A realist evaluation of the implementation of a national tobacco control programme and policy in India

Abstract

There is a growing interest in studying and unpacking the implementation of policies and programmes as it provides an opportunity to reduce the policy translation time lag taken for research findings to translate into policies and be implemented and to understand why policies may fail. Realist evaluation is a theory-driven approach that embraces complexity and helps to identify the mechanisms generating the observed policy outcomes in a given context. We aimed to study facilitators and barriers while implementing the Cigarettes and Other Tobacco Products Act 2003 (COTPA), a comprehensive national tobacco control policy, and the National Tobacco Control Programme 2008 (NTCP), using realist evaluation. We developed an initial programme theory (IPT) based on a realist literature review of tobacco control policies in low- and middle-income countries. Three diverse states—Kerala, West Bengal and Arunachal Pradesh—with varying degrees of implementation of tobacco control laws and programmes were chosen as case studies. Within the three selected states, we conducted in-depth interviews with 48 state and district-level stakeholders and undertook non-participant observations to refine the IPT. Following this, we organized two regional consultations covering stakeholders from 20 Indian states for a second iteration to further refine the programme theory. A total of 300 intervention–context–actor–mechanism–outcome configurations were developed from the interview data, which were later synthesized into state-specific narrative programme theories for Kerala, West Bengal and Arunachal Pradesh. We identified five mechanisms: collective action, felt accountability, individual motivation, fear and prioritization that were (or were not) triggered leading to diverse implementation outcomes. We identified facilitators and barriers to implementing the COTPA and the NTCP, which have important research and practical implications for furthering the implementation of these policies as well as implementation research in India. In the future, researchers could build on the refined programme theory proposed in this study to develop a middle-range theory to explain tobacco control policy implementation in India and other low- and middle-income countries.

Key messages.

• The implementation of tobacco control policies varies across settings. In a country like India each state and the districts within them it different tobacco products and practices, implementation machinery and expertise, contextual issues that hinder or facilitate tobacco control.

• The differing implementation outcomes may be generated by different mechanisms, such as prioritization of tobacco control, fear or enforcement, collective action on tobacco control, felt accountability and individual motivation of implementers.

• Policymakers and implementers could focus on the following gaps to strengthen tobacco control efforts—adapting the fines (or the policy instrument), updating the monitoring system to improve transparency, and tackling emerging context-specific issues such as hookah usage, over-the-top advertising, availability and usage of electronic cigarettes (despite a ban).

Introduction

Policy and programme implementation is a crucial phase of a policy lifecycle as it determines how policies may or may not work as intended (Howlett and Giest, 2015). Three schools of thought have dominated the literature on the implementation of public policies: (1) top-down theories of implementation, (2) bottom-up approaches and (3) hybrid theories of implementation (Hill and Hupe, 2014). In the top-down theories the statutory language that has been formally written down as per law is taken as a starting point of the policy process and implementation is then construed as an administrative process ignoring political aspects and emphasizing the role of policy framers, whereas the bottom-up theories emphasize the role of the local autonomy of implementers and local actors. Hence there has been an effort for a synthesis that has led to contributions that balance both perspectives. The advocacy coalition framework is one example of such a hybrid model that includes system elements and policy subsystems with a multiplicity of actors. Over the years, there has been a growing interest and effort to study policy implementation and unpack the black box or the underlying complexity between policy intent and outcomes (Hudson et al., 2019). Theory-driven evaluation or ‘white box evaluation’, where one unpacks and investigates the tangible and intangible components of a programme, has been developed and shaped by several researchers, including Weiss, Chen, Pawson and Tilley (Astbury and Leeuw, 2010), and is gaining significant interest in the field of health policy and systems research (Ramani et al., 2022). Evaluating and learning from policy processes helps in anticipating challenges and incorporating them into the policy process to improve future policies and their outcomes. This is certainly the case for tobacco control policies that have been in existence in diverse settings across the world and can benefit from incorporating learnings from evaluation.

Globally, tobacco consumption kills eight million people and >1.3 million adults in India each year (Ministry of Health & Family Welfare Government of India, 2017). According to the Global Adult Tobacco Survey (GATS), 28.6% (∼267 million) of Indian adults use tobacco in some form (Ministry of Health & Family Welfare Government of India, 2017). Additionally, India’s economic burden of tobacco-related diseases was INR 1773.4 billion (US$27.5 billion) for the year 2017–18 which was 1.04% of the country’s gross domestic product (John et al., 2021). Several tobacco control policies have been enacted and adopted in India in the last few decades such as the Cigarettes and Other Tobacco Products Act, 2003 (COTPA) and its amendments, (Saxena et al., 2020). The COTPA and the National Tobacco Control Program (NTCP) are two important tobacco control measures undertaken by the Government of India. COTPA is a comprehensive Act applicable to all states and covers all varieties of tobacco products including smoked and smokeless products. It mandates prohibition on smoking in public places, advertisements of tobacco products, sale of tobacco to and by minors, sale of tobacco products without prescribed pictorial health warnings and around educational institutions, among other tobacco control provisions. The NTCP is a national programme that provides a structure and finances for the effective implementation of COTPA and other tobacco control efforts in the country. The programme also focuses on raising awareness on tobacco harms and offering tobacco cessation services (Gazette-of-India, 2003; Nazar et al., 2020). In the field of tobacco control, evaluation of policy implementation is at a nascent stage. One such example in the past decade, attempting to rigorously evaluate tobacco control policy in diverse settings [31 countries across all World Health Organization (WHO) regions], is the International Tobacco Control (ITC) Project. The ITC Project measures the psychosocial and behavioural impact of key policies of the WHO Framework Convention on Tobacco Control (WHO-FCTC), but does not directly examine implementation (Fong et al., 2006). Recent studies evaluating COTPA and NTCP in India include compliance assessments of selected provisions of COTPA in sub-national jurisdictions, such as smoke-free provision, direct and indirect advertising, vendor compliance, quantitative assessments, studying policy processes, the Ministry of Health and Family Welfare’s financial and coverage assessments, and cross-country comparative evaluation (Goel et al., 2016; Rath et al., 2018; WHO, 2018; Department of Health and Family Welfare, G. of I., 2019; Ali et al., 2020; Nazar et al., 2020; Pradhan et al., 2020; Anu et al., 2021; Mondal and Van Belle et al., 2022), but detailed explanations of why policies work in certain settings and fail in others is lacking.

Realist evaluation is a theory-driven evaluation method that embraces complexity and seeks to explain the mechanisms that operate in relation to the context in order to generate outcomes (Gilmore, 2019). This method has been applied in several contexts and for the evaluation of a variety of programmes and policies (Marchal et al., 2012). Pawson and Tiley conceptualized mechanisms as a combination of resources offered by the programme or policy under study and the stakeholder’s (e.g. implementer’s) reasoning in response to the resources offered by the programme or policy. Recent definitions have explicitly disaggregated mechanisms into (1) resources and (2) reasoning, thereby capturing mechanisms that are possible at several levels, such as at the relational (meso) and structural (macro) levels. This is further explained by (Dalkin et al., 2015) as ‘Interventions resources are introduced in a context, in a way that enhances a change in reasoning. This alters the behaviour of participants which leads to outcomes’, which describes how the programme changes the way stakeholders perceive and reason about an issue, and whether the change in reasoning actually led to a change in the way a stakeholder behaves or reacts. The purpose of realist evaluation is to test and refine programme theories by identifying the contexts (C) in which the intervention was introduced, the mechanisms (M) that are likely to be triggered (or not), and how this combination generates outcomes (O) (Pawson and Tilley, 1997). The field of realist evaluation is under development, evolving and growing in clarity through methodological discussions amongst realist researchers. Recent studies have employed a variety of quantitative and qualitative methods to deepen the understanding of tobacco control policy implementation, such as two recent realist reviews, which theorize and evaluate the implementation of tobacco control policies (Mlinaric et al., 2019; Hebbar et al., 2022). Tobacco control policies could originate from and target actors/policies in different sectors (e.g. health, environment, finances, education etc.) and typically engage a wide variety of stakeholders from the policy formulation to the implementation phases, such as the departments of police, education, transport, food safety, excise, panchayat raj (local self-government), judiciary, and non-governmental and private agencies such as media businesses (Bhojani and Soors, 2015). Drawing attention to the important role of various actors and their interactions, social network analysis has been used by Mondal et al. to explain local district-level multisectoral governance for tobacco control in India (Mondal et al., 2022). In India, the application of realist methods is in the early stages and is seeing an upward trend, with studies in the health sector in evaluation of programmes related to diabetes and depression, school eye health, tobacco control policy implementation and district managers capacity-building programme evaluation (Marchal et al., 2024).

Through this realist evaluation, we aim to study the facilitators and barriers to implementing the COTPA, a comprehensive national tobacco control policy, and the NTCP. The NTCP was initially piloted in two districts per state in 21 states of India (42 districts). Currently, the programme covers 612 of the 766 districts across all 36 states/union territories of India (Ministry of Health and Family Welfare Government of India, 2023). We evaluate four major and notified provisions of COTPA: prohibition of (1) public smoking, (2) advertisement of cigarettes and other tobacco products, (3) sale of tobacco to and by minors and near educational institutes, and (4) on mandating pictorial health warnings on tobacco products. We also evaluate the priority areas of NTCP where the government plans to focus on: (1) training and capacity building of health and social workers, non-governmental organizations (NGOs), school teachers and enforcement officers, (2) information, education and communication activities for awareness regarding tobacco harms, (3) school programmes regarding tobacco control, (4) coordination and monitoring of tobacco control laws, and (5) setting up and strengthening cessation facilities. The implementation of COTPA and NTCP has been varied in India as reported in the literature, with low compliance to smoke-free provisions and signage requirements, and no updating of amounts of fines, sale and advertisement of tobacco products near educational institutes across different cities and districts in India (Nayak et al., 2010; Tripathy et al., 2013; Goel et al., 2014; 2016; Rath et al., 2018; Pradhan et al., 2020; Anu et al., 2021). The main research question we address in this study to better understand the drivers of implementation is: how did the selected (three) Indian states under study implement COTPA and NTCP between 2003 and 2018? What were the facilitators and barriers to effective policy implementation?

Methods

Study design

We used a realist evaluation study design that allowed us to investigate important national and state-level contextual factors, mechanisms underlying the implementation process, and the facilitators and barriers during the implementation of COTPA and NTCP in the three selected Indian states. We used the Realist And Meta-narrative Evidence Syntheses: Evolving Standards (RAMESES) II reporting standards (see supplementary file S1, see online supplementary material) for reporting this study. The study is divided into three steps, as depicted in Figure 1. The institutional ethics committee at the first author’s institute approved the study via letter IEC-FR/01/2020.

Figure 1.

Updated realist cycle of inquiry followed in this study initially published in the study protocol in BMJ Open reproduced with permission from BMJ Open. http://dx.doi.org/10.1136/bmjopen-2021-050859 (Hebbar et al., 2021)

Step 1: developing the initial programme theory

Broadly, programme theories are ideas about how programmes cause the intended or observed outcomes and provide plausible explanations on why interventions work or not in specific settings (Shearn et al., 2017). An initial programme theory (IPT) was developed based on theory gleaning from: implementation frameworks, the authors’ decade-long experience working in tobacco control policy implementation in India, and discussions with experts in the field. This IPT was refined through a realist literature review that identified the context–mechanism–outcome (CMO) configurations that would help explain the facilitators and barriers to implementing tobacco control policies in low- and middle-income countries (LMICs) (Hebbar et al., 2022). The programme theory developed at the end of the realist literature review focused on four major strategies: awareness, enforcement, intersectoral coordination and tobacco industry interference. This served as our IPT for this realist evaluation. We refer to Hebbar et al. (2022) for an overview of IPT development.

Study setting

India is a union of 28 states and eight union territories. We utilized a nationally representative survey, the GATS, which uses standardized tools and methodology and is conducted periodically in several countries globally to monitor tobacco use and tobacco control measures. We ranked all Indian states and union territories by examining the GATS India rounds 1 (2009–2010) and 2 (2016–2017), using proxy indicators of implementation of COTPA and NTCP. The proxy indicators of implementation of COTPA and NTCP include exposure to second-hand smoke, exposure to direct and indirect tobacco product advertisements, exposure to pictorial health warnings on tobacco products and awareness of the harms of tobacco use. More information about these analyses can be found elsewhere (Hebbar et al., 2023). Each state and union territory was placed into one of three categories: (1) states that showed an improvement in implementation of proxy indicators up to 2018, (2) average performing states that showed no significant change, and (3) states where the implementation had worsened between 2003 and 2018. A shortlist of three states in each category was circulated to 10 tobacco control experts from India, including both internal and external to the project, who helped select one state from each category. Kerala, West Bengal and Arunachal Pradesh were selected as the cases for this study.

Step 2: refining the IPT

Sampling and participant recruitment

The study population consisted of individuals directly or indirectly involved in the implementation of COTPA and NTCP, such as state-level programme designers, implementers, state- and district-level programme staff, officers from police, education, finance, municipalities, gram panchayat (local self-government bodies or municipalities), NGOs working on tobacco control and members of the media in each of the three states. The index participants meeting both the inclusion criteria were purposively selected and we later snowballed to other relevant participants.

Inclusion criteria were as follows.

1. Participants must be designated decision-makers/managers or implementers of tobacco control or individuals from the media, academia and civil societies.

2. Participants must be currently working or must have worked in the past in tobacco control within the selected states.

Interviews

We interviewed 48 individuals from several government and non-government organizations. (see Table 1). Most of the participants were interviewed via telephone or online (via Zoom) by P.B.H., V.D. and one trained local data collector in Arunachal Pradesh and in West Bengal. Most interviews were performed during the COVID-19 pandemic between September 2020 and January 2021, and some interviews were performed face-to-face by P.B.H. and V.D. between October 2021 and December 2021. The research team used a semi-structured interview guide prepared using a realist interviewing approach (Manzano, 2016) and explored potential mechanisms and key elements of the IPT (tobacco control-related awareness, enforcement, intersectoral coordination and tobacco industry interference) (supplementary file 2, see online supplementary material). As data collection took place during the subsequent COVID-19 waves and many of the officials we interviewed were charged with additional COVID-19 relief duties, we included specific challenges faced to implement COTPA and NTCP during the COVID-19 pandemic to help initiate conversation and build rapport during the interview. Interview questions were tailored based on stakeholder profile to clarify, reject or refine key elements of the draft IPTs. Informed consent was taken before every interview. Detailed notes were taken after the interviews capturing the researcher’s reflections. Each interview usually lasted for ∼45–60 min and was audio recorded. Interviews and notes were digitally transcribed and managed in NVIVO software V.12 where they were coded line by line and later categorized into broad themes relevant to the study.

Table 1.

Characteristics of data collected in three Indian states

Characteristics	Kerala	West Bengal	Arunachal Pradesh
Population	33 406 061	91 276 115	1 383 727
Number of administrative districts	14	23	25
Literacy rate	93.91%	77.08%	66.96%
Urban–rural population distribution	52.29% rural	68.1% rural	77.06% rural
Tobacco usage (GATS 2017)	12.70%	33.50%	45.50%
Change in tobacco usage (GATS 2010 to GATS 2017)	8.7% reduction	2.8% reduction	2.2% reduction
Financial resources sanctioned for NTCP versus utilized for 2016–2019 (Department of Health and Family Welfare, G. of I., 2019)	Sanctioned—53.19 Lakhs Utilized—32.57 Lakhs Utilization—61.2%	Sanctioned—238 Lakhs Utilized—48.53 Lakhs Utilization—20.39%	Sanctioned—140 Lakhs Utilized—51.13 Lakhs Utilization—36.52%
Proxy indicators of COTPA and NTCP implementation (GATS 2017 data) – Exposure to second-hand smoke – Exposure to direct or indirect advertisements – Exposure to pictorial health warnings – Awareness about the harms of smoking tobacco use	17.1% 10.3% 78% 93.9%	24.4% 39.8% 74% 96.3%	22.8% 38.3% 49% 90.5%
In-depth interview data characteristics
Men	8	14	15
Women	2	5	4
State-level departments
Health	4	8	6
Police	-	1	2
Education	-	1	3
General administration	1	-	-
NGO	2	3	4
Academia/research	-	4	-
Clinicians	1	-	2
Media	1	1	1
Women and child development	-	-	1
Excise	1	-	-
Municipality/local government	-	1	-
Total interviews completed (n = 48)	10	19	19
Non-participant observation data characteristics
COTPA-related observations	36 observations (360 min)	32 observations (365 min)	26 observations (345 min)
Government office observations	5 observations (50 min)	6 observations (390 min)	1 observation (10 min)
Total observations completed	41 observations (435 min)	38 observations (755 min)	27 observations (355 min)

Observations

Besides interviews, a total of 106 non-participant observations were performed in the three states (Arunachal Pradesh, West Bengal and Kerala) by research team members P.B.H., V.D., and data collectors Praveen Rao and Kumaran P. trained for data collection (Liu and Maitlis, 2013). Due to the pandemic, the team was initially unable to travel. When the cases of Covid-19 subsided, between October 2021 and December 2021, the team visited each state, and all four members used a predesigned piloted observation guide and collected data; P.B.H. led daily de-brief meetings after reviewing all the filled data files. The team conducted non-participant observations in government offices and public places where enforcement and violations frequently take place of e.g. smoking in public places, the presence of tobacco product advertising in public places and the sale of tobacco products near educational institutions. The aim of the observations was to understand the contextual factors, corroborate the impressions that arose during the in-depth interviews and take forward the process of sensemaking (Liu and Maitlis, 2013). The process of sensemaking is to be able to understand, interact, interpret and present the collected data in order to meaningfully complete the research process (Liu and Maitlis, 2013). Using multiple data sources like we did in this study, we are able to better situate our understanding after comparing and contrasting observations with coded interview data. Additional sources of data included detailed field notes and photographs to record, triangulate and reflect on the observational data. Data collection in the field was planned in keeping with gender-specific norms of tobacco use in India. The tea shops and other locations in public places where smoking and chewing tobacco is common in most settings has male users as well as sellers, and a female data collector observing and taking notes does not blend in and raises suspicion/curiosity amongst bystanders. Hence data collection was undertaken in pairs by three male data collectors and the first author, a female.

In public health research, photographs serve as a vital tool for recording, triangulating and reflecting on observational data, as well as for documentation. (Alenichev et al., 2024) (Graham et al., 2019) The study team took photographs relevant to tobacco control implementation to capture the broader environment where the non-participant observations were made, and to triangulate data from the literature and interviews along with field realities. These included close-up views of objects (tobacco products) to understand signage, packet warnings and illicit sale of tobacco; location and place-based characteristics (city versus suburb, areas near schools, markets and government offices, and public spaces like streets, bus stops and parks); and the conditions in which tobacco consumption and sale occurs (solitary smoking or communal smoking around tea shops, people waiting for the bus, buying groceries, etc.). The team avoided capturing identifiable features of people to protect their identity in the photographs. In addition, the team also used anonymizing techniques such as blurring if needed. Additional analyses such as thematic and content analysis were not conducted.

Data analysis

P.B.H. and V.D. analysed the interview transcripts and observation field notes including photographs and cross-checked the data to ensure quality and accuracy. Based on a preliminary coding analysis (in Microsoft Excel), P.B.H. and V.D. developed intervention—context—actor—mechanism—outcome (ICAMO) configurations (Marchal et al., 2018). The first iteration of the programme theory built upon the IPT which was structured in the CMO configurations, and since the ICAMO provided more clarity for each of the actor groups, the second iteration of the refined programme theory (RPT) was developed using the ICAMO version. The ICAMO version was preferred over the typically used CMO configurations to capture the different roles actors involved played in the process and for the different interventions under study (Marchal et al., 2018). P.B.H., V.D. and U.B. reviewed the ICAMO configurations, observations, photographs and field notes and developed a case study and if-then-because propositions for each state. The thick if-then-because propositions were a step in furthering our data analysis and are a narrative formulation of the CMOs. The ‘if’ component elaborates the context part, the ‘then’ component elaborates the outcome part and ‘because’ elaborates the mechanisms part. Visualizing/tabulating as CMOs and developing if-then-because statements are all ways of data analysis after coding. The if-then-because propositions explain either facilitators or barriers. This was discussed within the research team and further refined with the inputs received. Regular reflexive interactions took place within the research team to aid the data analysis process while also identifying issues in the conduct of the study.

Step 3: testing and further refinement of the programme theories

To refine the programme theories (PTs), two regional consultations were organized covering 20 Indian states with 55 participants. The participants included state-level government officials from departments of health, food safety, police and education, and civil society members and academics involved in tobacco control policy implementation. There was no overlap between the participants selected for in-depth interviews and invitees of the regional consultation. The first consultation occurred online in March 2022 while the second consultation was organized in-person in the Indian state of Chhattisgarh in April 2022. We examined the best practices and challenges these states face in implementing COTPA and NTCP and performed group discussions to validate our findings. Each state representative presented the state context, achievements and challenges related to COTPA and NTCP implementation. The group discussions moderated by P.B.H., V.D. and U.B. focused on the strategies of awareness, enforcement, intersectoral coordintion and tobacco industry interference strategies.

After the regional consultations, we used the three data sources (in-depth interviews, non-participant observations and regional consultations) to triangulate the data, refine our analysis and examine coherence in our interpretations. The ICAMOs, which were mainly developed from the in-depth interview data, were checked against observation and photographs to confirm patterns and interpretations of the data. The regional consultation and focus group discussions helped us understand several best practices and contextual factors that facilitated or hindered probable mechanisms. The state level if-then-because propositions were then synthesized into one programme theory for each of the three states.

The analysis is structured using an actor-oriented approach (Long, 2001), as the data presented insights related to four major categories of actors:

• law enforcers, which included health, police, panchayat raj and other departments authorized to enforce COTPA

• civil society organizations/non-governmental organizations

• policy makers such as the state-level heads of departments who are not involved in direct enforcement but rather in a supervision, planning and monitoring role

• citizens (the public) on whom the law is being implemented

A total of 300 ICAMO configurations were developed from the 48 in-depth interview transcript data. These were further synthesized to six if-then-because formulations for Kerala, nine if-then-because formulations for West Bengal and five if-then-because formulations for Arunachal Pradesh (supplementary file 3, see online supplementary material).

Results

The three selected Indian states differ in terms of their geographic size, topography, location within the country (South, East and Northeast) and population. Half of the interviewees from these states worked at the district level, while the other half worked at the state level; one interviewee worked at the state and national levels. Table 1 provides defining characteristics of the contextual factors, an overview of the data collected and their usage in this study.

Below we describe each of the three cases, zooming-in on state contexts, actors and legislative and programmatic platforms available to the actors. One of the commonly raised policy barrier across the states was the lack of an upward revision in the amounts of the fines in COTPA, which have become increasingly affordable over the years.

Case study of Kerala

Tobacco control activities have been initiated by the state health services department since 2008 and situated under the NCD control programme since 2010. In Kerala, where the level of implementation is comparatively better, NTCP was piloted in four districts in 2014–2015 and scaled up to all 14 districts in 2019. As early as July 1999, Kerala had already banned smoking in public places based on a public interest litigation by a concerned citizen. Kerala is a progressive state with leading health and development indices nationally and with a relatively low tobacco prevalence of 12.7% (Govt of India, 2019). Kerala has lower prevalence compared to the other two states and the third lowest in all of India, preceded only by Goa and Pudducherry. Kerala reports a high literacy rate, and families tend to have a matriarchal structure. It also had long-standing history or path dependency of tobacco control efforts, even before COTPA there were state-level initiatives, e.g. active citizenry demanded the regulation of smoking on public transport. According to a clinician active in tobacco control,

“…there were a lot of these civil society groups and others were actively doing all this. In fact, I think, here the World No Tobacco Day was started in Kerala around the late 80s. So that awareness part, it was going on, and then I think a lot of other NGOs also took up decisions, various associations, and all, even Resident Associations, so I think it gradually picked up automatically” (KT021 male clinician).

One of the contextual barriers is the influence of communism as a political ideology because of the symbolism of smoking in public and the change in social norms. In addition, reforms and unionization of the bidi (Indian traditional leaf rolled tobacco) sector with extensive manual labour was championed by leaders from this political ideology promoting the symbolism of bidi smoking in public (Franke et al., 1998), thus further shaping the social norms around smoking of bidis and embedding it within the routines. Another barrier is the influx of migrant labourers from other states where chewing tobacco prevalence is high. These contextual factors were confirmed during our observations and in the interviews with various stakeholders from the state.

“I would say culturally because like it has been very strongly entrenched with the Communist movement of the state. Bidi… So, policies are kind of, you know… there are policies which are made to kind of attune to those industries. It’s a small-scale industry, but still a large amount of protection is given to bidi manufacturing in the state. So, bidi use and you know, even in the small tea shops and these shacks, people would gather in the morning. They’d read the newspaper, and they’d also be smoking bidi. So, it is kind of associated with a normal life of a person in the state” (KT01 media professional).

In Kerala, tobacco control is seen as collective action by multiple stakeholders, as indicated in interviews and the regional consultation. Supported by media, civil society, district administration, police, education, local self-government, regional cancer centres, sensitized politicians and external donors, the health department was able to play an effective stewardship role. The shared vision of tobacco control, lack of ego issues, rapport and trust amongst stakeholders were key to facilitating such collective action.

“Success in tobacco control has been because of collective action. Whether it be the government or a private party or an NGO, whenever somebody would initiate a good activity, all the other tobacco control enthusiasts would pitch in without any kind of an ego issue or something. They feel that collective action is more important” (KT01 media professional).

“It’s a movement actually. We believe that it’s a movement, because in the sense only health department cannot do anything. NTCP general guidelines, it is very clearly said that it is not a health department activities, every interdepartmental collaboration should be there”(KT02 male civil society representative).

Synthesizing the ICAMO configurations focusing on several mechanisms, interventions and actors related to COTPA and NTCP we present an example of a programme theory that attempts to capture the implementation of COTPA and NTCP in Kerala in a CMO format (see Box 1).

Box 1. CMO configuration explaining implementation of COTPA and NTCP in Kerala.

In a state with high literacy and leading health and development indices due to a robust public health infrastructure and a path dependency on tobacco control laws and civil society action, in spite of communist politics embedding bidi use into daily routine (C), multiple stakeholders from the government departments like health, excise, police, administration, education, local self-government, tourism etc. and private stakeholders like regional cancer centres, media and civil society organizations can engage in ‘collective action’ without ego or coordination challenges due to underlying trust, shared vision and effective resource pooling (M), leading to strict enforcement-led behaviour change of compliance to laws, a widely accepted movement of tobacco control and reduction in tobacco use (O).

Case study of West Bengal

In West Bengal, tobacco control activities are led by the health department and supported by police, education and limited civil society engagement. In the 12th 5-year plan for development (2012–17), two districts were initially taken up under NTCP, followed by scaling-up to all districts. The West Bengal Prohibition of smoking and spitting and Protection of the health of Non-smokers and minors Act, 2001 was enacted in a context of high prevalence of chewing tobacco in the state before COTPA and NTCP. Historically, smoking cigarettes/cigars has been depicted in West Bengal as a habit of the elite or intellectuals, influenced by communist politics. There is a strong presence of the tobacco industry in West Bengal, population density is high and there is a high social acceptance of tobacco use. Tobacco products are easily accessible not only in tobacco-specific shops but in all types of shops selling other products and tobacco in urban and rural settings, as confirmed through our observations.

“It is not about religion, or it is not about any other thing. It is basically that the communists, they used to be bidi smoking, and it has become very prominent that the ideologues, the intellectuals, are having the ‘jhola (bag)’ with a bidi in his or her hand. So, it signifies that you are intellectual.”(WB01 male civil society representative).

A movement or mobilization for tobacco control is lacking as observed by officials from the health department.

“In COTPA, it seems to me that COTPA is too much of an administrative protocol driven by law. Which has nothing to do with the practical situation. There is much emphasis on legal provisions. The law does not say much about social reform, about social mobilization, about community mobilization” (WB032 district-level health official).

The significant role of individual motivation for enforcement in an environment or system that does not demand accountability is seen in the following quote.

“It can be done only on the basis of personal interest, what I found actually. It was my personal interest that I insisted or I instructed as a nodal officer. So, it differs from person to person” (WB040 state-level police official).

Violations of the provisions of COTPA attract compoundable (that can be settled with the offender paying money in lieu of his/her prosecution) and non-compoundable offences. Hence, the fear from enforcement authority about COTPA generally pertains to paying monetary penalties and a possibility of facing charges in court and typically a longer litigation period. There is also a negative social cost to being perceived as a law offender. In India, not all legal provisions are actively enforced on a routine basis and hence, it is the perception of whether a particular law is being enforced or not that drives the fear related to enforcement. Fear of enforcement was observed in Kerala as well as confirmed in interview data; such fear was lacking in West Bengal and Arunachal Pradesh as per the observation data and depicted in a quote below.

“Oh, you need two hundred, okay, I’ll give you 200 don’t disturb me all the day. This is the attitude so penalty thing I don’t think this is fine” (WB045 district-level official).

Consolidation of a programme theory in a CMO format that attempts to capture the implementation of COTPA and NTCP in West Bengal is as shown in Box 2.

Box 2. CMO configuration explaining implementation of COTPA and NTCP in West Bengal.

In a state with high population density and tobacco use prevalence, very high social acceptance of chewing tobacco and spitting, and normalization of tobacco use by communist politics and intellectual or elite portrayal of smoking, sparse civil society engagement for tobacco control, and non-existent citizen participation, inertia and the siloed nature of work through vertical programmes within the health system, low awareness of the law and harms of tobacco use, strong presence of the tobacco industry (C), the ‘felt accountability’ of officials to the public as well as to the system, ‘individual motivation’ to act and ‘fear’ amongst the public of sanctioning (enforcement/fines) is not triggered (M), leading to implementation being defined very narrowly by officials as a few fines levied, continued tobacco usage and flouting of laws by the public and the lack of momentum or movement for tobacco control, ultimately leading to continued tobacco usage (O).

Case study of Arunachal Pradesh

The tobacco control activities in Arunachal Pradesh are led by the health department and supported by police, education and limited civil society engagement. In the 12th 5-year plan (2012–17), two districts were initially taken up under NTCP, followed by scaling up to all districts. Activation of the programme in the state was comparatively delayed.

“Although NTCP was launched in our state many years ago, it was in a dormant state. Now the state tobacco control cell has actively taken it up in the past couple of years”(AR010 male clinician).

Arunachal Pradesh is one of the Northeastern states of India with a hilly and difficult to access terrain. There are international borders and security concerns in the state. Arunachal Pradesh has a majority tribal population with comparatively poor health and development indicators. The society is largely matriarchal with women leading homes and shops, there is a lack of parental supervision of children and family structures are not closely knit (who might condone tobacco use amongst youth). There is a high social acceptance of tobacco use.

“Increasingly parents are spending lesser and lesser time with children, not bothering about what they are up to and have very independent lifestyles”(AR02 female civil society representative).

On being questioned about the availability of tobacco products and an increasing trend in tobacco usage, an official reflected the following.

“Availability of the thing, as you can go to any goomti (petty shops selling tobacco products), any corner you can stand and get those products very easily. So I think availability is also added to the increase and again there is a peer pressure as well. So if you find some of your elder brothers, sisters taking it, and they feel it’s normal and acceptable. So I think this all contributes to the increase in the consumption as well”(AR043 woman and child development department official).

Some respondents shared about the local context and how people feared to report violations or take action to avoid repercussions.

“not many teachers are from here so whenever we train them also because they are not authorized to take the challan address (enforcement receipts) also, they are scared to do it. Because the people out here are violent… Not every case but yes, there’s a few. Especially the principals comes from outside state. So, they are scared to deal with the shopkeepers” (AP037 district-level official).

Consolidation of a programme theory in a CMO format that attempts to capture the implementation of COTPA and NTCP in Arunachal Pradesh is as shown in Box 3.

Box 3. CMO configuration explaining implementation of COTPA and NTCP in Arunachal Pradesh.

In a state with a high prevalence of tobacco use, high social acceptance and normalization of tobacco use within the family unit, which is geographically remote with international borders and security issues, illicit trade and drug usage concerns, sparse civil society attention to tobacco control, weak and non-transparent public systems, with a history of reliance on the central government for policy guidance and support (C), implementation of tobacco control law remains a non-issue due to ‘lack of prioritization’ and demand for accountability amongst officials, whereas the public has ‘no fear’ of sanctioning, but fears repercussions of reporting violations to authorities (M), leading to largely dormant policy status with sporadic implementation efforts, no community mobilization, minimal intersectoral coordination and continued high prevalence of tobacco use (O).

Using the interview and observation data from the three selected Indian states, we tested the IPT. This first iteration helped further contextualize the IPT from a LMIC perspective to Indian state settings. There were some deletions and additions made to the IPT indicated by the blue font in Figure 2.

Figure 2.

First iteration of RPT I

The observation data from the three states improved our understanding of the contextual factors and confirmed the patterns described above that were picked up in the interviews.

RPT

The focus group discussions with 55 state-level implementers from health, police, education, civil society and academia from 20 Indian states helped us confirm our observation and understanding of implementation facilitators and barriers. In the two consultations, eight focus group discussions were conducted with six to seven participants on thematic areas of the RPT: awareness, enforcement, intersectoral coordination, tobacco industry interference and review systems. The most accepted mechanism was the prioritization of tobacco control or lack thereof, which was relatable to several stakeholders, almost all of whom were state-level decision makers/implementers in their varying contexts. The stakeholders also agreed to the five major strategies identified in the RPT; there was no further addition of strategy beyond the ones specified in our RPT. The mechanism of fear was also widely accepted amongst stakeholders in the consultations as nudging compliance from the citizens (public). The mechanism of collective action identified in Kerala was reiterated by participants from Rajasthan, Himachal Pradesh and Karnataka in the regional consultations.

Some of the major barriers shared by the participants included lack of awareness and knowledge of COTPA amongst other department stakeholders beyond the health department, tobacco industry interference, low and affordable fines, and lack of periodic review at district/state and national levels. The state-level findings were confirmed in the group discussions and led to the second iteration of the RPT.

Figure 3, showing RPT II, highlights some enabling and disabling contextual factors, mechanism resources provided by the interventions (COTPA and NTCP), the actor groups where specific mechanisms were most likely to be seen and the mechanism responses that can (or not) be triggered leading to different outcomes. The ICAMO version of the heuristic helped articulate specific mechanisms for specific actor groups. The plus signs indicate situations when the mechanism is triggered and minus signs when the mechanism is not triggered.

Figure 3.

Second iteration of RPT II

Discussion

The major mechanisms explaining the diverse tobacco control policy implementation outcomes in the selected three Indian states were: collective action, felt accountability, individual motivation, fear and prioritization. These mechanisms were triggered in certain contexts and not others, leading to varying implementation outcomes. The major facilitators and barriers to COTPA and NTCP implementation were identified and discussed in the interviews and the regional consultations. These discussions helped validate our programme theories and the theory refinement process. The multiple sources of data utilized in this study, such as in-depth interviews, non-participant observations, field notes and photographs, and focus group discussions in regional consultations, helped deepen our understanding of contextual factors and mechanisms across multiple settings.

The mechanisms for collaboration in various settings have been subject to study in recent times. A study on organizational collaboration in healthcare theorized and explained determinants for the mechanism of collective action, which include foundational relationships, shared vision, values, structures and processes, and views about the nature of the collaboration and implementation (Rycroft-Malone et al., 2016). A recent realist study examined how collaborations worked, emphasizing and building upon the mechanisms of trust, faith and confidence to explain collaborative behaviour in interorganizational collaborations in healthcare settings (Aunger et al., 2021). The mechanism of collective action seen in Kerala and reiterated by the states of Rajasthan, Karnataka and Himachal Pradesh during the regional consultations encompasses certain elements such as the initial trust and effective pooling of resources across programmes seen in the web of causality of collaborative behaviour identified by Aunger et al. (2021). The key mechanisms for collective action in a study of the ‘Collaborations for Leadership in Applied Health Research and Care’, with the potential to close the ‘know–do gap’, include facilitation, motivation, review and reflection, which were similar to those in the Kerala ICAMO configurations (Rycroft-Malone et al., 2015).

Felt accountability has been defined as ‘an implicit or explicit expectation that one’s decisions or actions will be subject to evaluation by some salient audience(s), with the belief that there exists the potential for one to receive either rewards or sanctions based on this expected evaluation’ (Hall et al., 2017). The relatively unfavourable contextual conditions in West Bengal, such as relatively high prevalence and high social acceptance of tobacco use, and tobacco industry presence and interference leading to sub-optimal implementation outcomes, can be explained partly by the mechanism of felt accountability and individual motivation. This was evident as some government officials who had successfully implemented COTPA and NTCP in a campaign mode in certain districts with supportive contexts chose not to prioritize tobacco control when they were moved to another district as there was no external motivation or review by officials higher up to whom they were accountable. There are several theories on motivation and accountability in the literature; a recent article from public administration literature connects the theory of accountability to individuals beyond exisiting theories for organizations, framing felt accountability as dependent on whether the officials think they will be accountable to the higher-ups in the future (Diem Vo et al., 2022; Overman and Schillemans, 2022). The mechanism of individual motivation linked with felt accountability was identified as not being triggered in the data from West Bengal interviews and in Gujarat. These mechanisms were triggered in Rajasthan, Andhra Pradesh and Chattisgarh states, possibly leading to increased programme activities.

The mechanism of fear plays out differently in each of the three study states. While in Kerala, the fear of strict enforcement led to positive behaviour change amongst the public and the vendors, in West Bengal, owing to the sporadic enforcement of COTPA, the public had no fear of flouting the law. The fear prompted by felt accountability amongst the officials facilitated activities in some settings. In contrast, in Arunachal Pradesh, the fear of repercussion from the violators or their kin prevented citizens from reporting violations to the authorities. Policy instruments or types of interventions have been defined and broadly categorized by theorists into three possibilities: persuasion (sermons), incentives (carrots) and controls (sticks) (Bemelmans-Videc et al., 1998; Salazar-Morales, 2018). Several stakeholders from Kerala, Karnataka, Rajasthan, Himachal Pradesh and other states echoed the need and effectiveness of persuasion through information and educational campaigns and complementary controls or strict enforcement. The sequence and correct balance between awareness and enforcement was a matter of intense discussion in the regional consultations. The differences in opinion stemmed from their positionality of implementing COTPA and NTCP from different departments such as health, education and police. As one of the Kerala interviewees put it, awareness and enforcement are like two wheels of a bicycle, both necessary to take forward the journey of tobacco control.

The process of prioritization of health issues is deeply complex and political at the national and sub-national levels, with competing priorities in India and beyond (Heller et al., 2019; Bhojani et al., 2022; Connell et al., 2022). Lack of prioritization of tobacco control policy implementation against several state-specific competing issues was seen in Arunachal Pradesh and was the most widely accepted mechanism in the regional consultations. Lack of prioritization was also seen as a dominant mechanism in a recent realist review on implementation of smoke-free provisions (Mlinaric et al., 2019).

Strenghts and limitations

The strengths of this study include the diverse nature of the research team, large domain knowledge, and expertise of some of the research team members. This is the first attempt at theorizing policy implementation using realist and implementation research approaches in India and studying tobacco control implementation at the state level. The study fills an important knowledge gap with policy recommendations to improve practices. It builds on a previous realist review of tobacco control policy implementation in LMIC settings and deepens the knowledge on factors specific to tobacco control policy implementation in Indian settings. COTPA was enacted in 2003 and the NTCP in 2007–08 and this evaluation after two decades provides a balanced and detailed status of current implementation facilitators and barriers. Engaging with stakeholders through interviews and regional consultation provided an opportunity to co-produce knowledge guided by implementers’ experiential wisdom.

The study also has some limitations, such as limited time to capture the complex political economy at the state level, as India is a large country with varying contexts in each state. Because of pandemic-related restrictions, we had to shorten our travel times to the states, but working closely with local organizations who have an in-depth understanding of the local context helped the research team to address this limitation. There are several other tobacco control policies in each of the states and at the national level that are in force simultaneously and may have an impact on the implementation of COTPA and NTCP in those particular states, but examining these interactions were out of the scope of this study. It is also beyond the scope of this study to provide micro-level mechanisms for each category of stakeholder and sub-system or provision of COTPA and NTCP. The sociocultural contexts of the three states were extremely diverse, and especially in the North-eastern state of Arunachal Pradesh, the team found it difficult to blend in for the observations, as the population was sparse and the research team aroused suspicion among bystanders. This may have impacted the initial days of data collection as we were acclimatizing to the new setting. But through team reflections we understood the shared phenomena and discussed probable solutions for countering it, such as selecting peak hours and locations with relatively more population density. Instead of individually conducting the observations, we went in pairs of two to alert one another if any issues were spotted. These measures helped us conduct the observations as planned.

Implications for practice and future research

In this study, we identified gaps and policy recommendations to improve COTPA and NTCP implementation. For example, the fines for violations have not been updated since the Act was enacted in 2003 and are too low in the current context to act as a deterrent. With access to the internet and over-the-top (OTT) platforms, exposure to smoking and tobacco use is on the rise as these are currently not regulated under COTPA. Advertising of tobacco products via social media channels and other platforms on the internet is another unregulated space. Within the NTCP structure, legal personnel at the district or state level to support enforcement activities and counter any litigations or legal challenges is lacking and could support enforcement activities. Periodic national- and state-level multidepartment review meetings to monitor COTPA implementation will help prioritize implementation at the local levels and increase programme activities. Adapting the fines (or the policy instrument), updating the monitoring system to improve transparency, and tackling emerging context-specific issues such as hookah usage, OTT advertising, availability and usage of electronic cigarettes (despite a ban) are some areas that need to be focused on in the future. States would be able to identify various relatable contextual factors, strategies and mechanisms shared in RPT II while attempting to improve implementation efforts in their jurisdiction.

It is important that states improve the existing monitoring systems to document implementation milestones and progress that could serve as data points for future research. The RPT I and RPT II provide health policy and systems researchers a starting point for theorizing tobacco control policy implementation and testing and further refining it in different settings. Longer-term localized research to better capture the political economy and generate contextual data to feed into the policy process is required. Future research could build on our RPT to develop a middle-range theory related to policy implementation in LMIC settings. In the Indian context, developing and validating tools and indicators for measuring implementation outcomes is an area that needs to be focused on. Researchers could examine and contrast the implementation of different policies with similar intended public behaviour change, such as policies on physical inactivity, tobacco control and road safety, and identify how the mechanisms generate the intended outcomes in each of these cases.

Conclusion

Our study highlights the importance of understanding the role of contextual factors in the policy implementation process in a diverse setting of Indian states. Also, the insights from all the states emphasize the limitations of legal provisions and administrative protocols, such as COTPA, in addressing such complex issues effectively. A comprehensive approach that focuses on social reform, community mobilization and social mobilization is necessitated to support and strengthen implementation. Focusing on increasing priority for tobacco control, strengthening review systems and fostering collective action across key actors is essential for the successful implementation of the tobacco control programme in India.

Supplementary data

Supplementary data is available at HEAPOL Journal online.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

This work was supported by the DBT/Wellcome Trust India Alliance Early Career Fellowship (IA/CPHE/17/1/503338) awarded to P.B.H., also supporting V.D.’s time, and a DBT/Wellcome Trust India Alliance Intermediate Career Fellowship (IA/CPHI/22/1/506537) awarded to P.B.H. U.B. was supported for his time through the DBT/Wellcome Trust India Alliance Intermediate and senior fellowships awarded to him (IA/CPHI/17/1/503346 and IA/CPHS/22/1/506533). N.S.P. was supported for his time through the DBT/Wellcome Trust India Alliance intermediate fellowship awarded to him (IA/ CPHI/16/1/502648) and a centre grant (IA/CRC/20/1/600007).

Author contributions

P.B.H., U.B. and G.E.N. conceptualized the study and are overall guarantors of the study. P.B.H. and V.D. collected, extracted and analysed the data, N.S.P. and U.B. steered the data analysis workshops, S.vB. guided data analysis by pointing to relevant literature, G.R.B. and O.C.P.vS. provided inputs at every stage of the study and helped with data interpretation. P.B.H. prepared primary drafts and all authors critically reviewed and refined the paper over several iterations and approved the final version of the manuscript. V.D. helped with designing and editing the figures.

Reflexivity statement

This paper was conceptualized within the doctoral track of P.B.H. with inputs from supervisors in India and in The Netherlands. The authors include three women and five men spread across two high-income countries and one middle-income country. The team consisted of one early career researcher, three mid-career researchers and four senior researchers. Five of the eight authors reside and work in India and steered the design, conduct and writing of the study with reflexive discussions and inputs from their high-income country colleagues. Authors have diverse expertise including realist methods, public administration and governance, non-communicable diseases and public policies, and working with vulnerable populations and in LMICs

Ethical approval

The Institutional Ethics Committee, at the Institute of Public Health (Bengaluru), approved the study on 28 April 2020 (ref: IEC-FR/01/2020)

Conflict of interest

None declared.