Measuring injury magnitude and patterns in a low-income country : experiences from Nicaragua

Abstract: About 16,000 people around the world die every day from injuries. For every person that dies, several thousands more are injured, many suffering with permanent disabilities. However, like many other health problems, the magnitude and pattern of injuries in a certain population are often difficult to assess. While in many high-income countries, regular health registries may provide accurate figures, other approaches are often needed in low-income countries. Therefore, the overall aim of this thesis is to measure the magnitude and pattern of injuries in a defined population (mainly the municipality of Leon, Nicaragua) through the use of different methodologies. The first step in this study was to conduct an exploratory analysis (Paper 1) to compare the existing nationwide data sources on injury surveillance with respect to validity and prevention-relevance. However, these results were discouraging due to low validity and the lack of information for prevention purposes. In part due to the above, it was considered important to collect injury data in the municipality of León in other ways. One of the objectives of this second step was to describe the development of a hospital-based injury surveillance system aimed at prevention and to study the incidence of injuries based on data obtained from this data collection system (Paper 2). About 16% of emergency room visits were due to injuries in this low-income country context. For every death due to injury, 31 inpatients and 253 outpatients were reported. Homes and traffic areas were the main arenas where injuries occurred. The main causes were falls, traffic accidents and violence. The underreporting rate was 6%, and in 20.3% of the cases, no E-code was recorded. This study shows that hospital-based injury surveillance is an effective and potential means available for the prevention and control of injuries. However, its low coverage is a concern due to people s limited access to hospital services. Next, a study was conducted to provide a reliable estimation of traffic-related injuries in the same catchment area by capture-recapture analysis (using hospital and traffic police records) (Paper 3). This study demonstrated that neither police records nor hospital records nor the aggregate database provide acceptable coverage of traffic-related injuries. Limitations of coverage in both these data sources justified the realization of a survey in a representative sample in Leon municipality (Paper 4). Ninety-three percent of injuries were minor and seven percent were moderate or severe. The overall incidence rate was 414.2 per 1,000 inhabitants per year, but decreased to 27.6 per 1,000 when minor injuries were excluded. Most of the injuries were unintentional and only 1.2 percent were intentional. The main places of injury occurrence were homes and streets. Nine percent of all injured persons sought hospital treatment. The main causes of nonfatal cases were falls, traffic, and cuts, whereas fatalities were associated with intentional injuries. For every death due to injury, there was one permanent disability, 25 moderate/severe injuries, and 354 minor injuries. The seeking of hospital treatment depended on the severity of the injury. A complementary study on mortality and disability shows that the leading causes of mortality overall were non-communicable diseases (176/100,000), injuries (55/100,000) and communicable diseases (55/100,000). The incidence of disability-related injuries was 75/100,000, and the main types were skeletal, disfiguring, and blindness. Their causes were falls, traffic, and violence. Around three-quarters of disabilities and half of injury deaths received hospital attention. The leading causes of years of potential life lost (YPLL) and disability-adjusted life year (DALY) were traffic, falls, drowning, and violence (Paper 5). Based on these findings, the ascertainment and validity of data sources and injury indicators must be evaluated carefully when planning injury prevention measures. When results from these studies are compared, the distribution of causes of injury vary by severity and source. The main causes of injury deaths among different data sources were similar; however, these causes differ in non-fatal injuries. A methodological issue in hospital-based surveillance is that it often fails to capture most of the extreme outcomes on the injury spectrum. On the other hand, household surveys can capture most of the injury spectrum, which sometimes is missing in traditional data sources, but they cannot establish temporal variations of injuries, especially in fatal cases. Due to the cost associated with carrying out hospital or population-based studies, capture-recapture methods represent a good option for measuring the magnitude and pattern of injuries, especially in low-income countries where resources are scarce. In conclusion, methodological issues involving the sources of injury data, injury severity, and research methods must be assessed carefully to be able to measure injuries properly as a basis for effective interventions.