Getting shops to voluntarily stop selling cheap, strong beers and ciders: a time-series analysis evaluating impacts on alcohol availability and purchasing

Getting shops to voluntarily stop selling cheap, strong beers and ciders: a time-series analysis... Abstract Background ‘Reducing the Strength’ (RtS) is a public health initiative encouraging retailers to voluntarily stop selling cheap, strong beers/ciders (≥6.5% alcohol by volume). This study evaluates the impact of RtS initiatives on alcohol availability and purchasing in three English counties with a combined population of 3.62 million people. Methods We used a multiple baseline time-series design to examine retail data over 29 months from a supermarket chain that experienced a two-wave, area-based role out of RtS: initially 54 stores (W1), then another 77 stores (W2). We measured impacts on units of alcohol sold (primary outcome: beers/ciders; secondary outcome: all alcoholic products), economic impacts on alcohol sales and substitution effects. Results We observed a non-significant W1 increase (+3.7%, 95% CI: −11.2, 21.0) and W2 decrease (−6.8%, 95% CI: −20.5, 9.4) in the primary outcome. We observed a significant W2 decrease in units sold across all alcohol products (−10.5%, 95% CI: −19.2, −0.9). The direction of effect between waves was inconsistent for all outcomes, including alcohol sales, with no evidence of substitution effects. Conclusions In the UK, voluntary RtS initiatives appear to have little or no impact on reducing alcohol availability and purchase from the broader population of supermarket customers. alcohol policy, alcohol purchasing, alcohol retail sales data, reducing the strength, time–series analysis Introduction Modifying the availability of commercial products (e.g. alcohol, food) is a widely advocated public health strategy.1,2 The World Health Organization has proposed a number of interventions and policies to reduce availability including interventions reducing the alcoholic strength of products.3 Research from North America, Australia and Europe has examined ways in which modifying local food availability impacts on health outcomes,4–7 but there are relatively fewer evaluations of local alcohol availability interventions.1,6,8–14 Alcohol is a causal factor in more than 200 disease and injury conditions accounting for 5.9% of deaths worldwide.2 Social costs attributable to alcohol, including crime and disorder, representing 1.3–3.3% of gross domestic product globally.2 Interventions modifying alcohol availability have been seen to reduce both alcohol consumption and alcohol-related harm.2,15–19 In many countries, including the UK, attempts to modify availability through national government regulation, such as minimum unit pricing, have been met with political and legal barriers. Regulating the sale and consumption of alcohol products often takes place at sub-national levels.6,8,20 Local government initiatives to reduce alcohol availability have been implemented, involving both statutory and voluntary approaches, the latter often targeting specific population groups.15,21–24 Evaluative research of natural policy experiments is important because innovative practices can diffuse to new settings, including across national boundaries, sometimes before they have been robustly evaluated.25,26 Reducing the strength of alcoholic products or modifying high-strength product availability have been proposed as ‘best practices’ to regulate physical availability.3,27 This, however, stems from an interpretation of availability theory rather than a synthesis of empirical evidence assessing impacts of reducing availability of high-strength beers and ciders (so-called ‘superstrength’ products) and the evidence base around this is under-developed. Superstrength products and their marketing have been said to encourage alcohol misuse and harmful behaviours among vulnerable populations.28 In the UK, the term ‘Reducing the Strength’ (RtS) is now widely used to refer to area-based public health initiatives that involve removing low price, superstrength alcoholic products from sale in stores through voluntary agreements with local retailers and off-licences. RtS has been originally designed to tackle problems associated with alcohol social harms, often focused on street drinking.22 Suffolk was the first UK area to adopt the initiative in 2012 as part of a multi-intervention approach to tackling street drinking. Since then at least 30 schemes have been implemented in the UK.29 The approach varies, but most RtS initiatives tend to target alcohol products above 6.5% alcohol by volume (ABV), although some have focused on a slightly lower ABV or lower cost products.22 In this RtS, the products targeted were lower cost products above 7.5% ABV. Superstrength products vary by price, brand and strength. The least expensive products (e.g. ‘white ciders’) are amongst the lowest cost per unit alcohol products in UK stores, purchased for as little as 11.1 pence per unit.30,31 UK local and regional governments have complained to the alcohol industry that specific superstrength products sold in 500 ml cans encourage rapid consumption of high quantities of alcohol causing population harms; although this is disputed by the industry.32 It has been argued that targeted interventions, such as RtS, offer local and regional government authorities a potential means of tackling publicly visible social and health problems associated with alcohol consumption.21,22,29 Retailers and the alcohol industry have raised concerns about RtS that have included questioning its evidence base, legal status (in terms of competition law) and its potential financial impact.22,33–35 On the other hand, some retailers arguably demonstrate a degree of support for RtS by voluntarily participating in initiatives, although their reasons for doing so may vary. For example, some retailers saw street drinking as a problem in their area and hoped that participation would reduce anti-social behaviour within their own shops while others saw this as an opportunity to co-operate with the licensing authorities.35 An intervention that is designed to deter anti-social customers could potentially improve shops’ image with the wider customer base in addition to licensing authorities and other stakeholders.22,33,36,37 From a public health perspective, it remains unclear to what extent local-level voluntary interventions, such as RtS, can play an effective role in reducing alcohol consumption at the population level.12 Retail sales data routinely collected by shops provides one means of measuring the impact of alcohol interventions. Such data can provide an objective and accurate estimate of alcohol purchase and proxy consumption, particularly in the case of larger supermarket and shop chains that have invested heavily in data collection.38 However, shop-level data are hard to obtain due to commercial sensitivity.39 There are few published evaluations of alcohol interventions in the UK using retail data to assess changes in physical and economic availability of specific alcohol products for health improvement.18,40 The RtS studied here was originally launched as a joint initiative between Suffolk Police, Ipswich Borough Council, Suffolk County Council and the National Health Service (Suffolk) in September 2012.41 Following interviews with local practitioners and policymakers who designed and implemented the RtS in Suffolk, we hypothesized several possible mechanisms for RtS impacts on alcohol availability and sales. These include a potential ‘nudge’ effect where the impact of reducing physical availability of alcohol products by removing superstrength products helped discourage and denormalise the practice of purchasing cheap products. The RtS was also theorized as an economic availability intervention: customers with finite resources wishing to purchase low cost per unit superstrength products may, on finding those products removed, substitute for products with lower alcohol content or for different alcohol products.29,35 This study aims to evaluate the impact of the introduction of a RtS initiative on alcohol availability in the form of overall availability of alcohol units and purchasing in one national retail chain across three English counties using time-series analyses of retail sales data. Methods Setting and intervention A major supermarket chain (East of England Co-operative Society, known commonly as ‘Co-op’) voluntarily joined RtS in Suffolk and consequently ensured that its stores in that county cleared their stock of all their low-priced brands of high-strength beers/lagers and ciders in the month leading up to September 2012. These consisted of four superstrength products (7.5–9.0% ABV) but did not include any more expensive ‘craft’ or ‘premium’ high-strength products as the implementers did not associate such products with street drinking (Table 1). The same chain required stores in Essex and Norfolk to begin a similar process of withdrawing those products from sale by September 2013. Every shop from the chain participated in the intervention although a minority of stores, 6% from wave 1 and 36% from wave 2, took longer than 1 month to stop selling superstrength products (Appendix S1). Table 1 List of beer and cider products over 6.5% ABV sold during the ‘Reducing the Strength’ initiative EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 aRange of values during the period of study. bSuperstrength products (over 7.5% ABV) removed as part of the Reducing the Strength initiative. cHigh strength premium products (over 7.5% ABV) not removed as part of the Reducing the Strength initiative. All other products are high-strength premium products (over 6.5% but below 7.5% ABV) still available during the study period. EAN, European Article Number (also called International Article Number); ABV, Alcohol by volume (ABV) (%). Recommended weekly limit of 14 units of alcohol for men and women55. View Large Table 1 List of beer and cider products over 6.5% ABV sold during the ‘Reducing the Strength’ initiative EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 aRange of values during the period of study. bSuperstrength products (over 7.5% ABV) removed as part of the Reducing the Strength initiative. cHigh strength premium products (over 7.5% ABV) not removed as part of the Reducing the Strength initiative. All other products are high-strength premium products (over 6.5% but below 7.5% ABV) still available during the study period. EAN, European Article Number (also called International Article Number); ABV, Alcohol by volume (ABV) (%). Recommended weekly limit of 14 units of alcohol for men and women55. View Large Data Monthly retail sales data were provided for the period January 2012 to May 2014 obtained for 131 stores in one retail chain in the three English counties. We used the full range of data that East of England Co-operative Society provided us with for this analysis: the researchers did not have direct access to the company’s internal data systems but rather were sent data pertaining only to the intervention period and localities so that the researchers could analyse them independently. Shop-level characteristics and sales data were available including prices, quantities, product brands, alcohol content, and sales for the following drink categories: beer/lager and cider, wines, affordable sparkling and low alcohol wines, and spirits. Our primary outcome was units of alcohol sold for beer/lager and cider. Secondary outcomes included units of alcohol sold for two high-strength premium products (ABV over 7.5%) not removed as part of the RtS (Table 1), the remaining drink categories and for all products in order to examine substitution effects and in line with qualitative findings on drinkers’ responses to RtS. We looked at sales value to assess the potential economic impact of RtS on stores. Stores in Suffolk (n = 54) were regarded as stores participating in wave 1 (W1) of the intervention and stores in Norfolk and Essex (n = 77) as stores participating in wave 2 (W2) a year later. Statistical analysis We used a quasi-experimental multiple baseline time-series design42 to study changes in units of alcohol sold and sales value for beer/lager and cider, wines, sparkling and low alcohol wines, spirits and for total alcohol products after the introduction of the RtS initiative. The RtS was introduced in a staggered approach, implemented at two different time points (W1 and W2) across three different geographical areas with a combined population of 3.62 million people.43 We examined the impact of implementing RtS separately for the two waves in order to identify whether the intervention produced similar effects in the entire population of interest (i.e. whether the impact of the intervention was consistent in the two waves).42,44 The repeated pattern of a reduction in the measured outcome following the implementation of the intervention in each geographical area (i.e. wave) would suggest that the intervention is having an effect.42 An appropriate statistical approach to evaluate such impacts is the use of segmented linear regression, which divides a time series into pre- and post-intervention segments,44 with panel-corrected standard errors.45,46 We took autocorrelation into account by means of a common autoregressive first order (AR(1)) model and we included the calendar month as a term to adjust for seasonality.44,47 Details of the assumptions and model specification are available in Appendix S2. The intervention effect was assumed to occur immediately after implementation, so no transition period was taken into account in the analysis. We log-transformed our dependent variables as these were highly skewed. For ease of interpretation, regression coefficients (β) were converted into per cent change in sales and units of alcohol sold using the formula [exp(β) − 1] × 100. This approach was used to ensure data confidentiality when using commercially sensitive information, such as sales of specific alcohol products and brands. We therefore examined substitution effects at a product category level and for high-strength premium products that were not removed rather than at the level of specific products or brands. Analysis was carried out in Stata 14.1. Results Stores in W1 and W2 were similar in terms of size, area-level deprivation score and urban versus semi-urban location. Stores in W1 were open on average for fewer hours compared to those in W2 (Appendix S3). Mean units of alcohol sold per store per month were lower in W1 compared to W2 stores in all products. Overall, beer/lager and cider accounted for 32.4% of total units of alcohol sold during the study period. Superstrength products removed had previously accounted for 6.5 and 3.6% of total units sold for beer/lager and cider in W1 and W2 stores, respectively (Table 2). In terms of sales, these four products accounted for 2.1 and 1.3% of total revenue for W1 and W2 stores, respectively, before the intervention (data not shown). Table 2 Summary statistics for units of alcohol sold per store per month Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 aOnly for the period up until September 2012 for wave 1 and September 2013 for wave 2. View Large Table 2 Summary statistics for units of alcohol sold per store per month Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 aOnly for the period up until September 2012 for wave 1 and September 2013 for wave 2. View Large Our analysis indicates that the impact of RtS on units of alcohol sold for beer/lager and cider was not significant in the two waves (Fig. 1 and Appendix S4). More specifically, following RtS implementation, W1 stores experienced a non-significant increase (3.7%, 95% confidence intervals (CI): −11.2 to 21.0, P = 0.647) whereas W2 stores experienced a non-significant decrease (−6.8%, 95% CI: −20.5 to 9.4, P = 0.390) (Fig. 1). In terms of all alcohol products, the introduction of RtS was associated with a non-significant increase in W1 stores (8.0%, 95% CI: −1.3 to 18.3, P = 0.094). In contrast, a significant decrease (−10.5%, 95% CI: −19.2 to −0.9, P = 0.034) was observed in W2 stores (Fig. 2 and Appendix S4). Similar patterns for beer/cider and lager were observed for sales value (Fig. 2). Fig. 1 View largeDownload slide % Change in units of alcohol sold after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 1 View largeDownload slide % Change in units of alcohol sold after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 2 View largeDownload slide % Change in sales value after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 2 View largeDownload slide % Change in sales value after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. In order to examine substitution effects we repeated the analysis for high-strength premium products, spirits, affordable sparkling and low alcohol wines and wines. We found that all product categories experienced similar changes in units of alcohol sold and sales value during this time period in W1 and W2 to those observed for beer/lager and cider. None of them were significant except for units of alcohol sold for wines, which appeared to drive the significant decrease observed in units of alcohol sold for all products. We found no evidence of substitution effects for high-strength premium products (Fig. 1 and Appendix S4). Discussion Main findings of this study We used retail sales data to evaluate the introduction of RtS, a public health initiative targeted at supermarkets and off-licences to remove low cost, superstrength beers and ciders from sale in three English counties. Our results show that this RtS had no significant impact on total units of alcohol sold and sales value for beer/lager and cider. We also found no observable substitution effects of alcohol products attributable to the RtS intervention in the 131 stores. What is already know on the topic Only a small number of previous studies have used retail sales data in quasi-experimental designs to evaluate alcohol interventions. Evaluation of the Scottish Alcohol Act 2010 showed that banning alcohol multi-buy promotions did not reduce alcohol purchasing at the household level,18 and the introduction of the Alcohol Act was not associated with any changes in off-trade beer sales.40 In our study, the majority of results were non-significant. The small significant decrease in units and value of alcohol sales of all products in W2 stores appears to be driven by declining wine (rather than beer/cider) sales.48 Furthermore, the changes observed in the two waves were not consistent and so the overall findings showed no intervention attributable impact.42 An Australian evaluation of local alcohol availability restrictions (cask wines and products over 2.7% ABV) found that some participants travelled further to access non-participating shops.13,14 In our study we theorize that overall alcohol purchases could be influenced by whether or not customers changed where they purchased alcohol (i.e. shops not participating in RtS), or if they substituted products within participating stores.14 Our study focused on one retail chain which maintained compliance with RtS22 and we found no substitution effects between categories of alcohol products within study stores attributable to the intervention. Customers in the study areas had the ability to access other local stores that did not participate in the RtS but we did not detect any sudden or sustained loss of income in participating stores that might be expected if substantial numbers of customers had started shopping elsewhere for alcohol. The availability of alternative stores not participating may vary within and between the three counties studied. Limitations of this study The retail data we had available related to one retail supermarket chain and the data available could not be used to consider overall area effects, shop-level or brand/product-level substitution effects, individual or sub-group level purchasing or consumption.14,18,37 Our results cannot be generalized to RtS initiatives that have removed products with >6.5% or lower ABV. We did not have the data to measure long term impacts on purchasing and consumption, although we theorised that RtS should impact on availability as soon as shops stopped selling superstrength products.13,14 The confidence intervals for our findings were wide and statistical precision might have been improved with inclusion of a greater number of stores, and/or time points.44,46 Stores in W1 and W2 had different rates of compliance, which may compromise internal validity.42 In addition, RtS is only one intervention targeting alcohol consumption and harms, and we are aware that there are a range of local alcohol policies routinely implemented in local government which we were unable to adjust for. Such unmeasured events may introduce confounding and compromise internal validity.49 Finally, segmented regression analysis has its own limitations, allowing only linear trends to be examined but changes may follow non-linear patterns.44 What this study adds Our study makes an important contribution to the evidence base for local voluntary retail alcohol interventions.18,40 The use of retail data is novel for evaluating alcohol initiatives and it has been advocated as an important means to monitor alcohol consumption40,50 despite the limitations.38 In this study, we used a retail sales time series panel data set, that contains far more information than single cross-sectional data allowing for an increased precision in estimation.46 Panel difference-in-differences analysis has been used in a previous study,18 but we opted to use panel-corrected standard errors within a regression framework, because ignoring possible correlation of regression disturbances over time and between panels may lead to overly optimistic standard errors and lead to biased statistical inference.46 The RtS initiative21 was originally developed as part of a strategy that also involved alcohol and drug treatment services and street policing to tackle street drinking and anti-social behaviour due to excess alcohol consumption, and there is some evidence that this targeted, multi-intervention approach led to reductions in police call outs and other indicators of social problems related to street drinking.21,41 This evaluation does not test RtS’s impact on wider aims of tackling alcohol social harms including street drinking. The RtS was not originally expected to have impacts on reducing overall population alcohol consumption. Potential secondary effects of RtS on the broader population of alcohol consumers are of interest to the public health community. Voluntary agreements between governments and the private sector have previously been used to encourage businesses to take actions.36 However, there is little evidence to suggest such approaches are more (cost-) effective, particularly if they are unaccompanied by monitoring, and appropriate incentives and sanctions.36 The alcohol industry and retail sector may be more willing to participate in voluntary initiatives targeting selected population groups (i.e. street drinkers) that have minimal impact on their profits. Our analysis suggests that RtS had no impact on revenues. Addressing alcohol-related harms and drinking behaviours in ‘high-risk’ groups is important but our analysis suggests that RtS may not be effective for addressing alcohol harms across the whole population. The evidence base recommends regulatory or statutory enforcement interventions restricting alcohol availability are more effective than local non-regulatory or voluntary approaches targeting specific groups.12,51–54 Conclusion This evaluation did not specifically test impacts on target groups, such as street drinkers, but examined impacts on all consumers’ alcohol purchasing patterns from one retail supermarket chain. Our findings suggest that voluntary RtS initiatives, have little or no impact on reducing alcohol availability and purchase amongst a broader population of customers. The research literature suggests that more effective regulatory public health interventions will be required to achieve substantial population health benefits in reducing alcohol consumption and alcohol-related harms. Supplementary data Supplementary data are available at the Journal of Public Health online. Acknowledgements We would like to acknowledge the invaluable input of and the helpful comments of Professor Mark Petticrew. We would also like to thank Mark Lewenz of East of England Co-operative Society who provided the retail sales data. Antonio Gasparrini provided expert advice on our analysis. The evaluation of the Reducing the Strength initiative is part of the programme of the School for Public Health Research (http://sphr.lshtm.ac.uk/). This is an independent research unit based at the London School of Hygiene and Tropical Medicine, funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR). Sole responsibility for this research lies with the authors and the views expressed are not necessarily those of the NHS, the NIHR SPHR or the Department of Health. The East of England Co-operative Society supplied the retail sales data but played no role in the funding of the study. The East of England Co-operative Society, the NHS, the NIHR and the Department of Health played no role in the design of the study, the interpretation of the findings, the writing of the article or the decision to submit. The views expressed are those of the author(s) and not necessarily those of the East of England Co-operative Society, NHS, the NIHR or the Department of Health. Funding This study was funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR). Two co-authors (AJ and SA) contributed as part of their normal salaried work for Suffolk County Council. The East of England Co-operative Society supplied the retail sales data but played no role in the funding of the study. The East of England Co-operative Society, the NHS, the NIHR SPHR and the Department of Health played no role in the design of the study, the interpretation of the findings, the writing of the paper or the decision to submit. References 1 World Health Organization . Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013−2020 . Geneva : WHO , 2013 . 2 World Health Organization . Global Status Report on Alcohol and Health . Geneva : WHO , 2014 . 3 World Health Organization . Global Strategy to Reduce the Harmful Use of Alcohol . Geneva : WHO , 2010 . PubMed PubMed 4 Morland KB . An evaluation of a neighborhood-level intervention to a local food environment . Am J Prev Med 2010 ; 39 : e31 – 8 . Google Scholar CrossRef Search ADS PubMed 5 Gase LN , McCarthy WJ , Robles B et al. . Student receptivity to new school meal offerings: assessing fruit and vegetable waste among middle school students in the Los Angeles Unified School District . Prev Med 2014 ; 67 ( Suppl 1 ): S28 – 33 . Google Scholar CrossRef Search ADS PubMed 6 Soler RE , Whitten KL , Ottley PG . Communities putting prevention to work: local evaluation of community-based strategies designed to make healthy living easier . Prev Med 2014 ; 67 ( Suppl 1 ): S1 – 3 . Google Scholar CrossRef Search ADS PubMed 7 Valdivia Espino J , Guerrero N , Rhoads N et al. . Community-based restaurant interventions to promote healthy eating: a systematic review . Prev Chronic Dis 2015 ; 12 : 140455 . Google Scholar CrossRef Search ADS 8 Nelson TF , Xuan Z , Babor TF et al. . Efficacy and the strength of evidence of U.S. alcohol control policies . Am J Prev Med 2013 ; 45 : 19 – 28 . Google Scholar CrossRef Search ADS PubMed 9 Rammohan V , Hahn RA , Elder R et al. . Effects of dram shop liability and enhanced overservice law enforcement initiatives on excessive alcohol consumption and related harms: two community guide systematic reviews . Am J Prev Med 2011 ; 41 : 334 – 43 . Google Scholar CrossRef Search ADS PubMed 10 Campbell CA , Hahn RA , Elder R et al. . The effectiveness of limiting alcohol outlet density as a means of reducing excessive alcohol consumption and alcohol-related harms . Am J Prev Med 2009 ; 37 : 556 – 69 . Google Scholar CrossRef Search ADS PubMed 11 Stockwell T , Zhao J , Giesbrecht N et al. . The raising of minimum alcohol prices in Saskatchewan, Canada: impacts on consumption and implications for public health . Am J Public Health 2012 ; 102 : e103 – e10 . Google Scholar CrossRef Search ADS PubMed 12 Martineau F , Tyner E , Lorenc T et al. . Population-level interventions to reduce alcohol-related harm: an overview of systematic reviews . Prev Med 2013 ; 57 : 278 – 96 . Google Scholar CrossRef Search ADS PubMed 13 Kinnane S , Farringdon F , Henderson-Yates L et al. . Fitzroy Valley Alcohol Restriction Report: An Evaluation of the Effects of a Restriction on Take-away Alcohol Relating to Measurable Health and Social Outcomes, Community Perceptions and Behaviours After a Two Year Period, 2010 . http://www.dao.health.wa.gov.au/DesktopModules/Bring2mind/DMX/Download.aspx?EntryId=680&Command=Core_Download&PortalId=0&TabId=211 (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dhBDi59v on 11 December 2015. 14 National Drug Research Institute . Restrictions on the Sale and Supply of Alcohol: Evidence and Outcomes. Perth: National Drug Research Institute, Curtin University of Technology, 2007 . http://ndri.curtin.edu.au/local/docs/pdf/publications/R207.pdf (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dhAzukmG on 11 December 2015. 15 Maclennan B , Kypri K , Room R et al. . Local government alcohol policy development: case studies in three New Zealand communities . Addiction 2013 ; 108 : 885 – 95 . Google Scholar CrossRef Search ADS PubMed 16 Ahern J , Margerison-Zilko C , Hubbard A et al. . Alcohol outlets and binge drinking in urban neighborhoods: the implications of nonlinearity for intervention and policy . Am J Public Health 2013 ; 103 : e81 – 7 . Google Scholar CrossRef Search ADS PubMed 17 Fone D , Dunstan F , White J et al. . Change in alcohol outlet density and alcohol-related harm to population health (CHALICE) . BMC Public Health 2012 ; 12 : 428 . Google Scholar CrossRef Search ADS PubMed 18 Nakamura R , Suhrcke M , Pechey R et al. . Impact on alcohol purchasing of a ban on multi-buy promotions: a quasi-experimental evaluation comparing Scotland with England and Wales . Addiction 2014 ; 109 : 558 – 67 . Google Scholar CrossRef Search ADS PubMed 19 Purshouse RC , Meier PS , Brennan A et al. . Estimated effect of alcohol pricing policies on health and health economic outcomes in England: an epidemiological model . Lancet 2010 ; 375 : 1355 – 64 . Google Scholar CrossRef Search ADS PubMed 20 Han D , Gorman DM . Evaluating the effects of the introduction of off-sale alcohol outlets on violent crime . Alcohol Alcohol 2013 ; 48 : 370 – 4 . Google Scholar CrossRef Search ADS PubMed 21 Alcohol Concern . Guide to Alcohol for Councillors . London : Alcohol Concern , 2014 . http://www.alcoholconcern.org.uk/wp-content/uploads/woocommerce_uploads/2014/12/Guide-to-Alcohol-160813-APPROVED.pdf (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh8ZwRom. on 11 December 2015. PubMed PubMed 22 Local Government Association . Reducing the Strength Schemes: Guidance for Councils Considering Setting up a Scheme, 2014 . Available at: http://www.local.gov.uk/regulatory-services-and-licensing/-/journal_content/56/10180/6793034/ARTICLE (accessed 10 December 2015). Archived at: http://www.webcitation.org/6dh8jYS00 on 11 December 2015. 23 Nicholls J . The Government Alcohol Strategy 2012: alcohol policy at a turning point? Drugs Educ Prev Policy 2012 ; 19 : 355 – 9 . Google Scholar CrossRef Search ADS 24 Hughes K , Bellis MA , Leckenby N et al. . Does legislation to prevent alcohol sales to drunk individuals work? Measuring the propensity for night-time sales to drunks in a UK city . J Epidemiol Community Health 2014 ; 68 : 453 – 6 . Google Scholar CrossRef Search ADS PubMed 25 McGill E , Egan M , Petticrew M et al. . Trading quality for relevance: non-health decision-makers’ use of evidence on the social determinants of health . BMJ Open 2015 ; 5 : e007053 . Google Scholar CrossRef Search ADS PubMed 26 Dolowitz D , Marsh D . Learning from abroad: the role of policy transfer in contemporary policy‐making . Governance 2000 ; 13 : 5 – 23 . Google Scholar CrossRef Search ADS 27 Alcohol, Public Policy G . Alcohol: no ordinary commodity—a summary of the second edition . Addiction 2010 ; 105 : 769 – 79 . CrossRef Search ADS PubMed 28 Goodall T . White Cider and Street Drinkers: Recommendations to Reduce Harm . London : Alcohol Concern UK , 2011 . http://www.thamesreach.org.uk/publications/research-reports/white-cider-and-street-drinkers/ (25 June 2015, date last accessed). Archived at: http://www.webcitation.org/6ZYIecFEw on 25 June 2015. 29 McGill E , Marks D , Sumpter C et al. . Consequences of removing cheap, super-strength beer and cider: a qualitative study of a UK local alcohol availability intervention . BMJ Open 2016 ; 6 : e010759 . Google Scholar CrossRef Search ADS PubMed 30 Sweeting H . Alcohol units at pocket money prices [e-letter] . Br Med J 2006 ; 333 : 825 . -b. 31 Gornall J . Doctors and the alcohol industry: an unhealthy mix? Br Med J 2013 ; 346 : f1889 . Google Scholar CrossRef Search ADS 32 Portman Group . K Cider, Complaint Decision Details, 2017 . http://www.portmangroup.org.uk/complaints/complaint-decisions/complaint-decision-details/2017/01/12/k-cider (24 January 2017, date last accessed). Archived at: http://www.webcitation.org/6nkoGttF8 on 24 January 2017. 33 Association of Convenience Stores . Reducing the Strength: Retailer Guidance, 2015 . http://www.acs.org.uk/advice/reducing-the-strength/ (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh7Gcc5S on 11 December 2015. 34 Competition and Markets Authority . High-Strength Alcohol Schemes: Competition Law Advice for Retailers, 2015 . https://www.gov.uk/government/publications/high-strength-alcohol-schemes-competition-law-advice-for-retailers (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh7NK9Rb on 11 December 2015. 35 Sumpter C , McGill E , Dickie E et al. . Reducing the Strength: a mixed methods evaluation of alcohol retailers’ willingness to voluntarily reduce the availability of low cost, high strength beers and ciders in two UK local authorities . BMC Public Health 2016 ; 16 : 448 . Google Scholar CrossRef Search ADS PubMed 36 Bryden A , Petticrew M , Mays N et al. . Voluntary agreements between government and business—a scoping review of the literature with specific reference to the Public Health Responsibility Deal . Health Policy 2013 ; 110 : 186 – 97 . Google Scholar CrossRef Search ADS PubMed 37 Community Alcohol Partnerships . About Community Alcohol Partnerships, 2015 . http://www.communityalcoholpartnerships.co.uk/ (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh622qSP on 11 December 2015. 38 Robinson M , Thorpe R , Beeston C et al. . A review of the validity and reliability of alcohol retail sales data for monitoring population levels of alcohol consumption: a Scottish perspective . Alcohol Alcohol 2013 ; 48 : 231 – 40 . Google Scholar CrossRef Search ADS PubMed 39 Egan M , Pliakas T , Grace D et al. . Use of retail data in the assessment of natural experiments: the case of Reducing the Strength, an intervention to reduce alcohol availability . Lancet 2014 ; 384 : S30 . Google Scholar CrossRef Search ADS 40 Robinson M , Geue C , Lewsey J et al. . Evaluating the impact of the alcohol act on off-trade alcohol sales: a natural experiment in Scotland . Addiction 2014 ; 109 : 2035 – 43 . Google Scholar CrossRef Search ADS PubMed 41 Suffolk Constabulary . Reducing the Strength Ipswich—One Year on, 2012 . http://www.suffolk.police.uk/safetyadvice/personalsafety/alcoholawareness/reducingthestrength/oneyearon.aspx (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh6n05NZ on 11 December 2015. 42 Hawkins NG , Sanson-Fisher RW , Shakeshaft A et al. . The multiple baseline design for evaluating population-based research . Am J Prev Med 2007 ; 33 : 162 – 8 . Google Scholar CrossRef Search ADS PubMed 43 Office for National Statistics . Population Estimates for UK, England and Wales, Scotland and Northern Ireland, Mid-2014, 2015. Available at: http://www.ons.gov.uk/ons/publications/re-reference-tables.html?edition=tcm%3A77-368259 (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh6SoR4s on 11 December 2015. 44 Wagner AK , Soumerai SB , Zhang F et al. . Segmented regression analysis of interrupted time series studies in medication use research . J Clin Pharm Ther 2002 ; 27 : 299 – 309 . Google Scholar CrossRef Search ADS PubMed 45 Beck N , Katz JN . What to do (and not to do) with time-series cross-section data . Am Polit Sci Rev 1995 ; 89 : 634 – 47 . Google Scholar CrossRef Search ADS 46 Hoechle D . Robust standard errors for panel regressions with cross-sectional dependence . Stata J 2007 ; 7 : 281 – 312 . 47 Linden A . Conducting interrupted time-series analysis for single- and multiple-group comparisons . Stata J 2015 ; 15 : 480 – 500 . 48 The Drinks Report . UK Alcohol Market Returns to Growth, 2015 . http://www.thedrinksreport.com/news/2015/16101-uk-alcohol-market-returns-to-growth.html (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh702Jz8 on 11 December 2015. 49 Craig P , Cooper C , Gunnell D et al. . Using natural experiments to evaluate population health interventions: new Medical Research Council guidance . J Epidemiol Community Health 2012 ; 66 : 1182 – 6 . Google Scholar CrossRef Search ADS PubMed 50 Robinson M , Shipton D , Walsh D et al. . Regional alcohol consumption and alcohol-related mortality in Great Britain: novel insights using retail sales data . BMC Public Health 2015 ; 15 : 1 . Google Scholar CrossRef Search ADS PubMed 51 de Vocht F , Heron J , Angus C et al. . Measurable effects of local alcohol licensing policies on population health in England . J Epidemiol Community Health 2016 ; 70 : 231 – 7 . Google Scholar CrossRef Search ADS PubMed 52 de Vocht F , Heron J , Campbell R et al. . Testing the impact of local alcohol licencing policies on reported crime rates in England . J Epidemiol Community Health 2017 ; 71 : 137 – 45 . Google Scholar CrossRef Search ADS PubMed 53 Knai C , Petticrew M , Durand MA et al. . The Public Health Responsibility deal: has a public-private partnership brought about action on alcohol reduction? Addiction 2015 ; 110 : 1217 – 25 . Google Scholar CrossRef Search ADS PubMed 54 de Vocht F , Tilling K , Pliakas T et al. . The intervention effect of local alcohol licensing policies on hospital admission and crime: a natural experiment using a novel Bayesian synthetic time-series method . J Epidemiol Community Health 2017 ; 71 : 912 – 8 . Google Scholar CrossRef Search ADS PubMed 55 Department of Health . Health Risks From Alcohol: New Guidelines, 2016 . https://www.gov.uk/government/consultations/health-risks-from-alcohol-new-guidelines (2 August 2017, date last accessed). Archived at: http://www.webcitation.org/6sQ3CIIWX on 2 August 2017. © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Public Health Oxford University Press

Getting shops to voluntarily stop selling cheap, strong beers and ciders: a time-series analysis evaluating impacts on alcohol availability and purchasing

Loading next page...
 
/lp/ou_press/getting-shops-to-voluntarily-stop-selling-cheap-strong-beers-and-WJLjr9WAwI
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1741-3842
eISSN
1741-3850
D.O.I.
10.1093/pubmed/fdy003
Publisher site
See Article on Publisher Site

Abstract

Abstract Background ‘Reducing the Strength’ (RtS) is a public health initiative encouraging retailers to voluntarily stop selling cheap, strong beers/ciders (≥6.5% alcohol by volume). This study evaluates the impact of RtS initiatives on alcohol availability and purchasing in three English counties with a combined population of 3.62 million people. Methods We used a multiple baseline time-series design to examine retail data over 29 months from a supermarket chain that experienced a two-wave, area-based role out of RtS: initially 54 stores (W1), then another 77 stores (W2). We measured impacts on units of alcohol sold (primary outcome: beers/ciders; secondary outcome: all alcoholic products), economic impacts on alcohol sales and substitution effects. Results We observed a non-significant W1 increase (+3.7%, 95% CI: −11.2, 21.0) and W2 decrease (−6.8%, 95% CI: −20.5, 9.4) in the primary outcome. We observed a significant W2 decrease in units sold across all alcohol products (−10.5%, 95% CI: −19.2, −0.9). The direction of effect between waves was inconsistent for all outcomes, including alcohol sales, with no evidence of substitution effects. Conclusions In the UK, voluntary RtS initiatives appear to have little or no impact on reducing alcohol availability and purchase from the broader population of supermarket customers. alcohol policy, alcohol purchasing, alcohol retail sales data, reducing the strength, time–series analysis Introduction Modifying the availability of commercial products (e.g. alcohol, food) is a widely advocated public health strategy.1,2 The World Health Organization has proposed a number of interventions and policies to reduce availability including interventions reducing the alcoholic strength of products.3 Research from North America, Australia and Europe has examined ways in which modifying local food availability impacts on health outcomes,4–7 but there are relatively fewer evaluations of local alcohol availability interventions.1,6,8–14 Alcohol is a causal factor in more than 200 disease and injury conditions accounting for 5.9% of deaths worldwide.2 Social costs attributable to alcohol, including crime and disorder, representing 1.3–3.3% of gross domestic product globally.2 Interventions modifying alcohol availability have been seen to reduce both alcohol consumption and alcohol-related harm.2,15–19 In many countries, including the UK, attempts to modify availability through national government regulation, such as minimum unit pricing, have been met with political and legal barriers. Regulating the sale and consumption of alcohol products often takes place at sub-national levels.6,8,20 Local government initiatives to reduce alcohol availability have been implemented, involving both statutory and voluntary approaches, the latter often targeting specific population groups.15,21–24 Evaluative research of natural policy experiments is important because innovative practices can diffuse to new settings, including across national boundaries, sometimes before they have been robustly evaluated.25,26 Reducing the strength of alcoholic products or modifying high-strength product availability have been proposed as ‘best practices’ to regulate physical availability.3,27 This, however, stems from an interpretation of availability theory rather than a synthesis of empirical evidence assessing impacts of reducing availability of high-strength beers and ciders (so-called ‘superstrength’ products) and the evidence base around this is under-developed. Superstrength products and their marketing have been said to encourage alcohol misuse and harmful behaviours among vulnerable populations.28 In the UK, the term ‘Reducing the Strength’ (RtS) is now widely used to refer to area-based public health initiatives that involve removing low price, superstrength alcoholic products from sale in stores through voluntary agreements with local retailers and off-licences. RtS has been originally designed to tackle problems associated with alcohol social harms, often focused on street drinking.22 Suffolk was the first UK area to adopt the initiative in 2012 as part of a multi-intervention approach to tackling street drinking. Since then at least 30 schemes have been implemented in the UK.29 The approach varies, but most RtS initiatives tend to target alcohol products above 6.5% alcohol by volume (ABV), although some have focused on a slightly lower ABV or lower cost products.22 In this RtS, the products targeted were lower cost products above 7.5% ABV. Superstrength products vary by price, brand and strength. The least expensive products (e.g. ‘white ciders’) are amongst the lowest cost per unit alcohol products in UK stores, purchased for as little as 11.1 pence per unit.30,31 UK local and regional governments have complained to the alcohol industry that specific superstrength products sold in 500 ml cans encourage rapid consumption of high quantities of alcohol causing population harms; although this is disputed by the industry.32 It has been argued that targeted interventions, such as RtS, offer local and regional government authorities a potential means of tackling publicly visible social and health problems associated with alcohol consumption.21,22,29 Retailers and the alcohol industry have raised concerns about RtS that have included questioning its evidence base, legal status (in terms of competition law) and its potential financial impact.22,33–35 On the other hand, some retailers arguably demonstrate a degree of support for RtS by voluntarily participating in initiatives, although their reasons for doing so may vary. For example, some retailers saw street drinking as a problem in their area and hoped that participation would reduce anti-social behaviour within their own shops while others saw this as an opportunity to co-operate with the licensing authorities.35 An intervention that is designed to deter anti-social customers could potentially improve shops’ image with the wider customer base in addition to licensing authorities and other stakeholders.22,33,36,37 From a public health perspective, it remains unclear to what extent local-level voluntary interventions, such as RtS, can play an effective role in reducing alcohol consumption at the population level.12 Retail sales data routinely collected by shops provides one means of measuring the impact of alcohol interventions. Such data can provide an objective and accurate estimate of alcohol purchase and proxy consumption, particularly in the case of larger supermarket and shop chains that have invested heavily in data collection.38 However, shop-level data are hard to obtain due to commercial sensitivity.39 There are few published evaluations of alcohol interventions in the UK using retail data to assess changes in physical and economic availability of specific alcohol products for health improvement.18,40 The RtS studied here was originally launched as a joint initiative between Suffolk Police, Ipswich Borough Council, Suffolk County Council and the National Health Service (Suffolk) in September 2012.41 Following interviews with local practitioners and policymakers who designed and implemented the RtS in Suffolk, we hypothesized several possible mechanisms for RtS impacts on alcohol availability and sales. These include a potential ‘nudge’ effect where the impact of reducing physical availability of alcohol products by removing superstrength products helped discourage and denormalise the practice of purchasing cheap products. The RtS was also theorized as an economic availability intervention: customers with finite resources wishing to purchase low cost per unit superstrength products may, on finding those products removed, substitute for products with lower alcohol content or for different alcohol products.29,35 This study aims to evaluate the impact of the introduction of a RtS initiative on alcohol availability in the form of overall availability of alcohol units and purchasing in one national retail chain across three English counties using time-series analyses of retail sales data. Methods Setting and intervention A major supermarket chain (East of England Co-operative Society, known commonly as ‘Co-op’) voluntarily joined RtS in Suffolk and consequently ensured that its stores in that county cleared their stock of all their low-priced brands of high-strength beers/lagers and ciders in the month leading up to September 2012. These consisted of four superstrength products (7.5–9.0% ABV) but did not include any more expensive ‘craft’ or ‘premium’ high-strength products as the implementers did not associate such products with street drinking (Table 1). The same chain required stores in Essex and Norfolk to begin a similar process of withdrawing those products from sale by September 2013. Every shop from the chain participated in the intervention although a minority of stores, 6% from wave 1 and 36% from wave 2, took longer than 1 month to stop selling superstrength products (Appendix S1). Table 1 List of beer and cider products over 6.5% ABV sold during the ‘Reducing the Strength’ initiative EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 aRange of values during the period of study. bSuperstrength products (over 7.5% ABV) removed as part of the Reducing the Strength initiative. cHigh strength premium products (over 7.5% ABV) not removed as part of the Reducing the Strength initiative. All other products are high-strength premium products (over 6.5% but below 7.5% ABV) still available during the study period. EAN, European Article Number (also called International Article Number); ABV, Alcohol by volume (ABV) (%). Recommended weekly limit of 14 units of alcohol for men and women55. View Large Table 1 List of beer and cider products over 6.5% ABV sold during the ‘Reducing the Strength’ initiative EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 EAN ABV Description Size Units Price (£)a Price per unit (£)a 5010079105150 7.5 White Starb 2 l 15.0 2.50–5.23 0.17–0.35 5010153737048 9.0 Carlsberg special brewb 4 × 440 ml 15.8 1.52–9.75 0.10–0.62 5000128393041 7.5 Co-op superstrength lagerb 4 × 440 ml 13.2 1.39–7.25 0.11–0.55 5010017012526 9.0 Tennent’s super strong lagerb 4 × 440 ml 15.8 2.08–9.59 0.13–0.61 5014201655414 8.2 Special vintage ciderc 500 ml 4.1 1.73–2.13 0.42–0.52 5012845198120 8.2 Imperial ciderc 500 ml 4.1 2.15–2.61 0.52–0.64 5016878000207 6.7 Adnams Jack brand innovation 500 ml 3.4 1.42–2.94 0.42–0.88 5012845172809 7.0 Aspall dry Suffolk cider premier cru 500 ml 3.5 1.31–2.84 0.37–0.81 5012845177101 7.0 Aspall premier cru Suffolk cider pack 4 × 330 ml 9.2 1.78–6.17 0.19–0.67 5012845172830 7.0 Aspall organic Suffolk cider 500 ml 3.5 1.15–2.79 0.33–0.80 8594403110159 7.4 Budweiser Budvar Czech premium lager 330 ml 2.4 0.88–2.28 0.36–0.93 5014201203554 6.5 Westons—Wyld Wood Classic cider 500 ml 3.3 1.88–2.52 0.58–0.78 609722874786 7.0 NSB dry cider 750 ml 5.3 1.80–3.78 0.34–0.72 609722874793 7.0 NSB medium cider 750 ml 5.3 1.40–3.78 0.27–0.72 609722874809 7.0 NSB 7sweet cider 750 ml 5.3 0.90–3.78 0.17–0.72 5020628002809 7.4 Thatchers Katy cider 500 ml 3.7 1.78–2.51 0.48–0.68 5020628006685 7.4 Thatchers vintage cider 500 ml 3.4 1.82–2.37 0.49–0.64 5010327658544 6.6 Innis & Gunn original oak aged beer 330 ml 2.2 1.00–2.11 0.46–0.97 5410228102762 6.6 Leffe blonde 750 ml 5.0 2.94–4.49 0.59–0.91 5410228190424 6.6 Leffe blonde pack 4 × 330 ml 8.7 1.46–7.83 0.16–0.85 609224793127 7.0 Carter’s Essex cider 7% 500 ml 3.5 1.25–2.49 0.36–0.71 5011348010953 7.4 Banks’s Barley Gold 4 × 330 ml 9.8 4.42–5.70 0.48–0.62 5000264004184 7.3 McEwans champion ale 500 ml 3.7 2.02–2.14 0.55–0.58 5010549302348 6.5 Old crafty hen 500 ml 3.3 1.93–2.40 0.59–0.74 aRange of values during the period of study. bSuperstrength products (over 7.5% ABV) removed as part of the Reducing the Strength initiative. cHigh strength premium products (over 7.5% ABV) not removed as part of the Reducing the Strength initiative. All other products are high-strength premium products (over 6.5% but below 7.5% ABV) still available during the study period. EAN, European Article Number (also called International Article Number); ABV, Alcohol by volume (ABV) (%). Recommended weekly limit of 14 units of alcohol for men and women55. View Large Data Monthly retail sales data were provided for the period January 2012 to May 2014 obtained for 131 stores in one retail chain in the three English counties. We used the full range of data that East of England Co-operative Society provided us with for this analysis: the researchers did not have direct access to the company’s internal data systems but rather were sent data pertaining only to the intervention period and localities so that the researchers could analyse them independently. Shop-level characteristics and sales data were available including prices, quantities, product brands, alcohol content, and sales for the following drink categories: beer/lager and cider, wines, affordable sparkling and low alcohol wines, and spirits. Our primary outcome was units of alcohol sold for beer/lager and cider. Secondary outcomes included units of alcohol sold for two high-strength premium products (ABV over 7.5%) not removed as part of the RtS (Table 1), the remaining drink categories and for all products in order to examine substitution effects and in line with qualitative findings on drinkers’ responses to RtS. We looked at sales value to assess the potential economic impact of RtS on stores. Stores in Suffolk (n = 54) were regarded as stores participating in wave 1 (W1) of the intervention and stores in Norfolk and Essex (n = 77) as stores participating in wave 2 (W2) a year later. Statistical analysis We used a quasi-experimental multiple baseline time-series design42 to study changes in units of alcohol sold and sales value for beer/lager and cider, wines, sparkling and low alcohol wines, spirits and for total alcohol products after the introduction of the RtS initiative. The RtS was introduced in a staggered approach, implemented at two different time points (W1 and W2) across three different geographical areas with a combined population of 3.62 million people.43 We examined the impact of implementing RtS separately for the two waves in order to identify whether the intervention produced similar effects in the entire population of interest (i.e. whether the impact of the intervention was consistent in the two waves).42,44 The repeated pattern of a reduction in the measured outcome following the implementation of the intervention in each geographical area (i.e. wave) would suggest that the intervention is having an effect.42 An appropriate statistical approach to evaluate such impacts is the use of segmented linear regression, which divides a time series into pre- and post-intervention segments,44 with panel-corrected standard errors.45,46 We took autocorrelation into account by means of a common autoregressive first order (AR(1)) model and we included the calendar month as a term to adjust for seasonality.44,47 Details of the assumptions and model specification are available in Appendix S2. The intervention effect was assumed to occur immediately after implementation, so no transition period was taken into account in the analysis. We log-transformed our dependent variables as these were highly skewed. For ease of interpretation, regression coefficients (β) were converted into per cent change in sales and units of alcohol sold using the formula [exp(β) − 1] × 100. This approach was used to ensure data confidentiality when using commercially sensitive information, such as sales of specific alcohol products and brands. We therefore examined substitution effects at a product category level and for high-strength premium products that were not removed rather than at the level of specific products or brands. Analysis was carried out in Stata 14.1. Results Stores in W1 and W2 were similar in terms of size, area-level deprivation score and urban versus semi-urban location. Stores in W1 were open on average for fewer hours compared to those in W2 (Appendix S3). Mean units of alcohol sold per store per month were lower in W1 compared to W2 stores in all products. Overall, beer/lager and cider accounted for 32.4% of total units of alcohol sold during the study period. Superstrength products removed had previously accounted for 6.5 and 3.6% of total units sold for beer/lager and cider in W1 and W2 stores, respectively (Table 2). In terms of sales, these four products accounted for 2.1 and 1.3% of total revenue for W1 and W2 stores, respectively, before the intervention (data not shown). Table 2 Summary statistics for units of alcohol sold per store per month Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 aOnly for the period up until September 2012 for wave 1 and September 2013 for wave 2. View Large Table 2 Summary statistics for units of alcohol sold per store per month Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 Product categories Mean (SD) Median Min–Max Mean (SD) Median Min–Max Mean (SD) Median Min–Max Beer/lager & cider 11 641 (8364) 9189 2566–61 692 14 159 (9330) 11 646 884–71 467 13 120 (9029) 10 489 884–71 467 Of which super-strength products removeda 761 (680) 547 13–4782 512 (614) 305 13–5165 Of which high-strength premium products (over 7.5% ABV) not removed 334 (273) 246 4–1816 388 (344) 279 4–2325 365 (317) 258 4–2325 Spirits 9002 (8261) 6602 1984–62 816 9903 (8279) 7280 334–72 664 9531 (8282) 6967 334–72 664 Affordable sparkling and low alcohol wines 951 (1047) 643 66–13 151 1080 (1089) 711 35–9819 1026 (1074) 680 35–13 151 Wines 16 280 (16 722) 11 334 2485–133 557 17 147 (15 134) 12 786 668–102 783 16 790 (15 812) 12 087 668–133 557 All products 37 873 (33 311) 28 273 10 314–262 238 42 277 (32 390) 33 023 1920–221 608 40 462 (32 840) 30 944 1920–262 238 aOnly for the period up until September 2012 for wave 1 and September 2013 for wave 2. View Large Our analysis indicates that the impact of RtS on units of alcohol sold for beer/lager and cider was not significant in the two waves (Fig. 1 and Appendix S4). More specifically, following RtS implementation, W1 stores experienced a non-significant increase (3.7%, 95% confidence intervals (CI): −11.2 to 21.0, P = 0.647) whereas W2 stores experienced a non-significant decrease (−6.8%, 95% CI: −20.5 to 9.4, P = 0.390) (Fig. 1). In terms of all alcohol products, the introduction of RtS was associated with a non-significant increase in W1 stores (8.0%, 95% CI: −1.3 to 18.3, P = 0.094). In contrast, a significant decrease (−10.5%, 95% CI: −19.2 to −0.9, P = 0.034) was observed in W2 stores (Fig. 2 and Appendix S4). Similar patterns for beer/cider and lager were observed for sales value (Fig. 2). Fig. 1 View largeDownload slide % Change in units of alcohol sold after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 1 View largeDownload slide % Change in units of alcohol sold after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 2 View largeDownload slide % Change in sales value after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. Fig. 2 View largeDownload slide % Change in sales value after the introduction of the ‘Reducing the Strength’ initiative. Wave 1 stores started implementation by September 2012. Wave 2 stores started implementation by September 2013. In order to examine substitution effects we repeated the analysis for high-strength premium products, spirits, affordable sparkling and low alcohol wines and wines. We found that all product categories experienced similar changes in units of alcohol sold and sales value during this time period in W1 and W2 to those observed for beer/lager and cider. None of them were significant except for units of alcohol sold for wines, which appeared to drive the significant decrease observed in units of alcohol sold for all products. We found no evidence of substitution effects for high-strength premium products (Fig. 1 and Appendix S4). Discussion Main findings of this study We used retail sales data to evaluate the introduction of RtS, a public health initiative targeted at supermarkets and off-licences to remove low cost, superstrength beers and ciders from sale in three English counties. Our results show that this RtS had no significant impact on total units of alcohol sold and sales value for beer/lager and cider. We also found no observable substitution effects of alcohol products attributable to the RtS intervention in the 131 stores. What is already know on the topic Only a small number of previous studies have used retail sales data in quasi-experimental designs to evaluate alcohol interventions. Evaluation of the Scottish Alcohol Act 2010 showed that banning alcohol multi-buy promotions did not reduce alcohol purchasing at the household level,18 and the introduction of the Alcohol Act was not associated with any changes in off-trade beer sales.40 In our study, the majority of results were non-significant. The small significant decrease in units and value of alcohol sales of all products in W2 stores appears to be driven by declining wine (rather than beer/cider) sales.48 Furthermore, the changes observed in the two waves were not consistent and so the overall findings showed no intervention attributable impact.42 An Australian evaluation of local alcohol availability restrictions (cask wines and products over 2.7% ABV) found that some participants travelled further to access non-participating shops.13,14 In our study we theorize that overall alcohol purchases could be influenced by whether or not customers changed where they purchased alcohol (i.e. shops not participating in RtS), or if they substituted products within participating stores.14 Our study focused on one retail chain which maintained compliance with RtS22 and we found no substitution effects between categories of alcohol products within study stores attributable to the intervention. Customers in the study areas had the ability to access other local stores that did not participate in the RtS but we did not detect any sudden or sustained loss of income in participating stores that might be expected if substantial numbers of customers had started shopping elsewhere for alcohol. The availability of alternative stores not participating may vary within and between the three counties studied. Limitations of this study The retail data we had available related to one retail supermarket chain and the data available could not be used to consider overall area effects, shop-level or brand/product-level substitution effects, individual or sub-group level purchasing or consumption.14,18,37 Our results cannot be generalized to RtS initiatives that have removed products with >6.5% or lower ABV. We did not have the data to measure long term impacts on purchasing and consumption, although we theorised that RtS should impact on availability as soon as shops stopped selling superstrength products.13,14 The confidence intervals for our findings were wide and statistical precision might have been improved with inclusion of a greater number of stores, and/or time points.44,46 Stores in W1 and W2 had different rates of compliance, which may compromise internal validity.42 In addition, RtS is only one intervention targeting alcohol consumption and harms, and we are aware that there are a range of local alcohol policies routinely implemented in local government which we were unable to adjust for. Such unmeasured events may introduce confounding and compromise internal validity.49 Finally, segmented regression analysis has its own limitations, allowing only linear trends to be examined but changes may follow non-linear patterns.44 What this study adds Our study makes an important contribution to the evidence base for local voluntary retail alcohol interventions.18,40 The use of retail data is novel for evaluating alcohol initiatives and it has been advocated as an important means to monitor alcohol consumption40,50 despite the limitations.38 In this study, we used a retail sales time series panel data set, that contains far more information than single cross-sectional data allowing for an increased precision in estimation.46 Panel difference-in-differences analysis has been used in a previous study,18 but we opted to use panel-corrected standard errors within a regression framework, because ignoring possible correlation of regression disturbances over time and between panels may lead to overly optimistic standard errors and lead to biased statistical inference.46 The RtS initiative21 was originally developed as part of a strategy that also involved alcohol and drug treatment services and street policing to tackle street drinking and anti-social behaviour due to excess alcohol consumption, and there is some evidence that this targeted, multi-intervention approach led to reductions in police call outs and other indicators of social problems related to street drinking.21,41 This evaluation does not test RtS’s impact on wider aims of tackling alcohol social harms including street drinking. The RtS was not originally expected to have impacts on reducing overall population alcohol consumption. Potential secondary effects of RtS on the broader population of alcohol consumers are of interest to the public health community. Voluntary agreements between governments and the private sector have previously been used to encourage businesses to take actions.36 However, there is little evidence to suggest such approaches are more (cost-) effective, particularly if they are unaccompanied by monitoring, and appropriate incentives and sanctions.36 The alcohol industry and retail sector may be more willing to participate in voluntary initiatives targeting selected population groups (i.e. street drinkers) that have minimal impact on their profits. Our analysis suggests that RtS had no impact on revenues. Addressing alcohol-related harms and drinking behaviours in ‘high-risk’ groups is important but our analysis suggests that RtS may not be effective for addressing alcohol harms across the whole population. The evidence base recommends regulatory or statutory enforcement interventions restricting alcohol availability are more effective than local non-regulatory or voluntary approaches targeting specific groups.12,51–54 Conclusion This evaluation did not specifically test impacts on target groups, such as street drinkers, but examined impacts on all consumers’ alcohol purchasing patterns from one retail supermarket chain. Our findings suggest that voluntary RtS initiatives, have little or no impact on reducing alcohol availability and purchase amongst a broader population of customers. The research literature suggests that more effective regulatory public health interventions will be required to achieve substantial population health benefits in reducing alcohol consumption and alcohol-related harms. Supplementary data Supplementary data are available at the Journal of Public Health online. Acknowledgements We would like to acknowledge the invaluable input of and the helpful comments of Professor Mark Petticrew. We would also like to thank Mark Lewenz of East of England Co-operative Society who provided the retail sales data. Antonio Gasparrini provided expert advice on our analysis. The evaluation of the Reducing the Strength initiative is part of the programme of the School for Public Health Research (http://sphr.lshtm.ac.uk/). This is an independent research unit based at the London School of Hygiene and Tropical Medicine, funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR). Sole responsibility for this research lies with the authors and the views expressed are not necessarily those of the NHS, the NIHR SPHR or the Department of Health. The East of England Co-operative Society supplied the retail sales data but played no role in the funding of the study. The East of England Co-operative Society, the NHS, the NIHR and the Department of Health played no role in the design of the study, the interpretation of the findings, the writing of the article or the decision to submit. The views expressed are those of the author(s) and not necessarily those of the East of England Co-operative Society, NHS, the NIHR or the Department of Health. Funding This study was funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR). Two co-authors (AJ and SA) contributed as part of their normal salaried work for Suffolk County Council. The East of England Co-operative Society supplied the retail sales data but played no role in the funding of the study. The East of England Co-operative Society, the NHS, the NIHR SPHR and the Department of Health played no role in the design of the study, the interpretation of the findings, the writing of the paper or the decision to submit. References 1 World Health Organization . Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013−2020 . Geneva : WHO , 2013 . 2 World Health Organization . Global Status Report on Alcohol and Health . Geneva : WHO , 2014 . 3 World Health Organization . Global Strategy to Reduce the Harmful Use of Alcohol . Geneva : WHO , 2010 . PubMed PubMed 4 Morland KB . An evaluation of a neighborhood-level intervention to a local food environment . Am J Prev Med 2010 ; 39 : e31 – 8 . Google Scholar CrossRef Search ADS PubMed 5 Gase LN , McCarthy WJ , Robles B et al. . Student receptivity to new school meal offerings: assessing fruit and vegetable waste among middle school students in the Los Angeles Unified School District . Prev Med 2014 ; 67 ( Suppl 1 ): S28 – 33 . Google Scholar CrossRef Search ADS PubMed 6 Soler RE , Whitten KL , Ottley PG . Communities putting prevention to work: local evaluation of community-based strategies designed to make healthy living easier . Prev Med 2014 ; 67 ( Suppl 1 ): S1 – 3 . Google Scholar CrossRef Search ADS PubMed 7 Valdivia Espino J , Guerrero N , Rhoads N et al. . Community-based restaurant interventions to promote healthy eating: a systematic review . Prev Chronic Dis 2015 ; 12 : 140455 . Google Scholar CrossRef Search ADS 8 Nelson TF , Xuan Z , Babor TF et al. . Efficacy and the strength of evidence of U.S. alcohol control policies . Am J Prev Med 2013 ; 45 : 19 – 28 . Google Scholar CrossRef Search ADS PubMed 9 Rammohan V , Hahn RA , Elder R et al. . Effects of dram shop liability and enhanced overservice law enforcement initiatives on excessive alcohol consumption and related harms: two community guide systematic reviews . Am J Prev Med 2011 ; 41 : 334 – 43 . Google Scholar CrossRef Search ADS PubMed 10 Campbell CA , Hahn RA , Elder R et al. . The effectiveness of limiting alcohol outlet density as a means of reducing excessive alcohol consumption and alcohol-related harms . Am J Prev Med 2009 ; 37 : 556 – 69 . Google Scholar CrossRef Search ADS PubMed 11 Stockwell T , Zhao J , Giesbrecht N et al. . The raising of minimum alcohol prices in Saskatchewan, Canada: impacts on consumption and implications for public health . Am J Public Health 2012 ; 102 : e103 – e10 . Google Scholar CrossRef Search ADS PubMed 12 Martineau F , Tyner E , Lorenc T et al. . Population-level interventions to reduce alcohol-related harm: an overview of systematic reviews . Prev Med 2013 ; 57 : 278 – 96 . Google Scholar CrossRef Search ADS PubMed 13 Kinnane S , Farringdon F , Henderson-Yates L et al. . Fitzroy Valley Alcohol Restriction Report: An Evaluation of the Effects of a Restriction on Take-away Alcohol Relating to Measurable Health and Social Outcomes, Community Perceptions and Behaviours After a Two Year Period, 2010 . http://www.dao.health.wa.gov.au/DesktopModules/Bring2mind/DMX/Download.aspx?EntryId=680&Command=Core_Download&PortalId=0&TabId=211 (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dhBDi59v on 11 December 2015. 14 National Drug Research Institute . Restrictions on the Sale and Supply of Alcohol: Evidence and Outcomes. Perth: National Drug Research Institute, Curtin University of Technology, 2007 . http://ndri.curtin.edu.au/local/docs/pdf/publications/R207.pdf (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dhAzukmG on 11 December 2015. 15 Maclennan B , Kypri K , Room R et al. . Local government alcohol policy development: case studies in three New Zealand communities . Addiction 2013 ; 108 : 885 – 95 . Google Scholar CrossRef Search ADS PubMed 16 Ahern J , Margerison-Zilko C , Hubbard A et al. . Alcohol outlets and binge drinking in urban neighborhoods: the implications of nonlinearity for intervention and policy . Am J Public Health 2013 ; 103 : e81 – 7 . Google Scholar CrossRef Search ADS PubMed 17 Fone D , Dunstan F , White J et al. . Change in alcohol outlet density and alcohol-related harm to population health (CHALICE) . BMC Public Health 2012 ; 12 : 428 . Google Scholar CrossRef Search ADS PubMed 18 Nakamura R , Suhrcke M , Pechey R et al. . Impact on alcohol purchasing of a ban on multi-buy promotions: a quasi-experimental evaluation comparing Scotland with England and Wales . Addiction 2014 ; 109 : 558 – 67 . Google Scholar CrossRef Search ADS PubMed 19 Purshouse RC , Meier PS , Brennan A et al. . Estimated effect of alcohol pricing policies on health and health economic outcomes in England: an epidemiological model . Lancet 2010 ; 375 : 1355 – 64 . Google Scholar CrossRef Search ADS PubMed 20 Han D , Gorman DM . Evaluating the effects of the introduction of off-sale alcohol outlets on violent crime . Alcohol Alcohol 2013 ; 48 : 370 – 4 . Google Scholar CrossRef Search ADS PubMed 21 Alcohol Concern . Guide to Alcohol for Councillors . London : Alcohol Concern , 2014 . http://www.alcoholconcern.org.uk/wp-content/uploads/woocommerce_uploads/2014/12/Guide-to-Alcohol-160813-APPROVED.pdf (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh8ZwRom. on 11 December 2015. PubMed PubMed 22 Local Government Association . Reducing the Strength Schemes: Guidance for Councils Considering Setting up a Scheme, 2014 . Available at: http://www.local.gov.uk/regulatory-services-and-licensing/-/journal_content/56/10180/6793034/ARTICLE (accessed 10 December 2015). Archived at: http://www.webcitation.org/6dh8jYS00 on 11 December 2015. 23 Nicholls J . The Government Alcohol Strategy 2012: alcohol policy at a turning point? Drugs Educ Prev Policy 2012 ; 19 : 355 – 9 . Google Scholar CrossRef Search ADS 24 Hughes K , Bellis MA , Leckenby N et al. . Does legislation to prevent alcohol sales to drunk individuals work? Measuring the propensity for night-time sales to drunks in a UK city . J Epidemiol Community Health 2014 ; 68 : 453 – 6 . Google Scholar CrossRef Search ADS PubMed 25 McGill E , Egan M , Petticrew M et al. . Trading quality for relevance: non-health decision-makers’ use of evidence on the social determinants of health . BMJ Open 2015 ; 5 : e007053 . Google Scholar CrossRef Search ADS PubMed 26 Dolowitz D , Marsh D . Learning from abroad: the role of policy transfer in contemporary policy‐making . Governance 2000 ; 13 : 5 – 23 . Google Scholar CrossRef Search ADS 27 Alcohol, Public Policy G . Alcohol: no ordinary commodity—a summary of the second edition . Addiction 2010 ; 105 : 769 – 79 . CrossRef Search ADS PubMed 28 Goodall T . White Cider and Street Drinkers: Recommendations to Reduce Harm . London : Alcohol Concern UK , 2011 . http://www.thamesreach.org.uk/publications/research-reports/white-cider-and-street-drinkers/ (25 June 2015, date last accessed). Archived at: http://www.webcitation.org/6ZYIecFEw on 25 June 2015. 29 McGill E , Marks D , Sumpter C et al. . Consequences of removing cheap, super-strength beer and cider: a qualitative study of a UK local alcohol availability intervention . BMJ Open 2016 ; 6 : e010759 . Google Scholar CrossRef Search ADS PubMed 30 Sweeting H . Alcohol units at pocket money prices [e-letter] . Br Med J 2006 ; 333 : 825 . -b. 31 Gornall J . Doctors and the alcohol industry: an unhealthy mix? Br Med J 2013 ; 346 : f1889 . Google Scholar CrossRef Search ADS 32 Portman Group . K Cider, Complaint Decision Details, 2017 . http://www.portmangroup.org.uk/complaints/complaint-decisions/complaint-decision-details/2017/01/12/k-cider (24 January 2017, date last accessed). Archived at: http://www.webcitation.org/6nkoGttF8 on 24 January 2017. 33 Association of Convenience Stores . Reducing the Strength: Retailer Guidance, 2015 . http://www.acs.org.uk/advice/reducing-the-strength/ (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh7Gcc5S on 11 December 2015. 34 Competition and Markets Authority . High-Strength Alcohol Schemes: Competition Law Advice for Retailers, 2015 . https://www.gov.uk/government/publications/high-strength-alcohol-schemes-competition-law-advice-for-retailers (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh7NK9Rb on 11 December 2015. 35 Sumpter C , McGill E , Dickie E et al. . Reducing the Strength: a mixed methods evaluation of alcohol retailers’ willingness to voluntarily reduce the availability of low cost, high strength beers and ciders in two UK local authorities . BMC Public Health 2016 ; 16 : 448 . Google Scholar CrossRef Search ADS PubMed 36 Bryden A , Petticrew M , Mays N et al. . Voluntary agreements between government and business—a scoping review of the literature with specific reference to the Public Health Responsibility Deal . Health Policy 2013 ; 110 : 186 – 97 . Google Scholar CrossRef Search ADS PubMed 37 Community Alcohol Partnerships . About Community Alcohol Partnerships, 2015 . http://www.communityalcoholpartnerships.co.uk/ (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh622qSP on 11 December 2015. 38 Robinson M , Thorpe R , Beeston C et al. . A review of the validity and reliability of alcohol retail sales data for monitoring population levels of alcohol consumption: a Scottish perspective . Alcohol Alcohol 2013 ; 48 : 231 – 40 . Google Scholar CrossRef Search ADS PubMed 39 Egan M , Pliakas T , Grace D et al. . Use of retail data in the assessment of natural experiments: the case of Reducing the Strength, an intervention to reduce alcohol availability . Lancet 2014 ; 384 : S30 . Google Scholar CrossRef Search ADS 40 Robinson M , Geue C , Lewsey J et al. . Evaluating the impact of the alcohol act on off-trade alcohol sales: a natural experiment in Scotland . Addiction 2014 ; 109 : 2035 – 43 . Google Scholar CrossRef Search ADS PubMed 41 Suffolk Constabulary . Reducing the Strength Ipswich—One Year on, 2012 . http://www.suffolk.police.uk/safetyadvice/personalsafety/alcoholawareness/reducingthestrength/oneyearon.aspx (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh6n05NZ on 11 December 2015. 42 Hawkins NG , Sanson-Fisher RW , Shakeshaft A et al. . The multiple baseline design for evaluating population-based research . Am J Prev Med 2007 ; 33 : 162 – 8 . Google Scholar CrossRef Search ADS PubMed 43 Office for National Statistics . Population Estimates for UK, England and Wales, Scotland and Northern Ireland, Mid-2014, 2015. Available at: http://www.ons.gov.uk/ons/publications/re-reference-tables.html?edition=tcm%3A77-368259 (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh6SoR4s on 11 December 2015. 44 Wagner AK , Soumerai SB , Zhang F et al. . Segmented regression analysis of interrupted time series studies in medication use research . J Clin Pharm Ther 2002 ; 27 : 299 – 309 . Google Scholar CrossRef Search ADS PubMed 45 Beck N , Katz JN . What to do (and not to do) with time-series cross-section data . Am Polit Sci Rev 1995 ; 89 : 634 – 47 . Google Scholar CrossRef Search ADS 46 Hoechle D . Robust standard errors for panel regressions with cross-sectional dependence . Stata J 2007 ; 7 : 281 – 312 . 47 Linden A . Conducting interrupted time-series analysis for single- and multiple-group comparisons . Stata J 2015 ; 15 : 480 – 500 . 48 The Drinks Report . UK Alcohol Market Returns to Growth, 2015 . http://www.thedrinksreport.com/news/2015/16101-uk-alcohol-market-returns-to-growth.html (10 December 2015, date last accessed). Archived at: http://www.webcitation.org/6dh702Jz8 on 11 December 2015. 49 Craig P , Cooper C , Gunnell D et al. . Using natural experiments to evaluate population health interventions: new Medical Research Council guidance . J Epidemiol Community Health 2012 ; 66 : 1182 – 6 . Google Scholar CrossRef Search ADS PubMed 50 Robinson M , Shipton D , Walsh D et al. . Regional alcohol consumption and alcohol-related mortality in Great Britain: novel insights using retail sales data . BMC Public Health 2015 ; 15 : 1 . Google Scholar CrossRef Search ADS PubMed 51 de Vocht F , Heron J , Angus C et al. . Measurable effects of local alcohol licensing policies on population health in England . J Epidemiol Community Health 2016 ; 70 : 231 – 7 . Google Scholar CrossRef Search ADS PubMed 52 de Vocht F , Heron J , Campbell R et al. . Testing the impact of local alcohol licencing policies on reported crime rates in England . J Epidemiol Community Health 2017 ; 71 : 137 – 45 . Google Scholar CrossRef Search ADS PubMed 53 Knai C , Petticrew M , Durand MA et al. . The Public Health Responsibility deal: has a public-private partnership brought about action on alcohol reduction? Addiction 2015 ; 110 : 1217 – 25 . Google Scholar CrossRef Search ADS PubMed 54 de Vocht F , Tilling K , Pliakas T et al. . The intervention effect of local alcohol licensing policies on hospital admission and crime: a natural experiment using a novel Bayesian synthetic time-series method . J Epidemiol Community Health 2017 ; 71 : 912 – 8 . Google Scholar CrossRef Search ADS PubMed 55 Department of Health . Health Risks From Alcohol: New Guidelines, 2016 . https://www.gov.uk/government/consultations/health-risks-from-alcohol-new-guidelines (2 August 2017, date last accessed). Archived at: http://www.webcitation.org/6sQ3CIIWX on 2 August 2017. © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Journal of Public HealthOxford University Press

Published: Feb 13, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off