br Efforts to address malnutrition maternal and child mortal

Efforts to address malnutrition, maternal and child mortality, and other global health priorities are heavily reliant on behaviour change, including adoption and correct use of health technology, and following treatment recommendations. One long-standing concern is the limited effectiveness of many behaviour change interventions, even when exposure to intervention activities is adequate, and activities are implemented with high fidelity. A second concern is low coverage of proven behaviour change interventions, under conditions of routine programme implementation. Those people who are not reached might be the poorest or most vulnerable, and therefore the ones who stand to gain the most from the intervention. Target of emotional drivers such as affiliation (social inclusion), nurture, and disgust, has been identified as one way to increase the effectiveness of behaviour change interventions, with disgust considered appropriate specifically for promoting avoidance of sources of infection. The investigation by Katie Greenland and colleagues of a behaviour change intervention for diarrhoea control in Zambia represents an attempt to operationalise this concept of harnessing emotional drivers to increase uptake of desired behaviours under routine conditions. The authors categorise their trial as a proof of concept, and lay out their theory of change to demonstrate how intervention activities bring about changes in protective and treatment-related health behaviours. The intervention promotes four very different behaviours. Intervention uptake goals vary, from establishing a new and permanent habit (handwashing with soap), to a temporary habit (exclusive breastfeeding), to the immediate response to a sick child (preparation of oral rehydration salt solution and zinc treatment for childhood diarrhoea). Furthermore, the emotional drivers the intervention aims to encourage might apply to the four behaviours in different ways. Implicit in the authors\' theory of change is a concern for intervention fidelity, “the degree to which an intervention is delivered as intended”. Kim and colleagues identify four aspects of implementation fidelity that warrant attention in a glucocorticoid receptors based study: adherence, dosage and exposure, quality of delivery, and participant responsiveness. Greenland and colleagues assert that their study “demonstrate[ed] that a model based on emotional drivers may prompt change in exclusive breastfeeding behaviour”. Using disgust for promotion of handwashing as one example, Greenland and colleagues do not present evidence that the change agents (Komboni housewives) received training or guidelines on how to create feelings of disgust when interacting with the community, or followed these guidelines (adherence). Dosage and exposure to activities eliciting feelings of disgust are not presented, nor do we learn if people actually felt disgusted when exposed to these activities (participant responsiveness). Further, we question whether the change agents operationalised disgust to a sufficient degree as so to elicit a behavioural response. The concept of sufficient quantities is crucial for interventions targeting emotions such as disgust. Nicholson and colleagues indicate that Heterochromatin is not necessarily an individual\'s tendency (propensity) to feel disgusted that leads to effects on behaviour, but rather the degree (sensitivity) to which that disgust is felt. Greenland and colleagues present no data on intervention fidelity in this paper, so we cannot judge if this study constitutes a fair test of the effectiveness of emotional drivers in increasing uptake of these four behaviours. When we fail to observe changes in behaviour, we are not in a position to know if the problem lies with the limited technical effectiveness of the intervention, or insufficient exposure to intervention activities for change to occur. The low coverage achieved in this study is another major concern. High coverage promotes equity, and evidence shows that rapid increases in coverage of maternal and child health interventions bring about gains in equity. Greenland and colleagues compare clusters with high exposure and low exposure, and note better results for some indicators in the high exposure clusters. We do not learn if there were pre-existing differentials in wealth, education or access to information between the high-exposure and low-exposure clusters. If such differentials do exist, then they may explain the better results in the high exposure clusters. Reaching the poorest and the marginalised, and delivering behaviour change interventions with high quality and intervention fidelity, is a crucial challenge for the field of implementation science, particularly in low-density rural populations in Africa.