Observe before you leap: Why observation provides critical insights for formative research and intervention design that you'll never get from focus groups, interviews, or KAP surveys

Formative research is essential to designing both study instruments and interventions in global health. While formative research may employ many qualitative methods, focus group discussions and in-depth interviews are the most common. Observation is less common but can generate insights unlikely to emerge from any other method. This article presents 4 case studies in which observation revealed critical insights: corralling domestic poultry to reduce childhood diarrhea, promoting insecticide-treated bed nets (ITNs) to prevent malaria, evaluating skilled birth attendant competency to manage life-threatening obstetric and neonatal complications, and assessing community health worker (CHW) ability to use malaria rapid diagnostic tests (RDTs). Observation of Zambian CHWs to design malaria RDT training materials revealed a need for training on how to take finger-stick blood samples, a procedure second nature to many health workers but one that few CHWs had ever performed. In Lima, Peru, study participants reported keeping their birds corralled "all the time," but observers frequently found them loose, a difference potentially explained by an alternative interpretation of the phrase "all the time" to mean "all the time (except at some specific seemingly obvious times)." In the Peruvian Amazon, observation revealed a potential limitation of bed net efficacy due to the built environment: In houses constructed on stilts, many people sleep directly on the floor, allowing mosquitoes to bite from below through gaps in the floorboards. Observation forms and checklists from each case study are included as supplemental files; these may serve as models for designing new observation guides. The case studies illustrate the value of observation to clearly understanding clinical practices and skills, details about how people carry out certain tasks, routine behaviors people would most likely not think to describe in an interview, and environmental barriers that must be overcome if an intervention is to succeed. Observation provides a way to triangulate for social desirability bias and to measure details that interview or focus group participants are unlikely to recognize, remember, or be able to describe with precision.