Systematic review

Objective: To assess the effect of “hospital in the home” (HITH) services that significantly substitute for inhospital time on mortality, readmission rates, patient and carer satisfaction, and costs.Data sources: MEDLINE, Embase, Social Sciences Citation Index, CINAHL, EconLit, PsycINFO and the Cochrane Database of Systematic Reviews, from the earliest date in each database to 1 February 2012.Study selection: Randomised controlled trials (RCTs) comparing HITH care with ...

Gideon A Caplan MB BS, MD, FRACP · Nur S Sulaiman MB BS · Dee A Mangin MB ChB · Nicoletta Aimonino Ricauda MD · Andrew D Wilson MD, FRCGP · Louise Barclay BN

Objective: To evaluate the evidence for and against the effectiveness of homeopathy.Data sources: The Cochrane Database of Systematic Reviews (generally considered to be the most reliable source of evidence) was searched in January 2010.Study selection: Cochrane reviews with the term “homeopathy” in the title, abstract or keywords were considered. Protocols of reviews were excluded. Six articles met the inclusion criteria.Data extraction: Each of the six reviews was examined for specific subject matter; number of clinical trials reviewed; total number of patients involved; and authors’ conclusions. The reviews covered the following conditions: cancer, attention-deficit hyperactivity disorder, asthma, dementia, influenza and induction of labour.Data synthesis: The findings of the reviews were discussed narratively (the reviews’ clinical and statistical heterogeneity precluded meta-analysis).Conclusions: The findings of currently available Cochrane reviews of studies of homeopathy do not show that homeopathic medicines have effects beyond placebo.

Edzard Ernst FMedSci, FRCP, FRCPEd

Objective: To evaluate the evidence for and against the effectiveness of reflexology for treating any medical condition.Data sources: Six electronic databases were searched from their inception to February 2009 to identify all relevant randomised controlled trials (RCTs). No language restrictions were applied.Study selection and data extraction: RCTs of reflexology delivered by trained reflexologists to patients with specific medical conditions. Condition studied, study design and controls, primary outcome measures, follow-up, and main results were extracted.Data synthesis: 18 RCTs met all the inclusion criteria. The studies examined a range of conditions: anovulation, asthma, back pain, dementia, diabetes, cancer, foot oedema in pregnancy, headache, irritable bowel syndrome, menopause, multiple sclerosis, the postoperative state and premenstrual syndrome. There were > 1 studies for asthma, the postoperative state, cancer palliation and multiple sclerosis. Five RCTs yielded positive results. Methodological quality was evaluated using the Jadad scale. The methodological quality was often poor, and sample sizes were generally low. Most higher-quality trials did not generate positive findings.Conclusion: The best evidence available to date does not demonstrate convincingly that reflexology is an effective treatment for any medical condition.

Edzard Ernst MD, PhD, FMedSci

Objective: To evaluate current evidence in support of therapies for preventing and treating cardiogenic shock (CS) after acute myocardial infarction that can be initiated in hospitals without invasive cardiac facilities.Study design: Systematic review.Data sources: MEDLINE and PubMed were searched from January 1985 to May 2008 using the MeSH terms “myocardial infarction”, “thrombolytic therapy”, “shock, cardiogenic”, “angioplasty, transluminal, percutaneous coronary”, “intra-aortic balloon pumping” and “platelet aggregation inhibitors”. Additional keyword and reference list searches were performed. Articles in English relating to adults were included.Study selection: Meta-analyses and comparative studies were included if they reported mortality or prevention of CS as an endpoint. In total, 35 articles were analysed (four meta-analyses, eight randomised controlled trials and 23 cohort studies).Data extraction: Studies were summarised by the first author and the level of evidence graded. Each study was checked by the second author and consensus was reached about inclusion and levels of evidence.Data synthesis: In the management and prevention of CS, the following are supported by high-level evidence: prehospital thrombolysis, transfer for emergency revascularisation (patients aged < 75 years) and thrombolysis for older patients (patients aged ≥ 75 years). In established CS, evidence supporting inhospital thrombolysis and intra-aortic balloon pump use in patients aged < 75 years and emergency revascularisation in older patients is limited to subgroup analyses and observational studies.Conclusions: In regional centres, prevention of CS is achieved with early fibrinolysis, preferably before hospital arrival. Patients of all ages should be considered for thrombolysis, early transfer for coronary revascularisation, and intra-aortic balloon pump insertion unless contraindicated. Glycoprotein inhibitors have no role in the management of CS in non-tertiary hospitals.

Enda O’Connor MRCPI, FJFICM · John F Fraser FRCA, FFARCSI, FJFICM

Objective: To establish the nature and efficacy of Australian school-based prevention and early intervention programs for anxiety and depression.Data sources: Cochrane, PsychInfo and PubMed databases, and the Primary Mental Health Care Australian Resource Centre database, were searched in June 2006. Additional materials were obtained from program websites, reference lists and authors.Study selection: Programs that were developed in Australia or trialled in Australia and addressed anxiety, depression, or resilience were included.Data synthesis: 24 efficacy or effectiveness trials of 9 intervention programs were identified. Most were based on cognitive behaviour therapy, interpersonal therapy or psychoeducation. Six were universal interventions, two were indicated programs and one was a treatment program. Most were associated with short-term improvements or symptom reduction at follow-up.Conclusions: A number of schools programs produce positive outcomes. However, even well established programs require further evaluation to establish readiness for broad dissemination as outlined in the standards of the Society for Prevention Research.

Alison L Neil BAppPsych(Hons) · Helen Christensen PhD

Objective: To assess the evidence of an association between periconceptional folic acid (FA) supplementation or fortification of foods with FA and the risk of twinning, using the Food Standards Australia New Zealand (FSANZ) framework for assessing evidence when substantiating nutrition, health and related claims on foods.Data sources: The Cochrane Library Database, MEDLINE, MEDLINE in Process, EMBASE, PubMed National Library of Medicine, and CINAHL were searched to identify systematic reviews and primary intervention and observational studies published from 1 July 1994 to 7 July 2006.Study selection: One prospective and five retrospective cohort studies that assessed the rate of twinning in populations exposed to FA through supplementation, and six retrospective registry-based cohort studies examining twinning rates after fortification of foods with FA.Data extraction: Two reviewers appraised eligible studies and evaluated data independently.Data synthesis: The best maximal risk estimates of twinning after FA supplementation were an adjusted odds ratio (adjOR) of 1.26 (95% CI, 0.91–1.73) for preconceptional supplementation and dizygotic twinning and an adjOR of 1.02 (95% CI, 0.85–1.24) for overall twinning. Data from four FA fortification studies in the United States that allowed for calculation of an annual percentage increase showed a maximal annual increase in twinning rates of 4.6%.Conclusions: Overall, under the FSANZ framework, there is possible evidence for a relationship between periconceptional FA intake and increased twinning. To support this tentative relationship, more well designed, long-term follow-up studies are needed in places where fortification with FA has been introduced, focusing on dose–response and obtaining accurate data on infertility treatments.

Evelyne E Muggli MPH · Jane L Halliday PhD

Objective: Exposure to non-steroidal anti-inflammatory drugs (NSAIDs) is associated with increased risk of serious gastrointestinal (GI) events compared with non-exposure. We investigated whether that risk is sustained over time.Data sources: Cochrane Controlled Trials Register (to 2002); MEDLINE, EMBASE, Derwent Drug File and Current Contents (1999–2002); manual searching of reviews (1999–2002).Study selection: From 479 search results reviewed and 221 articles retrieved, seven studies of patients exposed to prescription non-selective NSAIDs for more than 6 months and reporting time-dependent serious GI event rates were selected for quantitative data synthesis. These were stratified into two groups by study design.Data extraction: Incidence of GI events and number of patients at specific time points were extracted.Data synthesis: Meta-regression analyses were performed. Change in risk was evaluated by testing whether the slope of the regression line declined over time. Four randomised controlled trials (RCTs) provided evaluable data from five NSAID arms (aspirin, naproxen, two ibuprofen arms, and diclofenac). When the RCT data were combined, a small significant decline in annualised risk was seen: − 0.005% (95% CI, − 0.008% to − 0.001%) per month. Sensitivity analyses were conducted because there was disparity within the RCT data. The pooled estimate from three cohort studies showed no significant decline in annualised risk over periods up to 2 years: − 0.003% (95% CI, − 0.008% to 0.003%) per month.Conclusions: Small decreases in risk over time were observed; these were of negligible clinical importance. For patients who need long-term (> 6 months) treatment, precautionary measures should be considered to reduce the net probability of serious GI events over the anticipated treatment duration. The effect of intermittent versus regular daily therapy on long-term risk needs further investigation.

Delia Schaffer MMedSci, BSc · Timothy Florin FRACP, MB BS(Hons), MSc · Craig Eagle FRACP, FRCPA, MB BS · Ian Marschner PhD · Gurkirpal Singh MD · Mendel Grobler MBA, BPharm · Chris Fenn PhD, MB BS · Manjula Schou MSc · Kathleen M Curnow PhD

Objective: To determine whether the adoption of the CONSORT checklist is associated with improvement in the quality of reporting of randomised controlled trials (RCTs).Data sources: MEDLINE, EMBASE, Cochrane CENTRAL, and reference lists of included studies and of experts were searched to identify eligible studies published between 1996 and 2005.Study selection: Studies were eligible if they (a) compared CONSORT-adopting and non-adopting journals after the publication of CONSORT, (b) compared CONSORT adopters before and after publication of CONSORT, or (c) a combination of (a) and (b). Outcomes examined included reports for any of the 22 items on the CONSORT checklist or overall trial quality.Data synthesis: 1128 studies were retrieved, of which 248 were considered possibly relevant. Eight studies were included in the review. CONSORT adopters had significantly better reporting of the method of sequence generation (risk ratio [RR], 1.67; 95% CI, 1.19–2.33), allocation concealment (RR, 1.66; 95% CI, 1.37–2.00) and overall number of CONSORT items than non-adopters (standardised mean difference, 0.83; 95% CI, 0.46–1.19). CONSORT adoption had less effect on reporting of participant flow (RR, 1.14; 95% CI, 0.89–1.46) and blinding of participants (RR, 1.09; 95% CI, 0.84–1.43) or data analysts (RR, 5.44; 95% CI, 0.73–36.87). In studies examining CONSORT-adopting journals before and after the publication of CONSORT, description of the method of sequence generation (RR, 2.78; 95% CI, 1.78–4.33), participant flow (RR, 8.06; 95% CI, 4.10–15.83), and total CONSORT items (standardised mean difference, 3.67 items; 95% CI, 2.09–5.25) were improved after adoption of CONSORT by the journal.Conclusions: Journal adoption of CONSORT is associated with improved reporting of RCTs.

Amy C Plint MD · David Moher PhD · Andra Morrison BSc · Kenneth Schulz PhD, MBA · Douglas G Altman PhD · Catherine Hill MB BS, MSc · Isabelle Gaboury PhD(c)

Objectives: To compare the safety, effectiveness and cost-effectiveness of drug-eluting coronary stents used in Australia with bare-metal stents and determine whether the benefits are greater for high-risk subgroups.Data sources: MEDLINE, Pre-Medline, EMBASE, Current Contents, CINAHL and the Cochrane Library database were searched to identify eligible randomised controlled trials and systematic reviews published in English between January 1966 and June 2004.Study selection: Seven randomised controlled trials that assessed polymer-based paclitaxel- or sirolimus-eluting stents versus bare-metal stents in patients with coronary atherosclerosis and reported on stent thrombosis, mortality, myocardial infarction, coronary artery bypass grafting or target lesion revascularisation.Data extraction: Two independent reviewers appraised eligible studies and extracted data. Relative risks (RRs) were calculated for each outcome and pooled using the Mantel–Haenszel method.Data synthesis: Rates of stent thrombosis, mortality, myocardial infarction and bypass grafts did not differ by stent type. Drug-eluting stents (DESs) resulted in a 71%–80% lower risk of revascularisation at 12 months (RR 0.29 [95% CI, 0.20–0.43] for paclitaxel-eluting stents [n = 1593 patients]; RR 0.20 [95% CI, 0.13–0.29] for sirolimus-eluting stents [n = 1296 patients]). Similar benefits were seen in several high-risk subgroups of patients: those with diabetes, lesion length > 20 mm and target-vessel diameter ≤ 2.5 mm. The benefits of DESs in these high-risk groups over lower-risk groups were inconclusive because of low numbers. The cost per revascularisation avoided by using DESs was A$3750–$6100, with an estimated cost per quality-adjusted-life-year (QALY) gained of A$46 829–$76 467. In sensitivity analyses, estimates varied from DESs being cost-saving to costing an additional $314 385 per QALY gained.Conclusions: DESs are effective in reducing revascularisation. Estimates of cost-effectiveness are very sensitive to changes in estimates of their true effects in clinical practice, market price and the number of stents used per patient. Decisions to limit DESs to only patients at the highest risk of restenosis may improve their cost-effectiveness but will need to be reassessed when evidence is available to compare absolute benefits between patient groups.

Sarah J Lord MB BS, MS(Epi), FRACGP · Felicity Allen BVSc(Hons), MPH · Luke Marinovich BA(Hons) · David C Burgess BMed, FRACP · Kirsten Howard MAppSc(Biopharm), MPH, MHlthEcon · Richard King MB BS, FRACP · John J Atherton MB BS, PhD, FRACP

Objective: To estimate the risk of death from healthcare-associated (nosocomial) bacteraemia caused by methicillin-resistant Staphylococcus aureus (MRSA), and compare it with that of nosocomial bacteraemia caused by methicillin-sensitive S. aureus (MSSA), by meta-analysis of selected studies. Data sources: Medline, EMBASE, Current Contents and Cochrane Library were searched for the period January 1978 (or earliest date of the database, if later than 1978) to December 2000. Study selection: Studies which compared mortality of nosocomial MRSA and MSSA bateraemia. Data synthesis: Nine studies were analysed. All but one found an increased relative risk (RR) of death from MRSA bacteraemia, with RR ranging from 0.89 to 4.94. Meta-analysis showed that patients with MRSA bacteraemia have an RR of death, compared with patients with MSSA bacteraemia, of 2.21 (95% CI, 1.76–2.57) using the fixed-effect method, and 2.03 (95% CI, 1.55–2.65) using the random-effect method. Conclusion: MRSA bacteraemia is associated with a real increase in risk of death, further justifying ongoing MRSA surveillance and control in healthcare facilities.

Michael Whitby · Mary-Louise McLaws · Geoffrey Berry

Objective: To conduct a systematic review of the evidence for the effectiveness of a range of possible treatments for depression in older people.Data sources: Literature search using the PubMed, PsycInfo and Cochrane Library databases.Data synthesis: Treatments that have been suggested to be effective for depression were grouped under three categories: medical treatments, psychological treatments, and lifestyle changes/alternative treatments. We describe each treatment, review the studies of its effectiveness in people aged ≥ 60 years, and give a rating of the level of evidence.Conclusions: The treatments with the best evidence of effectiveness are antidepressants, electroconvulsive therapy, cognitive behaviour therapy, psychodynamic psychotherapy, reminiscence therapy, problem-solving therapy, bibliotherapy (for mild to moderate depression) and exercise. There is limited evidence to support the effectiveness of transcranial magnetic stimulation, dialectical behaviour therapy, interpersonal therapy, light therapy (for people in nursing homes or hospitals), St John’s wort and folate in reducing depressive symptoms.

Cathy J Frazer PhD · Helen Christensen PhD · Kathleen M Griffiths PhD

Objective: To investigate species-specific envenoming rates and spectrum of severity of funnel-web spider bites, and the efficacy and adverse effects of funnel-web spider antivenom.Data sources: Cases were identified from a prospective study of spider bite presenting to four major hospitals and three state poisons information centres (1999–2003); museum records of spider specimens since 1926; NSW Poisons Information Centre database; MEDLINE and EMBASE search; clinical toxinology textbooks; the media; and the manufacturer’s reports of antivenom use.Data extraction: Patient age and sex, geographical location, month, expert identification of the spider, clinical effects and management; envenoming was classified as severe, mild–moderate or minor/local effects.Data synthesis: 198 potential funnel-web spider bites were identified: 138 were definite (spider expertly identified to species or genus), and 77 produced severe envenoming. All species-identified severe cases were attributed to one of six species restricted to NSW and southern Queensland. Rates of severe envenoming were: Hadronyche cerberea (75%), H. formidabilis (63%), Atrax robustus (17%), Hadronyche sp. 14 (17%), H. infensa (14%) and H. versuta (11%). Antivenom was used in 75 patients, including 22 children (median dose, 3 ampoules; range, 1–17), with a complete response in 97% of expertly identified cases. Three adverse reactions were reported, all in adults: two early allergic reactions (one mild and one with severe systemic effects requiring adrenaline), and one case of serum sickness.Conclusions: Severe funnel-web spider envenoming is confined to NSW and southern Queensland; tree-dwelling funnel webs (H. cerberea and H. formidabilis) have the highest envenoming rates. Funnel-web spider antivenom appears effective and safe; severe allergic reactions are uncommon.

Geoffrey K Isbister BSc, MD, FACEM · Michael R Gray MSc, PhD · Corrine R Balit BPharm, MB BS · Robert J Raven BSc, PhD · Barrie J Stokes BSc, MMath · Kate Porges FACEM · Alan S Tankel BSc, FACEM · Elizabeth Turner MAIBiol · Julian White MD, FACTM · Malcolm McD Fisher MD, FANZCA, FJFICM

Objective: To conduct a systematic review of studies on clinically relevant digital mobile phone electromagnetic interference with medical equipment.Data sources: MEDLINE and SUMSEARCH were searched for the period 1966–2004. The Cochrane Library and Database of Abstracts of Reviews of Effects were also searched for systematic reviews.Study selection: Studies were eligible if published in a peer-reviewed journal in English, and if they included testing of digital mobile phones for clinically relevant interference with medical equipment used to monitor or treat patients, but not implantable medical devices.Data synthesis: As there was considerable heterogeneity in medical equipment studied and the conduct of testing, results were summarised rather than subjected to meta-analysis.Results: Clinically relevant electromagnetic interference (EMI) secondary to mobile phones potentially endangering patients occurred in 45 of 479 devices tested at 900 MHz and 14 of 457 devices tested at 1800 MHz. However, in the largest studies, the prevalence of clinically relevant EMI was low. Most clinically relevant EMI occurred when mobile phones were used within 1 m of medical equipment.Conclusions: Although testing was not standardised between studies and equipment tested was not identical, it is of concern that at least 4% of devices tested in any study were susceptible to clinically relevant EMI. All studies recommend some type of restriction of mobile phone use in hospitals, with use greater than 1 m from equipment and restrictions in clinical areas being the most common.

Nathan Lawrentschuk MB BS · Damien M Bolton FRACS, MD

Objective: To update evidence-based guidelines for the treatment of proximal femoral fractures published in the Journal in 1999.Data sources: Systematic literature search of MEDLINE, CINAHL and EMBASE from January 1996 to September 2001 and the Cochrane Database of Systematic Reviews (most recent issue searched — Issue 2, 2002).Study selection: Randomised controlled trials and meta-analyses of all aspects of acute-care hospital treatment and rehabilitation of proximal femoral fractures among subjects aged 50 years and over with proximal femoral fractures not associated with metastatic disease or multiple trauma.Data extraction: All studies were read independently by two reviewers. Reviewers recorded individual study results, and an assessment of study quality and treatment conclusions according to Cochrane Collaboration protocols. If necessary, a third review was performed to reach consensus.Results: 93 new studies were identified and 82 met our inclusion criteria. Recommendations for thromboprophylaxis, anaesthesia, surgical fixation of fractures and nutritional status have been altered to incorporate new evidence. Recommendations have been added regarding postoperative blood transfusion, the management of subtrochanteric fractures, and the type of surgical swabs which should be used.Conclusions: Although there have been few significant changes to the previous recommendations, updating the guidelines has required substantial effort. The common clinical problem of hip fracture should be treated according to the most up-to-date evidence to achieve the best possible outcomes and optimal utilisation of limited resources. Guideline updates also require resourcing.

Michael N Chilov MB BS, BOptom · Ian D Cameron MB BS, PhD, FAFRM(RACP) · Lyn M March MB BS, PhD, FRACP

Objectives: To conduct a systematic review of how short-stay observation units (SOUs) affect the efficiency of healthcare delivery and the quality of services provided.Data sources: MEDLINE, CINAHL, Best Evidence and The Cochrane Library were searched for the period 1 January 1960 to 31 July 2000.Study selection: Studies were eligible if published in English and rated at National Health and Medical ...

Sue Daly RN, MN, FRCNA · Donald A Campbell MD, FRACP · Peter A Cameron MB BS, FACEM

Objective: To systematically review randomised controlled trials (RCTs) of computer-generated medication reminders or feedback directed to healthcare providers or patients.Data sources: Extensive computerised and manual literature searches identified 76 English-language reports of RCTs reported before 1 January 2002. Searches were conducted between June 1998 and April 2002.Study selection: 26 papers making 29 comparisons (two papers reported on multiple interventions) of computer-supported medication management to a control group.Data extraction: The quality of the RCTs was systematically assessed and scored independently by two reviewers. Rates of compliance with (potential) reminders for the control and intervention groups were extracted.Data synthesis: Heterogeneity of studies prevented a meta-analysis. Where possible, rates were calculated using the intention-to-treat principle. The comparisons were grouped into five areas. Reminders to providers in outpatient settings: six of 12 comparisons demonstrated positive effects (relative rates [RRs: intervention rates/control rates], 1.0 to 42.0). Provider feedback in outpatient settings: five of seven comparisons showed improved clinician behaviour (RRs, 1.0 to 2.5). Combined reminders and feedback in outpatient settings: the single comparison found no improvement. Reminders to providers in inpatient settings: three of five comparisons showed improvements (RRs, 1.0 to 2.1). Patient-directed reminders: two of four comparisons showed improvements in patient compliance.Conclusion: Reminders are more effective than feedback in modifying physician behaviour related to medication management. Patient-directed reminders can improve medication adherence.

John W Bennett MB BS FRACGP · Paul P Glasziou MB BS PhD

Objective: To define the evidence for doses of inhaled corticosteroids in asthma and describe this in clinically meaningful, evidence-based terms.Data source: Cochrane Database of Systematic Reviews.Study selection and data extraction: We identified systematic reviews of randomised controlled trials of dosing of inhaled corticosteroids in asthma. Data on efficacy and safety of different doses were extracted from meta-analyses and summarised as the number needed to treat (NNT) and number needed to harm (NNH).Data synthesis: Inhaled corticosteroids were highly efficacious, with a relatively flat dose–response curve. Three patients needed to be treated with fluticasone 100 μg daily to prevent worsening asthma (NNT 3), and for fluticasone 1000 μg the NNT was 2.1 patients. The dose–response curve for side effects was steep. For a dose of fluticasone 100 μg, oral candidiasis developed in one of every 90 subjects treated (NNH 90). In contrast, the NNH for fluticasone 1000 μg and 2000 μg daily were 23 and 6, respectively.Conclusion: Level 1 evidence supports the use of low-dose inhaled corticosteroids in asthma. Clinicians should review doses of inhaled corticosteroids used for treating patients with asthma.

Heather Powell RN, MMedSci (ClinEpid) · Peter G Gibson MB BS, FRACP

Objectives: To conduct a systematic review of the effects of genetic counselling and testing for familial breast cancer on women's perception of risk and psychological morbidity.Data sources: MEDLINE, PsychLIT and EMBASE were searched for the period 1980–2001.Study selection: Studies were eligible if published in a peer-reviewed journal in English, included women with a family history of breast cancer who underwent genetic counselling or testing and had either a randomised controlled trial or prospective design, with a pre- and at least one post-counselling assessment.Data synthesis: As there was considerable heterogeneity in populations and measures, results were summarised rather than subjected to meta-analysis.Results: Overall, genetic counselling and testing appear to produce psychological benefits and to improve accuracy of risk perception. Carriers of mutations in cancer predisposition genes did not experience significant increases in depression and anxiety after disclosure of their mutation status, while non-carriers experienced significant relief. Women who were tested but declined to learn their results seemed to be at greater risk of a worse psychological outcome.Conclusions: To date, the data on psychological outcomes after genetic counselling and testing are reassuring. However, few studies used a randomised trial design, limiting the strength of the conclusions. Follow-up to date has been short, and we know little about the long-term impact of testing on patient behaviours, perceptions and psychological state.

Phyllis N Butow PhD, MPH, MClinPsych · Elizabeth A Lobb PhD, MAppSci, BAdEd · Alexandra Barratt PhD, MB BS(Hons) · Bettina Meiser PhD, BAppSci · Katherine M Tucker FRACP

Objective: To systematically review surveys of clinicians' attitudes to clinical practice guidelines.Data sources: MEDLINE, HealthStar, Embase and CINAHL were searched electronically for English-only surveys published from 1990 to 2000.Study selection: We included surveys with responses to one or more of seven propositions (see below). Studies were excluded if they had fewer than 100 respondents or if the response rate was less than 60%.Results: Thirty studies included responses to one or more of the seven items, giving a total of 11 611 responses. The response rate for the included studies was 72% (95% confidence interval [CI], 69%–75%). Clinicians agreed that guidelines were helpful sources of advice (weighted mean, 75%; 66%–83%), good educational tools (71%; 63%–79%) and intended to improve quality (70%; 60%–80%). However, clinicians also considered guidelines impractical and too rigid to apply to individual patients (30%; 23%–36%), that they reduced physician autonomy and oversimplified medicine (34%; 22%–47%), would increase litigation (41%; 32%–49%) and were intended to cut healthcare costs (52.8%; 39%–66%).Conclusions: Surveys of healthcare providers consistently report high satisfaction with clinical practice guidelines and a belief that they will improve quality, but there are concerns about the practicality of guidelines, their role in cost-cutting and their potential for increasing litigation.

Cynthia M Farquhar MB ChB, MD, FRANZCOG · Emma W Kofa BA · Jean R Slutsky PA, MSPH

Objective: To summarise recent evidence from case reports (published January 1995 – September 2001) of adverse events after cervical spine manipulation.Data sources: Five computerised literature searches (MEDLINE – Pubmed; EMBASE, the Cochrane Library, AMED [Allied and Complementary Medicine Database], and CISCOM [Centralised Information Service for Complementary Medicine]) were performed. No language restrictions were applied.Study selection: All case reports containing original data of adverse events after cervical spine manipulation were included.Data extraction: All articles were evaluated and key data extracted according to pre-defined criteria: patient's age, sex and diagnosis; type of therapist; type of treatment; nature of adverse event; method of diagnosis; and clinical outcome.Data synthesis: Thirty-one case reports (42 individual cases) were found. The patients were equally distributed between the sexes (21 male, 20 female, one unknown) and mostly middle-aged (range, 3 months to 87 years). Most were treated by chiropractors. Arterial dissection causing stroke was reported in at least 18 cases.Conclusions: Serious adverse events after cervical spine manipulation continue to be reported. As the incidence of these events is unknown, large and rigorous prospective studies of cervical spine manipulation are needed to accurately define the risks.

Edzard Ernst MD, PhD, FRCP(Edin)

Objective: To assess the (i) benefits, (ii) harms and (iii) costs of continuing mammographic screening for women 70 years and over.Data sources and synthesis: (i) We conducted a MEDLINE search (1966 – July 2000) for decision-analytic models estimating life-expectancy gains from screening in older women. The five studies meeting the inclusion criteria were critically appraised using standard criteria. We estimated relative benefit from each model's estimate of effectiveness of screening in older women relative to that in women aged 50–69 years using the same model. (ii) With data from BreastScreen Queensland, we constructed balance sheets of the consequences of screening for women in 10-year age groups (40–49 to 80–89 years), and (iii) we used a validated model to estimate the marginal cost-effectiveness of extending screening to women 70 years and over. Results: For women aged 70–79 years, the relative benefit was estimated as 40%–72%, and 18%–62% with adjustment for the impact of screening on quality of life. For women over 80 years the relative benefit was about a third, and with quality-of-life adjustment only 14%, that in women aged 50–69 years. (ii) Of 10 000 Australian women participating in ongoing screening, about 400 are recalled for further testing, and, depending on age, about 70–112 undergo biopsy and about 19–80 cancers are detected. (iii) Cost-effectiveness estimates for extending the upper age limit for mammographic screening from 69 to 79 years range from $8119 to $27 751 per quality-adjusted life-year saved, which compares favourably with extending screening to women aged 40–49 years (estimated at between $24 000 and $65 000 per life-year saved).Conclusions: Women 70 years and over, in consultation with their healthcare providers, may want to decide for themselves whether to continue mammographic screening. Decision-support materials are needed for women in this age group.

Alexandra L Barratt MB BS, PhD · Les M Irwig MB BCh, PhD · Glenn P Salkeld B Business, PhD · Paul P Glasziou MB BS, PhD · Nehmat Houssami MB BS, MPH