Research methods used, sampling and data collection procedures
| Method | Sample and sampling | Data collection |
| Desk review of key textual sources to analyse official published documents about user association mandate and functionality. | Documents (n=72) including legal codes, regulatory and judicial decisions (25), content from insurance companies’ web sites (40), and government background documents (7). Identified through internet searches and consultation with key stakeholders. | Documents were read by authors (EA-S, MB-V, JB). Relevant content on user association mandate, envisioned mechanisms of action and actual functionality was summarised in a data extraction spreadsheet (MS Excel). |
| Insurance beneficiary survey to assess beneficiaries’ knowledge and use of user associations, as well as their health and insurance status and their overall participation and empowerment (see online supplemental annex 1) | A nationally representative sample of beneficiaries 18 and older (n=1311), in the contributory scheme (n=666) and subsidised scheme (n=645) Probabilistic sampling, stratified in two stages. Statistical strata were formed for six geographical regions and two types of municipalities (considering their weight in total population size, municipalities were segmented in two types; capitals and the rest of municipalities). In the first sampling stage, municipalities were randomly selected within each stratum and in the second stage, telephone numbers from the sampling frame were randomly selected, surveying one adult person of the selected household. The sample for the first stage comprised 50 municipalities, including 5 capitals (Bogotá, Medellín, Cali, Barranquilla, Bucaramanga). To account for the two-stage sampling design, expansion factors were calculated based on the inverse of the inclusion probabilities in the sample of the final sampling units, to obtain unbiased estimates for the parameters of interest. | The 53-item questionnaire was administered via telephone by a local firm specialised in surveys and data collection. The interviews started by describing the study and asking for consent to participate. Then filters were applied (≥18 years old, enrolled in either contributory or subsidised scheme) and the full questionnaire applied to eligible individuals. The questionnaire was programmed in a specialised software that administers the flow of the questionnaire. The interviewer entered the responses in the system. The interviews were also recorded for quality control and to codify open questions. |
| User association member survey to collect data on members’ characteristics and opinions about their user association | A non-representative sample (n=27) of user association members found through emailing a description of the study and link to the survey on user association member listservs. | The 22-item questionnaire was self-administered online and included questions on the demographic characteristics of user association members, geographical location, occupation, their role and the functioning of the user associations. |
| Secondary data analysis to assess the number and distribution of user associations | Quantitative analysis of a directory of user associations and databases of insurance beneficiaries for each insurer. | Data were requested to Superintendencia Nacional de Salud (SNS, National Health Oversight Agency) and then redirected to the publicly available data in the SNS’s web page. The directory of user associations was last retrieved in January 2021, and includes the name of the association, the municipality where it is located, the EPS to which it is linked and contact information of the representatives of the user association. |
| Semi-structured interviews to qualitatively understand stakeholder experiences with user associations and their perspectives on how user associations operate, their effectiveness, and enabling and hindering factors related to their functioning | Interviews (n=19) with members of users associations, government officials from national and sub-national governments, representatives of health insurance companies, pharmaceutical industry, patients associations and other members of civil society. Purposive sampling to select respondents across the following criteria: recognition by their peers and among the stakeholders in the health sector, representation of the different regions of the country, representation of the different entities, and public and private organisations of the health sector. | Interviews were conducted between September 2020 and March 2021on Zoom and ranged from 1.5 to 2.5 hours in duration. Each interview was led by a single member of the research team and other members of the team were also able to listen to the interview and pose clarifying questions or points for further elaboration. Interviews covered the following topics: (a) an introductory section to tease out characteristics of the informant and the institutions they represent, (b) questions on the informant’s involvement in the initiative, (c) questions on the functioning of the user associations and the enabling and limiting factors involved, (d) questions on the perspectives of the initiative for the future. |
| Focus group discussions to obtain rich details on the most salient factors identified via interviews | Focus groups (n=6) with 37 participants: members of insurer user associations (n=2), insurer citizen representatives and patient associations (n=2), members of the medical association (n=1) and people working in the pharmaceutical industry (n=1). Purposive sampling across the informant types and geographical variability. | Focus group were held on Zoom in June 2021 and ranged from 1.5 to 2 hours in duration. Each focus group was led by a facilitator (JB, MB-V) and other members of the team were able to participate. Focus groups covered the following topics: How the user associations function, categories of associations, major limiting factors and policy alternatives to improve the initiative. |
| Policy discussions to get feedback on preliminary results and recommendations, identify key factors not sufficiently considered and discuss the feasibility of and prospects for effectiveness of the policy recommendations | Sessions (n=6), with between two and six participants, one each for: representatives from government (Ministry of Health, local health authorities and oversight agencies), the Constitutional Court, associations of health insurance companies, insurance companies, leaders in the health sector and community leaders. Purposive sampling to cover the most relevant decision-makers in the health system or related to the protection of human rights in relation to healthcare access. | Discussions were held on Zoom in May and June 2021 and lasted for approximately 2 hours each. After presenting preliminary results and policy recommendations, the principal investigator (MB-V) opened the floor for participants and facilitated a discussion about the results (Do you agree with the results? are there any factors not sufficiently considered?) and the policy recommendation (do you agree with the recommendations?, are they feasible to implement?). |
EPS, Entidades Promotoras de Salud.