Inconsistency in Diarrhea Measurements when Assessing Intervention Impact in a Non-Blinded Cluster-Randomized Controlled Trial

To explore the consistency in impact evaluation based on reported diarrhea, we compared diarrhea data collected through two different surveys and with observed diarrhea-associated hospitalization for children aged ≤ 5 years from a non-blinded cluster-randomized trial conducted over 2 years in urban Dhaka. We have previously reported that the interventions did not reduce diarrhea-associated hospitalization for children aged ≤ 5 years in this trial. We randomly allocated 90 geographic clusters comprising > 60,000 low-income households into three groups: cholera vaccine only, vaccine plus behavior change (cholera vaccine and handwashing plus drinking water chlorination promotion), and control. We calculated reported diarrhea prevalence within the last 2 days using data collected from two different survey methods. The "census" data were collected from each household every 6 months for updating household demographic information. The "monthly survey" data were collected every month from a subset of randomly selected study households for monitoring the uptake of behavior change interventions. We used binomial regression with a logarithmic link accounting for clustering to compare diarrhea prevalence across intervention and control groups separately for both census and monthly survey data. No intervention impact was detected in the census (vaccine only versus control: 2.32% versus 2.53%; P = 0.49; vaccine plus behavior change versus control: 2.44% versus 2.53%; P = 0.78) or in the vaccine only versus control in the monthly survey (3.39% versus 3.80%; P = 0.69). However, diarrhea prevalence was lower in the vaccine-plus-behavior-change group than control in the monthly survey (2.08% versus 3.80%; P = 0.02). Although the reasons for different observed treatment effects in the census and monthly survey data in this study are unclear, these findings emphasize the importance of assessing objective outcomes along with reported outcomes from non-blinded trials.

Impact of a Large-Scale Handwashing Intervention on Reported Respiratory Illness: Findings from a Cluster-Randomized Controlled Trial

We assessed the impact of handwashing promotion on reported respiratory illness as a secondary outcome from among > 60,000 low-income households enrolled in a cluster-randomized trial conducted in Bangladesh. Ninety geographic clusters were randomly allocated into three groups: cholera-vaccine-only; vaccine-plus-behavior-change (handwashing promotion and drinking water chlorination); and control. Data on respiratory illness (fever plus either cough or nasal congestion or breathing difficulty within previous 2 days) and intervention uptake (presence of soap and water at handwashing station) were collected through monthly surveys conducted among a different subset of randomly selected households during the intervention period. We determined respiratory illness prevalence across groups and used log-binomial regression to examine the association between respiratory illness and presence of soap and water in the handwashing station. Results were adjusted for age, gender, wealth, and cluster-randomized design. The vaccine-plus-behavior-change group had more handwashing stations with soap and water present than controls (45% versus 25%; P < 0.001). Reported respiratory illness prevalence was similar across groups (vaccine-plus-behavior-change versus control: 2.8% versus 2.9%; 95% confidence interval [CI]: -0.008, 0.006; P = 0.6; cholera-vaccine-only versus control: 3.0% versus 2.9%; 95% CI: -0.006, 0.009; P = 0.4). Irrespective of intervention assignment, respiratory illness was lower among people who had soap and water present in the handwashing station than among those who did not (risk ratioadjusted: 0.82; 95% CI: 0.69-0.98). With modest uptake of the handwashing intervention, we found no impact of this large-scale intervention on respiratory illness. However, those who actually had a handwashing station with soap and water had less illness. This suggests improving the effectiveness of handwashing promotion in achieving sustained behavior change could result in health benefits.

Impact of adding hand-washing and water disinfection promotion to oral cholera vaccination on diarrhoea-associated hospitalization in Dhaka, Bangladesh: evidence from a cluster randomized control trial

Background

Information on the impact of hygiene interventions on severe outcomes is limited. As a pre-specified secondary outcome of a cluster-randomized controlled trial among >400 000 low-income residents in Dhaka, Bangladesh, we examined the impact of cholera vaccination plus a behaviour change intervention on diarrhoea-associated hospitalization.

Methods

Ninety neighbourhood clusters were randomly allocated into three areas: cholera-vaccine-only; vaccine-plus-behaviour-change (promotion of hand-washing with soap plus drinking water chlorination); and control. Study follow-up continued for 2 years after intervention began. We calculated cluster-adjusted diarrhoea-associated hospitalization rates using data we collected from nearby hospitals, and 6-monthly census data of all trial households.

Results

A total of 429 995 people contributed 500 700 person-years of data (average follow-up 1.13 years). Vaccine coverage was 58% at the start of analysis but continued to drop due to population migration. In the vaccine-plus-behaviour-change area, water plus soap was present at 45% of hand-washing stations; 4% of households had detectable chlorine in stored drinking water. Hospitalization rates were similar across the study areas [events/1000 person-years, 95% confidence interval (CI), cholera-vaccine-only: 9.4 (95% CI: 8.3–10.6); vaccine-plus-behaviour-change: 9.6 (95% CI: 8.3–11.1); control: 9.7 (95% CI: 8.3–11.6)]. Cholera cases accounted for 7% of total number of diarrhoea-associated hospitalizations.

Conclusions

Neither cholera vaccination alone nor cholera vaccination combined with behaviour-change intervention efforts measurably reduced diarrhoea-associated hospitalization in this highly mobile population, during a time when cholera accounted for a small fraction of diarrhoea episodes. Affordable community-level interventions that prevent infection from multiple pathogens by reliably separating faeces from the environment, food and water, with minimal behavioural demands on impoverished communities, remain an important area for research.

Explaining low rates of sustained use of siphon water filter: evidence from follow-up of a randomised controlled trial in Bangladesh

OBJECTIVE

To assess sustained siphon filter usage among a low-income population in Bangladesh and study relevant motivators and barriers.

METHODS

After a randomised control trial in Bangladesh during 2009, 191 households received a siphon water filter along with educational messages. Researchers revisited households after 3 and 6 months to assess filter usage and determine relevant motivators and barriers. Regular users were defined as those who reported using the filter most of the time and were observed to be using the filter at follow-up visits. Integrated behavioural model for water, sanitation and hygiene (IBM-WASH) was used to explain factors associated with regular filter use.

RESULTS

Regular filter usage was 28% at the 3-month follow-up and 21% at the 6-month follow-up. Regular filter users had better quality water at the 6-month, but not at the 3-month visit. Positive predictors of regular filter usage explained through IBM-WASH at both times were willingness to pay >US$1 for filters, and positive attitude towards filter use (technology dimension at individual level); reporting boiling drinking water at baseline (psychosocial dimension at habitual level); and Bengali ethnicity (contextual dimension at individual level). Frequently reported barriers to regular filter use were as follows: considering filter use an additional task, filter breakage and time required for water filtering (technology dimension at individual level).

CONCLUSION

The technological, psychosocial and contextual dimensions of IBM-WASH contributed to understanding the factors related to sustained use of siphon filter. Given the low regular usage rate and the hardware-related problems reported, the contribution of siphon filters to improving water quality in low-income urban communities in Bangladesh is likely to be minimal.

Microbiological evaluation of the efficacy of soapy water to clean hands: a randomized, non-inferiority field trial

We conducted a randomized, non-inferiority field trial in urban Dhaka, Bangladesh among mothers to compare microbial efficacy of soapy water (30 g powdered detergent in 1.5 L water) with bar soap and water alone. Fieldworkers collected hand rinse samples before and after the following washing regimens: scrubbing with soapy water for 15 and 30 seconds; scrubbing with bar soap for 15 and 30 seconds; and scrubbing with water alone for 15 seconds. Soapy water and bar soap removed thermotolerant coliforms similarly after washing for 15 seconds (mean log10 reduction = 0.7 colony-forming units [CFU], P < 0.001 for soapy water; mean log10 reduction = 0.6 CFU, P = 0.001 for bar soap). Increasing scrubbing time to 30 seconds did not improve removal (P > 0.05). Scrubbing hands with water alone also reduced thermotolerant coliforms (mean log10 reduction = 0.3 CFU, P = 0.046) but was less efficacious than scrubbing hands with soapy water. Soapy water is an inexpensive and microbiologically effective cleansing agent to improve handwashing among households with vulnerable children.

Risk factors for community-based reports of gastrointestinal, respiratory, and dermal symptoms: findings from a cohort study in Australia

Background

Although gastrointestinal (GI), respiratory, and dermal symptoms are common, few studies have conducted concurrent and comparative prospective analyses of risk factors for these 3 morbidity outcomes.

Methods

We used data from a community-based randomized controlled trial among 277 South Australian families to analyze GI (diarrhea, vomiting), respiratory (sore throat, runny nose, cough) and dermal (rash, generalized itch, dermal infection) symptoms.

Results

Log-binomial regression analysis revealed similar risks of GI (adjusted risk ratio [RR], 1.65; 95% CI, 1.05–2.58) and respiratory (RR, 1.68; 95% CI, 1.31–2.15) symptoms among childcare/kindergarten attendees. Swimming in public pools/spas in the current or previous week was associated with all 3 symptom complexes, conferring similar risk for each (RR for GI: 1.33; 95% CI, 0.99–1.77; respiratory: 1.20; 95% CI, 1.04–1.38; dermal: 1.41; 95% CI, 1.08–1.85). Pet ownership was not associated with symptoms. Household clustering of GI and respiratory symptoms was common, and clustering of respiratory symptoms correlated with number of individuals per household.

Conclusions

This simultaneous examination of risk factors for 3 health outcomes yielded new comparative data that are useful for developing prevention strategies.

Community based study to compare the incidence and health services utilization pyramid for gastrointestinal, respiratory and dermal symptoms

Background

Gastrointestinal (GI), respiratory and dermal symptoms are common and cause substantial morbidity, although the information on their exact incidence and comparative burden is limited. The aim of this study was to describe the epidemiology and rate these three major symptom complexes in order to improve our understanding of the health burden imposed by these symptoms.

Methods

We used data from a community based randomised control trial conducted from June 2007 to August 2008 among 277 South Australian families consuming rainwater. Using weekly health diaries, we prospectively collected information on GI (diarrhoea or vomiting), respiratory (sore throat, runny nose or cough) and dermal (rash, generalised itch or dermal infection) symptoms, as well as on relevant GP visits, time off work and/or hospitalisation due to these symptoms. Data were analysed using generalized estimating equations approach taking into account the variable number of weeks of follow-up of each individual and within-family clustering of responses.

Results

Over one year, at least one episode of GI symptoms was reported by 54% of participants (95% CI 50%-58%), at least one respiratory episode by 91% (95% CI 88%-93%) and at least one episode of dermal symptoms by 27% (95% CI 24%-30%). The average number of weeks per year during which respiratory symptoms occurred was four times greater than for GI or dermal symptoms (4.9, 1.2 and 1.2 weeks, respectively, p<0.001), with an average number of GP visits per person per year being twice as frequent (0.48, 0.26, 0.19 respectively, p<0.001). However, on a per episode basis, a higher proportion of people saw a GP or were hospitalised for GI symptoms.

Conclusions

This first comparative study of three different symptom complexes showed that although respiratory symptoms are most common, GI symptoms cause a greater per episode burden on healthcare resources. Measuring and comparing the community based burden of these symptom complexes will assist evidence-based allocation of resources.

Learning to dislike safe water products: results from a randomized controlled trial of the effects of direct and peer experience on willingness to pay

Low-cost point-of-use (POU) safe water products have the potential to reduce waterborne illness, but adoption by the global poor remains low. We performed an eight-month randomized trial of four low-cost household water treatment products in Dhaka, Bangladesh. Intervention households (n = 600) received repeated educational messages about the importance of drinking safe water along with consecutive two-month free trials with each of four POU products in random order. Households randomly assigned to the control group (n = 200) did not receive free products or repeated educational messages. Households' willingness to pay for these products was quite low on average (as measured by bids in an incentive-compatible real-money auction), although a modest share was willing to pay the actual or expected retail price for low-cost chlorine-based products. Furthermore, contrary to our hypotheses that both one's own personal experience and the influence of one's peers would increase consumers' willingness to pay, direct experience significantly decreased mean bids by 18-55% for three of the four products and had no discernible effect on the fourth. Neighbor experience also did not increase bids. Widespread dissemination of safe water products is unlikely until we better understand the preferences and aspirations of these at-risk populations.

Inequalities in care-seeking for febrile illness of under-five children in urban Dhaka, Bangladesh

Fever is an easily-recognizable primary sign for many serious childhood infections. In Bangladesh, 31% of children aged less than five years (under-five children) die from serious infections, excluding confirmed acute respiratory infections or diarrhoea. Understanding healthcare-seeking behaviour for children with fever could provide insights on how to reduce this high rate of mortality. Data from a cross-sectional survey in the catchment areas of two tertiary-level paediatric hospitals in Dhaka, Bangladesh, were analyzed to identify the factors associated with the uptake of services from trained healthcare providers for under-five children with reported febrile illness. Health and demographic data were collected in a larger study of 7,865 children using structured questionnaires. Data were selected from 1,290 of these under-five children who were taken to any healthcare provider for febrile illness within two months preceding the date of visit by the study team. Certified doctors were categorized as 'trained', and other healthcare providers were categorized as 'untrained'. Healthcare-seeking behaviours were analyzed in relation to these groups. A wealth index was constructed using principal component analysis to classify the households into socioeconomic groups. The odds ratios for factors associated with healthcare-seeking behaviours were estimated using logistic regression with adjustment for clustering. Forty-one percent of caregivers (n=529) did not seek healthcare from trained healthcare providers. Children from the highest wealth quintile were significantly more likely [odds ratio (OR)=5.6, 95% confidence interval (CI) 3.4-9.2] to be taken to trained healthcare providers compared to the poorest group. Young infants were more likely to be taken to trained healthcare providers compared to the age-group of 4-<5 years (OR=1.6, 95% CI 1.1-2.4). Male children were also more likely to be taken to trained healthcare providers (OR=1.5, 95% CI 1.2-1.9) as were children with decreased level of consciousness (OR=5.3, 95% CI 2.0-14.2). Disparities across socioeconomic groups and gender persisted in seeking quality healthcare for under-five children with febrile illness in urban Dhaka. Girls from poor families were less likely to access qualified medical care. To reduce child mortality in the short term, health education and behaviour-change communication interventions should target low-income caregivers to improve their recognition of danger-signs; reducing societal inequalities remains an important long-term goal.