Table 1 Factors affecting children’s right to health in cluster randomized trials in Kenya, January 2002 to February 2014
Author, Year of publication | Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | |
|---|---|---|---|---|---|---|---|---|---|
Intervention arm | Control arm | ||||||||
Ayieko et al., 2011 [25] | Reported reasons; logistically convenient and intervention targeted groups (pediatric teams) | Both interventions based on the same evidence-based best-practice guidelines. | Partial intervention: | All interventions above usual Government approach and practice, which is to adapt and distribute printed materials to staff and provide ad hoc training and opportunistic seminars | Intervention design based on 2004 study assessing delivery of care at first level referral hospitals in Kenya as well as theories from five publications suggesting approaches for improving pediatric care. Baseline data and several reports from Low Income | Outcome measures including 14 process and assessment tasks, and structure indicators. Most, not all indicators showed improvement, to varying degrees. For instance, higher completion of admission task, mean 0.94 versus 0.65, adjusted difference 0.54 (95% CI 0.05 to 0.29) | Conclusion: multifaceted intervention was effective but study advised against scaling up without further work on cost benefit analysis and because the system and resources may not immediately support intervention. For instance, no current role for facilitators or training program. | Guidelines are still available in public domain. | Strategy developed in partnership with the government |
Full intervention: | 6 months surveys with only written feedback, provision of guidelines, job aides and 1.5 day seminar for 40 hospital staff. | None of the hospitals had explicit procedure for taking up new clinical care guidelines | Countries show low uptake of guidelines despite their availability slow uptake of guidelines common, sometimes linked to organization culture and change | Adoption of once daily gentamicin 89.2% versus 74.4 %; 17.1 (8.04 to 26.1%); loading dose quinine (91.9% versus 66.7%, 26.3 % (-3.66 to 56.3%); adequate prescriptions of intravenous fluids for severe dehydration 67.2 versus 40.6%; 29.9% (10.9 to 48.9%). Children receiving inappropriate drug doses lower in intervention hospitals quinine dose > 40%mg/kg/day (1.0% versus 7.5%; -6.5% (12.9% to 0.20%)). inadequate gentamicin dose (2.2% versus 9.0%; - 6.8% (-11.9% to -1.6%) | No changes announced on government approach. | Ministry of Health and Public Sanitation, district hospitals and pediatric teams involved in implementation | |||
a) three hospital assessments or survey done regularly every six months, | Similar job aids provided including a structured admission record to replace free standard notes | Admission records written as non-standard free text | |||||||
b) Written and face to face feedback, | Other studies show financial incentives improve malaria specific outcome and implementation of guidelines improves pediatric emergency triage assessment, hospital care and outcomes for severe malnutrition. | ||||||||
c) three to five days of training aimed at hospital workers of all cadres, provision of guidelines, job aides, | Used alternative interventions (that is no placebo) for ethical reasons and because of uncertainty in literature about effect of multicomponent interventions to implement guidelines | ||||||||
d) An external supervisory process every two to three months | |||||||||
e) Identification of full time local facilitator to promote guidelines and onsite problem solving. One job aide for both groups was a structured admission record to replace previous free standing notes | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post-trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Brooker et al., 2010 [26] | Reported reasons: school clusters avoided contamination. | Arm 3 received both interventions | Arm 2) received only the literacy intervention, | Several unpublished surveys show 50% prevalence in malaria and 21 to 38% anemia in schoolchildren in region | Seven studies show school based intervention improves student achievement but few studies assess multifaceted interventions. Sri Lankan RCT showed malaria treatment improved exam scores. Previous Kenyan cluster trial showed reduced anemia and sustained attention but no effect on educational achievement. Multicentre study in eight sub Saharan African and Asian countries showed 50% of children in developing countries suffer from anemia. One systematic review showed anemia affected cognitive ability. Little experimental data for hypothesized causal chain: malaria prevention - reduced anemia - sustained attention - educational achievement | Of 88.3% children in intervention schools screened at baseline and follow-up. 17.5% tested positive | IST as implemented had no effect on health or education | In original protocol, schools in control arm were to receive literacy and malaria intervention after two years at end of study | The drug AL used in intervention identified through stakeholder consultations |
Inferred reasons: logistic convenience. | a) intermittent screening and treatment including rapid diagnostic testing every school term, AL treatment for positive tests | Arm 1) received only the malaria intervention | |||||||
Study sought information on interactive effects of two interventions. | b) Literacy enhancement, training and support of class one teachers, teacher manual, workshops including weekly interactive texts and monthly written communiqués | Arm 4) did not receive any intervention | No school based malaria program. Educational achievement very low in region- district with worst mean scores in Kenya Certificate of Primary Education- national primary school matriculation exam and few children proceeding to secondary school. | 81.8% in control and 83.0% in intervention group followed | Possible reasons: variable testing, reinfection, and geographical heterogeneity in transmission | However, researchers’ communication that final results were not positive, thus intervention will not be rolled out | Public randomization event involving community, stakeholders and researchers. Testing and treatment done by district health workers supported by the division of malaria control, MoHPS | ||
Both interventions school based intermittent malaria screening and treatment, and enhanced literacy instruction, required group delivery or use | |||||||||
No impact of malaria IST intervention observed on prevalence of anemia or malaria at either 12 or 24 months or on scores of classroom attention. | |||||||||
No effect of IST observed on educational achievement in older class, but an apparent negative effect noted on spelling scores in the younger class at 9 and 24 months and on arithmetic scores at 24 months | |||||||||
In 2009, GoK changed drug policy for malaria from quinine-based therapy to artemisinin combination therapy (ACT) | |||||||||
Rationale for cluster randomization design | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Clarke et al., 2008 [27] | Not reported. | Interventions included: i) school based malaria prevention treatment- sulfadoxine pyrimethamine and amodiaquinine tablets 3 times a year (IPT) | Children received placebo tablets- no effect on malaria parasitemia, similar in size and shape but different in taste to intervention tablets | No information on school based programs malaria prevention and treatment of malaria. | Sri Lankan RCT showed school performance is related to cumulative effect of malaria attacks and reported weekly malaria treatment improved exam scores | Malaria prevalence in IPT group 6.3% and in placebo 12.6% adjusted risk ratio (0.51, 95% CI 0.29 to 0.93 P = 0.28). | School based IPT reduced prevalence of anemia and asymptomatic malaria in semi immune children and improved attention for school children but had no discernable effect on educational performance | School intervention not implemented | Primary schools and schoolteachers implemented study |
Inferred reasons: intervention delivered through a group-school | |||||||||
No Information on malaria and anemia: 2 studies showed 50% of children in region have asymptomatic malaria | |||||||||
Soon after this study implementation and before publications, GoK policy on first line malaria treatment changed from chloroquinine based drugs to ACT drugs | One review and one publication show lack of evidence to show malaria affects cognition and performance in school children. Three other African studies showed treating malaria in schoolchildren improved their cognitive abilities | Protective effect of IPT on anemia 48% (95% CI 8 to71, P = 0.028), Plasmodium prevalence lower in intervention than in placebo group 89% protective efficacy with 95% CI 73 to 95, P = 0.0002. Adjusted risk difference 0.35 (0.30-0.41). Cognitive tests and sustained attention | Effect on anemia was comparable or larger than effects produced through anti-helminthic or iron supplementation. | Government policy for malaria treatment changed and phased out the drug used for IPT in study | |||||
Logistical convenience | ii) passive surveillance of schools and local health systems for adverse reactions | Control group received items ii) to viii) | Study provided evidence of link between asymptomatic malaria and cognition | ||||||
One report and one study say malaria chemoprophylaxis given to school children was associated with lower rates of malaria parasitemia and severe anemia, fewer clinical attacks and malaria deaths and reduced malaria related school absenteeism. | scores higher in IPT group | ||||||||
More work needed to replicate effect in different epidemiological settings, on operational feasibility, and on any long-term educational benefits of this approach | |||||||||
iii) one-day teacher training in social studies | Counting test adjusted mean difference 1.80 (0.19 to 3.41; P = 0.03), effect size 1.22 (95% CI 0.003 to 0.35) Code transmission test adjusted mean difference 6.05 (2.83 to 9.27; P = 0.0007). | ||||||||
iv) Finger prick blood samples; v) stool samples; vi) cognitive tests administered by psychologist; vii) social studies classroom tests administered by teachers; viii) end-of-term pupil assessment in social studies | No mean difference noted in hyper activity test or in educational achievement test | ||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence off efficacy and effectiveness | Results | Conclusions | Access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Crump et al., 2005 [28] | Not reported. | Household based treatment of drinking water with flocculant disinfectant involving technology that removes suspended matter | Comparison group received Intervention 2: sodium hypochlorite solution | Source waters in study area have high fecal contamination and turbid water. | Guatemala RCT showed flocculant disinfection, (incorporating chemical used in municipal water treatment) can be used in household based water treatment | Flocculant disinfectant associated with significant reduction in diarrhea for children under two years old compared to local treatment practices | Further research needed to better understand association between water treatment and reduction in mortality | No reported provisions or arrangements for control arm or for all other arms after study | Not reported |
Inferred reasons: | Control group received intervention 3: Control - usual water collection treatment and storage, the use of safe storage containers not promoted | No health guideline for turbidity but WHO policy is five TBU for drinking water. | Bolivian, Zambian, Guatemalan and Uzbekistani studies show household-based water treatment reduces diarrheal diseases by 20 to 8%. | No significant difference in prevalence of diarrhea between interventions in all ages | |||||
intervention (a small water treatment facility) designed for use by households | |||||||||
Reduced contamination and | Study area-rural western Kenya had high turbidity ranging from 100 to 1000 TBU | Kenya study showed sodium hypochlorite less effective compared to flocculant disinfectant in highly turbid water and other study showed intervention less effective for pathogens resistant to chlorine | Data pooled from both arms showed a significant reduction in mortality in children < two years. | ||||||
logistically convenient | |||||||||
Among people who have highly turbid source water, flocculant disinfectant can provide water that looks cleaner and reduces risk of diarrhea | |||||||||
Study reported most families collected water from ponds, rivers and springs using plastic containers and stored water in clay pots Families either treated water by | |||||||||
In children < two years, absolute difference in diarrhea prevalence compared to control: -25% in flocculant arm (95% CI -40 to -5) and -17% in sodium hypochlorite arm | |||||||||
decanting or cloth filtration, boiling, or chemical treatment of water not common | |||||||||
Effect in all age groups, absolute difference compared to control -19% in flocculant arm (-34 to 4) and -26% in sodium hypochlorite arm (-39 to -9). Fewer deaths in intervention compounds compared to control, relative risk ratio of death 0.58 | |||||||||
No reported GoK standards for turbidity | P = 0.036 | Sodium hypochlorite was commercially available in the market | |||||||
E. coli concentration < 1 CFU/100ml 14% in intervention arm, 82% in flocculant disinfectant arm and 78% in sodium, hypochlorite arm. Mean turbidity 8 NTU, 55 NTU in other arms (P < 0.001) | At time of study, flocculant disinfectant for household use was not commercially available | ||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Desai et al., 2004 [29] | Not reported. | Arm I: daily supervised iron supplementation | Arm II: daily unsupervised iron supplementation | Study reported lack of standardized guidelines for iron supplementation Previous study reported iron supplementation was not routine for children with mild to moderate anemia and was prescribed in only 12% of children diagnosed with severe anemia who also received presumptive antimalarial treatment. | Several studies showed weekly or biweekly supplementation is as efficacious in prevention and treatment of mild to moderate anemia as conventional daily supplementation despite 30 to 70 % reduction in cumulative dose | In supervised groups hemoglobin concentration was higher in the daily iron group than in the twice-weekly iron group. | For this region, and after initial antimalarial treatment, daily iron supplementation is better than twice weekly iron supplementation for hemoglobin level regardless of whether adherence can be ensured | No information | Not reported |
Inferred reasons: logistically convenient to randomize children by the compounds they lived in, | |||||||||
Mean difference at 6 weeks 4.2 g/L (2.1 versus 6.4) and at 12 weeks 4.4 g/L (1.8 versus7.0) with 95% CI | |||||||||
Arm III: twice weekly supervised iron supplementation | Clinics that prescribed iron for severe anemia use short courses involving two weeks of relatively high doses of iron, which is combined with a presumptive antimalarial treatment | However, a 1999 meta- analysis showed daily iron supplementation compared to intermittent supplementation was more efficacious for pregnant women, the beneficial effect in preschool children is inconclusive and indicated that the degree of supervision affected anemia prevalence. A 1999 study with aboriginal communities showed six weeks of unsupervised twice-weekly iron supplementation superior to unsupervised daily supplementation for increasing hemoglobin in children | |||||||
Among unsupervised groups, hemoglobin concentration did not differ at 6 weeks mean difference 0.86 g/L (-1.4, 3.1) but were higher at 12 weeks for daily iron group, mean difference 3.4 g/L (0.79 versus 6.0) P = 0.02 95% CI | |||||||||
reduced contamination between study arms | Arm IV: twice weekly unsupervised iron supplementation | Additional benefits of participation - health passports issued to all children in trials to allow free healthcare at local clinics and at hospital | |||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Freeman et al., 2012 [30] | Not reported. | Full Intervention only given to Arm 2: hygiene promotion (HP), water treatment supplies (WT), and three-day training of teachers in hygiene promotion (HP), behavior change and water treatment with regular follow-up visits in school year | Arm 1: hygiene promotion and water treatment supplies (hand washing and drinking water containers and one time supply of disinfectant known as water guard). | GoK ratio for latrines in schools is 25:1, one latrine for every 25 girls and 30:1, and 1 latrine for every 30 boys | Three studies show school based hand washing or water treatment in low-income settings reduces school absence by between 21 and 42% | HP and WT intervention no significant impact on absence, OR 0.81, 95% CI 0.50 to 1.35 even when sanitation was added | Improved WASH access did not mitigate absence on boys, | Arm 3 received all interventions after study | Program embedded within a larger research and learning project run by CARE (an international NGO). |
Inferred reasons: school based water treatment, sanitation and hygiene promotion intervention intended for delivery through groups-schools. | |||||||||
Three-day training of teachers on hygiene promotion, water treatment, but no additional latrines | |||||||||
Intervention in Arm 3: control group received no intervention during study, | OR 0.63, CI 0.31 to 1.27. | ||||||||
Schools received hand washing and drinking water containers and one time one-year supply of water guard (1.2% chlorine based water disinfectant promoted by Population Services International (PSI) | One systematic review shows WASH cluster trials in middle-income countries had limitations (small sample sizes/non-adjustment for school level clustering/utilization of non- equivalent group design). Most frequently cited evidence on impact of WASH is from a non-experimental study with 11% reduction in girls’ absence in Bangladesh and included monetary subsidies. | Several factors affected study; heterogeneity among schools, election violence and displacements resulted in schools closing for four months | |||||||
One systematic review, a combined systematic review and meta-analysis, one quantitative systematic review shows WASH at home improves health of children under five years old | |||||||||
Provision of latrines to GoK standard to maximum of seven latrines, not more than seven latrines proved to any one school. | |||||||||
No significant difference between single and multiple WASH interventions | |||||||||
No impact noted on enrollment or test scores | |||||||||
Schools selected for study had to exceed GoK ratio. | |||||||||
All groups received deworming after baseline and once more during study | Only 29% of schools in study area met the government ratio | However, both interventions reduced absence in girls although the study did not know how. | Study area chosen through rapid assessment by study partners, study partners not named likely includes the Kenya National Bureau of Statistics. | ||||||
WASH/HP and WT alone reduced the odds of 2 week school absence in girls by 58% (OR 0.42, 95% CI 0.21 to 0.085 but no effect on boys (OR 0.88, 0.45 to 1.71). | |||||||||
WASH, HP and WT and sanitation showed comparable benefit for girls (OR 0.47, .21 to 1.05) but no benefit for boys (OR 0.98 95% 0.52 to 1.87). | |||||||||
For HP and WT 0.34 days of absence were avoided per pupil per two week recall period and | |||||||||
For HP, WT and sanitation 0.38 days | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusion | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Gewa et al., 2013 [31] | Not reported. | Arm 1: vegetarian | Arm 4: control group no food supplementation | Not reported. | Guatemalan study showed that in household based programs, foods can be redistributed to non-targeted household members, and studies in Bangladesh, India and Philippines showed daily energy intake rose about 75 to 100% on school feeding days. No studies measuring effect of school based feeding program on distribution of food in households | Children in meat group had increased energy and protein intakes at follow- up but these changes were not significantly different from changes experienced by control group. Children in vegetarian, milk and control group did not experience significant changes in energy and protein intake. | No evidence that school children who received supplementary feeding at school experienced reduced intakes at home or that food intake for other family members was increased at the expense of school children | Not reported | Project implemented by Child Nutrition Program in turn supported by a grant by United States Agency for International Development (USAID) |
Inferred reasons: cluster randomization logistically convenient and reduced contamination between arms | |||||||||
Intervention intended for delivery through groups-schools | |||||||||
A serving of maize and bean (gither i) | No official school feeding programs to cover all primary school going children but there may have been independently sponsored programs by parties such as churches, self help groups, parent-groups or non-governmental organizations (NGOs) | ||||||||
Parents: significant decline in protein intake reported for parents of children in vegetarian group. Other three groups did not experience significant decline | |||||||||
Arm 2: githeri and meat serving Arm 3: milk and githeri dish | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusion | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Kangwana et al., 2011 [32] | Not reported. | Retail outlets in sub locations in intervention arm received subsidized pediatric Artemether Lumenfantrine (AL) and one day training for retail staff. | No intervention provided, retailers did not receive subsidized AL | AL first line recommended treatment for uncomplicated malaria, available only through prescriptions and at no cost in public facilities | Pilot programs in Tanzania and Uganda show rapid uptake of subsidized ACT, decreased use of old regimens and good adherence to prices. Senegal study showed irregular availability, Cambodia study showed high prices. Multi country study using community distributors showed increased uptake of AL | Percentage of children receiving AL in same or following day in intervention arm was 35.9%, higher than those receiving AL in the control arm 14% by percentage difference of 25% points 95% CI, P = 0.0002, adjusted P = 0.0001. | While increase in prompt ACT coverage rose to 44.9% it is still well below roll back malaria targets, and more work is needed to understand ACT subsidy mechanisms and effects, on effect of enhanced diagnosis and cost and effectiveness studies on public sector strategies versus community driven approaches | Study did not roll out intervention. | Pharmacy and Poisons Board of Kenya (PPB), Population Services International staff collected complete referral forms from retail outlets. PSI is an international NGO with presence in many developing countries |
In 2006 to 2007 the government conducted AL awareness campaigns, so both arms received some general info on current malaria policy | |||||||||
Community awareness: nine community leader workshops, ten small group discussions targeting two hundred people and outreaches by community based organization targeted 21,000 people, branded t-shirts, pens and scarves given to community, shops given posters or painted for advertisement | |||||||||
In practice often sold without prescription, private sector retail price 500 Kenyan shillings or 6.16 USD | |||||||||
Tibamal, the intervention drug, constituted 63% of all ACT used. | |||||||||
Old antimalarial regime was still being sold at 0.37 USD on average. | |||||||||
At time of study, ACT subsidy mechanism AMF-m was in development Global Fund to make direct payments to prescribed manufactures to lower import cost to public and private retailers | |||||||||
Inferred reasons: | Frequents stock outs of AL reported in public facilities. | 93 % (SD 5.9%) of Tibamal was bought at recommended retail price 0.25 USD no significant difference in accuracy of dosage was recorded between intervention and control group | Study highlighted likely differences between this intervention and the roll back malaria (RBM) program where AL is usually availed through prescriptions and in health facilities. | Increased coverage of AL achieved through other means, at follow -up: 65% of GoK health facilities stocked AL. | Implemented by Division of Malaria, MoHPS and PPB | ||||
to avoid contamination | WHO report: only 15% of children with fever < 5 years in malaria endemic Africa promptly received Artemisinin- based combination therapy (ACT) | Over 30% accessed AL through retail sector but | |||||||
among participants. | increased public sector coverage noted during study. | ||||||||
Involved group level delivery of intervention | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusion | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Opondo et al., 2011 [33] | Not reported. | Intervention comprised of multifaceted implementation of MoH adapted guidelines to reduce inappropriate use of antibiotics: training on guidelines, supervision, and support, including face-to-face feedback. | Control arm received partial intervention which included adapted guidelines, didactic training, | Guidelines are usually adapted by MoH and sent to hospitals in written format, or charts, ad hoc training is provided. | Guideline well established and recommended by Integrated Management of Childhood Illness and WHO guidelines. | Multifaceted implementation of guidelines reduced the inappropriate use of antibiotics in non-bloody diarrhea. | Intervention to reduce inappropriate use of antibiotics when nested within a large integrated approach will not only significantly reduce inappropriate use but also improve pediatric care in other areas. Intervention effects may cut across other areas in care. | No information provided but parent study did not roll out intervention. | Study based on analysis of data previously collected by Ayieko et al., 2011 which examined uptake of 14 point pediatric admission guidelines for rural district hospitals. |
Guidelines for diarrhea alone include taking history, assessing the patients for shock and dehydration, classification of severity of dehydration and appropriate rehydration therapy as well as other four points. | |||||||||
Inferred reasons: study emphasized intervention is intended for hospital level, logistically convenient and group effects useful | Written feedback, facilitation and supervision | ||||||||
9,549 admission records reviewed, 4,232 diagnosed with diarrhea and 1,160 diagnosed with non-bloody diarrhea. 750 children received antibiotics inappropriately, 313 in intervention hospitals and 437 in intervention hospitals. Odds of children receiving inappropriate antibiotics for non-bloody diarrhea in intervention hospitals was 0.30 that in control hospitals (95% CI 0.09 to 1.02) | Study implemented in district hospitals, Staff involved, cooperation from MoH, MoHPS | ||||||||
No information provided on adherence to or uptake of guideline | |||||||||
(95% CI 0.09 to 1.02) | |||||||||
Guidelines: for patients with diarrhea and non-bloody stool, manage with fluids and for patients with bloody stool, give antibiotics | |||||||||
Botswana study shows bloody diarrhea is often caused by Shigella and made worse by antibiotics. One study showed antibiotic misuse increases antibiotic resistance. Botswana study showed non-adherence to guideline is common. | |||||||||
Literature from Niger, South America, Pakistan and Armenia show implementation of diarrhea management guideline difficult | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusion | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Patel et al., 2012 [34] | Not reported | Thirty villages | Thirty villages 405 students 331 households 22 schools | Study did not publish information on standard of care. Many primary schools in the area may not have access to treated water and study did not provide information on government arrangements to provide soap or water in public primary schools. There may be some hygiene education within primary school curriculum. | Two systematic reviews showed handwashing with soap reduced risk to both illnesses and household based water treatment also shown to reduce diarrhea risk. | Mixed findings: percentage of students who could demonstrate proper handwashing technique was higher in intervention arm at mid point (46% versus 14% EDM 32%, 90% CI 10 to 46%), but similar at second follow-up (54% versus 50% EDM 0% 90% CI -17 to 11%). | Handwashing stations and hygiene education may have contributed to reduced rates of illness in primary school children but changes did not result to reduced diarrhea rates, study authors attributed this to the effect of other interventions being run in that region | One year later intervention rolled out to control schools | Project situated within a wider Nyando Integrated child health and education program in the region |
Inferred reasons: logistically convenient to randomize by schools, reduced contamination and measurements of group effects were useful | |||||||||
378 students, | At first round reduction in percentage of students with any illness (5% versus 7% EDM -3%, 90% CI 4 to 1%) and with ARI (2% versus 3%, EDM -2%, 90% CI -3% to -1%) but similar rates at second round for ARI (0.8% versus 0.7%, EDM 0%, 90% CI -1 to 1%) or any illness (3% versus 2%, EDM 1%, 90 CI -1 to 1%). | ||||||||
Studies in Ghana, Oudomxay and western Kenya showed school based programs increase hygiene knowledge among students and changed practices. Four studies in Kenya and one systematic review showed WASH programs reduce absenteeism but studies have not shown a direct impact on health. | 'No difference noted in diarrheal illness between intervention and comparison arm during either year of surveillance (year I: 0% versus 0.3% EDM (0% 90 CI 0 to 0%; year II 0% versus 0% EDM 90% CI 0 to 0%) | ||||||||
312 households | |||||||||
21 schools. | |||||||||
Teachers in intervention schools received handwashing and water treatment training and teachers provided instructional material for students. Schools received drinking water stations and handwashing stations as well as three month starter supply of soap and water treatment solution. Home visits and school visits in both arms to collect data | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusion | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Philips-Howard et al., 2003 [35] | Not reported | 113 households were given | No bed-nets for duration of study | Not reported | RCTs in Ghana, Gambia, Kenya and Burkina Faso showed insecticide treated bed- nets to be effective in reducing all cause mortality in children under five years by 17% | Crude mortality rates were 51.9 in control village compared to 43 in intervention villages per 1,000. The protective efficacy with 95% CI was 16% (6 to 25%). | Treated nets prevent approximately one out of four deaths in areas with high perennial malaria transmission but are not as effective if retreatment is delayed beyond six months | Villages in control arm received intervention after study | Study funded by USAID |
Inferred reasons: study wished to measure effect of the intervention on the group. | |||||||||
Intervention intended for public health level administration, | Protective efficacy of retreated nets overall 20% (10 to 29%), and 26% (12 to 37%) in children aged between one and eleven months old children and 14% (-1 to 26%) in children aged between twelve and fifty-nine months old | ||||||||
Study conducted in between 1997 to 1999, did not report information on access to bed-nets for young children within national malaria program, but treated bed-nets may not have been part of malaria program | |||||||||
A large-scale social marketing program in Tanzania estimated a 27% increase in survival of children under five but there is scientific disagreement whether treated bed-nets reduce mortality in very young children in areas with high transmission pressure | |||||||||
Treated bed-nets to cover sleeping space and the nets retreated after six months | |||||||||
logistically convenient, | |||||||||
reduced likelihood of contamination among participants | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Skarbinski et al., 2009 [36] | Not reported | Study provided rapid diagnostic tests (RDTs) to all health facilities | Training: one-off three day training to all health facilities. | GoK to implement new malaria policy between April and October 2006: deliver AL to all health facilities, provide training on new malaria guidelines and on use of AL and diagnostic tests at least one per facility | Three studies in Zambia and one study in Kenya showed health workers under-prescribed AL and over- prescribed other antimalarial drugs | Treatment: RDTs significantly reduced recommended treatment by 63% (P = 0.04) | Providing RDTs significantly reduced recommended treatment without reducing overtreatment | No arrangement in protocol to provide RDTs to comparison arm | No information. |
Inferred reasons: intervention, rapid diagnostic tests, intended for groups - pediatric teams and health facilities receiving, diagnosing and treating pediatric malaria | |||||||||
At baseline, 25% of health facilities in study region received RDTs from GoK | |||||||||
Artemether Lumenfantrine | TGS alone increased recommended treatment by 41% (P = 0.05). | Increase independent of trial. By time of study implementation, 50% of patients in each arm were seen in facilities where RDT was available | |||||||
Neither the RDTs or microscopy significantly reduced overtreatment with non-recommended antimalarial (prescribed old antimalarial for patients who tested positive) | |||||||||
Drug SOC: | AL - new but expensive first line treatment for un- complicated malaria | Testing: in intervention arm use of RDTs increased from 35% to 46%, use of blood slides decreased from 38% to 8% but overall use of tests (RDTs or microscopy) did not change | |||||||
At time of study most facilities received AL but routine in service training incomplete. | |||||||||
Testing SOC: limited lab capacity & only 25% of health facilities had received RDTS through the Government | June 2006 - GoK introduced AL as recommended first line treatment of uncomplicated malaria, revised national malaria treatment guidelines and a new treatment algorithm | ||||||||
Training: one-off three day training to all health facilities. | A copy of the revised national malaria treatment guidelines, | Adherence to treatment guidelines: RDT significantly reduced use of clinical diagnosis of malaria to prescribe AL by 36% (P = 0.03) | Because RDT use replaced microscopy and provision of RDT reduced proportion of AL prescriptions for patients who were not tested | Implemented in government health facilities- hospitals, health centers, dispensaries, | |||||
Artemether Lumenfantrine | Supervision (TGS) | More work is needed on health worker approach to RDT use and antimalarial use. Several reasons put forward by study: conflicting messages in TGS | |||||||
A copy of the revised national malaria treatment guidelines | |||||||||
Supervision | |||||||||
(TGS) | |||||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Suchdev et al., 2012 [37] | Not reported. | In intervention villages, sprinkles sachets (branded MNPs) socially marketed and sold to households with children 6 to 59 months by women trained by safe water and AIDS project (SWAP) | Control arm: no MNPs sold. Other SWAP products such as soap, insecticide treated nets, water disinfectant and condoms sold in all arms | Public health authorities distribute vitamin A in health facilities | One Cochrane review, one systematic review and meta -analysis show MNPs reduce anemia, have higher uptake and fewer side effects than iron drops | 27.2% absolute reduction and 40.9% relative reduction in prevalence of anemia in intervention group and 20.1% absolute and 29.9% relative reduction in control group and P = 0.10 | Community based distribution of MNPs sprinkles improved recovery rates for anemia and some measures of iron and vitamin A deficiency even where children received less than recommended doses of iron | After follow-up survey, community based distribution of MNPs expanded to cover control villages | |
Inferred reasons: intervention (micronutrients powders-MNPs) delivered through a group mechanism - a community program that that trained women to distribute MNPs. | |||||||||
Decrease in prevalence of iron deficiency by 19.3% in intervention group compared to 5.3% in control, P = 0.001 | |||||||||
Study a pilot as sponsors and partners intended to increase coverage within the wider community project | |||||||||
Vitamin A deficiency decreased by 7.5% in intervention and 2.5% increase in control (P = 0.01) | |||||||||
One double blind placebo-control trial compared doses and offered free MNPs found anemia recovery rate of 53% | |||||||||
MNPs also supplied by authorities, oral dosing may also be used | |||||||||
Cost two Kenya shillings per sachet (approximately 2.7 US cents at the time) | Government of Kenya (GoK) changed policy in 2007 from mass distribution of vitamin A with 80% cover in study region, to provision in health facilities resulting to 22% cover | Bangladesh, Mongolia and Bolivia have national program, other countries planning large-scale programs | Sprinkles global health initiative provided MNPs | ||||||
One community based cluster randomized cluster trial showed many disadvantages and low adherence to prophylactic micronutrient supplements | Study had recovery rates similar to previous clinical trials where MNPs distributed at no cost, thus approach is potentially self-sustaining | SWAP: Safe Water and AIDS Program trained the vendors who were women from the community | |||||||
Rationale for cluster randomization | Intervention | Standard of care pre trial | Evidence of efficacy and effectiveness | Results | Conclusions | Post trial access to intervention | Other involvement | ||
Intervention arm | Control arm | ||||||||
Zurovac et al., 2011 [38] | Not reported. | Between Monday to Friday, one-way text messages about malaria case management sent to health workers’ personal mobile phones for period of six months | No intervention provided. | No previous information on health workers’ adherence. Studies in Angola, Uganda and Tanzania and one publication on sub Saharan Africa found despite simple guidelines for managing malaria in children with fevers, there is little health worker adherence to case management, prescription, dosing, counseling or dispensing practices | Researchers conducted a systematic review and assessed strategies to improve health worker performance and found an overall median improvement of 9% (IQR 3 to 23%) | When all 4 treatment tasks and 5 of 6 dispensing and counseling tasks were measured, improvement of 21.4% points in short term (95% CI 9.0 to 33.7, P = 0.0007), and 23.7% (11.6 to 35.7, P = 0.0001) after six months. | Intervention worked just as well and better than more complex interventions, and achieved modest adherence of 9 % | No reports of arrangements | Division of Malaria at Ministry of Health and Public Sanitation assisted in intervention development |
Inferred reasons: intervention targeted health workers in health facilities, | |||||||||
Messages extracted from national guidelines and training manuals | At baseline, study found 9 % adherence to treatment guidelines for outpatient pediatric malaria | ||||||||
logistical convenience | |||||||||
Effect size smaller when study measured all 4 treatment tasks and all 10 dispensing and counseling tasks at short term 10.3% (4.0 to 16.6, P = 0.0013) at 6 months and 11.3% (5.1 to 17.6, P = 0.0004) at 6 months | |||||||||
Two Kenyan RCTS show concept works, texts sent from health | |||||||||
Texts included short motivating/entertaining quotes | Surveys conducted at baseline and follow-up to measure adherence to treatment guidelines. | workers to HIV/AIDS patients improves patient adherence to treatment regimen | Study did not understand why/how text messaging health workers worked | Study undertaking a cost and operational assessment of their trial with the government looking at how to replicate on national scale | |||||
Cost of single text 0.01 USD, full exposure in intervention cost 2.96 per health worker | All arms received three rounds of malaria case management training sessions for health workers as well as national guidelines, drug management wall charts for ACT | Kenyan RCT using in-service training and passive job aide distribution to improve health workers’ malaria case management showed slight improvement | More work on cost effectiveness of study. | ||||||
Two RCTS in Angola and Uganda investigated texts to health workers and looked at short-term effects | Despite low cost of texts more funds may be needed to establish and maintain the system. | ||||||||
Researchers recommended intervention to supplement other adherence strategies | |||||||||