How can community participation strengthen a health insurance system? The case of health insurer’s user associations in Colombia

Study framework

This study is part of the research programme ‘Making Health Insurance Programmes Responsive to Citizens’, supported by the Alliance for Health Policy and Systems Research, WHO. We followed an overarching framework developed for the research programme that uses Molyneux’s framework2 to assess how content, process and context factors, and their interactions, influence whether an accountability initiative is able to increase citizen empowerment and the responsiveness of health insurance programmes. To analyse the functionality and effectiveness of Colombia’s user associations as a health insurance accountability initiative, the framework addresses three domains: (1) content: what is the design of the user association approach and how is it supposed to be implemented? (2) Process: how does the user association actually work? (3) Context: what are the wider contextual factors at the health system, national and community levels that might influence how these user associations work?

Colombia’s health insurance system

Health insurance in Colombia is administered by multiple health insurance companies called Entidades Promotoras de Salud (EPS, meaning Health Promotion Entities in Spanish). While most EPSs are shareholder-based private for profit companies, a few large public (government owned) and mixed (includes government shareholders) EPSs insure 18% of the population. In Colombia, there are two major insurance schemes. About half of all citizens are enrolled in ‘contributory scheme’ and the other half in ‘subsidised scheme’. The contributory scheme is for people working in the formal sector and their spouses and dependents, who contribute with a payroll tax of 12.5% or equivalent. The subsidised scheme is for people without ability to pay (identified through a proxy means test) and it is financed by general government revenues and a cross-subsidy of 1.5% percentage points from the payroll taxes collected for the contributory scheme.

All EPSs must provide the same benefit package, which is identical for both the contributory and subsidised schemes. It includes primary care, hospitalisation, surgery, other procedures, drugs and catastrophic diseases. EPSs have the responsibility to ‘organize and guarantee, directly or indirectly, the access to and provision of the benefit package to affiliates’ and ‘to establish procedures to ensure efficient, timely and quality care in the services provided by healthcare providers’ (Art. 178, Law 100, 1993). They exercise this responsibility partly by configuring the network of providers using a mix of public and private sector clinics and hospitals.

Colombia is divided into 32 departments and 1117 municipalities. Few health insurance companies operate country-wide, most operate regionally in several departments (and the municipalities within). The various insurance companies compete for members in the market, who are free to choose their insurer. Insurers cannot compete on contribution rates or benefits, because these are fixed, and the same services are covered. Thus, insurers are expected to attract enrollees based on quality of their interactions with members, their network of providers, waiting times for care and hospital ward amenities (eg, private rooms), among other factors, several of which are captured by the official ranking of EPSs published by the Ministry of Health. In practice, however, competition seems to be limited and markets tend to be concentrated.3 4

Access and utilisation have improved since the 1993 reform: for example, between 1997 and 2019, the percentage of people visiting a physician for prevention increased from 59% to 78% for the contributory scheme and from 35% to 65% for the subsidised scheme.5 6 Inequalities have reduced for most indicators, including morbidity and access. For example, the percentage of people having access to all medicines in a prescription increased from 25% in 2003 to 67% in 2013 for the lowest income quintile, and from 53% to 63% for the highest quintile. Nonetheless, there are persistent inequities in access, healthcare utilisation and quality of care, disproportionately affecting rural areas and the lowest wealth quintiles.7 Likewise, catastrophic health expenditure remains an issue. In 2016, 8% of households experienced catastrophic health expenditure at the threshold of spending 10% or more of their total household consumption on healthcare. While problematic, this value is lower than regional estimates for upper-middle income countries (16.7%) and comparable with the regional estimates for Latin America (8.7%).8

Continued barriers to access have resulted in widespread and persistent judicial litigation in health; lawsuits are a last-resort mechanism used by enrollees.9 In 2020, enrollees filed more than 80 000 lawsuits to request healthcare services with 89% of these related to services already included in the benefit package and that should have been provided.10 Lawsuits have proven to be an effective mechanism to get access to healthcare when it has been unjustly denied: historically, the vast majority of lawsuits have been fully or partially won by litigants; for example, in 2020, only 9% of lawsuits were rejected.

Study design

We followed a mixed methods approach using a sequential explanatory design11 drawing from document review, quantitative surveys, qualitative interviews and focus group discussions, secondary analysis of existing databases and policy discussions (table 1). We use qualitative data from interviews and focus groups primarily to explain the findings from quantitative data (survey and secondary data) that measure people’s knowledge use and overall participation in the initiative. All data sources contributed to understanding the content, process and context of Colombia’s user associations initiative.

Data analysis

Quantitative data from insurance beneficiaries survey and secondary data sources were descriptively analysed, including calculating descriptive statistics and measures of association between indicators. We report point estimates and confidence intervals of key indicators for the study, including public awareness, use and participation in user associations. Further details of the results from quantitative data can be consulted in the online supplemental annex. Data from the health insurance beneficiaries survey were analysed in R using the {survey} package.12–14

Qualitative data were used primarily to explain the findings from the quantitative analysis. In particular, qualitative data collection aims to explain the functioning of user associations to understand why the initiative has progressed to the observed results (in terms of awareness, use, participation) and what are the key factors driving these findings. Interviews, focus group and policy discussions sessions were recorded and transcribed. Qualitative data were analysed thematically,15 beginning with deductive–inductive open coding of the data from in-depth interviews. For this we used the theoretical framework to define initial categories and codes to assign to the data and we let other codes emerge during the coding process. For the deductive process, we used a set of codes that underlined enabling and limiting factors for the user associations, classified using the process–content–context structure. These codes comprised information, motivation, cost, purpose, power-relationships, regulation, culture, politics, geography. Influential factors for the initiative that did not fit within any of the initial codes were added to the coding matrix. We then applied our coding framework to focus group discussion transcripts and policy discussion meeting notes. Afterwards, we identified relationships among categories and aggregated them to build an overarching understanding of how the user association initiative was designed (content), how it was actually functioning (process) and what are the most important reasons for the way the user associations function (context).

Patient and public involvement statement

There were no patients involved in the study. Members of the public were involved as research participants, and their experiences, priorities and preferences shaped the policy recommendations arising from this research. While the intent of the research is to improve health insurance responsiveness to the general public, the public was not involved in setting the research questions or designing the study.

See also the reflexivity statement in the online supplemental appendix.

Content: user associations as per the legal provisions

The origins of health insurance user associations can be traced back to Article 49 of the 1991 Constitution that states that the community will participate in the organisation of healthcare. This participation was further developed in the 1993 health reform that gives health insurance enrollees the right to join or create user associations (Art. 156, lit. h). These user associations would ‘represent them [insured people] before the EPSs [health insurance companies] and the health care providers’ (Art. 156, lit. h) and ‘strengthen the negotiating capacity, the protection of rights and community participation of the affiliates’ (Art. 157, par. 3). Moreover, in 1994, the national government mandated that health insurance companies guarantee community participation (Art. 9, Decreto 1757/1994) through user associations that ‘will look after the quality of services and defend users’ (Art. 10).

Although the 1993 law mandated that it was the government’s responsibility to ‘promote’ user associations (par. 3. Art. 157), the government later delegated this responsibility to the health insurance companies. In particular, health insurance companies have to ensure there is at least one user association in every department in which they operate, and they must ‘promote and strengthen’ participation in user associations by inviting the insured to create or join them (Circular externa 008, SNS). However, the law does not define how the government or health insurance companies are to promote the initiative. In particular, beyond the mandate to promote and support user associations, the law does not define specific mandates on education, resources or technical support that the government or insurance companies should provide.

Legal provisions have made it straightforward for the insured to form user associations: it requires as few as two affiliates signing creation minutes and submitting them to the Superintendencia Nacional de Salud (SNS, National Health Oversight Agency). Multiple user associations can be formed for any health insurance company (Art. 12, Decreto 1757/1994) and the insured can create an association even if they have not been ‘invited’ to do so by the health insurance company (Decreto 780/2016). It remains the users’ responsibility to create and operate the associations. The regulatory provisions do not endow user associations with financial support, but associations are allowed to charge membership fees (Art. 157, Par. 3°, Law 100/1993).

User associations are supposed to provide advice to citizens on how to navigate the health system and access health insurance benefits, and solicit citizen needs, concerns and complaints relating to their health insurance company (Decreto 1757/1994), thereby empowering citizens exercise their insurance entitlements. The user associations are in turn supposed to share this feedback with the health insurance companies and suggest improvements, They are also envisioned as participants in decision-making within the health insurance companies. Through these functions, it was believed that the user associations would enhance the responsiveness of health insurance companies (figure 1).

Request permission

Although health insurance companies are legally required to provide ‘adequate and timely processing of the concerns and requests of their enrollees’ (Circular externa 008, SNS), the user associations do not have any special means to enforce this. Thus, in the event of a breach by the health insurance company, the user association would need to use the regular complaint mechanisms of the health system (operated by SNS), as any other enrollee could.

The regulation dictates that there must be one seat for a user association representative on the board of directors of publicly-owned and mixed-ownership insurance companies (Decreto 1757/1994). After a legal battle led by user association representatives, a 2012 ruling from the Constitutional Court enforced that and mandated the EPSs to comply with the regulation.16 Yet, the regulation does not define the details of how the EPSs should adhere to user association input.

Process: how user associations actually function, and their effect on the health insurance scheme

Health insurance affiliates have created at least 1289 user associations for the 40 different health insurance companies currently operating across the 32 departments of Colombia.17 There is compliance with the regulation that there must be at least one user association in each department where a health insurance company operates. For 23 EPSs in 27 different departments, there is more than one user association (from 2 to 116 user associations) in a single department. While there is no requirement that user associations be formed in each municipality, there is at least one user association in 69% of the municipalities in the country. This is no minor achievement because there are only a handful of social interventions that reach so many municipalities. Yet, that means that there are 348 municipalities out of the 1117 (31%) without any user association. User association density—defined as the percentage of insurance beneficiaries that have a user association in their EPS and municipality—ends to be lowest in the south-east of the country (towards Orinoquía-Amazonía), which also has a lower population density, is marked by poorer health system performance, poorer access to public services, lower household incomes and weaker governance (see online supplemental figure A for a geographical density map).

Despite the high number of user associations, the health insurance beneficiaries survey shows that just a small fraction of the population is aware of and engages with user associations (figure 2). Overall, only 10% of people with health insurance know of any user associations, just 3% have had any communication with a user association, only 2% have joined at least one user association meeting and less than 1% reported finding them useful.

Request permission

Knowledge about and engagement with user associations was not significantly different between users in the contributory and subsidised schemes. Awareness of user associations was significantly higher among those who used their health insurance benefits (11% vs 4%, p<0.05) and those who had a medical visit (15% vs 2%, p<0.05) in the last year versus those who had not, and those who had a chronic disease or disability versus those who did not (15% vs 8%, p<0.05).

User association members and patients interviewed reported insufficient promotion of user associations by governments (both national and local) and health insurance companies. While some people reported receiving letters informing them about user associations, most reported that health insurance companies limited their publicisation efforts to a single informational page on their websites. Not only does this require active searching, but it also excludes the 30% of the population who lack routine access to the Internet.18

The EPS [health insurance company], what does it contribute to inviting participation? […] A little web page for everything about user associations. It’s well presented but not everyone goes online to dig, sometimes the user doesn't even know [that user associations exist]. (Interview 46, User associations and civil society)

I think that only those who are involved in these user associations know the potential of this participation mechanism, but the rest of the people do not. […] The vast majority of EPS users are unaware of this participation mechanism. (Interview 12, National Government)

The number of people participating in user associations is usually small. Of the 27 respondents in the user association member survey (likely a sample of highly engaged members with easy access to Internet), 7 reported that their user association was composed of a single active member and none of them reported more than 10 active members.

User associations leaders work on a voluntary basis and had to pay for their own transportation, communications, meeting venue and any other costs, including costs associated with reaching out to the insured. A few health insurance companies were reported to provide some resources (eg, an email address, a venue for meetings), but support was limited and up to the discretion of the specific company.

Training for user association members was irregular and non-standardised. While some user associations received training or strategy development support from the SNS or health insurance companies, others received nothing. Respondents from user associations, government and civil society agreed that members’ lack of knowledge or technical skills was an issue, including on the complexities of the healthcare system and health insurance financing.

The volunteer (unpaid) nature of the role, lack of financial support for activities and limited training has made membership in user associations primarily feasible to retirees and individuals seeking to solve specific issues that they or their family members are facing with healthcare; among the 27 members who completed our survey, 20 reported participating to try to solve a personal issue, most were over 65 years in age, and none had full time employment.

Despite some exceptions and success stories, overall, user associations were reported to have minimal influence on health insurance companies and played a marginal role in supporting citizens in accessing and navigating the health insurance system. Some user associations were entirely non-functional, and were reported as having been created to meet the regulatory requirement without any genuine membership.

There are some that are created simply to comply with the norm that requires them to have an association, but it is an association that is constituted and is never seen again. (FG315, User associations and civil society)

User association members reported limited outreach to help other insurance participants/enrollees navigate the health system or to solicit patient complaints or suggestions. Even the most active user associations reported that they could reach (eg, provide information, receive a complaint) hundreds to a few thousand users in a year. This is a lot given the limited resources they have, but little in relation to the overall population. Most user associations were said to be limited to working as ‘complaints processors’ (Interview 33, EPS representative) wherein they passed on individual requests and complaints to the insurance companies or the SNS when approached by someone with a specific issue.

Among associations whose members attempted to take on further activities, respondents explained that health insurance companies at times actively resisted user association involvement in decision-making and were free to disregard user association suggestions and recommendations. User associations lacked mechanisms to compel health insurance companies to seriously consider their concerns and respond accordingly. User associations’ lack of influence and power over health insurance companies was discussed as a fundamental problem.

We do not have the right to influence what they [the EPS] consider important. They are clearly limited to receiving requests, but these are not going to be a strong reason for them to make a decision. (Interview 45, User associations and civil society)

User associations have the right to ‘democratically elect’ one representative to participate in the EPS Board of Directors, for public or mixed EPSs. Being part of the board of directors should help compelling the EPS to respond to the associations’ feedback and not only respond to their shareholders, who typically have a seat on the board. Yet, interviewees reported that EPSs tried to impede such participation altogether or circumvented it, by managing to place on the board a representative from a friendly user association who would do little to try and convey user’s needs and expectations to the board.

Although interviewees characterise most user associations as non-functional, complaints-processors, struggling for resources or contending against a reluctant EPS, we identified success stories where there was a supportive relationship between the EPS and the user association. In one of these cases, interviewees reported a close collaboration between the EPS and the user association, who were said to have ‘an open door, a direct line, and a permanent interaction’ (Interview 33, EPS employee) with regional health insurance managers. User association members took part in the microlevel decision-making of some committees, but not in the Board of Directors, where more strategic high-level decisions were taken. A member of this user association reflected that they gained this influence because the EPS wanted to learn about and rectify user complaints, in order to retain users.

The user association [is very important] for the operation of the EPS. Why? Because through the user association, they can collect all the disagreements, the complaints […]. And that’s beneficial for them because in the event that there is a user or many users who relocate [switch to another insurance company] due to bad service provision—oh mamita!—there once and for all, rest assured that the financial part of the EPS goes down because that talk goes everywhere (interview 46, User associations and civil society).

The key driver of the success story was ultimately the EPS’s willingness to support and collaborate with the user association. Without such a willingness, even empowered leaders end up struggling to make their voices heard due to lack of resources. But the factors leading to a favourable stance of the EPSs towards user associations may be complex. Sometimes even EPS that may find some value in user associations, may not be willing to support them due to a combination of factors that include the cost of doing so, perceived legal barriers to do it and other idiosyncratic factors such as who is leading the EPS or the history of the insurance company. For example, in the case discussed above, the Chief Executive Officer of the insurance company was reported to truly believe in community participation and therefore, designated a top-ranking official within the EPS to lead community participation issues and assigned her a budget. Other success stories were reported in EPSs with roots in the community (mutuales) and among EPSs that took part in one of the EPS guilds in the subsidised scheme, which also engages in some efforts to promote community involvement in health insurance.

Context: factors affecting the functioning of user associations

The functionality of user associations is grounded in the Colombian health system context, broader perspectives on participation in Colombian society, and the Colombian legal and judicial context. While the first two contextual features hinder user association effectiveness, the legal and judicial context is a potentially significant enabler.

In the Colombian health system model, health insurance companies have an incentive to contain costs while consumers are expected to demand quality through market competition. Interviewees reported that EPSs may fear that an empowered user association could lead to increased health expenditure through increased demand or by requesting improvements in services.

I think that the last thing that the EPS wants to see [in the user associations] is the involvement of people who have made ourselves known […] Because they know [we can guide the] new patients [to get the services they need]. I really feel that they put up a great barrier so that we [activists] […] cannot be leaders in the user association. (Informant 35, Focus Group 3, User associations and civil society)

The user association tries to be a bridge between the [health insurance] administration and the community, and fight for better user services, but it becomes very difficult when the administration doesn’t want this. (Informant 311, Focus Group 3, User associations and civil society)

Theoretically citizens can switch health insurance companies when dissatisfied with their services. In reality however this regulatory mechanism falls short. The insured may feel that all EPSs offer similar services, may find the bureaucratic requirements to switch burdensome, may lack the technical knowledge necessary to assess quality or may live in a geographical area without another EPS option. Thus, health insurance companies support the creation of user associations in order to comply with regulations but have little incentive to empower user associations to take on a meaningful grievance redressal or oversight function.

In terms of the broader social context, the population has low awareness of the existence of user associations and low motivation to engage in participatory activities. Low motivation is grounded partially in the design features of the user association (eg, lack of resources to support user association activities and a lack of influence over EPSs as discussed above) and in broader contextual features of Colombian society. In Colombia, people see a risk of violence against social leaders and there is, to some extent, a lack of belief in participatory and democratic instruments.

Unfortunately this violence, …, so many community leaders murdered, that people do not want to participate in these processes anymore; fear has invaded us (Interview 21, Sub-national government).

I think that Colombia has in general, like the countries of Latin America, […] an apathy of citizen participation and I think that we are in a context of apathy for democratic instruments. (Interview 41, User associations and civil society)

In terms of the legal and judicial context, there is a human rights approach underlying the user association initiative that could facilitate future improvements. The right to participate is a key part of the right to health, which makes the initiative enforceable through the courts, using expedited rights protection mechanism called ‘tutela’. A tutela is a lawsuit that can be filed without the need for a lawyer or any legal assistance and must be resolved in 10 days. The population is generally aware of this mechanism and its effectiveness. It is highly regarded and typically portrayed as one of the most important social achievements of the 1991 Constitution. This context has enabled user associations for public EPSs to demand a position on the board of directors. Although this positive ruling has so far proven insufficient to overcome the numerous other barriers facing user associations, future lawsuits may strengthen user association power and support—for example, they could try to get representation in the decision-making process of private insurance companies.

Limitations

A number of limitations of this study arise from the fact that it was undertaken during the COVID-19 pandemic and the governments’ response with general lockdown and social distancing measures. This forced us to change the primary data collection approach. In particular, we could not do face-to-face interviews as originally planned. Instead, we had to rely on telephone interviews for the insurance beneficiaries survey and online meetings via Zoom for the in-depth interviews and focus groups.

As a result of the lack of reliable contact information for all user association members, the survey for them uses a non-random sample. Therefore, one cannot generalise those results to the whole population of user association members and thus, we use those results rather qualitatively. Yet, it is likely that those that responded the survey are some of the most active user association members.

Further research

This study identified several challenges that health insurance user associations face and also outlined a set of policy measures to strengthen user associations. This leaves interesting questions for future research, such as, how to best sensibilise health insurers to support and take advantage of community participation. In which decision-making bodies are user associations more useful. How to involve a broader population in the user associations. In addition, even though they were not part of the study, during the data collection we kept receiving information of other types of associations that operate in the health system. In particular, patients associations seem to be a powerful force in the health system and apparently can be frequently more effective than health insurance user associations. Key factors such as motivation to participate are different, but it would be worth exploring in more detail those other kind of citizens participation in the health sector to see what the health insurance user associations can learn from those other experiences.