From knowledge uncertainty to adaptive policy implementation: knowledge governance and integration in policy responses

Policy implementation under knowledge uncertainty requires governments to respond to evolving scientific evidence, professional knowledge, and implementation conditions. Yet, how institutional arrangements shape the use of diverse knowledge and its relationship with adaptive implementation remains insufficiently understood. This article examines these relationships in Indonesia’s COVID-19 response through a predominantly qualitative study combining a survey of 505 health workers in Depok, Surakarta, and Medan, six semi-structured interviews, document analysis, and policy process tracing. The analysis focuses on adjustments to diagnostic testing policies and workforce and healthcare service capacity. The survey identified multiple channels through which health workers contributed suggestions to policy discussions, including direct submissions to government (88 respondents), professional associations (87), and academic forums or conferences (61), with some respondents using more than one channel. Among the 138 respondents reporting contributions through these channels, 126 provided substantive follow-up responses, while 12 provided no substantive follow-up response. Of these 126 respondents, 42 (33.3%) reported no follow-up, 35 (27.8%) reported that their proposals remained under discussion, 22 (17.5%) reported incorporation into regulations, and 27 (21.4%) reported implementation. These self-reported findings indicate variation in follow-up but do not establish whose contributions shaped particular decisions. The findings indicate that, at the national level, professional organizations were formally involved in developing collaborative clinical guidance, providing an institutional basis for responding to evolving knowledge. However, local implementation depended on institutional capacity, resources, and socioeconomic and contextual conditions. Thus, knowledge integration may support adaptive implementation but does not automatically produce it.