Prospective policy analysis: how an epistemic community informed policymaking on intentional self poisoning in Sri Lanka

Suicide in Sri Lanka is a major public health problem. In 1995 the country had the highest rate of suicide worldwide (approximately 47 per 100,000) [1]. Recent analysis of the incidence of suicide has shown a substantial decline from the peak of 47 in 1995 (male 80 and female 28) to 24 per 100,000 in 2005 (male 37 and female 10) [2]. Young people are over-represented in admission rates following self-harm. In 2000, 60% of all self-harm admissions were aged 16-25 [3]. Of all the suicide deaths recorded, between 60 - 80% are caused by the ingestion of pesticides [4]. In many rural districts of Sri Lanka the most common cause of death from any cause is from the intentional ingestion of pesticides [5]. While deaths from suicide overall have been falling, there remains a considerable burden of intentional self poisoning [6‐9].

Policy responses to the problem of suicide in Sri Lanka included establishing a Presidential Committee, legislative changes, and improved clinical management. The Presidential Committee (formed in 1996) developed a National Suicide Prevention Strategy in 1998 with the goals of:

Since the inception of the Control of Pesticides Act 1980 the Department of Agriculture has embarked on a concerted programme to regulate the most toxic pesticides. These restrictions have been associated with considerable success, as described above, in reducing the overall mortality from intentional self-poisoning [10, 11]. In 2008 the Department of Agriculture announced a phased withdrawal of three pesticides (Paraquat, Dimethoate and Fenthion) based on strong evidence of the high death rates associated with their misuse in rural communities. Following these decisions there was concern from stakeholders about which interventions should, at that stage, be pursued and how best to continue to reduce mortality associated with the intentional ingestion of pesticides.

Health policy analysis is a reasonably developed field in industrialized nations but only a relatively small body of work has emerged from low and middle income countries [12, 13]. Lessons to date have recently been synthesized [14, 15]. Since the early 1990's there has been a growing interest in why policies do not achieve their anticipated impact and much research has moved away from analysis of the content of policy to focus more on the political dimensions of policymaking and implementation [16, 17]. Most reported health policy analyses have been retrospective case studies looking at the success or failure of past policies [15], yet policymakers need prospective tools to help analyse strategies before they are implemented.

Health policy analysis draws on two prominent discourses, based on different disciplinary influences. While not mutually exclusive their traditions differ slightly with public policy analysis rooted in the social and political economy frameworks of public policy while evidence-informed policymaking is rooted in the scientific and health care domains and draws from, but refines, the evidence-based medicine traditions.

Within the health field a popular approach to examining policy is to utilise the 'policy triangle' - an elegant and simple tool which has been used extensively in low and middle income countries and identifies the content of policy, contextual factors, processes and actors as all deserving of attention [13]. Walt and Gilson suggest that these factors are intertwined and need to be systematically considered to achieve an holistic understanding [18]. Buse, an advocate of prospective policy analysis [19], has gone further to develop a policy engagement framework to prospectively analyse policy that incorporates strategies for change (see Figure 1). The policy engagement framework "offers a systematic approach to the ongoing collection, analysis and use of political information (e.g. concerning actors, their interests, institutions, ideas, and policy processes and context) that can alter the balance of power between those in support of and those resisting change by enabling pro-reformers to intervene more effectively in the policy process" [20]. This framework, like the political mapping approach of Reich [12], identifies the key domains of interest in analyzing policy prospectively.

In addition to these domains developed through the policy sciences traditions, there is increasing focus on the pathways to promote evidence-informed policy-making [21]. Recent publication of the SUPPORT tools for evidence-informed policymaking seeks to guide policymakers and the policy community to access better knowledge to support policy making. The 18 papers in the SUPPORT suite address four broad areas related to policymaking: supporting evidence-informed policymaking, identifying needs, finding and assessing information and implementation [22]. They propose that well informed policy decisions are made based on using systematic reviews and knowledge of local conditions in order to assess benefits, harms and costs and make judgements about trade-offs, as seen in Figure 2. While these tools are specifically geared towards policymakers' uses of evidence to inform policy, they recognise that evidence is only one factor in a multi-faceted process and acknowledge that this framework could be used to complement other strategies.

Lindblom argued that a good policy is one that commands consensus among policy makers and interest groups [23]. All the above frameworks and tools identify the importance of actors and stakeholders, reflecting the view of policy as the outcome of continuing interaction between diverse agendas, multiple stakeholders, and contested and ambiguous issues [24]. One important type of stakeholder network that is at times influential is an epistemic community - "a network of professionals with recognised expertise and competence in a particular domain and an authoritative claim to policy-relevant knowledge in that domain or issue area" [25]. This group is tightly integrated, limited in number, with high continuity and members having roughly equal status and power [26]. Thus, working with an epistemic community to generate a range of options for further analysis may itself yield important insights into policy, as well as into how the epistemic community operates, perceives the problem(s) and proposed solutions.

Eight options were generated in the first round of discussion and described in Table 2. All of these options were supported from the literature however the generation of a novel suggestion that incorporated several ideas together, repackaging into non-lethal doses, was an innovative outcome from the discussion.

The first round of voting generated a high level of support for six of the 8 options: regulation, safe storage, Integrated Pesticide Management (IPM), repackaging, advertising and using bio-pesticides. There were mixed levels of support and opposition to the two remaining policy options, taxation and dealer training programmes. In round one the level of consensus was high in seven of the eight options, with only price and tax incentives being contentious. The discussion between the panellists generated views about the strengths and weaknesses of the various options and has been summarised in Table 3. The second round of voting generated a greater level of differentiation between the options and reflected changes in response to the discussion regarding the desirability of each option. A high level of consensus was only achieved in five of the options. The policy options of regulation and repackaging were more strongly supported than others, with consensus growing for repackaging. This change observed between round 1 and 2 voting can be seen in Figure 3 where support was seen to drop away for IPM, bio-pesticides, price/tax, safe storage, dealer training and advertising. A final ranking of the options seen in Table 3 was initiated by the panellists and identified further regulation and repackaging as the preferred options, followed by IPM, safe storage, advertising, using bio-pesticides, training dealers and taxation.

Through the discussion several themes about policy emerged: options were identified and framed in terms of their strengths and weaknesses, local conditions and contextual factors were highlighted and factors that could modify policy were identified. The panellists went through a subconscious but deliberate process of weighing up the various benefits, harms and costs of each option as seen in Table 4. This process was highlighted by the discussion surrounding price incentives and weighing the impact of political involvement.

Consideration by the panellists of the benefits, harms and costs through the rounds of discussion helped them to make final assessments of their levels of support or opposition to the various options.

Important contextual issues were highlighted by the panellists in relation to the policy process. The delegated authority through the Control of Pesticides Act 1980 to the Pesticide Technical and Advisory Committee (PeTAC) depoliticised many aspects of pesticide control and regulation. This allowed a technical focus for decision making, free of interference from local manufacturing commercial interests.

In addition to the decision making strength of the PeTAC committee the other perceived strengths of the legislation was its ability to have an impact at field level. PeTAC's decisions to withdraw registration for particular products were seen as important previous successes.

Strong links between decision making, research and technical advice assisted policymaking. In addition the relationships within agriculture between the public and private sector were strong reflecting close kinship ties generated from links made in the higher education system. These links remained important and the culture of the industry was seen as having been supportive of policy changes. Finally there was a strong belief that policymaking should be informed by evidence, and ongoing surveillance of poisoning was important to continuing the technical basis for decision making.

Several factors were identified as posing threats to some of the potential options. Risks of political interference resulting from pressure from commercial interests and political patronage structures within farming communities were identified. The lack of engagement in pesticide control of the farming communities has strengthened the ability of policymaking but highlighted a potential threat if the issue were to become politicised. Lack of evidence or insufficient evidence of impact was a key factor in the lower ratings accorded to some options, as illustrated by the following quote.

The multidimensional context of suicide and the intersectoral action required for solutions highlights the sensitivity of policy to the conflicting interests of each sector. It also made more complex how the problem was defined and what solutions were being proposed. At times this complexity was seen as an issue that those in the health and agriculture sectors could avoid:

Difficulties of implementation were often related to weak systems, weak linkages between the sectors and the likely costs. The acceptability of the policy options to the community and the need for behaviour change to accompany policy responses were highlighted as difficulties of implementation at the grass root level. In addition the commission and marketing system for pesticide dealers created market incentives that challenged safe use initiatives and programmes designed to reduce pesticide use.

A limitation of this study was the number and range of panellists available to participate. However, a number of studies have shown that panel size is not always a good indicator of the success of a consensus exercise [29, 30]. The consideration of panel size can only be understood in relation to the ability of the exercise to produce innovative ideas. In this case study, the process produced a new policy option that blended together, in an innovative way, several previously considered options. The advantages of having the face to face group process were that it enabled a robust discussion of the options and a thorough examination of the strengths and weaknesses of each. The limited range of experts meant that the panellists were familiar with the background literature and the main proposed solutions and so the discussion was able to be more intensive and focussed, with limited defensiveness on the part of the participants given the high levels of continuity and trust that operated within the community.