**2. Methods**

Estonia lies in the North-Eastern corner of Europe. Out of a population of 1.3 million, 35.5% are nationalities other than Estonian, including Russians (25%). One million people (77.2% of the population) live in urban areas [40]. Income disparities are relatively large at 33.2% [40]. The health inequalities among different age, education, and ethnicity groups are considerably larger in Estonia than in other Nordic countries [41]. For example, life expectancy among Estonians is 78.5 years, and non-Estonians 76.5, and total life expectancy is 77.6 years in Estonia, compared with 81.1 years in Finland [42].

For analysing the health systems' amenability to internal and contextual pressures, Estonian legislative acts, directives, strategic documents, such as activity plans, and their implementation were examined. In addition, 21 semi-structured interviews were conducted with key experts connected to health system adaptation in Estonia, ranging from policy designers to officials to scientists to social workers (Appendix A). The interviews clarified the expert's observations and experiences with the functioning of the Estonian health system and the actors and processes shaping its effectiveness. The interviews addressed the questions of the capability of the health and rescue system, operability of implemented measures, and the importance of various factors in shaping health system adaptation (Appendix B). The interviews were conducted in the spring of 2015 when compiling an adaptation strategy for the reducing the health risks of climate change and when health impact assessments were in their infancy [43].

To address public salience as an external driver of adaptation governance, we use a survey on environmental health risk perception and coping conducted in Estonia in 2015 of persons aged 18–75 years, stratified by age, sex, and geographical location [44]. The survey invited 2207 participants (administered by IBP Saar Poll), of which 1000 agreed (45.3% response rate). We used a semi-structured questionnaire constructed to assess climate change-related issues, such as perceived exposure to extreme weather events, demand for state support for coping with extreme events, beliefs about state institutions' efficacy in taking care of the healthfulness of the environment, and concerns about health risks from the environment (Appendix C). Additionally, the instrument had entries for respondents' demographic data and self-rated health status.

We used a logistic regression analysis to estimate which factors were associated with perceived needs for measures for coping with the health risks of climate change. For statistical modelling of covariates of perceived need, we collapsed perceived need ratings into dichotomous groups by combining the answers high to total agreement (scores 4–5) into a group perceiving need, and the other categories as not. We recoded worry about health risk to self and family arising from the environment, as scores 1 and 2 = group 1, score 3 = group 2, and scores 4 and 5 = group 3.
