Simple sample size calculation for cluster-randomized trials

Provision of mouth-care in long-term care facilities: an educational trial

OBJECTIVES

This randomized clinical trial aimed to assess the effectiveness of a pyramid-based education for improving the oral health of elders in long-term care (LTC) facilities.

METHODS

Fourteen facilities matched for size were assigned randomly to an active or control group. At baseline in each facility, care-aides in the active group participated with a full-time nurse educator in a seminar about oral health care, and had unlimited access to the educator for oral health-related advice throughout the 3-month trial. Care-aides in the control group participated in a similar seminar with a dental hygienist but they received no additional advice. The residents in the facilities at baseline and after 3 months were examined clinically to measure their oral hygiene, gingival health, masticatory potential, Body Mass Index and Malnutrition Indicator Score, and asked to report on chewing difficulties.

RESULTS

Clinical measures after 3 months were not significantly different from baseline in either group, indicating that education neither influenced the oral health nor the dental hygiene of the residents.

CONCLUSIONS

A pyramid-based educational scheme with nurses and care-aides did not improve the oral health of frail elders in this urban sample of LTC facilities.

Stroke-unit care for acute stroke patients: an observational follow-up study

BACKGROUND

Large numbers of stroke patients arrive at hospital at a very early stage, and effective treatments for the acute phase of the disease are available. However, evidence that patients with acute stroke benefit from stroke-unit care is scarce. Our aim was to determine whether admission to a stroke unit, rather than a conventional ward, affected the outcome of patients with acute stroke.

METHODS

We did an observational follow-up study of 11 572 acute stroke patients hospitalised within 48 h of the onset of symptoms either in a stroke unit (n=4936) or in a conventional ward (6636). Patients were identified retrospectively from discharge records from 260 Italian hospitals. The primary outcome was mortality or disability (Rankin score greater than two), assessed prospectively by independent, masked assessors 2 years after admission. Analyses were adjusted for patient characteristics and clustered at the hospital level.

FINDINGS

Overall, 1576 patients died in hospital; 2169 died during the follow-up period. 347 patients were lost to follow-up. Compared with conventional-ward care, stroke-unit care was associated with a reduced probability of death or being disabled at the end of follow-up (odds ratio 0.81, 95% CI 0.72-0.91; p=0.0001). The potential benefit was significant across all age ranges and clinical characteristics, except for unconsciousness. No specific elements of setting, organisation, or process of care were associated with outcome.

INTERPRETATION

Admission to a stroke-unit ward with dedicated beds and staff within 48 h of onset should be recommended for all patients with acute stroke.

Home safety education and provision of safety equipment for injury prevention

BACKGROUND

In industrialised countries injuries are the leading cause of childhood death and steep social gradients exist in child injury mortality and morbidity. The majority of injuries in pre-school children occur at home, but there is little meta-analytic evidence that child home safety interventions improve a range of safety practices or reduce injury rates and little evidence on their effect by social group.

OBJECTIVES

We evaluated the effectiveness of home safety education, with or without the provision of low cost, discounted or free equipment in increasing home safety practices or reducing child injury rates and whether the effect varied by social group.

SEARCH STRATEGY

We searched The Cochrane Library, MEDLINE, EMBASE, CINAHL, DARE, ASSIA, Psychinfo and Web of Science, plus a range of relevant web sites, conference proceedings and bibliographies. We contacted authors of included studies and surveyed a range of organisations.

SELECTION CRITERIA

Randomised controlled trials (RCTs), non-randomised controlled trials and controlled before and after studies where home safety education with or without the provision of safety equipment was provided to those aged 19 years and under, which reported safety practices, possession of safety equipment or injury.

DATA COLLECTION AND ANALYSIS

Two authors independently assessed study quality and extracted data. We attempted to obtain individual participant level data (IPD) for all included studies and summary data and IPD were simultaneoulsy combined in meta-regressions by social and demographic variables.

MAIN RESULTS

Eighty studies were included; 37 of which were included in at least one meta-analysis. Twenty-three (62%) were RCTs and 12 (32%) included in the meta-analysis provided IPD. Home safety education was effective in increasing the proportion of families with safe hot tap water temperatures (OR 1.35, 95% CI 1.01 to 180), functional smoke alarms (OR 1.85, 95% CI 1.24 to 2.75), storing medicines (OR 1.58, 95% CI 1.18 to 2.13) and cleaning products (OR 1.63, 95% CI 1.22 to 2.17) out of reach, syrup of ipecac (OR 3.34, 95% CI 1.50 to 7.44) and poison control centre numbers accessible (OR 3.66, 95% CI 1.84 to 7.27), fitted stair gates (1.26, 95% CI 1.05 to 1.51), socket covers on unused sockets (OR 3.73, 95% CI 1.48 to 9.39) and storing sharp objects out of reach (OR 1.52, 95% CI 1.01 to 2.29). There was a lack of evidence that interventions reduced rates of thermal injuries, poisoning or a range of injuries. There was no consistent evidence that interventions were less effective in families whose children were at greater risk of injury.

AUTHORS' CONCLUSIONS

Home safety education provided most commonly as one-to-one, face-to-face education, in a clinical setting or at home, especially with the provision of safety equipment is effective in increasing a range of safety practices. There is a lack of evidence regarding its impact on child injury rates. There was no consistent evidence that home safety education, with or without the provision of safety equipment was less effective in those at greater risk of injury.

Adding zinc to supplemental iron and folic acid does not affect mortality and severe morbidity in young children

Studies have found a substantial reduction in diarrhea and respiratory morbidity in young children receiving zinc supplementation. The impact of daily zinc supplementation administered with iron plus folic acid (IFA) in young children on all-cause hospitalizations and mortality in comparison with IFA alone was evaluated. In a double blind cluster-randomized controlled trial, 94,359 subjects aged 1-23 mo were administered a daily dose of zinc plus IFA or IFA alone for a duration of 12 mo after enrollment. The intervention group tablet contained 10 mg of elemental zinc, 12.5 mg of iron, and 50 microg of folic acid. The control group tablets were similar except that they contained a placebo for zinc. Infants aged <6 mo were administered half a tablet, and those older received 1 tablet dissolved in breast milk or water. Hospitalizations were captured by trained study physicians through the surveillance of 8 hospitals. Deaths and hospitalizations were ascertained through visits to households by study supervisors once every 2 mo. The overall death rates did not differ significantly between the 2 groups when adjusted for cluster randomization (hazard ratio = 1.02, 95% CI 0.87, 1.19). Zinc and IFA supplementation compared with IFA alone did not affect adjusted hospitalization rates (overall rate ratio = 1.08, 95% CI 0.98, 1.19; diarrhea-specific rate ratio = 1.15, 95% CI 0.99, 1.34; or pneumonia-specific rate ratio = 1.09, 95% CI 0.94, 1.25). The lack of impact of zinc on mortality and hospitalization rates in this study may have been due to the use of lower daily zinc dosing than used in some of the morbidity prevention trials or from an interaction between zinc and iron, where the addition of iron may have adversely affected potential effects of zinc on immune function and morbidity. Future research should address iron and zinc interaction effects on important functional outcomes.

Effect of a structural intervention for the prevention of intimate-partner violence and HIV in rural South Africa: a cluster randomised trial

BACKGROUND

HIV infection and intimate-partner violence share a common risk environment in much of southern Africa. The aim of the Intervention with Microfinance for AIDS and Gender Equity (IMAGE) study was to assess a structural intervention that combined a microfinance programme with a gender and HIV training curriculum.

METHODS

Villages in the rural Limpopo province of South Africa were pair-matched and randomly allocated to receive the intervention at study onset (intervention group, n=4) or 3 years later (comparison group, n=4). Loans were provided to poor women who enrolled in the intervention group. A participatory learning and action curriculum was integrated into loan meetings, which took place every 2 weeks. Both arms of the trial were divided into three groups: direct programme participants or matched controls (cohort one), randomly selected 14-35-year-old household co-residents (cohort two), and randomly selected community members (cohort three). Primary outcomes were experience of intimate-partner violence--either physical or sexual--in the past 12 months by a spouse or other sexual intimate (cohort one), unprotected sexual intercourse at last occurrence with a non-spousal partner in the past 12 months (cohorts two and three), and HIV incidence (cohort three). Analyses were done on a per-protocol basis. This trial is registered with ClinicalTrials.gov, number NCT00242957.

FINDINGS

In cohort one, experience of intimate-partner violence was reduced by 55% (adjusted risk ratio [aRR] 0.45, 95% CI 0.23-0.91; adjusted risk difference -7.3%, -16.2 to 1.5). The intervention did not affect the rate of unprotected sexual intercourse with a non-spousal partner in cohort two (aRR 1.02, 0.85-1.23), and there was no effect on the rate of unprotected sexual intercourse at last occurrence with a non-spousal partner (0.89, 0.66-1.19) or HIV incidence (1.06, 0.66-1.69) in cohort three.

INTERPRETATION

A combined microfinance and training intervention can lead to reductions in levels of intimate-partner violence in programme participants. Social and economic development interventions have the potential to alter risk environments for HIV and intimate-partner violence in southern Africa.

Measuring improved targeting of health interventions to the poor in the context of a community-randomised trial in rural India

In spite of growing interest in socioeconomic differentials in health outcomes and access to health services, little has been written about methodologies for assessing the impact of equity-enhancing policies or programs. This paper describes three methodological challenges involved in designing a randomised trial with an equity outcome, and how these were met in a trial of alternative strategies to improving the uptake of benefits of a health insurance scheme among its poorest members. The Vimo SEWA trial is nested within a community-based insurance scheme in rural India. While conducting this trial, three methodological problems were encountered: (i) measuring poverty (or "wealth", or "socioeconomic status") (ii) assessing beneficiaries against an appropriate reference standard population and (iii) settling on an appropriate equity measure as an outcome indicator. These problems are likely to arise in any policy or program assessment that has an equity outcome. In the Vimo SEWA trial, the socioeconomic status of beneficiaries (claimants) is assessed relative to that of all scheme members living in same sub-district by applying a rapid assessment questionnaire--which reduces to an integrated index of socioeconomic status--to both a random sample of members in each sub-district, and to all claimants. The results are used to estimate the full distribution of socioeconomic status of members in each sub-district, with each member given a rank score between 0 and 100. Interpolation is used to estimate the rank scores of claimants relative to the membership base. The primary outcome measure for the trial is the mean socioeconomic rank score of claimants. In developing country settings, using an index of socioeconomic status is simpler than assessing household income or the value of household consumption. It is also relatively straightforward to compare the socioeconomic status of health program beneficiaries with a relevant reference population, although two independent surveys are required. Expressing relative wealth on a scale from zero to 100 is conceptually appealing, and the mean value of this rank score provides an equity-specific outcome measure readily integrated into the usual analytic framework for cluster-randomised trials.

Sample size for cluster randomized trials: effect of coefficient of variation of cluster size and analysis method

BACKGROUND

Cluster randomized trials are increasingly popular. In many of these trials, cluster sizes are unequal. This can affect trial power, but standard sample size formulae for these trials ignore this. Previous studies addressing this issue have mostly focused on continuous outcomes or methods that are sometimes difficult to use in practice.

METHODS

We show how a simple formula can be used to judge the possible effect of unequal cluster sizes for various types of analyses and both continuous and binary outcomes. We explore the practical estimation of the coefficient of variation of cluster size required in this formula and demonstrate the formula's performance for a hypothetical but typical trial randomizing UK general practices.

RESULTS

The simple formula provides a good estimate of sample size requirements for trials analysed using cluster-level analyses weighting by cluster size and a conservative estimate for other types of analyses. For trials randomizing UK general practices the coefficient of variation of cluster size depends on variation in practice list size, variation in incidence or prevalence of the medical condition under examination, and practice and patient recruitment strategies, and for many trials is expected to be approximately 0.65. Individual-level analyses can be noticeably more efficient than some cluster-level analyses in this context.

CONCLUSIONS

When the coefficient of variation is <0.23, the effect of adjustment for variable cluster size on sample size is negligible. Most trials randomizing UK general practices and many other cluster randomized trials should account for variable cluster size in their sample size calculations.

Analysis of clustered data in multicentre trauma studies

INTRODUCTION

In order to design multicentre studies an estimate of the correlation of the observations within each centre is necessary. A standard measure of the correlation between observations within each centre is the Intraclass Correlation Coefficient (ICC).

METHOD

We used the National Trauma Data Bank (NTDB). By 2004, 448 trauma centres (including 110 level I and 123 level II trauma centres) from 43 states and US territories contributed over 1.2 million records to the NTDB. Data of patients directly transported from the scene of injury to level I or II trauma centres were used to calculate the ICC of in-hospital trauma fatality and emergency department (ED) shock rate.

RESULTS

The ICCs of ED shock and in-hospital fatality rate were 0.010 (95% confidence interval (CI): 0.003-0.018) and 0.039 (95% CI: 0.028-0.050), respectively. The ICC of shock in the ED was the highest for penetrating injuries (0.017, 95% CI: 0.003-0.032) and the lowest for women (0.008, 95% CI: 0.002-0.013) although the observed difference between men and women was not statistically significant. The ICC of trauma fatality was the highest for penetrating injuries (0.073, 95% CI: 0.047-0.098), and the lowest for blunt injuries (0.029, 95% CI: 0.020-0.037).

DISCUSSION

Although the calculated ICCs might seem so small as to be ignored, the required sample size in studies with exclusively exposed or non-exposed clusters depends on the ICC and the average number of subjects within clusters. Therefore, investigators should be aware of the influence that these ICCs might have on sample size and power of their studies.

Uptake of workplace HIV counselling and testing: a cluster-randomised trial in Zimbabwe

BACKGROUND

HIV counselling and testing is a key component of both HIV care and HIV prevention, but uptake is currently low. We investigated the impact of rapid HIV testing at the workplace on uptake of voluntary counselling and testing (VCT).

METHODS AND FINDINGS

The study was a cluster-randomised trial of two VCT strategies, with business occupational health clinics as the unit of randomisation. VCT was directly offered to all employees, followed by 2 y of open access to VCT and basic HIV care. Businesses were randomised to either on-site rapid HIV testing at their occupational clinic (11 businesses) or to vouchers for off-site VCT at a chain of free-standing centres also using rapid tests (11 businesses). Baseline anonymised HIV serology was requested from all employees. HIV prevalence was 19.8% and 18.4%, respectively, at businesses randomised to on-site and off-site VCT. In total, 1,957 of 3,950 employees at clinics randomised to on-site testing had VCT (mean uptake by site 51.1%) compared to 586 of 3,532 employees taking vouchers at clinics randomised to off-site testing (mean uptake by site 19.2%). The risk ratio for on-site VCT compared to voucher uptake was 2.8 (95% confidence interval 1.8 to 3.8) after adjustment for potential confounders. Only 125 employees (mean uptake by site 4.3%) reported using their voucher, so that the true adjusted risk ratio for on-site compared to off-site VCT may have been as high as 12.5 (95% confidence interval 8.2 to 16.8).

CONCLUSIONS

High-impact VCT strategies are urgently needed to maximise HIV prevention and access to care in Africa. VCT at the workplace offers the potential for high uptake when offered on-site and linked to basic HIV care. Convenience and accessibility appear to have critical roles in the acceptability of community-based VCT.

Effective control of dengue vectors with curtains and water container covers treated with insecticide in Mexico and Venezuela: cluster randomised trials

OBJECTIVES

To measure the impact on the dengue vector population (Aedes aegypti) and disease transmission of window curtains and water container covers treated with insecticide.

DESIGN

Cluster randomised controlled trial based on entomological surveys and, for Trujillo only, serological survey. In addition, each site had a non-randomised external control.

SETTING

18 urban sectors in Veracruz (Mexico) and 18 in Trujillo (Venezuela).

PARTICIPANTS

4743 inhabitants (1095 houses) in Veracruz and 5306 inhabitants (1122 houses) in Trujillo.

INTERVENTION

Sectors were paired according to entomological indices, and one sector in each pair was randomly allocated to receive treatment. In Veracruz, the intervention comprised curtains treated with lambdacyhalothrin and water treatment with pyriproxyfen chips (an insect growth regulator). In Trujillo, the intervention comprised curtains treated with longlasting deltamethrin (PermaNet) plus water jar covers of the same material. Follow-up surveys were conducted at intervals, with the final survey after 12 months in Veracruz and nine months in Trujillo.

MAIN OUTCOME MEASURES

Reduction in entomological indices, specifically the Breteau and house indices.

RESULTS

In both study sites, indices at the end of the trial were significantly lower than those at baseline, though with no significant differences between control and intervention arms. The mean Breteau index dropped from 60% (intervention clusters) and 113% (control) to 7% (intervention) and 12% (control) in Veracruz and from 38% to 11% (intervention) and from 34% to 17% (control) in Trujillo. The pupae per person and container indices showed similar patterns. In contrast, in nearby communities not in the trial the entomological indices followed the rainfall pattern. The intervention reduced mosquito populations in neighbouring control clusters (spill-over effect); and houses closer to treated houses were less likely to have infestations than those further away. This created a community effect whereby mosquito numbers were reduced throughout the study site. The observed effects were probably associated with the use of materials treated with insecticide at both sites because in Veracruz, people did not accept and use the pyriproxyfen chips.

CONCLUSION

Window curtains and domestic water container covers treated with insecticide can reduce densities of dengue vectors to low levels and potentially affect dengue transmission.

Lessons and implications from a mass immunization campaign in squatter settlements of Karachi, Pakistan: an experience from a cluster-randomized double-blinded vaccine trial [NCT00125047]

Objective

To determine the safety and logistic feasibility of a mass immunization strategy outside the local immunization program in the pediatric population of urban squatter settlements in Karachi, Pakistan.

Methods

A cluster-randomized double blind preventive trial was launched in August 2003 in 60 geographic clusters covering 21,059 children ages 2 to 16 years. After consent was obtained from parents or guardians, eligible children were immunized parenterally at vaccination posts in each cluster with Vi polysaccharide or hepatitis A vaccine. Safety, logistics, and standards were monitored and documented.

Results

The vaccine coverage of the population was 74% and was higher in those under age 10 years. No life-threatening serious adverse events were reported. Adverse events occurred in less than 1% of all vaccine recipients and the main reactions reported were fever and local pain. The proportion of adverse events in Vi polysaccharide and hepatitis A recipients will not be known until the end of the trial when the code is broken. Throughout the vaccination campaign safe injection practices were maintained and the cold chain was not interrupted. Mass vaccination in slums had good acceptance. Because populations in such areas are highly mobile, settlement conditions could affect coverage. Systemic reactions were uncommon and local reactions were mild and transient. Close community involvement was pivotal for information dissemination and immunization coverage.

Conclusion

This vaccine strategy described together with other information that will soon be available in the area (cost/effectiveness, vaccine delivery costs, etc) will make typhoid fever control become a reality in the near future.

Family-member DOTS and community DOTS for tuberculosis control in Nepal: cluster-randomised controlled trial

BACKGROUND

A key component of the DOTS strategy for tuberculosis control (short-course chemotherapy following WHO guidelines) is direct observation of treatment. WHO technical guidelines recommend that health workers should undertake this part of the strategy, but will also accept direct observation of treatment in the community; WHO does not think that a family member should undertake this role. Supporting evidence for these recommendations is not available. The Nepal national tuberculosis programme asked us to develop and test a strategy of direct observation of treatment for the hill districts of Nepal, where direct observation of treatment by health workers is not feasible. We aimed to assess the success rates of two DOTS strategies developed for such areas.

METHODS

Between mid-July, 2002, and mid-July, 2003, we undertook a cluster-randomised controlled trial to compare two strategies-community DOTS and family-member DOTS--in ten hill and mountain districts of Nepal. Districts were used as the unit of randomisation. Primary outcome was success rate (proportion of registered patients who achieved cure or completed treatment), and analysis was by intention to treat.

FINDINGS

Five districts (549 patients) were allocated to community DOTS and five (358 patients) were allocated family-member DOTS. Community DOTS and family-member DOTS achieved success rates of 85% and 89%, respectively (odds ratio of success for community DOTS relative to family-member DOTS, 0.67 [95% CI 0.41-1.10]; p=0.09). Estimated case-finding rates were 63% with the community strategy and 44% with family-member DOTS.

INTERPRETATION

The family-member DOTS and community DOTS strategies can both attain international targets for treatment success under programme conditions, and thus are appropriate for the hill and mountain districts of Nepal. Both strategies might also be appropriate in other parts of the world where directly observed treatment by health workers is not feasible. Our findings lend support to adoption of this patient-responsive approach to direct observation of treatment within global tuberculosis control policy.

A cluster randomized-controlled trial to determine the effectiveness of Stepping Stones in preventing HIV infections and promoting safer sexual behaviour amongst youth in the rural Eastern Cape, South Africa: trial design, methods and baseline findings

OBJECTIVE

To describe the study design, methods and baseline findings of a behavioural intervention trial aimed at reducing HIV incidence.

METHOD

A cluster randomized-controlled trial (RCT) conducted in 70 villages in rural South Africa. A behavioural intervention, Stepping Stones, was implemented in 35 communities in two workshops of 20 men and 20 women in each community who met for 17 sessions (50 h) over a period of 3-12 weeks. Individuals in the control arm communities attended a single session of about 3 h on HIV and safer sex. Impact assessment was conducted through two questionnaire and serological surveys at 12-month intervals. The primary outcome was HIV incidence and secondary measures included changes in knowledge, attitude and sexual behaviours. Qualitative research was also undertaken with 10 men and 10 women from two sites receiving the intervention (one rural and one urban) and five men and five women from one village in the control arm. They were interviewed individually three times prior to the workshops and then 9-12 months later.

RESULTS

A total of 2776 participants (1409 intervention and 1367 control) were enrolled at baseline and had an interview, and HIV sero-status was established. HIV baseline prevalence rates in women were 9.8% in the intervention arm and 12.8% in the control arm. In men the prevalence was 1.7% in the intervention arm and 2.1% in the control arm. Demographic and behavioural characteristics were similar in the two arms. In the intervention groups 59.9% of participants attended more than 75% of the sessions. In the control group 66.3% attended the control session.

CONCLUSION

This is the third RCT to be conducted in sub-Saharan Africa evaluating a behavioural intervention using HIV incidence as a primary outcome. It is of particular interest as the intervention in question is used in many developing countries. There is good baseline comparability between the study arms and the process data on the workshops suggested that the interventions were feasible and adequately implemented.

A multi-country cluster randomized controlled effectiveness evaluation to accelerate the introduction of Vi polysaccharide typhoid vaccine in developing countries in Asia: rationale and design

Phase-III vaccine efficacy trials typically employ individually randomized designs intended to ensure that measurements of vaccine protective efficacy reflect only direct vaccine effects. As a result, decisions about introducing newly licensed vaccines into public health programmes often fail to consider the substantially greater protection that may occur when a vaccine is deployed in public health programmes, due to the combination of direct plus indirect vaccine protective effects. Vaccine total protection can be better evaluated with cluster randomized trials. Such a design was considered to generate policy relevant data to accelerate the rationale introduction of the licensed typhoid fever Vi polysaccharide (PS) vaccine in Asia by the Diseases of the Most Impoverished (DOMI) typhoid fever programme. The DOMI's programme multi-country study is one of the largest cluster randomized vaccine trials ever mounted in Asia, which includes approximately 200,000 individuals. Its main objective is to determine the effectiveness of a licensed Vi PS vaccine. The rationale and design of this study are discussed. Preliminary results are presented that determined the final planning of the trial before immunization. Important methodological and practical issues regarding vaccine cluster randomized designs are illustrated.

Ensayos clínicos aleatorizados comunitarios (CONSORT CLUSTER)

With the publication of the CONSORT statement there is now increased awareness of the need to adequately report the findings of randomised controlled trials. The CONSORT statement includes a checklist of items that should be addressed in the trial report. The original CONSORT statement was developed to ensure the appropriate reporting of parallel group randomised controlled trials in which individual participants are allocated to different intervention groups. In cluster randomised trials, however, groups of participants, rather than individuals, are randomly allocated to study groups. The process of allocating groups of participants raises additional reporting considerations and led to the publication of an extension to the CONSORT statement specifically for cluster randomised trials. In this paper we review the CONSORT extension to cluster randomised trials, outlining the special features of the cluster randomised trial which must be considered.

Case management for the treatment of patients with major depression in general practices--rationale, design and conduct of a cluster randomized controlled trial--PRoMPT (PRimary care Monitoring for depressive Patient's Trial) [ISRCTN66386086]--study protocol

BACKGROUND

Depression is a disorder with high prevalence in primary health care and a significant burden of illness. The delivery of health care for depression, as well as other chronic illnesses, has been criticized for several reasons and new strategies to address the needs of these illnesses have been advocated. Case management is a patient-centered approach which has shown efficacy in the treatment of depression in highly organized Health Maintenance Organization (HMO) settings and which might also be effective in other, less structured settings.

METHODS/DESIGN

PRoMPT (PRimary care Monitoring for depressive Patients Trial) is a cluster randomised controlled trial with General Practice (GP) as the unit of randomisation. The aim of the study is to evaluate a GP applied case-management for patients with major depressive disorder. 70 GPs were randomised either to intervention group or to control group with the control group delivering usual care. Each GP will include 10 patients suffering from major depressive disorder according to the DSM-IV criteria. The intervention group will receive treatment based on standardized guidelines and monthly telephone monitoring from a trained practice nurse. The nurse investigates the patient's status concerning the MDD criteria, his adherence to GPs prescriptions, possible side effects of medication, and treatment goal attainment. The control group receives usual care--including recommended guidelines. Main outcome measure is the cumulative score of the section depressive disorders (PHQ-9) from the German version of the Prime MD Patient Health Questionnaire (PHQ-D). Secondary outcome measures are the Beck-Depression-Inventory, self-reported adherence (adapted from Moriskey) and the SF-36. In addition, data are collected about patients' satisfaction (EUROPEP-tool), medication, health care utilization, comorbidity, suicide attempts and days out of work. The study comprises three assessment times: baseline (T0) , follow-up after 6 months (T1) and follow-up after 12 months (T2).

DISCUSSION

Depression is now recognized as a disorder with a high prevalence in primary care but with insufficient treatment response. Case management seems to be a promising intervention which has the potential to bridge the gap of the usually time-limited and fragmented provision of care. Case management has been proven to be effective in several studies but its application in the private general medical practice setting remains unclear.

A community-based study to examine the effect of a youth HIV prevention intervention on young people aged 15-24 in South Africa: results of the baseline survey

OBJECTIVES

To determine whether South African youths living in communities that had either of two youth human immunodeficiency virus (HIV) prevention interventions [(a) loveLife Youth Centre or (b) loveLife National Adolescent Friendly Clinic Initiative] would have a lower prevalence of HIV, sexually transmitted infections (STIs), and high risk sexual behaviours than communities without either of these interventions.

METHODS

In 2002 the baseline survey of a quasi-experimental, community-based study was conducted in South Africa. In total 33 communities were included in three study arms (11 communities per study arm). The final sample included 8735 youths aged 15-24 years. All participants took part in a behavioural interview and were tested for HIV, gonorrhoea (Neisseria gonorrhoeae) and Chlamydia (Chlamydia trachomatis).

RESULTS

HIV prevalence was 20.0% among females and 7.5% among males (OR 3.93 95% CI 2.51-6.15). There were no significant differences between study arms for HIV, NG or CT prevalence at baseline. In multiple regression analyses, HIV was significantly associated with NG infection (OR 1.96 95% CI 1.24-3.12) but not with CT infection. Youths who reported >1 lifetime partner were also significantly more likely to be infected with HIV (OR 1.98 95% CI 1.55-2.52), as were those who reported ever having engaged in transactional sex (OR 1.86 P = 0.02) or having had genital ulcers in the past 12 months (OR 1.71 P < or = 0.001).

CONCLUSIONS

HIV prevention programmes must ensure that gender inequities that place young women at greater risk for HIV infection are urgently addressed and they must continue to emphasize the importance of reducing the number of sexual partners and STI treatment.

Methodological issues in measuring the impact of interventions against female genital cutting

With increasing efforts being made to introduce systematic interventions for encouraging abandonment of female genital cutting (FGC) comes the need to better understand how such interventions work and what effects they have. Many interventions are based on theoretical models of behaviour change and so studies to evaluate them should develop indicators appropriate to the type of behaviour change anticipated. Systematic evaluations need also to use some form of quasi-experimental design to be able to attribute change to the intervention and not to any 'natural' change in FGC behaviour or other activities that may be concurrent. A sustained change in the prevalence of FGC is the ultimate indicator and there are several ways this can be measured, although with many limitations given the intimate nature of the practice. Moreover, appropriate sample sizes must be calculated and used to be able to draw valid conclusions. Many of those implementing FGC interventions are not familiar with such basic research principles and so there is an urgent need to ensure that projects are well designed so that valid conclusions concerning their effectiveness can be drawn.

Estimating sample size in critical care clinical trials

Estimating the required sample size for a study is necessary during the design phase to ensure that it will have maximal efficiency to answer the primary question of interest. Clinicians require a basic understanding of the principles underlying sample size calculation to interpret and apply research findings. This article reviews the critical components of sample size calculation, including the selection of a primary outcome, specification of the acceptable types I and II error rates, identification of the minimal clinically important difference, and estimation of the error associated with measuring the primary outcome. The relationship among confidence intervals, precision, and study power is also discussed.

A cluster randomized controlled trial of visual, cytology and human papillomavirus screening for cancer of the cervix in rural India

The impact of screening by visual inspection with acetic acid (VIA), cytology or HPV testing on cervical cancer incidence and mortality is investigated in a cluster randomized controlled trial in India. We report findings after the screening phase, when 52 clusters, with a total of 142,701 women aged 30-59 years in Osmanabad District, India, were randomized into 4 arms for a single round of screening by trained midwives with either VIA, cytology or HPV testing as well as a control group. All laboratory tests were done locally. Test-positive women underwent investigations (colposcopy/biopsy) and treatment in the base hospital. Data on participation, test positivity, positive predictive value and detection rates of cervical neoplasia were analyzed using cluster design methodology. Of the eligible women, 72-74% were screened. Test positivity rates were 14.0% for VIA, 7.0% for cytology and 10.3% for HPV. The detection rate of high-grade lesions was similar in all intervention arms (0.7% for VIA, 1.0% for cytology and 0.9% for HPV testing) (p = 0.06, Mann-Whitney test). While the detection rate for VIA dropped to 0.5% with declining test positivity during the course of the study, it remained constant for cytology and HPV testing. Over 85% of women with high-grade lesions received treatment. Our results show that a high level of participation and good-quality cytology can be achieved in low-resource settings. VIA is a useful alternative but requires careful monitoring. Detection rates obtained by HPV testing were similar to cytology, despite higher investments.

The MEMA kwa Vijana Project: Design of a community randomised trial of an innovative adolescent sexual health intervention in rural Tanzania

BACKGROUND

Effective interventions to reduce the incidence of HIV, other sexually transmitted infections (STIs) and unwanted pregnancy among adolescents in sub-Saharan Africa are urgently needed. This paper describes the rationale and design of a randomised trial of the impact of an innovative sexual health intervention among adolescents in rural Mwanza Region, Tanzania.

METHODS

The MEMA kwa Vijana intervention comprises a teacher-led, peer-assisted sexual health education programme for students in the last 3 years of primary school, training and supervision of health workers in the provision of youth-friendly health services, peer condom promotion and distribution, and wider community activities. Detailed process evaluation was conducted and the impact of the intervention was evaluated through a community-randomised trial in which a cohort of 9645 adolescents was followed up for 3 years. Both process and impact evaluation used multiple assessment methods. Impact measures included incidence and prevalence of HIV and other STIs, pregnancy rates, knowledge and reported attitudes and sexual behaviour, as well as qualitative assessments.

RESULTS

Results of the baseline survey of the cohort have been presented previously. The outcome of the trial will be reported separately.

CONCLUSIONS

Behaviour change interventions among adolescents have been widely advocated, but there have been few rigorously designed trials of their effectiveness, particularly in developing countries, and measurement of sexual behaviour is particularly problematic in this age group. The MEMA kwa Vijana trial was undertaken to address these problems and to collect rigorous evidence on the effectiveness of an innovative intervention, designed to be implemented on a very large scale.

Effect of screening on oral cancer mortality in Kerala, India: a cluster-randomised controlled trial

BACKGROUND

Oral cancer is common in men from developing countries, and is increased by tobacco and alcohol use. We aimed to assess the effect of visual screening on oral cancer mortality in a cluster-randomised controlled trial in India.

METHODS

Of the 13 clusters chosen for the study, seven were randomised to three rounds of oral visual inspection by trained health workers at 3-year intervals and six to a control group during 1996-2004, in Trivandrum district, Kerala, India. Healthy participants aged 35 years and older were eligible for the study. Screen-positive people were referred for clinical examination by doctors, biopsy, and treatment. Outcome measures were survival, case fatality, and oral cancer mortality. Oral cancer mortality in the study groups was analysed and compared by use of cluster analysis. Analysis was by intention to treat.

FINDINGS

Of the 96,517 eligible participants in the intervention group, 87,655 (91%) were screened at least once, 53,312 (55%) twice, and 29,102 (30%) three times. Of the 5145 individuals who screened positive, 3218 (63%) complied with referral. 95,356 eligible participants in the control group received standard care. 205 oral cancer cases and 77 oral cancer deaths were recorded in the intervention group compared with 158 cases and 87 deaths in the control group (mortality rate ratio 0.79 [95% CI 0.51-1.22]). 70 oral cancer deaths took place in users of tobacco or alcohol, or both, in the intervention group, compared with 85 in controls (0.66 [0.45-0.95]). The mortality rate ratio was 0.57 (0.35-0.93) in male tobacco or alcohol users and 0.78 (0.43-1.42) in female users.

INTERPRETATION

: Oral visual screening can reduce mortality in high-risk individuals and has the potential of preventing at least 37,000 oral cancer deaths worldwide.

Do rates of hospital admission for falls and hip fracture in elderly people vary by socio-economic status?

OBJECTIVE

To determine the relationship between hospital admissions for falls and hip fracture in elderly people and area characteristics such as socio-economic deprivation.

STUDY DESIGN

Ecological study of routinely collected hospital admissions data for falls and hip fracture in people aged 75 years or over for 1992-1997, linked at electoral ward level with characteristics from census data.

METHODS

In total, 42,293 and 17,390 admissions were identified for falls and hip fracture, respectively, from 858 electoral wards in Trent. Rate ratios (RRs) for hospital admissions for falls and hip fracture were calculated by the electoral wards' Townsend score divided by quintiles. RRs were estimated by negative binomial regression and adjusted for the ward characteristics of age, gender, ethnicity, rurality, proportion of elderly people living alone and distance from hospital.

RESULTS

There was a small but statistically significant association at electoral ward level between hospital admissions for falls and the Townsend score, with the most deprived wards having a 10% higher admission rate for falls compared with the most affluent wards (adjusted RR 1.10, 95% CI 1.01-1.19). No association was found between hospital admission for hip fracture and deprivation (adjusted RR 1.05, 95% CI 0.95-1.16).

CONCLUSION

There is some evidence of an association at electoral ward level between hospital admissions for falls and socio-economic deprivation, with higher rates in deprived areas. No such association was found for hip fracture. Further work is required to assess the impact of interventions on reducing inequalities in hospital admission rates for falls in elderly people.

Using baseline data to design a group randomized trial

Group randomized trials (GRT) are often designed with relatively little preliminary data available to estimate key parameters. In this paper, however, the opposite situation is considered-very good baseline data are available on the primary outcome of interest. These data can then be used to inform key design and analysis decisions such as (i) should the trial be designed as an unmatched or pair-matched study, or stratified in some other fashion; (ii) is analysis of "change from baseline" preferable to using end-of-study data alone; and (iii) what power might be expected by pursuing these various strategies. The results are applied to a GRT for sexually transmitted diseases prevention in Peru.

Effectiveness and cost of facility-based Integrated Management of Childhood Illness (IMCI) in Tanzania

BACKGROUND

The Integrated Management of Childhood Illness (IMCI) strategy is designed to address major causes of child mortality at the levels of community, health facility, and health system. We assessed the effectiveness of facility-based IMCI in rural Tanzania.

METHODS

We compared two districts with facility-based IMCI and two neighbouring comparison districts without IMCI, from 1997 to 2002, in a non-randomised study. We assessed quality of case-management for children's illness, drug and vaccine availability, and supervision involving case-management, through a health-facility survey in 2000. Household surveys were used to assess child-health indicators in 1999 and 2002. Survival of children was tracked through demographic surveillance over a predefined 2-year period from mid 2000. Further information on contextual factors was gathered through interviews and record review. The economic cost of health care for children in IMCI and comparison districts was estimated through interviews and record review at national, district, facility, and household levels.

FINDINGS

During the IMCI phase-in period, mortality rates in children under 5 years old were almost identical in IMCI and comparison districts. Over the next 2 years, the mortality rate was 13% lower in IMCI than in comparison districts (95% CI -7 to 30 or 5 to 21, depending on how adjustment is made for district-level clustering), with a rate difference of 3.8 fewer deaths per 1000 child-years. Contextual factors, such as use of mosquito nets, all favoured the comparison districts. Costs of children's health care with IMCI were similar to or lower than those for case-management without IMCI.

INTERPRETATION

Our findings indicate that facility-based IMCI is good value for money, and support widespread implementation in the context of health-sector reform, basket funding, good facility access, and high utilisation of health facilities.

Effect of a participatory intervention with women's groups on birth outcomes in Nepal: cluster-randomised controlled trial

BACKGROUND

Neonatal deaths in developing countries make the largest contribution to global mortality in children younger than 5 years. 90% of deliveries in the poorest quintile of households happen at home. We postulated that a community-based participatory intervention could significantly reduce neonatal mortality rates.

METHODS

We pair-matched 42 geopolitical clusters in Makwanpur district, Nepal, selected 12 pairs randomly, and randomly assigned one of each pair to intervention or control. In each intervention cluster (average population 7000), a female facilitator convened nine women's group meetings every month. The facilitator supported groups through an action-learning cycle in which they identified local perinatal problems and formulated strategies to address them. We monitored birth outcomes in a cohort of 28?931 women, of whom 8% joined the groups. The primary outcome was neonatal mortality rate. Other outcomes included stillbirths and maternal deaths, uptake of antenatal and delivery services, home care practices, infant morbidity, and health-care seeking. Analysis was by intention to treat. The study is registered as an International Standard Randomised Controlled Trial, number ISRCTN31137309.

FINDINGS

From 2001 to 2003, the neonatal mortality rate was 26.2 per 1000 (76 deaths per 2899 livebirths) in intervention clusters compared with 36.9 per 1000 (119 deaths per 3226 livebirths) in controls (adjusted odds ratio 0.70 [95% CI 0.53-0.94]). Stillbirth rates were similar in both groups. The maternal mortality ratio was 69 per 100000 (two deaths per 2899 livebirths) in intervention clusters compared with 341 per 100000 (11 deaths per 3226 livebirths) in control clusters (0.22 [0.05-0.90]). Women in intervention clusters were more likely to have antenatal care, institutional delivery, trained birth attendance, and hygienic care than were controls.

INTERPRETATION

Birth outcomes in a poor rural population improved greatly through a low cost, potentially sustainable and scalable, participatory intervention with women's groups.

Persistent itching nodules after the fourth dose of diphtheria–tetanus toxoid vaccines without evidence of delayed hypersensitivity to aluminium

Studies in Gothenburg, Sweden, reported an exceptionally high rate of persistent itching nodules at the site of injection of aluminium containing vaccines, usually with positive epicutaneous tests to aluminium. When a new booster diphtheria-tetanus vaccine was introduced we performed a prospective cluster randomised active surveillance in 25,232 10-year-olds. Parental reports 6 months after vaccination with Duplex or diTeBooster were collected for 22,365 (88%) pupils in 851 schools. We identified 3-6 children per 10,000 with a local itching nodule persisting for at least 2 months. There were no significant differences between the vaccine groups. Contact allergy to aluminium was not detected. The findings support the use of the vaccine presently available in the Swedish vaccination program. Continued surveillance of persistent itching nodules and aluminium contact allergy is, however, warranted for vaccines containing pertussis toxoid and aluminium.

Intracluster correlation coefficients in cluster randomized trials: empirical insights into how should they be reported

Background

Increasingly, researchers are recognizing that there are many situations where the use of a cluster randomized trial may be more appropriate than an individually randomized trial. Similarly, the need for appropriate standards of reporting of cluster trials is more widely acknowledged.

Methods

In this paper, we describe the results of a survey to inform the appropriate reporting of the intracluster correlation coefficient (ICC) – the statistical measure of the clustering effect associated with a cluster randomized trial.

Results

We identified three dimensions that should be considered when reporting an ICC – a description of the dataset (including characteristics of the outcome and the intervention), information on how the ICC was calculated, and information on the precision of the ICC.

Conclusions

This paper demonstrates the development of a framework for the reporting of ICCs. If adopted into routine practice, it has the potential to facilitate the interpretation of the cluster trial being reported and should help the development of new trials in the area.

Initial results from a randomized trial of cervical visual screening in rural south India

The impact of a single round of screening of visual inspection with acetic acid (VIA) on cervical cancer incidence and mortality was investigated in a cluster randomized trial in south India. Women 30-59 years of age in 113 clusters in Dindigul District were randomized to VIA screening (57 clusters, 48,225 women) by nurses and to a control group (56 clusters, 30,167 women). 30,577 eligible women were screened between May 2000 and April 2003; 2,939 (9.6%) screen-positive women were investigated with colposcopy by nurses and 2,777 (9.1%) women had biopsy. CIN 1 was diagnosed in 1,778 women, CIN 2-3 lesions were found in 222, and there were 69 screen detected invasive cervical cancers. The detection rates of lesions per 1,000 screened women were 58.2 for CIN 1, 7.3 for CIN 2-3, and 2.3 for invasive cancer. The detection rate of high-grade lesions in our study was 2-3-fold higher than those observed in repeatedly screened populations in developed countries. 71% of women with CIN 1 and 80% of those with CIN 2-3 lesions accepted cryotherapy provided by nurses and surgical treatment by mid-level clinicians. Overall, 97 and 34 incident cervical cancer cases were observed in the intervention and control arms, respectively. The intervention arm accrued 124,144 person years and the control arm accrued 90,172 during the study period. The age standardized cervical cancer incidence rates were 92.4/100,000 person-years in the intervention and 43.1/100,000 in the control arms. In the screened arm, 35.0% of cases were in Stage I as opposed to none in the control arm. The preliminary findings from our study indicate that not only is a VIA-based screening programme feasible, safe and acceptable to a population in rural settings, it also results in early detection of cervical neoplasia.

Comparing completely and stratified randomized designs in cluster randomized trials when the stratifying factor is cluster size: a simulation study

Stratified randomized designs are popular in cluster randomized trials (CRTs) because they increase the chance of the intervention groups being well balanced in terms of identified prognostic factors at baseline and may increase statistical power. The objective of this paper is to assess the gains in power obtained by stratifying randomization by cluster size, when cluster size is associated with an important cluster level factor which is otherwise unaccounted for in data analysis. A simulation study was carried out using a CRT where UK general practices were the randomized units as a template. The results show that when cluster size is strongly associated with a cluster level factor which is predictive of outcome, the stratified randomized design has superior power results to the completely randomized design and that the superiority is related to the number of clusters.

Design and analysis of group-randomized trials: a review of recent methodological developments

We review recent developments in the design and analysis of group-randomized trials (GRTs). Regarding design, we summarize developments in estimates of intraclass correlation, power analysis, matched designs, designs involving one group per condition, and designs in which individuals are randomized to receive treatments in groups. Regarding analysis, we summarize developments in marginal and conditional models, the sandwich estimator, model-based estimators, binary data, survival analysis, randomization tests, survey methods, latent variable methods and nonlinear mixed models, time series methods, global tests for multiple endpoints, mediation effects, missing data, trial reporting, and software. We encourage investigators who conduct GRTs to become familiar with these developments and to collaborate with methodologists who can strengthen the design and analysis of their trials.

The effects of alternative study designs on the power of community randomized trials: evidence from three studies of human immunodeficiency virus prevention in East Africa

BACKGROUND

Randomized intervention trials in which the community is the unit of randomization are increasingly being used to evaluate the impact of public health interventions. In the design of community randomized trials (CRT), the power of the study is likely to be affected by two issues: the matching or stratification of communities, and the number and size of the communities to be randomized.

METHODS

Data from three East African community intervention trials, designed to evaluate the impact of interventions to reduce human immunodeficiency virus (HIV) incidence, are used to compare the efficiency of different trial designs.

RESULTS

Compared with an unmatched design, stratification reduced the between-community variation in the Mwanza trial (from 0.51 to 0.24) and in the Masaka trial (from 0.38 to 0.28). The reduction was smaller in the Rakai trial where the selected communities were more homogeneous (from 0.15 to 0.11). For all trials, individual matching of communities produced estimates of between-community variation similar to those from the stratified designs. The linear association between HIV prevalence and incidence was strong in the Mwanza trial (correlation coefficient R = 0.83) and the Masaka trial (R = 0.83), but weak in the Rakai trial (R = 0.28). Unmatched study designs that use smaller communities tend to increase between-community variation, but reduce the design effect and improve study power.

CONCLUSIONS

These empirical data suggest that selection of homogeneous communities, or stratification of communities prior to randomization, may improve the power of CRT.

A school-based randomized controlled trial of peer-led sex education in England

This article discusses the design of an ongoing cluster-randomized trial comparing two forms of school-based sex education in terms of educational process and sexual health outcomes. Twenty-nine schools in southern England have been randomized to either peer-led sex education or to continue with their traditional teacher-led sex education. The primary objective is to determine which form of sex education is more effective in promoting young people's sexual health. The trial includes an unusually detailed evaluation of the process of sex education as well as the outcomes. The sex education programs were delivered in school to pupils ages 13-14 years who are being followed until ages 19-20. Major trial outcomes are unprotected sexual intercourse and regretted intercourse by age 16 and cumulative incidence of abortion by ages 19-20. We discuss the rationale behind various aspects of the design, including ethical issues and practical challenges of conducting a randomized trial in schools, data linkage for key outcomes to reduce bias, and integrating process and outcome measures to improve the interpretation of findings.

Physical activity and nutrition programs for couples: a randomized controlled trial

Diet and physical activity habits may deteriorate after cohabitation, leading to weight gain and increased risk of lifestyle diseases. We carried out a 4-month, randomized controlled trial of a diet and physical activity program for couples with a 1-year follow-up, comparing two methods of delivery. The program used six modules, which, after an initial group session, were mailed to the low-level intervention group. In the high-level intervention group, half of the modules were mailed, and the others were delivered at interactive group sessions. A control group received no intervention. Postintervention and at follow-up, physical fitness improved in the high-level group, saturated fat intake decreased in both intervention groups, and low-density lipoprotein cholesterol fell in the high-level group. Fewer participants in the high-level group became overweight or obese. Health promotion for couples can improve health behaviors and potentially lower the risk of lifestyle diseases in participants and their future families.

The effect of mass administration of sulfadoxine-pyrimethamine combined with artesunate on malaria incidence: a double-blind, community-randomized, placebo-controlled trial in The Gambia

A double-blind, community-randomized, placebo-controlled trial was conducted in a rural area of The Gambia between June and December 1999 to test whether a reduction in the infectious reservoir can reduce malaria transmission. Overall 14,017 (85%) individuals living in the study area were treated with either placebo or sulfadoxine-pyrimethamine (SP) combined with a single dose of artesunate (AS). Following the mass drug administration (MDA) 1375 children aged 6 months to 10 years were kept under surveillance for clinical malaria in 18 villages throughout the 1999 malaria transmission season. During a 20-week surveillance period 637 episodes of malaria were detected. The mean incidence rate was 2.5/100 child-weeks in the placebo villages, and 2.3/100 child-weeks in villages that received SP + AS. The mean rate ratio, adjusted for individual and village-level covariates, was 0.91 (95% CI 0.68-1.22, P = 0.49). During the first 2 months of surveillance, the malaria incidence was lower in treated villages. After 2 months the incidence was slightly higher in the MDA group but this was not statistically significant. Overall, no benefit of the MDA could be detected. The reason for the absence of an impact on malaria transmission is probably the very high basic reproductive number of malaria, and the persistence of mature gametocytes, which are not affected by AS treatment.

Empirical modelling of population sampling: lessons for designing sentinel surveillance

BACKGROUND

In order to consider the practical viability of 1% sentinel area surveillance for health information in resource-poor settings where complete registration is unrealistic, the effects of different sampling procedures on the representativeness of 1% population samples have been investigated.

METHODS

Using the 1991 census of England as a basis for modelling, 20 1% samples, each incorporating seven key parameters, were drawn at random from the overall dataset by each of eight different sampling procedures. Each sample was compared with the 'gold standard' of the overall census results, enabling comparisons between the different sampling procedures.

RESULTS

Representativeness of the 1% samples varied considerably between parameters and sampling procedures. At one extreme, the proportion of males in the population was distributed such that different sampling methods had little effect. On the other hand, samples of a heterogeneous parameter such as the proportion of non-whites in the population depended greatly on the procedure used. Sampling smaller administrative units tended to be more accurate. However, sampling units using probability proportional to size generally gave less representative samples. Stratifying urban and rural populations in the samples had little effect. Multistage sampling, emulating typical demographic surveillance sites, also generally gave less representative samples.

CONCLUSIONS

It is possible to achieve representative data by taking 1% of a national population in a sentinel surveillance approach, but sampling design can have an important influence on the outcome. This modelling supports the concept of 1% sentinel surveillance for health information in poorer settings, where complete data are unavailable.

Intervention to promote appropriate blood use in India

The decision to transfuse should be guided by information on the risks and benefits of transfusion. Safer alternatives should be considered. Properly screened blood or components from a reputable source should be used. In this study, a simple, self-educating transfusion request form was developed, and its effects on transfusion practice were assessed, using a cluster-randomized trial. Transfusions at six study hospitals and six control hospitals at four locations in India (Delhi, Bangalore, Nasik and Imphal) were monitored over a 4-month pre-intervention period and a 5-month post-intervention period. During the trial, 56 171 units were transfused to in-patients at the participating hospitals. Among the six intervention hospitals, there was some evidence of a nonsignificant post-intervention reduction in all the three main outcome measures: number of transfusion requests per admission (P = 0.09), number of units transfused per admission (P = 0.11) and number of crossmatches per admission (P = 0.06). No such changes were seen at control sites over the same period. Simple interventions to promote good clinical practice can have an effect, but may be better placed within longer term, broad-based strategies that are able to consider some of the background factors. Lack of clinical training, the hospital environment and fragmented blood bank services influence the way blood is used in India. It is thought that the intervention was ultimately unsuccessful because these factors remained as detrimental influences. A focus on education, policy and infrastructure in line with the new National Blood Policy will be important in coming years.

Effect of insecticide-impregnated dog collars on incidence of zoonotic visceral leishmaniasis in Iranian children: a matched-cluster randomised trial

BACKGROUND

Deltamethrin-impregnated dog collars reduce sandfly bite rates on dogs, and are effective in killing sandflies that attempt to feed. Because domestic dogs are the principal reservoir hosts of zoonotic visceral leishmaniasis, we tested whether community-wide application of dog collars could protect children against infection with Leishmania infantum, the parasite that causes the disease.

METHODS

18 villages were paired, matched by preintervention child prevalence of L infantum infection. Within pairs, villages were randomly assigned to either control or intervention. All domestic dogs in intervention villages were provided with collars for the transmission season. The main outcome measure was incidence of L infantum infection after 1 year measured by seroconversion. Secondary outcomes were leishmanin skin test (LST) conversion and seroconversion in dogs.

FINDINGS

The seroconversion rate in children was 1.49% (17/1141) in the intervention villages and 2.41% (26/1078) in control villages (odds ratio 0.57, 95% CI 0.36-0.90, p=0.017). LST conversion was also lowered, but not significantly (odds ratio 0.66, 0.41-1.08, p=0.096). The seroconversion rate in dogs in intervention villages was also significantly reduced (0.46, 0.30-0.70, p=0.0003).

INTERPRETATION

Community-wide application of deltamethrin-impregnated dog collars not only protects domestic dogs from L infantum infections, but might also reduce the risk of L infantum infection in children. These dog collars could have a role in control of visceral leishmaniasis and replace controversial dog culling programmes in some countries. However, the effectiveness of dog collars will depend on the importance of wild versus domestic canids as reservoir hosts of L infantum.

Methods for the analysis of incidence rates in cluster randomized trials

BACKGROUND

The published literature on cluster randomized trials focuses on outcomes that are either continuous or binary. In many trials, the outcome is an incidence rate, such as mortality, based on person-years data. In this paper we review methods for the analysis of such data in cluster randomized trials and present some simple approaches.

METHODS

We discuss the choice of the measure of intervention effect and present methods for confidence interval estimation and hypothesis testing which are conceptually simple and easy to perform using standard statistical software. The method proposed for hypothesis testing applies a t-test to cluster observations. To control confounding, a Poisson regression model is fitted to the data incorporating all covariates except intervention status, and the analysis is carried out on the residuals from this model. The methods are presented for unpaired data, and extensions to paired or stratified clusters are outlined.

RESULTS

The methods are evaluated by simulation and illustrated by application to data from a trial of the effect of insecticide-impregnated bednets on child mortality.

CONCLUSIONS

The techniques provide a straightforward approach to the analysis of incidence rates in cluster randomized trials. Both the unadjusted analysis and the analysis adjusting for confounders are shown to be robust, even for very small numbers of clusters, in situations that are likely to arise in randomized trials.

Clinical trials of behavioural interventions with heterogeneous teaching subgroup effects

Behaviour modification is often delivered to teaching subgroups. For example, experimental and control smoking cessation programmes may be given to 15 classes (subgroups) with 10 (otherwise independent) individuals. We present general statistical tests and power estimates to compare continuous outcomes from two interventions in settings where the magnitude of teaching subgroup heterogeneity, number of subgroups and subgroup size can differ between intervention arms. An application is made to data from a trial to reduce disease-transmitting sexual behaviour. The statistical impact of teaching subgroup heterogeneity effect increases as the (a) number of participants in a subgroup increases, and (b) ratio of 'averaged experimental and control subgroup effect variance' to study subject variance increases. If plausible levels of subgroup teaching effect heterogeneity are ignored, the true sizes of tests with nominal 0.05 two-sided type I errors range from 0.055 to 0.47, while when planning studies, estimated sample sizes are only 11.1-95.2 per cent of the true requirements.

Cross sectional survey of socioeconomic variations in severity and mechanism of childhood injuries in Trent 1992-7

OBJECTIVE

To determine the relation between morbidity from injury and deprivation for different levels of injury severity and for different injury mechanisms for children aged 0-14 years.

DESIGN

Cross sectional survey of routinely collected hospital admission data for injury 1992-7.

SETTING

862 electoral wards in Trent Region.

SUBJECTS

21 587 injury related hospital admissions for children aged 0-4 years and 35 042 admissions for children aged 5-14.

MAIN OUTCOME MEASURES

Rate ratios for hospital admission for all injuries, all injuries involving long bone fracture, and all injuries involving long bone fracture requiring an operation; rate ratios for hospital admission for six types of injury mechanism divided by quintiles of the electoral wards' Townsend scores for deprivation. Rate ratios calculated by Poisson regression, with adjustment for distance from nearest hospital admitting patients with injuries, rurality, ethnicity, and percentage of males in each electoral ward.

RESULTS

Both total number of admissions for injury and admissions for injuries of higher severity increased with increasing socioeconomic deprivation. These gradients were more marked for 0-4 year old children than 5-14 year olds. In terms of injury mechanisms, the steepest socioeconomic gradients (where the rate for the fifth of electoral wards with the highest deprivation scores was > or =3 times that of the fifth with the lowest scores) were for pedestrian injuries (adjusted rate ratio 3.65 (95% confidence interval 2.94 to 4.54)), burns and scalds (adjusted rate ratio 3.49 (2.81 to 4.34)), and poisoning (adjusted rate ratio 2.98 (2.65 to 3.34)).

CONCLUSION

There are steep socioeconomic gradients for injury morbidity including the most common mechanisms of injury. This has implications for targeting injury prevention interventions and resources.

Antimicrobial use and antimicrobial resistance: a population perspective

The need to stem the growing problem of antimicrobial resistance has prompted multiple, sometimes conflicting, calls for changes in the use of antimicrobial agents. One source of disagreement concerns the major mechanisms by which antibiotics select resistant strains. For infections like tuberculosis, in which resistance can emerge in treated hosts through mutation, prevention of antimicrobial resistance in individual hosts is a primary method of preventing the spread of resistant organisms in the community. By contrast, for many other important resistant pathogens, such as penicillin-resistant Streptococcus pneumoniae, methicillin-resistant Staphylococcus aureus, and vancomycin-resistant Enterococcus faecium resistance is mediated by the acquisition of genes or gene fragments by horizontal transfer; resistance in the treated host is a relatively rare event. For these organisms, indirect, population-level mechanisms of selection account for the increase in the prevalence of resistance. These mechanisms can operate even when treatment has a modest, or even negative, effect on an individual host's colonization with resistant organisms.

Geographical variation in prevalence of hypertension within The Gambia

Hypertension has become an important public health problem for sub-Sahara Africa. In a previous nationwide study, we observed a high degree of geographical variation in the prevalence of diastolic hypertension. Geographical variation provides essential background information for the development of community randomised trials could suggest aetiological mechanisms, inform control strategies and prompt further research questions. We designed a follow-up study from the nine high-prevalence communities, and from 18 communities where hypertension was found least prevalent (controls). In each community, 50 households were randomly selected. In each household, an (unrelated) man and woman were enrolled. The risk for hypertension (blood pressure > or =160/95 mm Hg) was higher in the high prevalence communities compared to the control villages (adjusted OR = 1.7, 95% CI 1.3-2.2). The observed coefficient of variation in hypertension prevalence, k, was 0.30. Thus we confirmed significant geographical variation in prevalence of hypertension over time, which has implications for planning of interventions.

Do single handed practices offer poorer care? Cross sectional survey of processes and outcomes

OBJECTIVES

To determine whether there are important differences in performance between group practices and singlehanded general practitioners and the extent to which any differences are explained by practice characteristics such as deprivation.

DESIGN

Cross sectional survey.

SETTING

206 singlehanded practices and 606 partnerships in Trent region, United Kingdom.

METHOD

Comparison of process and outcome measures derived from routinely collected data on hospital admissions and target payments for singlehanded practices and partnerships. Multivariate analysis was used to adjust for the confounding effects of general practice characteristics-deprivation (Townsend score), percentage of Asian residents, percentage of black residents, proportion of men over 75 years, proportion of women over 75 years, rurality, presence of a female general practitioner, and vocational training status.

RESULTS

Differences in achievement of immunisation and cytology targets apparent on univariate analysis were not seen after adjustment for other general practice characteristics. Similarly, significant differences (>15%; P<0.01) for three types of hospital admission seen on univariate analysis were not present after adjustment for other practice characteristics.

CONCLUSIONS

This study provides no evidence that singlehanded general practitioners are underperforming clinically. Our results offer insight into the structural difference between the two types of practice and underline the importance of the effect of other practice characteristics on process and outcome measures.

The effectiveness of interventions aimed at increasing handwashing in healthcare workers - a systematic review

Handwashing is widely accepted as being key to the prevention of hospital-acquired infection but the frequency of handwashing by healthcare workers has been found to be low. A systematic critical literature review was conducted to establish the effectiveness of interventions aimed at increasing compliance with handwashing in healthcare workers. The results showed that one-off educational interventions have a very short-term influence on handwashing behaviour. Use of strategically placed reminders, or asking patients to remind staff of the need to conduct handwashing can have a modest but more sustained effect. Feedback of performance can increase levels of handwashing but if feedback is not repeated regularly, then this effect is not maintained over long periods. Automated sinks increase the quality of handwashing but healthcare workers can be discouraged from using these because of the additional time involved. Provision of moisturized soaps appears to make little difference to handwashing behaviour but providing 'dry' hand rubs near patient beds may lead to a minimal increase in the frequency with which staff decontaminate their hands. Multifaceted approaches which combine education with written material, reminders and continued feedback of performance can have an important effect on handwashing compliance and rates of hospital-acquired infection.

Vaccine trials

This article reviews some of the issues involved in evaluating vaccines in humans. Vaccine trials are required for licensure and are essential for demonstrating a vaccine's safety and protective efficacy. The formal framework of phase I, II, and III trials is described, with particular emphasis on the choice of hypotheses, trial design, and biases that arise in the context of vaccine trials. However, some aspects of a vaccine's performance cannot be evaluated in clinical trials owing to their relatively small size. Thus, vaccine evaluation must continue after licensure, for example, to evaluate the vaccine with respect to rare reactions, duration of protection, and ecological effects. The article reviews some of the methods commonly used for post-licensure studies of vaccine efficacy and safety.