controlled trials that compared surgical spinal fixation, with or without decompression, to any other treatment, in patients with a traumatic SCI. DATA COLLECTION AND ANALYSIS: Two reviewers independently selected studies. One reviewer assessed the quality of the studies and extracted data. MAIN RESULTS: No randomised controlled trials or controlled trials were identified that compared surgical spinal fixation surgery to other treatments in patients with a traumatic SCI. All of the studies identified were retrospective observational studies and of poor quality. AUTHORS' CONCLUSIONS: The current evidence does not enable conclusions to be drawn about the benefits or harms of spinal fixation surgery in patients with traumatic SCI. Well-designed, prospective experimental studies with appropriately matched controls are needed.

Risk factors for chronic fatigue syndrome/myalgic encephalomyelitis: a systematic scoping review of multiple predictor studies.

BACKGROUND: The aetiology of chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) is still unknown. The identification of risk factors for CFS/ME is of great importance to practitioners. METHOD: A systematic scoping review was conducted to locate studies that analysed risk factors for CFS/ME using multiple predictors. We searched for published and unpublished literature in 11 electronic databases, reference lists of retrieved articles and guideline stakeholder submissions in conjunction with the development of a forthcoming national UK guideline. Risk factors and findings were extracted in a concise tabular overview and studies synthesized narratively. RESULTS: Eleven studies were identified that met inclusion criteria: two case-control studies, four cohort studies, three studies combining a cohort with a case-control study design, one case-control and twin study and one cross-sectional survey. The studies looked at a variety of demographic, medical, psychological, social and environmental factors to predict the development of CFS/ME. The existing body of evidence is characterized by factors that were analysed in several studies but without replication of a significant association in more than two studies, and by studies demonstrating significant associations of specific factors that were not assessed in other studies. None of the identified factors appear suitable for the timely identification of patients at risk of developing CFS/ME within clinical practice. CONCLUSIONS: Various potential risk factors for the development of CFS/ME have been assessed but definitive evidence that appears meaningful for clinicians is lacking.

Systematic review of effectiveness of different treatments for childhood retinoblastoma.

OBJECTIVES: To assess the clinical effectiveness of treatments for childhood retinoblastoma. DATA SOURCES: Electronic databases were searched from inception to April 2004. REVIEW METHODS: Studies of participants diagnosed with childhood retinoblastoma, any interventions and all clinical outcomes were eligible for inclusion. Randomised and non-randomised controlled trials and cohort studies with clear comparisons between treatment groups were included. Methodological quality was assessed. A narrative synthesis was conducted. Where possible, studies assessing common interventions were grouped together, with prospective and retrospective studies grouped separately. Emphasis was placed on prospective studies. RESULTS: Thirty-one individual studies, from 42 publications, were included in the review. Apart from one non-randomised controlled trial, only comparative studies of observational design were available for any of the treatments. Four of the included studies were prospective and the remaining 27 were retrospective. Most of the studies were of radiotherapy or chemotherapy, with few studies available on enucleation or focal treatments such as brachytherapy, photocoagulation, cryotherapy and thermotherapy. The methodological quality was generally poor, with a high risk of bias in all included studies. The main problems were in relation to how treatment was allocated and lack of consideration of potentially confounding factors, such as initial disease severity, in the study design and data analysis. The evidence base for effectiveness of treatments for childhood retinoblastoma is extremely limited. Owing to the considerable limitations of the evidence identified, it was not possible to make meaningful and robust conclusions about the relative effectiveness of different treatment approaches for childhood retinoblastoma. CONCLUSIONS: In the authors' opinion, the evidence base for the effectiveness of treatments for childhood retinoblastoma is not sufficiently robust to provide clear guidance for clinical practice. Ideally, good-quality randomised controlled trials (RCTs) assessing the effectiveness of different treatment options for childhood retinoblastoma are required. Research is required on all the treatments currently used for this condition. Where RCTs are not feasible, for ethical or practical reasons, only high-quality, prospective, non-randomised studies should be given consideration, owing to the generally higher risk of bias in retrospective studies. To reduce the risk of confounding due to allocation by clinical indication, studies should compare patients with similar disease severity rather than compare patients of mixed disease severities. Standardised outcomes should be agreed for use in studies assessing the effectiveness of treatment. These outcomes should encompass potential important adverse effects of treatment such as loss of visual acuity and cosmetic outcome, as well as beneficial effects.

Sertindole for schizophrenia.

BACKGROUND: Sertindole is an atypical antipsychotic, which is thought to give a lower incidence of extrapyramidal side effects at clinically effective doses than typical antipsychotic drugs. In December 1998, Lundbeck Ltd., the manufacturers of sertindole, voluntarily suspended the availability of the drug due to concerns about cardiac arrhythmia and sudden cardiac death associated with its use. However, based on the advice of an appointed expert group, the Committee for Proprietary Medicinal Products (CPMP) lifted the suspension of sertindole in October 2001, a decision that was ratified by the European Commission on the 26th of June 2002. Lundbeck have committed to the CPMP to carry out two post-marketing surveillance (PMS) studies (which were initiated in July 2002) to provide additional epidemiological data under conditions of normal drug usage. Initial marketing of the product will be restricted and Lundbeck is currently in discussions with the US health authorities (FDA) to investigate whether, and if so when, it would be possible to launch Serdolect in the US market. OBJECTIVES: To determine the effects of sertindole compared with placebo, typical and other atypical antipsychotic drugs for schizophrenia and related psychoses. SEARCH STRATEGY: Our Initial searches included electronic searches of Biological Abstracts (1980-1999), The Cochrane Library (Issue 1, 1999), The Cochrane Schizophrenia Group's Register (August 2000), EMBASE (1980-1999), LILACS (1982-1996), MEDLINE (1966-1999), PSYNDEX (1977-1995) and PsycLIT (1974-1999). In addition, we searched pharmaceutical databases on the Dialog Corporation Datastar and Dialog services. We searched references of all identified studies for further trials. We contacted the manufacturer of sertindole and authors of trials. We updated the literature search by searching the Cochrane Schizophrenia Group's Trials Register in April 2003. SELECTION CRITERIA: All randomised controlled trials that compared sertindole to placebo or other antipsychotic (atypical or typical) drug treatments for patients with schizophrenia or related psychosis . DATA COLLECTION AND ANALYSIS: We independently inspected citations and, where possible abstracts; ordered papers for re-inspection and quality assessment and independently extracted data. For homogeneous dichotomous data, we calculated the risk ratio (RR), 95% confidence interval (CI) and, where appropriate, the number needed to treat (NNT) or number needed to harm (NNH) on an intention-to-treat basis. For continuous data, we calculated weighted mean differences (WMD). We inspected all data for heterogeneity. MAIN RESULTS: Currently the review includes three studies with a total of 1,104 participants. One was a medium term (eight weeks) placebo controlled study that examined three different doses of sertindole (8, 12 and 20mg/day). The remaining two studies compared the use of sertindole with haloperidol (10mg/day). One was a short term study (six weeks) that looked at four different doses of sertindole (8, 16, 20, 24mg/day) and the other was a long term study (one year) that evaluated the use of sertindole 24mg/day in participants attending outpatients. We excluded two large important studies because they did not report any usable data. (Both had greater than 50% loss to follow-up and data on 'leaving the study early' was inadequately reported). SERTINDOLE VERSUS PLACEBO: Sertindole at 20mg/day was found to be more effective than placebo in terms of BPRS total scores (1 study, n=78, MD 6.2, CI -11.8 to -0.6) and CGI total end point scores (1 study, n=78, MD -0.9, CI -1.6 to -0.2). A marginally statistically significantly greater number of participants that were treated with 20 mg of sertindole were reported to have been 'very much improved' as compared to those taking placebo (1 study, n=102, RR 7.6, CI 1.0 to 57.9, NNT 7.9, CI 4.3 to 41.1). There was no statistically significant difference between sertindole at 8 or 12 mg/day and placebo for these three outcome measures. There were no statistically significant differences between sertindole (8, 12 or 20 mg) and placebo for the incidence of extrapyramidal symptoms, extrapyramidal related events or use of medication to avoid extrapyramidal symptoms. There were no statistically significant differences found between sertindole and placebo for the movement disorders akathisia, cogwheel rigidity, hypertonia and tremor or somnolence. At eight weeks a statistically significant difference between placebo and all sertindole groups (8, 12 and 20 mg) for mean change from baseline in the QT and QTc intervals were observed (p values and SD were not reported). There was a statistically significant greater mean weight gain among participants taking sertindole (20 mg, mean weight gain of 3.3 kg) as compared to placebo (mean weight gain of 0.8 kg; p<0.05). SERTINDOLE VERSUS HALOPERIDOL: At one year, a greater number of participants who were treated with haloperidol as compared to sertindole (24mg/day) were leaving the study early due to any reason (1 study, n=282, RR 0.6, CI 0.4 to 1.0, NNH 8.8, CI 4.7 to 74.0) or non-compliance (1 study, n=282, RR 0.2, CI 0.0 to 0.7, NNH 12.8, CI 7.7 to 37.8). However, at six weeks, there was no statistically significant difference between sertindole (at 8, 16, 20, or 24mg) and haloperidol for this latter outcome. The incidence of EPS was higher among those treated with haloperidol than sertindole at 8, 16, 20 or 24mg/day (8mg: 1 study, n=245, RR 0.1, CI 0.0 to 0.7, NNH 11.4, CI 7.1 to 29.8; 16mg: 1 study, n=252, RR 0.3, CI 0.1 to 1.0, NNH 15.5, CI 8.0 to 217.9; and 20mg: 1 study, n=253, RR 0.2, CI 0.1 to 0.8, NNH 13.7, CI 7.7 to 68.3; 24mg: 2 studies, n=524, RR 0.6, CI 0.4 to 0.8, NNH 8.7, CI 5.4 to 23.0). More participants treated with haloperidol experienced akathisia, tremor and hypertonia than those treated with sertindole (Akathisia - 8mg: 1 study, n=245, RR 0.2, CI 0.1 to 0.5, NNH 6.0, CI 4.1 to 11.2; 16mg: 1 study, n=252, RR 0.1, CI 0.0 to 0.3, NNH 5.4, CI 3.9 9.0; 20mg: 1 study, n=253, RR 0.3, CI 0.2 to 0.7, NNH 7.3, CI 4.6 to 17.9; 24mg: 2 studies, n=524, RR 0.5, CI 0.3 to 0.7, NNH 8.6, CI 5.6 to 18.3. Tremor - 8mg: 1 study, n=245, RR 0.3, CI 0.1 to 0.7, NNH 8.5, CI 5.2 to 24.0; 16mg: 1 study, n=252, RR 0.2, CI 0.1 to 0.5, NNH 7.3, 4.8 to 15.6; 20mg: 1 study, n=253, RR 0.2, CI 0.1 to 0.6, NNH 7.8, CI 4.9 to 18.1; 24mg: 2 studies, n=524, RR 0.4, CI 0.2 to 0.6, NNH 8.2, CI 5.6 to 15.3. For Hypertonic - 24mg: 2 studies, n=524, RR 0.5, CI 0.3 to 0.8, NNH 12.4, CI 7.5 to 35.0; for sertindole 8, 16 and 20mg there was no statistically significant differences between the treatment groups). One study reported that at six weeks, there was a statistically significant greater increase from baseline to final value in mean QTc interval in the sertindole 16, 20 and 24mg groups (20, 26, and 24msec, respectively) than in the haloperidol group (0msec; p value was not reported), but no SD or any other measure of variance for the effect sizes were reported. For one long term study only one participant from the sertindole group (24mg) had a QT interval that exceeded 500msec (1 study, n=282, RR 3.0 CI 0.1 to 73.0), but 11participants treated with Sertindole had QTc intervals of at least 500msec, compared to none in the haloperidol treated group (1 study, n=282, RR 23.0, CI 1.4 to 386.6, NNH 12.8, CI 8.2 to 29.6). At six weeks, fewer participants treated with sertindole at 8mg or 24mg were affected by somnolence than those treated with haloperidol (sertindole 8mg: 1 study, n=245, RR 0.1, CI 0.0 to 0.7, NNH 11.4, CI 7.1 to 29.8; 24mg: 2 studies, n=524, RR 0.6, CI 0.4 to 1.0, NNH 14.8, CI 7.7 to 205.2). The incidence of rhinitis was found to be statistically significantly higher among those taking sertindole at 16 or 24mg as compared to haloperidol (16mg: 1 study, n=252, RR 10.8, CI 1.4 to 82.6, NNH 12.7, CI 7.7 to 36.7; 24mg: 2 studies, n= 524, RR 2.1, CI 1.4 to 3.1, NNH 8.7, CI 5.6 to 18.6). At one year, 33 participants treated with sertindole (24mg) had experienced the sexual adverse event of decreased ejaculatory volume, compared with six participants treated with haloperidol. However the number of included male participants was not reported and therefore the RR could not be calculated. At one year, more participants taking sertindole (24mg/day) had put on weight compared to those taking haloperidol (1 study, n=282, RR 6.3, CI 1.9 to 20.9, NNH 8.8, CI 5.7 to 19.1). At six weeks, all of the sertindole groups showed an increase in body weight from baseline to final evaluation ranging from 1.3kg to 1.9kg, all of which represented a statistically significantly different weight change than that recorded for the haloperidol treatment group (-0.1Kg). However, the actual weight gain for each sertindole dosage group was not reported and no SD or any other measure of variance was given. AUTHORS' CONCLUSIONS: Sertindole at a dose of 20mg/day was found to be more antipsychotic than placebo. When used at 8, 12 or 20mg/day it appears to be as acceptable as placebo (in terms of various adverse events including movement disorders and somnolence), but seems to be associated with more cardiac problems (8, 12 or 20mg/day) and an increase in weight gain (20mg/day) than placebo. Sertindole at a dose of 24mg/day was better tolerated than haloperidol (in terms of participants leaving the study early). It was also found to be was associated with fewer movement disorders (at 8, 16, 20 or 24mg/day) and sedation (8 or 24mg/day) than haloperidol. However, it was shown to cause more cardiac anomalies (16, 20 or 24mg/day), weight gain (all doses combined), rhinitis (16 or 24mg/day), and problems with sexual functioning (24mg/day) than haloperidol. One short term study reported that sertindole 16mg/day was the most optimal dose.

A rapid and systematic review and economic evaluation of the clinical and cost-effectiveness of newer drugs for treatment of mania associated with bipolar affective disorder.

OBJECTIVES: To evaluate the clinical and cost-effectiveness of quetiapine, olanzapine and valproate semisodium in the treatment of mania associated with bipolar disorder. DATA SOURCES: Electronic databases; industry submissions made to the National Institute for Clinical Excellence. REVIEW METHODS: Randomised trials and economic evaluations that evaluated the effectiveness of quetiapine, olanzapine or valproate semisodium in the treatment of mania associated with bipolar disorder were selected for inclusion. Data were extracted by one reviewer into a Microsoft Access database and checked for quality and accuracy by a second. The quality of the cost-effectiveness studies was assessed using a checklist updated from that developed by Drummond and colleagues. Relative risk and mean difference data were presented as Forest plots but only pooled where this made sense clinically and statistically. Studies were grouped by drug and, within each drug, by comparator used. Chi-squared tests of heterogeneity were performed for the outcomes if pooling was indicated. A probabilistic model was developed to estimate costs from the perspective of the NHS, and health outcomes in terms of response rate, based on an improvement of at least 50% in a patient's baseline manic symptoms derived from an interview-based mania assessment scale. The model evaluated the cost-effectiveness of the alternative drugs when used as part of treatment for the acute manic episode only. RESULTS: Eighteen randomised trials met the inclusion criteria. Aspects of three of the quetiapine studies were commercial-in-confidence. The quality of the included trials was limited and overall, key methodological criteria were not met in most trials. Quetiapine, olanzapine and valproate semisodium appear superior to placebo in reducing manic symptoms, but may cause side-effects. There appears to be little difference between these treatments and lithium in terms of effectiveness, but quetiapine is associated with somnolence and weight gain, whereas lithium is associated with tremor. Olanzapine as adjunct therapy to mood stabilisers may be more effective than placebo in reducing mania and improving global health, but it is associated with more dry mouth, somnolence, weight gain, increased appetite, tremor and speech disorder. There was little difference between these treatments and haloperidol in reducing mania, but haloperidol was associated with more extrapyramidal side-effects and negative implications for health-related quality of life. Intramuscular olanzapine and lorazepam were equally effective and safe in one very short (24 hour) trial. Valproate semisodium and carbamazepine were equally effective and safe in one small trial in children. Olanzapine may be more effective than valproate semisodium in reducing mania, but was associated with more dry mouth, increased appetite, oedema, somnolence, speech disorder, Parkinson-like symptoms and weight gain. Valproate semisodium was associated with more nausea than olanzapine. The results from the base-case analysis demonstrate that choice of optimal strategy is dependent on the maximum that the health service is prepared to pay per additional responder. For a figure of less than 7179 British pounds per additional responder, haloperidol is the optimal decision; for a spend in excess of this, it would be olanzapine. Under the most favourable scenario in relation to the costs of responders and non-responders beyond the 3-week period considered in the base-case analysis, the incremental cost-effectiveness ratio of olanzapine is reduced to 1236 British pounds. CONCLUSIONS: In comparison with placebo, quetiapine, olanzapine and valproate semisodium appear superior in reducing manic symptoms, but all drugs are associated with adverse events. In comparison with lithium, no significant differences were found between the three drugs in terms of effectiveness, and all were associated with adverse events. Several limitations of the cost-effectiveness analysis exist, which inevitably means that the results should be treated with some caution. There remains a need for well-conducted, randomised, double-blind head-to-head comparisons of drugs used in the treatment of mania associated with bipolar disorder and their cost-effectiveness. Participant demographic, diagnostic characteristics, the treatment of mania in children, the use of adjunctive therapy and long-term safety issues in the elderly population, and acute and long-term treatment are also subjects for further study.

Effectiveness and cost-effectiveness of acute hospital-based spinal cord injuries services: systematic review.

OBJECTIVES: To examine four key areas: (1) the effectiveness and cost-effectiveness of spinal fixation surgery, (2) the consequences of immediate versus delayed referral to a spinal injuries unit (SIU), (3) the number of people with a new spinal cord injury (SCI) who are discharged from hospital without ever being transferred to an SIU, and (4) the effectiveness and cost-effectiveness of steroids for people with SCI. DATA SOURCES: Searches were carried out on several databases and also on the Internet. Specialist SCI and spinal injury related websites were searched, specifically the Spinal Injuries Association, the British Association of Spinal Cord Injury Specialists and the National Spinal Cord Injury Association. REVIEW METHODS: Three separate search strategies were devised to find studies relating to the four key areas. Two reviewers independently screened all study citations for inclusion. The lists of all retrieved studies were scanned for additional studies. Quality of studies was assessed and data were extracted by one reviewer then checked by the second. Data from included studies were summarised within each key area. For dichotomous data, relative risks were calculated with 95% confidence intervals. Pooled relative risks were calculated as appropriate. For continuous data, mean differences with 95% confidence intervals were calculated and, if data were pooled, weighted mean differences were calculated. Searches were carried out to identify economic evaluations, details of these together with a critical appraisal of quality are presented in structured tables. Quality was assessed using a checklist supplemented with additional comments on the adequacy of methodology where appropriate. RESULTS: For spinal fixation versus no fixation, 68 retrospective observational studies were found that suggested some benefits of fixation surgery. Only four studies were found on fixation surgery in SIUs compared with non-SIU hospitals and no significant differences were seen. All 28 studies concerning delayed referral to a SIU were retrospective observational studies. In most, study details were poorly reported and there was doubt over the comparability of groups at baseline and on confounding factors. Times of referral and transfer were not reported separately. Evidence suggested an effect in favour of the SIU group for neurological improvement. No relevant published studies of any design were found regarding how many people with a new SCI are discharged from hospital without ever being transferred to an SIU. Two systematic reviews were found that assessed the effectiveness of steroids. No studies were identified that considered both costs and the impact on patient outcomes of a given intervention. CONCLUSIONS: Although there was evidence to suggest some benefits of fixation surgery and also a benefit of immediate referral to SIUs compared with delayed or no referral, owing to the limitations of the data these should be interpreted with caution. Not enough data were found to assess whether surgery is more beneficial when carried out in SIUs and further research is required in this area. Well-designed prospective observational studies with appropriately matched controls are needed. High-dose methylprednisolone steroid therapy may be effective in promoting some degree of neurological recovery if given within 8 hours of injury. There is a need for more randomised controlled trials (RCTs) of pharmacological therapy for acute SCI. No published studies of any design were found to answer the question of how many people with acute SCI are discharged from hospital without ever being transferred to an SIU. Primary research involving audit of selected hospital records should be commissioned and published. The search strategy did not identify any full economic evaluations. Future research should include full economic evaluations, possibly alongside a large RCT, which fully consider the costs and consequences of implementing interventions.

A rapid and systematic review of the clinical effectiveness and cost-effectiveness of topotecan for ovarian cancer.

BACKGROUND: Ovarian cancer is the most common gynaecological cancer with an annual incidence of 21.6 per 100,000 in England and Wales. Due to the often asymptomatic nature of the early stages of the disease, most cases are not detected until the advanced stages. Consequently, the prognosis after diagnosis is poor and the 5-year survival rate in the UK is only about 30%. Current recommendations suggest that first-line chemotherapy for ovarian cancer should involve paclitaxel and platinum (Pt)-based therapy (cisplatin/ carboplatin), however, most patients develop resistant or refractory disease and require second-line therapy. Patients may respond to re-challenge with Pt-agents if the treatment-free interval is > 6 months, but an alternative is often required. Topotecan is one of six drugs currently licensed in the UK for second-line therapy, and recent reviews suggest that it has modest efficacy in the treatment of advanced disease and performs favourably against paclitaxel. However, these reviews are based on a limited number of reports mainly consisting of non-randomised Phase I and II studies. OBJECTIVES OF THE REVIEW: To examine the clinical effectiveness and cost-effectiveness of oral and intravenous topotecan (Hycamtin, SmithKline Beecham, UK) for the treatment of all stages of ovarian cancer. SEARCH STRATEGY: Sixteen electronic databases from inception to September 2000 and Internet resources were searched, in addition to the bibliographies of retrieved articles and submissions from pharmaceutical companies. INCLUSION AND EXCLUSION CRITERIA: Two reviewers independently screened all titles/abstracts and included/excluded studies based on full copies of manuscripts. Any disagreements were resolved through discussion. Only randomised controlled trials (RCTs) and full economic evaluations comparing topotecan to non-topotecan regimens were included. All stages of therapy and disease were considered, and the outcomes included were survival, response, symptom relief, quality of life, adverse effects and costs. METHODS: DATA EXTRACTION STRATEGY: Data were extracted into an Access database by one reviewer and checked by a second. Any disagreements were resolved through discussion. METHODS: QUALITY ASSESSMENT STRATEGY: Two reviewers, using specified criteria, independently assessed the quality of the clinical effectiveness studies and the economic evaluations. Any disagreements were resolved through discussion. METHODS: ANALYSIS STRATEGY: Due to the limited number of studies included in the review and the fact that they compared topotecan with different comparators, the out-come data could not be pooled statistically. Clinical effectiveness data are discussed separately under the different outcome subheadings. For time-to-event data, hazard ratios with 95% confidence intervals are presented where available, and for the remaining outcomes, relative risks are reported or calculated where sufficient data were available. Relative risk data are also presented in the form of Forest plots without pooled estimates. Economic data are presented in the form of a summary and critique of the evidence, and a grading (A-I) assigned to each study indicating the direction and magnitude of the cost-effectiveness data. INCLUDED STUDIES: A total of 568 titles/abstracts were identified and screened for relevance. Full copies of 72 papers were assessed and seven published manuscripts reporting details of two studies of clinical effectiveness and one economic evaluation were included. Further details of the two clinical effectiveness studies and two new economic evaluations were identified from confidential company submissions. Overall, two international multicentre RCTs of effectiveness comparing topotecan with paclitaxel (trial 039) and topotecan with caelyx (trial 30-49) were included in the review. The three economic evaluations included in the review comprised one cost-minimisation analysis (CMA) comparing topotecan with caelyx, one cost-consequences analysis (CCA) comparing topotecan with paclitaxel, etoposide and altretamine and one cost-effectiveness analysis (CEA) comparing topotecan with paclitaxel. RESULTS: QUALITY OF CLINICAL EFFECTIVENESS DATA: Both clinical effectiveness studies (trial 30-49 and 039) were of reasonable quality, although it was unclear whether either performed valid intention-to-treat analyses. In addition, trial 30-49 failed to state whether the outcome assessors were blinded to treatment allocation. RESULTS --QUALITY OF ECONOMIC EVALUATIONS: The CCA (comparing topotecan with three comparators) was of poor quality and of little relevance to the UK NHS. The CMA and CEA were of reasonable quality overall and relevant to the UK NHS. However, both, in particular the CEA, suffered from methodological problems, and thus their findings should be interpreted with caution. RESULTS: ASSESSMENT OF CLINICAL EFFECTIVENESS: The assessment of clinical effectiveness was based on limited data. Only two trials with a total of 709 participants were identified. In general, with a few minor exceptions, there were no statistically significant differences between topotecan and paclitaxel, or topotecan and caelyx in survival, response rate, median time to response, median duration of response and quality of life. Significant differences that were reported were mainly identified in subgroup analyses (Pt-sensitive disease and disease without ascites) of questionable validity and their relevance to a general advanced ovarian cancer patient population undergoing second-line chemotherapy is unclear. However, statistically significant differences were observed in the incidence of adverse effects. Topotecan was associated with increased incidences of haematological toxicities (including neutropenia, leukopenia, anaemia and thrombocytopenia), alopecia, nausea and vomiting. Caelyx-treated patients suffered from significantly increased incidences of Palmar-Plantar erythrodysesthesia, stomatitis, mucous membrane disorders and skin rashes. Paclitaxel was associated with significant increases in alopecia, arthralgia, myalgia, neuropathy, paraesthesiae, skeletal pain and flushing. RESULTS: ASSESSMENT OF COST-EFFECTIVENESS: The assessment of cost-effectiveness was also based on limited data, with three evaluations identified, one of which was not relevant. The two remaining studies, comparing topotecan with paclitaxel (CEA) and topotecan with caelyx (CMA), both used effectiveness data from multicentre RCTs and based their costs on 1999/2000 UK sources. The evaluations were conducted from a UK NHS perspective and findings presented in GB pounds/Euros. Topotecan for the second-line treatment of advanced ovarian cancer was shown to be more cost-effective than paclitaxel (32,513 GB pounds versus 46,186 GB pounds per person in terms of any response (complete or partial), incremental cost-effectiveness = 3065 GB pounds) in all respects except cost per time without toxicity or symptoms, but less cost-effective than caelyx (14,023 GB pounds versus 9979 GB pounds per person regardless of whether the patient responded). However, direct comparisons of the cost findings between the two studies is difficult because they used different designs, different time horizons for the cost analyses and the findings were presented as costs per person for only patients who responded in one study (topotecan versus paclitaxel) and costs per person regardless of whether they responded in the other study (topotecan versus caelyx). CONCLUSIONS: This review indicates that there is little evidence in the form of RCTs on which to base an assessment of the effectiveness of topotecan as second-line therapy for advanced ovarian cancer. The evidence suggests there were no statistically significant differences overall between topotecan and paclitaxel, or topotecan and caelyx in clinical outcomes. However, statistically significant differences were observed in the incidence of adverse effects. The clinical significance of the findings is not discussed. Overall, the effects of topotecan could at best be described as modest, but the alternative agents offer no real advantages except fewer side-effects and possibly improved cost-effectiveness. Both of the clinical effectiveness studies on which this evidence is based had methodological flaws, the most serious being the lack of a blinded assessor in the topotecan versus caelyx trial, which is important for unbiased assessment of response outcomes. The economic evaluations also suffered from a number of potential problems. CONCLUSIONS: RECOMMENDATIONS FOR RESEARCH: Further good quality RCTs and CEAs are required comparing topotecan with other licensed and potentially useful (soon to be licensed) second-line treatments for ovarian cancer. At present, it is difficult to make any decisions about topotecan and other drugs for second-line therapy without good quality direct comparisons. In view of the ongoing studies identified, an update of the current review should be considered in approximately 18 months (Summer 2002) or possibly sooner if the recently commissioned National Institute for Clinical Excellence review of caelyx for ovarian cancer identifies additional data relevant to topotecan.

Interventions for the treatment and management of chronic fatigue syndrome: a systematic review.

CONTEXT: A variety of interventions have been used in the treatment and management of chronic fatigue syndrome (CFS). Currently, debate exists among health care professionals and patients about appropriate strategies for management. OBJECTIVE: To assess the effectiveness of all interventions that have been evaluated for use in the treatment or management of CFS in adults or children. DATA SOURCES: Nineteen specialist databases were searched from inception to either January or July 2000 for published or unpublished studies in any language. The search was updated through October 2000 using PubMed. Other sources included scanning citations, Internet searching, contacting experts, and online requests for articles. STUDY SELECTION: Controlled trials (randomized or nonrandomized) that evaluated interventions in patients diagnosed as having CFS according to any criteria were included. Study inclusion was assessed independently by 2 reviewers. Of 350 studies initially identified, 44 met inclusion criteria, including 36 randomized controlled trials and 8 controlled trials. DATA EXTRACTION: Data extraction was conducted by 1 reviewer and checked by a second. Validity assessment was carried out by 2 reviewers with disagreements resolved by consensus. A qualitative synthesis was carried out and studies were grouped according to type of intervention and outcomes assessed. DATA SYNTHESIS: The number of participants included in each trial ranged from 12 to 326, with a total of 2801 participants included in the 44 trials combined. Across the studies, 38 different outcomes were evaluated using about 130 different scales or types of measurement. Studies were grouped into 6 different categories. In the behavioral category, graded exercise therapy and cognitive behavioral therapy showed positive results and also scored highly on the validity assessment. In the immunological category, both immunoglobulin and hydrocortisone showed some limited effects but, overall, the evidence was inconclusive. There was insufficient evidence about effectiveness in the other 4 categories (pharmacological, supplements, complementary/alternative, and other interventions). CONCLUSIONS: Overall, the interventions demonstrated mixed results in terms of effectiveness. All conclusions about effectiveness should be considered together with the methodological inadequacies of the studies. Interventions which have shown promising results include cognitive behavioral therapy and graded exercise therapy. Further research into these and other treatments is required using standardized outcome measures.

Risperidone versus other atypical antipsychotic medication for schizophrenia.

BACKGROUND: Risperidone is one of a number of 'atypical antipsychotics' which are currently being marketed for the treatment of those with schizophrenia, largely on the basis of claims of improved tolerability and effectiveness compared to much cheaper conventional antipsychotics. The efficacy of risperidone has already been compared to conventional drugs, but it remains unclear how risperidone compares with other atypical antipsychotic drugs such as clozapine. OBJECTIVES: To determine the effects of risperidone compared with other atypical antipsychotic drugs for schizophrenia. SEARCH STRATEGY: Electronic searches of Biological Abstracts (1980-1999), The Cochrane Library (Issue 1, 2000), The Cochrane Schizophrenia Group's Register (January 1999), EMBASE (1980-1999), MEDLINE (1966-1999), LILACS (1982-1999), PSYNDEX (1977-1999) and PsycLIT (1974-1999) were undertaken. In addition, pharmaceutical databases on the Dialog Corporation Datastar and Dialog services were searched. References of all identified studies were searched for further trials. Pharmaceutical companies and authors of trials were contacted. SELECTION CRITERIA: All randomised controlled clinical trials that compared risperidone to other atypical antipsychotic treatments for schizophrenia and schizophrenia-like psychoses were included by independent assessment. DATA COLLECTION AND ANALYSIS: Citations and, where possible, abstracts were independently inspected by reviewers, papers ordered, re-inspected and quality assessed. Data were independently extracted. For homogeneous dichotomous data the risk ratio (RR), 95% confidence interval (CI) and, where appropriate, the number needed to treat (NNT) were calculated on an intention-to-treat basis. For continuous data, standardised and weighted mean differences were calculated (SMD, WMD). All data were inspected for heterogeneity. MAIN RESULTS: Nine studies were obtained, comparing risperidone with clozapine (five studies - largely amongst treatment resistant patients); olanzapine (three studies); and amisulpiride (one study). The research was beset by problems of high attrition rates and short term follow up. Clozapine does seem equally acceptable to risperidone in the short term (leaving the study early, n=466, RR 1. 00 CI 0.73-1.37). For most other outcomes wide confidence intervals were obtained, which meant that it was impossible to judge whether the two compounds were equally effective, or whether one was in fact superior to the other. Olanzapine and risperidone seem broadly similar according to numbers of patients responding to treatment (40% reduction in PANSS scores: n=339, RR 1.14, CI 0.99-1.32). Olanzapine caused fewer people to leave the study early (n=404, RR 1. 31 CI 1.06-1.60; NNT 8 CI 4-32) and fewer extrapyramidal side effects (n=339, RR 1.67 CI 1.14-2.46; NNH 8 CI 5-33), although comparative doses of risperidone were higher than those recommended in practice. In one single study (n=228) amisulpiride seemed broadly similar to risperidone in most respects. There were no useful data presented relating to service use and costs. Very few data relating to quality of life were presented. REVIEWER'S CONCLUSIONS: The equivalence of clozapine and risperidone for treatment resistant schizophrenia cannot yet be assumed and there seems to be little to chose between risperidone and both olanzapine and amisulpiride. The research is limited in many respects, and longer term studies measuring clinically important outcomes, including service use and quality of life are needed to judge the comparative value of the various atypical drugs.

Helicobacter pylori infection and mortality from ischaemic heart disease: negative result from a large, prospective study.

OBJECTIVE: To determine whether there is an independent association between Helicobacter pylori infection of the stomach and ischaemic heart disease. DESIGN: Prospective study with measurement of IgG antibody titres specific to H pylori on stored serum samples from 648 men who died from ischaemic heart disease and 1296 age matched controls who did not (nested case-control design). SUBJECTS: 21,520 professional men aged 35-64 who attended the British United Provident Association (BUPA) medical centre in London between 1975 and 1982 for routine medical examination. MAIN OUTCOME MEASURE: Death from ischaemic heart disease. RESULTS: The odds of death from ischaemic heart disease in men with H pylori infection relative to that in men without infection was 1.06 (95% confidence interval 0.86 to 1.31). In a separate group of 206 people attending the centre, plasma fibrinogen was virtually the same in those who were positive for H pylori (2.62 g/l) and those who were negative (2.64 g/l). CONCLUSIONS: A study that by its size and design minimised both random error and socioeconomic bias found no relation between H pylori infection and ischaemic heart disease. The validity of the study was shown by its confirmation of the recognised association between H pylori infection and stomach cancer (odds ratio 4.0 (1.9 to 8.2); P < 0.001). Eradication of H pylori infection may greatly reduce the incidence of stomach cancer, one of the most common causes of death from cancer worldwide, but it cannot be expected to have any effect in preventing ischaemic heart disease.