Resumen

Aunque en la bibliografía se ha reportado la relación entre el consumo de alcohol y los accidentes, se ha estudiado poco la asociación entre el lugar de ocurrencia de las lesiones y el consumo de alcohol. Por esta razón, nuestro objetivo es estimar la asociación entre los lugares de ocurrencia de la lesión y el consumo de alcohol.

Los datos usados en este estudio provienen de una sala de urgencias de la Ciudad de México y son parte de un estudio epidemiológico multinacional (Estudio Colaborativo de Alcohol y Lesiones) coordinado por la Organización Mundial de la Salud. La muestra se conformó por 705 pacientes mayores de 18 años admitidos para tratamiento por lesión por primera vez. Los datos fueron recolectados por entrevistadores entrenados usando un cuestionario estructurado y con el consentimiento informado de los participantes.

Los resultados que se presentan se basan en un diseño case-crossover, que se caracteriza por usar a los sujetos como sus propios controles pero en un periodo de tiempo diferente. Se calcularon razones de momios (RM) e intervalos de confianza al 95% (IC), para cada lugar de ocurrencia de lesiones (casa propia, casa de otra persona, calle/carretera, escuela, bar/lugar para beber y lugar de trabajo).

Se analizaron como posibles modificadores de efecto el sexo, la edad y el consumo de alcohol seis horas antes de la lesión, y se calculó una prueba de heterogeneidad para detectar diferencias estadísticamente significativas entre éstos. Además -y sólo para las lesiones en calle/carretera-, se obtuvieron estimadores por tipo de accidente (de tránsito, violencia o caídas/otro), consumo de alcohol antes (número de copas) y autopercepción de embriaguez.

La calle/carretera fue el único lugar en que el riesgo relativo de sufrir una lesión era elevado y estadísticamente significativo (RM=12.11; IC95%=7.5-19.6). Por otro lado, la casa propia (RM=0.25; IC95%=0.19-0.34) y el lugar de trabajo tuvieron un riesgo bajo.

No se encontraron diferencias significativas en modificadores de efecto como el sexo o la edad. Sin embargo, el consumo de alcohol incrementó el riesgo de sufrir una lesión en la calle/carretera RM=21.33, que contrasta con una RM de 10.27 para aquellos que no lo hicieron. No obstante, la prueba de heterogeneidad (p=0.260) sugiere que estas RM no son estadísticamente diferentes entre sí.

El riesgo de sufrir lesiones en casa de otra persona también aumentó con el uso de alcohol (RM=2.60; IC95%=0.93-7.29), y la prueba de heterogeneidad (p=0.017) muestra que dicho estimador sí difiere de aquel obtenido sin consumo (RM=0.63; IC95%=0.04-0.19).

Si se profundiza en el contexto de las lesiones en calle/carretera, el consumo de alcohol incrementó el riesgo de una lesión debida a la violencia (RM=40.0; IC95%=5.5-290.9), además de que el riego aumentó en función del número de copas y de la percepción de embriaguez del sujeto.

Estos resultados son los primeros en cuantificar el riesgo asociado a los lugares en que ocurren las lesiones, el consumo de alcohol y algunas otras variables asociadas usando un diseño de case-crossover. Se espera que éstos ayuden a entender el contexto del uso de alcohol y los problemas relacionados con la intención de informar a las personas y apoyar políticas públicas destinadas a reducir las consecuencias negativas del mismo.

Palabras clave: lugar de lesión; consumo de alcohol; sala de urgencia; case-crossover; violencia en la calle

Summary

The relationship between alcohol use and injuries (fatal and non-fatal) is well documented. Through the years, new methodological approaches have been developed, looking for validity in measurements and control groups in order to quantify this relationship. The case-crossover designs use the same subjects as cases and controls but in a different time period, overcoming the selection bias of case-control studies.

The risk associated to alcohol use and injuries using this design has been estimated at 4.0 and at 10.0 for alcohol and violence-related injuries. However, it is not well known the rol played by the context of drinking in this risk like companions, places of injury, places, frequency and quantity of drinking. In this paper we will estimate the relationship between places of injury and alcohol use before the event, as well as other relevant variables.

Data presented here comes from a Mexico City emergency department (ED) in the multi-country epidemiological study (Collaborative Study on Alcohol and Injuries) coordinated by the World Health Organization.

The sample consisted of all eligible patients (adults 18 years and older) who entered the ED for an injury and first treatment. The final sample totalled 705 respondents and was collected with informed consent by trained interviewers using a structured questionnaire lasting about 25 minutes. Among other items, respondents were asked questions about demographics, injury details (type, place, violence involvement), alcohol use prior to injury, typical drinking, place, and alcohol use during the previous week.

Case-crossover analyses were performed using STATA’s conditional logistic regression. Odds ratios (OR) and 95% confidence intervals (CI) were estimated for each of the six injury places revised: own home, someone else’s home, street/highway, school, drinking place, and workplace.

Sex, age, and alcohol use were analyzed as effect modifiers. OR’s and CI’s were estimated for each one, as well as a heterogeneity test. In addition -and only for street/highway-, estimates were obtained for injury type (traffic accident, violence, or falls/other), alcohol consumption before event (number of drinks), and drunkenness self-perception.

The most common injury places were street/highway (n=254; 36.1%) and own home (n=213; 30.3%). Also, there was a relatively high number of injuries at work places (15%). Likewise, the most common place where this patients were the same day of the week, same time, one week before, was their own home (n=378; 53.8%) and work place (n=164; 23.3%).

Street/highway was the place with higher injury risk (OR=12.11; CI95%=7.49-19.59). Own home (OR=0.25; CI95%=0.19-0.34) and work place had both a low injury risk.

No significant differences were found in effect modifiers as gender or age group. However, alcohol use before accident increased the risk for street/highway (OR=21.33). Although the OR for not drinking in the event was 10.27, the heterogeneity test (p=0.260) suggested no differences between them.

Risk for someone else’s home was also increased with alcohol use (OR=2.60; CI95%=0.93-7.29). Although the null value is within the CI, the heterogeneity test showed significant differences (p=0.017) between this risk and non-alcohol associated risks (OR=0.63; CI95%=0.37-1.07). In addition, alcohol related injuries at own home had an even lower risk (OR=0.08; CI95%=0.04-0.19).

Accident type in the street was also analyzed. It was not possible to estimate the OR for vehicle-related accidents (as a pedestrian, driver, or passenger) due to zero cell in our data. Violence related injuries (by a gun, knife, punch, etc.) (OR=10.22; CI95%=5.16-20.27) were more likely to occur in the street than non-violent injuries (such as falls) (OR=8.67; CI95%=4.35-17.28).

In this context, alcohol use greatly increased the risk of violence-related injury (OR=40.0; CI95%=5.5-290.9) while non-alcohol related injuries risk decreases (OR=6.5; CI=3.09-13.68).

For alcohol-related injuries in the street/highway, estimates for drunkenness self-perception and quantity of alcohol consumed six hours before injury increased; however, these estimates were not statically significant.

These results are the first to quantify risks associated to injury places, alcohol use, and some others variables linked to it, using a case-crossover design. It is clear that being injured at home (OR=0.25) or the workplace (OR=0.27) is less likely to occur than being injured in the street or highway (OR=12.11), and alcohol use usually increased this difference, especially in someone else’s home or in the street. Due to our small sample size many CI’s were extremely wide or it was not possible to estimate really important risks like traffic-related injuries. Future research may consider this, as well as asking for several control periods according to this design.

It is expected that this findings will help to understand alcohol use context and its related problems in order to inform people and support public policies addressed to reduce the negative consequences derived of episodic and acute alcohol use.

Key words: Injury place; alcohol use; emergency department; case-crossover; street violence