Table 1.

Implementation fidelity of the awareness programme.

Elements of fidelity	Kyrgyzstan	Vietnam
Adherence (was the programme implemented as it was designed?)
Content	We used the session content template addressing elements of the COM-B model (Appendix 3) in each training of HCWs in workstream 1. A concise version was used for the training of the health workers in workstream (2 and) 3
	The content displayed on flip-overs and posters (Appendix 4) was aligned with the session content template
		A tradition of constantly burning coal around a new-born during 1 month turned out to be also relevant, but was not addressed
Coverage	Direct reach: 10 HCWs were trained first. We had planned to train 50 health workers from different levels (e.g. CHWs and social workers). Due to high enthusiasm of trained health workers, we trained 90. Trained health workers reported to have been in contact with 80–160 community members each month, training ~15,000 community members within 6 months	Direct reach: 17 HCWs were trained first (one per health centre). Each centre covered 3–7 villages, resulting in 77 trained CHWs. Each CHW reported to have contact with 100–150 community members and so reached ~10,000 community members directly within 6 months
	Number of drop-outs was not registered
Frequency/duration	Initial training was 2 days shorter than in Uganda, due to experience facilitating the training in Uganda and because the materials were in a further development stage
	Initial group of HCWs was trained for 3 days, CHWs and social workers were then trained for half a day within 3 months after HCW training	Initial group of HCWs was trained for 3 days, new group of HCWs trained for 1 day within 3 months and CHWs trained for half a day within another 3 months
	Outside of the programme, the training was used to train HCWs from neighbouring countries during an international conference (IPCRG in Bishkek, 2018)
	Training of communities is ongoing to date. Using the materials, training continues to take place to patients and their families during visits to health facilities
Moderators (factors that have influenced the degree of fidelity)
Intervention complexity	Simplicity was enhanced by accompanying the training materials with short, explicit explanations and illustrations, e.g. specific instructions on the back of flip-overs with main messages to be addressed
	The module and training materials were translated in the local languages
	We co-created training materials together with health workers and other stakeholders to ensure easy understanding
Facilitation strategy	We strategically engaged stakeholders through collaboration meetings and hence enhanced (1) compatibility with the local context by co-developing the delivery strategy with them and (2) continuation of the programme through their support and ownership of the programme
	We adapted the strategy and programme materials to the local settings in collaboration with local stakeholders, HCWs, CHWs and the community. Key messages remained identical
	An active session was held on the national state TV channel, supplemented by messages on the radio and newspapers	The budget for a media campaign was exchanged for refresher courses of the trainers
Quality of delivery	HCWs and CHWs were trained on how to train. Training was supported by materials: both local FRESH AIR teams chose to use a PowerPoint for the health workers training, flip-overs for training the community, …
	… and brochures + posters to be distributed to health centres/public spaces	… and printed flip-overs instead of posters as the budget did not allow for printing additional brochures. The local team also delivered refresher courses for monitoring and feedback
Participant responsiveness	HCWs and CHWs reported and demonstrated to feel ownership due to the co-creating process. The enthusiastic participation of communities and observed behaviour change (e.g. adoption of changed cooking practices) motivated the health workers to continue the process
	CHWs (and social workers) reported high numbers of community members reached, which was confirmed by triangulation with the number of collected knowledge questionnaires
Recruitment	All participants were recruited within the existing health infrastructure
	District health managers with expert knowledge on the local context selected the first HCWs to be trained
	Some of our research team members participated in this first group of HCWs. The local FRESH AIR team explained that they were more easily available than regular HCWs to travel (which took relatively long in Kyrgyzstan due to the rough terrains). Also for the sake of travel time, these HCWs trained other health workers (CHWs and social workers) directly instead of via workstream 2 (Fig. 1)	One HCW per ward (the head of the health station) was selected for the initial training. They selected the next group based on convenience
	CHWs were purposely selected based on convenience (living in villages in vicinity of health centres), in collaboration with local HCWs
	Community members were recruited during regular health events	CHWs and the local team organised health sessions
Context	Local context was well known due to preliminary explorative FRESH AIR fieldwork, due to close collaboration with the stakeholders and because our team consisted of local and international team members
		Due to a miscommunication with the local and coordinating team, a costly pilot study was conducted assessing the frequency of biomass fuel use. However, the high frequency of use reassured the relevance to the selected setting
	Compatibility with the local context was enhanced by adapting interventions in collaboration with local stakeholders, HCWs and CHWs and by embedding the intervention within the local healthcare system

COM-B model capability, opportunity, motivation—behaviour model, HCW healthcare worker, CHW community health worker, structured by the modified Conceptual Framework for Implementation Fidelity.