Legal and regulatory instruments for NCD prevention: a scoping review and descriptive analysis of evaluations in OECD countries

There is a large range of regulatory approaches and instruments utilised for the prevention of NCDs. Our review found that government regulation was the most commonly evaluated approach and the most likely to be effective. This is compared to self-regulation and quasi-regulation which were almost always reported to be ineffective. This was particularly evident for unhealthy food, which was reported highly effective when government regulated and not effective when self- or quasi-regulated. The only effective self-regulatory instrument was designed and administered by a public health organisation, independent of industry, with the purpose of countering the alcohol industry’s ineffective voluntary scheme [116]. Another was reported to be partially effective, however the authors raised questions about the efficacy, as despite there being high compliance among signatories to the industry developed voluntary guidelines, the guidelines themselves did not align with the Attorney General’s recommendations [122].

There were few quasi-regulation instruments evaluated. This may be because this is not a well-defined or commonly used governance mechanism. Quasi-regulation typically features and has been well defined in the context of food policy [5, 6] but is often used interchangeably in broader regulatory contexts with other approaches [134]. However, of those included in our review, none were effective. For example, Magnusson & Reeve (2015) compared voluntary quasi-regulatory food reformulation instruments in Australia and the UK and reported both instruments to be ineffective. The evaluation of the Australian instrument found there had not been any nutritional improvement in the target products, there was incomplete coverage of industry with limited participation, and limitations in governance with no monitoring and self-reported compliance. The evaluation of the UK instrument found the salt reduction target had not been achieved, there was inconsistent coverage and industry participation, and a lack incentives and disincentives to encourage participating.

The results for co-regulatory approaches were unclear due to the small number identified in the included studies. Co-regulation is thought to be of value as it reduces the administrative burden and associated costs on government, however, due to the close relationships required between industry and government there is a higher risk of regulatory capture [134‐137]. Much of the theoretical literature discusses best practice principles of co-regulation, including transparency, accountability and concern for reputational risk [136, 138]. Yet, there appears to be a lack of empirical evaluations of co-regulatory approaches to NCD prevention making it difficult to draw conclusions regarding effectiveness in practice. More empirical evaluations of co-regulatory approaches could help to better understand the effectiveness of this model.

Government regulation was largely mandatory with a small proportion taking a voluntary or mixed approach. Voluntary and mixed government regulated approaches were reported to be effective half the time (50%) compared to majority of the time (87%) for mandated government regulation. Co-regulation utilised voluntary and mandatory approaches, quasi-regulation and self-regulation were all voluntary. Given the reported in-effectiveness of quasi- and self-regulation, and the mixed results for voluntary government and co-regulation, it appears that voluntary approaches are ineffective at achieving the desired outcome. For example, an evaluation of voluntary government nutrition guidelines for Mexican schools found the voluntary nature and lack of monitoring and compliance demonstrated poor uptake and low compliance, resulting in a lack of effectiveness [92]. Another evaluation of voluntary food reformulation in the UK and USA found a disproportionate level of influence on the governing bodies by industry, and that there was a lack of specific and time-bound commitments, lack of participation across the sector, inadequate monitoring mechanisms and lack of enforcement options [107]. Voluntary approaches are promoted by industry groups in an attempt to avoid explicit government regulation and as a mechanism to build trusted relationships with governments [139, 140]. There is a wealth of literature on the ineffectiveness of voluntary approaches in public health law [129, 141‐144]. Voluntary regulation has also been criticised in areas not included in this review including pharmaceuticals [145] and food safety [136] due to low uptake and limited accountability mechanisms.

The overall effectiveness of instruments by risk factor indicated that food and alcohol regulation were reported to be ineffective more often. However, when analysing studies by regulatory form, food and alcohol were more likely to be self-regulated or quasi-regulated compared to tobacco and environmental pollutants. This likely impacts their overall effectiveness. Our results indicate when food and alcohol are government regulated they appear to be more effective. This was particularly apparent for food regulation. Food laws and regulation are not necessarily less effective than tobacco but rather they are less effective due to their design and governance mechanisms being largely quasi- and self-regulated, and lacking independence [142, 146]. These regulatory approaches also lack transparency and accountability.

It is likely that food and alcohol are more frequently self- or quasi-regulated as they do not have the same extensive history of advocacy and research, nor the international laws that govern tobacco and environmental pollutants. International treaties such as the World Health Organization Framework Convention on Tobacco Control (2003), the United Nations Environment Programme Stockholm Convention on Persistent Organic Pollutants (2001) and the International Labor Organization Asbestos Convention (1990) are global treaties to protect human health that require Parties to take measures to prevent and eliminate exposure to harmful products. They have played an instrumental role in shaping local government laws and evolving the regulation of tobacco and environmental pollutants beyond voluntary approaches. Despite decades of evidence on the risks of unhealthy food and beverages including alcohol, some have suggested that these unhealthy commodities remain largely self-regulated or governed by voluntary agreements because these industries have successfully positioned themselves as part of the solution and have learnt from the ‘mistakes’ of the tobacco industry [146, 147]. Additionally, the approach taken is ultimately a decision of government policy and given the sheer effort and decades of research and advocacy needed to implement government regulation, it is likely that these risk factors have been secondary to tobacco and environmental pollutants.

The argument for regulating tobacco is also much clearer in that there is no health benefit to tobacco consumption. For food there is the claim that there may be some nutritional value and that eating small amounts of unhealthy food or drinking small amounts of alcohol are not as damaging compared to tobacco. Therefore, there may not be the political or public will to advocate for mandatory government regulation. The barrier to mandatory food regulation may be even higher as the health claims for alcohol have in recent years been largely rejected by evidence [148, 149]. This has brought about a perception of complexity when it comes to regulating unhealthy food and beverages. This perception may stem from concerns that such integration could introduce unfamiliar complexities, potentially leading to challenges in the implementation and enforcement, and that food is seen as a personal choice that doesn’t inherently harm others the way that tobacco and alcohol can. Additionally, fast-food outlets contribute to local economies by providing employment opportunities and generating revenue. Moreover, the challenges associated with obesity and dietary problems are recognised as multifaceted and extending beyond single factors that can be regulated (e.g. density of outlets) so policymakers are unclear about the potential effectiveness of these interventions. This has often led policy makers to, rather than regulating density, direct efforts towards increasing public awareness, promoting healthier food choices, empowering individuals to make informed dietary choices, and encouraging the food industry to offer healthier alternatives.

Public health researchers have previously stated that there is an inherent conflict of interest in allowing unhealthy commodity industries regulatory control over public health matters [141, 142]. The OECD principles of best practice regulation also prioritise independence, accountability, transparency, preventing undue influence and maintaining trust [10]. Yet, our review shows that regulatory approaches that do not align with these principles remain common practice. There is an industry preference for voluntary, self-regulatory or partnership approaches, providing businesses with more control and the ability to avoid impact on successful business models by utilising less effective regulatory approaches [142, 150]. This renders co-, quasi- and self-regulation problematic from a public health perspective and reinforces concerns around issues of regulatory capture with partnership approaches.

Interestingly the policy area being regulated also varied significantly by the different regulatory form and governance mechanisms. Government regulation (n = 90) focused largely on availability (78%); co-regulation focused mainly on marketing (75%); quasi-regulation focused exclusively on reformulation (100%); and self-regulation focused largely on marketing (88%). Again this is a product of policy decisions and the political economy. It could also be that availability of harmful and unhealthy products has been seen to be more in the public interest and therefore requiring explicit regulation for consumer protection [151]. While marketing and reformulation may be perceived by governments to be more amendable to co-ordination and co-operation, principles that could be achieved via ‘win-win’ partnerships through ostensibly softer regulatory approaches such as co-, quasi- or self-regulation [152]. Ideology also plays a role, with conservative and centrist governments viewing food and beverage choices as personal and individualistic, and therefore restricting marketing via government regulation is not seen as a valid solution [153]. However, as previously mentioned partnership and self-regulatory approaches are susceptible to regulatory capture, conflicts of interest and limited transparency and accountability; particularly as there are limited sanctions on industry players who continue to push back the boundaries on these regulations. Additionally, marketing regulation has become more complex with the rapid growth of new marketing techniques and the huge influence of digital marketing leaving many governments uncertain in how to regulate [154]. There is however examples of government marketing regulation such as new laws in Australia for influencer marketing of supplements and health care products [155], and in the UK for digital junk food marketing [156], which could be evaluated in future to support the public interest argument in other jurisdictions.

It was unclear whether the presence of monitoring or enforcement impacted the effectiveness of a regulatory instrument due to large inconsistencies in reporting. Approximately half of the included evaluations and reviews did not provide detail of whether monitoring or enforcement was present. This is critical design information that can help better understand effectiveness of interventions. Other literature has highlighted the importance of independent monitoring and enforcement as key design features for effective regulation. A comparative analysis of regulatory governance conditions identified comprehensive monitoring as a necessary (or required) condition for positive nutrition outcomes in 94% of cases [157]. They also identified a combination of conditions that when used together are sufficient at providing positive outcomes: absence of high industry involvement, strict regulatory design, best practice design and comprehensive monitoring and enforcement (96% of cases); or absence of high industry involvement, best practice design and comprehensive monitoring (82% of cases). It will be difficult to draw associations through future systematic reviews if evaluations are not reporting vital information regarding monitoring and enforcement. Included studies also infrequently reported on duration of implementation, important factors for assessing effectiveness.

Research in this area should provide more detail of the design and implementation features of the legal intervention they are evaluating. This is critical information to policy makers who may seek to replicate interventions in their own jurisdictions or learn lessons from ineffective attempts. It is also important for understanding the history of the law in question and how it may have evolved overtime following evaluations and government reviews.

There was also a spectrum of what was considered effective by authors and different outcome measures and methodologies used. While this scoping review was not intended as a rigorous quantitative synthesis of the rigor of legal evaluations, it does describe and highlight the diversity of the public health law evaluation literature. For example, Yorifuji 2011 conducted an outcome/ impact evaluation of a mandatory government regulated tobacco law using quantitative methods to measure compliance and correlations with behaviour (decreases in smoking rates) and health (decreases in lung cancer) outcomes. They found the instrument to be effective with increased compliance associated with decreased prevalence of tobacco smoking, and decreased tobacco smoking associated with decreased cancer mortality. While Hadfield 2015 undertook a process and outcome/ impact evaluation using mixed methods to examine implementation and compliance with legislation (mandatory government regulated) and a supporting corporate social responsibility initiative (voluntary self-regulated) for liquor licence premises. They found both the voluntary and statutory frameworks were ineffective and failing to drive good practice because the structures for monitoring and enforcing the legislation were expensive, resource intensive and required evidence that was difficult to obtain, and the self-regulated instrument was not being consistently adopted, lacked clear targets and had high rates of non-compliance.

To our knowledge this is the first study to map the diversity of regulatory evaluation literature and draw linkages between design features and reported effectiveness across a range of public health risk factors. We utilised a comprehensive search strategy across health and legal databases considering multiple risk domains and all legal instruments across OECD countries. The strength of our findings is a testament to this comprehensive approach, focusing on what does and does not work, thus providing a good distribution of effective and ineffective models. The results of this paper may be known anecdotally but this paper empirically reports it.

The assessment of effectiveness was based on what the authors of the papers reported. We did not independently verify the claims of each paper. Assessment of monitoring and enforcement was based on whether it was reported to be a design feature, not whether monitoring or enforcement was actually working in practice. Due to the size and nature of this scoping review it was not possible to analyse the effectiveness of monitoring or enforcement specifically. However, it is worth noting that some evaluations reporting monitoring or enforcement stated they were limited in effectiveness.

Another limitation is the small sample of co-regulated and quasi-regulated instruments. While quasi-regulation is less common, co-regulation is frequently used in public health and it is surprising only four regulatory instruments were identified in the included studies. There also appears to be some inconsistencies in how authors categorised legal instruments by regulatory form. Three quarters of evaluations examined a single regulatory instrument. While this is common, they can be difficult to evaluate if they sit as part of a framework as it is hard to isolate what components are effective.

There were few included studies that focused on pricing regulations. There is extensive literature on taxation and other pricing policies that has not been captured by our scoping review [158]. It is possible that our search strategy did not capture some of this literature that was published before the year 2000. There were also numerous studies on taxation that did not pass full text review as they had study designs outside our scope or they did not provide enough detail of the regulatory instrument being evaluated.

It is also likely that many evaluations may not have made it into peer-review literature due to weak study designs or lack of interest in submitting them to peer-reviewed journals (e.g. government evaluations). Our results found there may be limitations with some of the included study designs. Implementation was the major focus of the included evaluations and very few focused on health or economic outcomes, with no cost-benefit or economic evaluations identified. Stronger study design and the use of rigorous quantitative methods such as interrupted time series analyses should be encouraged to support qualitative implementation assessments. Regulatory instruments are regularly reviewed by government through internal or parliamentary process. It was not within our scope to include these in this review. Some regulations may have been updated following their evaluations, while others may not have. This scoping review is not able to look at changes to regulations following evaluations.

Last, nearly half (42%) of the evaluations are coming from the USA, a high-income country with a unique system of governance. This may also make the transferability of some findings difficult. The field of empirical public health law has originated from the USA, likely contributing to the high number of publications from that country. Many low and middle income countries (LMICs) were also excluded from this review due to the focus on the OECD. Another limitation of our scoping review is its focus on English only publications. We acknowledge that there is work occurring on NCD prevention regulation within countries outside of the OECD and that evaluations are occurring in non-English speaking languages, which have not been captured by our search and may offer lessons for future research.

Our search strategy yielded few studies on pricing when we know there is extensive literature in this area.

As we focused on the OECD, future work could look at countries outside of this group, in particular LMICs which are not well represented in our scoping review. For example, countries in Central and South America have shown leadership in new food and beverage regulation, such as Mexico’s sugar-sweetened beverage tax, which fell outside the scope of our review. LMICs may also provide more evidence on co-regulatory approaches, as they tend to have more limited regulatory capacity, rendering explicit government regulation unfeasible in some contexts. It will be interesting to see what approaches are taken in other contexts and how public health law evaluations are occurring in LMICs. More evaluations on co-regulatory approaches in countries of varying economic contexts would be beneficial.

The Hein Online legal database is not user friendly for systematic review searches in the way that medical and social science databases are. This is likely because systematic reviews are not a methodology used in legal scholarship. However, as the area of empirical legal evaluations grows this may become more common practice. Researchers wishing to conduct similar work in public health law should be aware of the limitations of legal databases.

The included evaluations in this review largely focused on the regulation of unhealthy food and tobacco. More evaluations could be conducted on alcohol and environmental pollutants (including asbestos and chemicals) as there are many lessons to learn from the regulation of these products that may currently be missing from the literature. There were also many studies on other public health risk areas such as pharmaceuticals, food safety and road safety, which were excluded from this review but could feature in future scoping or systematic reviews. Future systematic reviews could also look to draw associations between design and effectiveness, which was not able to be done through this scoping review.

Moving forward scientific research should provide details of the law being evaluated including whether it is voluntary or mandatory and whether monitoring or enforcement is in place. This can sometimes be difficult when publishing in scientific journals with limited word counts, however, this is important evidence that needs to be reported for better understanding and potential replication of the instrument being evaluated.