Response

We read with interest the results of the Suicide Prevention Implementation and Research Initiative (SPIRIT) trial that tested a multifaceted intervention for suicide prevention in the state of Gujurat, India, ^{Reference Vijayakumar, Pathare, Jain, Pandit, Lohumi and Kalha1} and congratulate the authors on completing this technically difficult study. However, the SPIRIT study findings do not support the abstract’s conclusion that the intervention reduced suicide deaths. No reduction in suicide deaths was observed in the intervention arm; the estimated lower risk reflects a non-significant relative difference driven by an increase in deaths in the control arm.

We are particularly concerned about the paper’s implications for global pesticide suicide prevention. The SPIRIT study evaluated three integrated interventions: a school-based programme, community pesticide storage and training of community health workers. While the first and third addressed suicide prevention broadly, the second focused on pesticide poisoning. Given the study design, the specific impact of pesticide storage on pesticide suicides cannot be isolated. ^{Reference Sethi and Eddleston2} Moreover, the study did not specifically assess pesticide-related suicides or attempts. Therefore, the findings should be interpreted with caution and not taken as evidence that ‘community [pesticide] storage offers a feasible short-term strategy in India’, as the authors argue.

We also have concerns about the quality of the surveillance data. The surveillance system set up by the study identified just 1.6 times as many non-fatal acts of self-harm as fatal acts (307 v. 192; fatal cases 192/499, 38.5%; ratio 1:1.6). This raises major questions about quality because each fatal act is usually matched by around 20 times as many non-fatal acts. ³ In our own study of self-harm in a similar rural region of Sri Lanka, fatal acts (n = 149) made up only 6.1% of all acts of self-harm identified through our surveillance system (n = 2437; ratio 1:15.4). ^{Reference Pearson, Metcalfe, Jayamanne, Gunnell, Weerasinghe and Pieris4} The authors do not report whether the extra suicides in the control arm were due to pesticides or other means.

It is unlikely that the marked increase in suicide attempts in the intervention arm can be related to an ‘improved surveillance [system] capturing data for the first time and greater help-seeking behaviours’, as stated by the authors, when it should not have changed between the pre- and post-intervention periods and was picking up so few cases. If there were ‘improvements’ in the surveillance system during the study duration, it is important to know whether they were of equal magnitude and timing across intervention and control areas. If not, this would cause bias.

Importantly, the actual community uptake and sustained use of the central pesticide storage facility remain uncertain. The paper noted: ‘Of the lockers provided, 88.83% were used at least once and 82.53% were used frequently (more than twice in 12 months) to store pesticides’. We observed similarly high initial uptake in our study of household-level storage boxes in Sri Lanka. ^{Reference Pearson, Metcalfe, Jayamanne, Gunnell, Weerasinghe and Pieris4} However, regular use declined to 55% after 36 months, highlighting concerns about long-term sustainability. Small-scale farmers typically use pesticides regularly (on a weekly or monthly basis) during cultivation seasons, taking them out of storage multiple times. Therefore, using a central storage facility only a few times per year suggests limited and irregular adoption. It is also unclear whether such infrequent use meaningfully reduced household access to pesticides at times of impulsive self-harm. No attempt was made to calculate how much time pesticides spent in the central locker versus in the household.

This lack of detailed data on usage of the central lockers, a lack of data on which episodes of self-harm were due to pesticide self-poisoning (in particular the increased number of suicides in the control arm), and the central lockers being just one part of a multifaceted intervention mean it is not possible to draw conclusions about the effectiveness of central pesticide lockers.

The authors’ feasibility study in four villages did not, as claimed, ‘[show] that storage facilities for pesticides reduced suicide rates’. Instead, it was small and likely showed a regression to the mean in suicide rates in just one intervention village that was atypically high at baseline. ^{Reference Vijayakumar, Jeyaseelan, Kumar, Mohanraj, Devika and Manikandan5} This paper also highlights another issue with central storage lockers – their low uptake (just 34.4%) because they are located in the village centre and away from the fields. ^{Reference Sethi and Eddleston2} It is inconvenient to walk into town to get stored pesticides when the fields are in the other direction, making central lockers a very active form of suicide prevention.

The authors finally argue that bans of acute toxic highly hazardous pesticides – the World Health Organization-recommended and most effective public health intervention in low- and middle-income country suicide prevention ^{Reference Lee, Chisholm, Eddleston, Gunnell, Fleischmann and Konradsen6,7} – are not (currently) practicable in India, despite major impact elsewhere. ^{Reference Rubbo, Tu, Barrass, Chang, Konradsen and Gunnell8} The Government of India has shown that this is not true. In 2018, it banned methyl parathion, phosphamidon and phorate; in 2023, it banned monocrotophos – pesticides that have killed hundreds of thousands of Indian citizens. These decisive actions demonstrate that large-scale bans are both feasible and politically achievable in India. Moreover, the rapid public health impact of such bans is well documented: restricting access to highly hazardous pesticides has led to substantial reductions in suicide rates in India and other countries where pesticide suicide is common. ^{Reference Rubbo, Tu, Barrass, Chang, Konradsen and Gunnell8}

Pesticide bans save lives. Promoting alternative interventions that have not yet demonstrated effectiveness risks diverting attention and resources from strategies with proven life-saving impact. ^{Reference Knipe and Eddleston9} In this context, the SPIRIT study’s conclusions do not advance the evidence base for suicide prevention and may inadvertently undermine ongoing efforts to eliminate highly hazardous pesticides – the single most effective global means of preventing suicide.