This is the first study in Sweden analysing ecological associations between the local alcohol environment and non-fatal injuries in children and young people. Except for the associations between alcohol access and alcohol consumption with injury rates in boys aged 13–17 years, no associations were observed between alcohol access or per capita alcohol consumption and nonfatal childhood injuries. Municipalities with high rates of crimes against alcohol laws, however, had high injury rates in both boys and girls aged 6–17 years. High rates of driving while intoxicated with alcohol were associated with non-fatal injuries in boys and girls aged 13–17 years. No association was found between any of the indicators of alcohol environment and nonfatal injuries in children aged 5 years or under or in young adults (18–24 years).
Since many studies have shown that alcohol consumption and alcohol-related problems have increased in the adult Swedish populations [23, 24, 27], positive associations were expected between the alcohol environment and injuries in young people in this study. However, except for the results obtained in boys aged 13–17 years no other association at the municipality-level was found between local alcohol access or per capita alcohol consumption and nonfatal childhood injuries. It has been suggested elsewhere that modifications in the local alcohol environment lead to individual changes in drinking behavior and in alcohol-related problems . Thus, one possible explanation for the lack of area-level associations between alcohol access or per capita alcohol consumption and nonfatal childhood injuries could be the restrictive alcohol policies in Sweden. Apart from policies and restrictions, local drinking patterns might also be influenced by local norms and social control in some of the smaller municipalities with high participation rates in sport clubs, social associations and churches. Unfortunately, variations in this respect between the municipalities were not possible to study due to the small number (n = 14) of municipalities involved in injury registrations with SLISS. Since this is the first time that the local alcohol access and per capita alcohol consumption is studied in Sweden in association to nonfatal childhood injuries, another explanation may be that these alcohol indicators and nonfatal childhood injuries are simply not associated in Sweden at all, or at least not in smaller semi-rural municipalities as those studied here.
In congruence with earlier studies [7, 8] we found a positive area-level association between per capita alcohol consumption and injuries in boys 13–17 years. One important difference between this ecological study and the earlier studies [7, 8] is that they used fatal injuries at national level, whereas this study used nonfatal injuries at the municipality-level. Another important difference is that earlier studies used populations aged 15 and over. Since there are studies suggesting that adolescent drinking patterns mirror those of adults  it is possible that adolescent alcohol consumption is an important determinant of the observed association. There were, however, no data available to control for that possibility.
The positive area-level associations in boys and the negative associations in girls between alcohol access, alcohol consumption, and injuries in children aged 6–17 years could have at least three explanations. One explanation, not explored here, is that boys may have higher access to alcohol  and consume more alcohol than girls even if the reported differences are higher in older age groups than in younger [15, 28, 29]. It is also possible that local gender norms and opportunities may influence not only access to and consumption of alcohol but also the access to and choice of leisure time activities . Boys, for example, may have had higher exposure to injury risks than girls due to higher access to sports arenas and higher mobility, but we were not able to control for this in the analysis. Another explanation is that boys may have higher outdoor injuries and intentional injuries than girls , which could also explain the observed gender differences in strengths and direction of the associations between alcohol access and per capita alcohol consumption and injuries in children aged 13–17 years. A reason for not including possible confounders in the analysis is twofold. First, even if Sweden has rich public statistics there is a lack of specific data at the municipality level. National statistics are more available but given the large variation in municipality characteristics in Sweden the national mean numbers cannot reflect local circumstances. Sweden is a vast country but the population density is low and local variations and high self-determinations in municipalities are high. Thus, a lack of relevant statistical information on possible confounders at the municipality level limited our confounding analysis. Second, we were limited also by the number of municipalities participating in the injury registration. The data is descriptive and constitutes a basis for generating hypotheses to be explored in future studies.
The two ecological indicators of local alcohol-related crimes (DWI and crimes against alcohol laws) showed positive area-level associations with the occurrence of nonfatal injuries in both boys and girls aged 13–17 years. No earlier study was found exploring these ecological indicators in association with childhood injuries, it is important to discuss these results in relation to the validity of the alcohol indicators. The Swedish legislation has severe penalties for DWI offences , and Hubicka, Bergman and Laurell  have suggested that the registration of DWI offences is strongly influenced by active control by the police. Since the studied municipalities are mainly rural, it is highly probable that the registered DWI offences in the studied region were associated with damage to a property or to violent injury. Hubicka and Bergman  also found an association between DWI and criminality and antisocial behaviours. Therefore, it is possible that the two ecological indicators of alcohol-related crimes are more closely related to the level of overall criminality than merely to the alcohol environment. Additionally, since mean annual rates for 2000–2005 were used to stabilize the data, low numbers of alcohol-related crimes might have produced some random difference in our data compared with Sweden as a whole. Finally, because the 14 municipalities share the same Police Administration and the same County Administrative Board, we do not assume large geographical differences in enforcement, policies or practices between the studied municipalities.
In this study none of the ecological indicators of the local alcohol environment was associated with the occurrence of nonfatal injuries in young adults (18–24 years). This result was unexpected since previous area-level studies [11, 12] have found a positive association between different indicators of alcohol use, alcohol behaviour and injuries in young adults. A possible explanation for the absence of area-level associations in this age group might be methodological and related to health care seeking behaviour for minor injuries compared to younger age groups. Young adults have the highest alcohol consumption of all age groups in Sweden and it is possible that they avoid seeking health care for minor injuries occurring while they are under intoxication. Self care might also be more common in this age group compared to the younger ones. Another possible explanation is that better coping strategies for avoiding injuries in hazardous situations have been developed with age resulting in lower injury prevalence rates.
It is well known that alcohol consumption and alcohol-related crimes are unevenly distributed between women and men. However, few studies have analysed gender in relation to risk exposure measured at the municipality level. The observed gender differences in the area-level associations between per capita alcohol consumption and nonfatal injuries in children aged 6–17 years show that gender is an essential determinant of social outcomes, including injuries, and further studies should take this factor into account. It is also important that statistics and other sources of information present data for both genders. Boys, and in particular young men, are not only consuming more alcohol and more often involved in alcohol-related crimes they also have a higher burden of injuries. This study suggests that the local alcohol environment might be an important factor to study further in the future. It would be of specific interest to be able to better separate individual-level exposure from the municipality-level of exposure and also to identify possible interactive effects. Another perspective is important to mention. The area-level association between crimes against alcohol laws and injuries was similar between girls and boys. In the injury register it is not possible to separate violent related injuries otherwise, it would have been interesting to separately analyse the association between crimes against alcohol laws and injuries in girls and boys contributing to the understanding of the influence of alcohol and violence against girls and young women. The underreporting of violence behind injuries is expected to be higher among girls and young women due to the stigma and shame associated with violence exposure in girls and women.