Long-term mortality following injury in working-age adults: a systematic review

Priorities for prevention activities and planning for services depend on comprehensive knowledge of the distribution of the injury-related burden in the community. The aim of this systematic review was to quantify the effect of being injured, compared with not being injured, on long-term mortality in working age adults. Cohort studies were selected that were population-based, measured mortality post-discharge from inpatient treatment, included a non-injured comparison group and related to working-age adults. Data synthesis was in tabular and text form with a meta-analysis not being possible because of the heterogeneity between studies. Eleven studies met the inclusion criteria. All studies found an overall positive association between injury and increased mortality. While the greatest excess mortality was evident during the initial period post-injury, increased mortality was shown in some studies to persist for up to 40 years after injury. Due to the limited number of injury types studied and heterogeneity between studies, there is insufficient published evidence on which to calculate population estimates of long-term mortality, where injury is a component cause. The review does suggest there is considerable excess mortality following injury that is not accounted for in current methods of quantifying injury burden, and is not used to assess quality and effectiveness of trauma care.     

Indicators of injury burden: which types are the most important?

Injury indicators are used for monitoring the impact of injury prevention initiatives on the population burden of injury. The object of the present study was to identify the types of injury responsible for the major component of the population health burden of injury in a large cohort in Manitoba, Canada. Injury cases (ICD-9-CM 800-995) aged 18-64 years were identified from all Manitoba hospital data between 1988 and 1991. Morbidity data were obtained from hospital discharge abstracts 12 months prior to date of injury and for 12 months post-injury. Outcomes for individuals were calculated as the difference pre- and post-injury in hospital inpatient days. Death outcomes in the 12 months post-injury were obtained by linking the cohort with the population registry. Summed outcomes across the population were stratified into injury types based on the International Code of Diseases (ICD) code of the index injury. Outcomes were also stratified by injury severity score categories where the injury severity score was obtained using ICDMAP-90. When ranked by contribution to the cohort's cumulative hospital inpatient days in the 12 months post-injury, the six most common ICD subchapter groups accounted for 65% of the total inpatient days. These six injury types also accounted for 62% of the total number of deaths in this cohort in 12 months after injury. The suggested injury types to use as indicators of burden include fracture of the lower limb, fracture of the head and neck, poisonings, intracranial injury, fracture of the upper limb, and fracture of skull.     

How well do anatomical-based injury severity scores predict health service use in the 12 months after injury?

There is an acknowledged need for valid and reliable injury scores, suitable for use at the population level, which can accurately predict the long-term outcome of injury. The objective was to quantify the extent to which the abbreviated injury severity score (AIS) and the functional capacity index score (FCI) predict use of health services in the 12 months following an injury event. A cohort of injured people (ICD-9-CM 800-995) aged 18 - 64 years was identified from Manitoba hospital discharge abstracts from January 1988 to December 1991. For each member of the cohort whose injuries could be mapped to an abbreviated injury scale unique identifier, a maximum AIS (maxAIS) and a maximum FCI (maxFCI) were obtained. The cohort was linked with hospital discharge abstracts, physicians' claims and deaths from the population registry for the 12 months following injury. Negative binomial regression was used to model the relationships between the severity scores and the three outcome measures, while controlling for potential confounding variables. In total, 20 677 (97%) eligible cases were identified, of which 16 834 (81%) could be assigned a maxAIS and 15 823 (77%) a maxFCI. MaxAIS and maxFCI were significantly associated with total days in hospital following injury, but explained little of the variation in any of the health service use outcome variables (maxAIS, partial pseudo r2 ranging from < 0.001 to 0.041; and maxFCI, partial pseudo r2 ranging from < 0.001 to 0.018). It was concluded that anatomical damage is only partly responsible for long-term injury outcome. Additional variables would need to be included in predictive models of health outcomes of injury before these models could be reliable.     

Airline Codeshare Alliances

The paper juxtaposes the challenges that airline codeshare alliances create for analytical information systems on the one hand and their motivation from a marketing perspective on the other. The authors review the state-of-the-art literature on potential marketing benefits and analyze the impact on airline planning systems. In this regard, revenue management systems are of particular interest. Based on a simulation study, the authors infer a severe impact of decentralized codeshare controls as currently widely implemented in the industry on revenue management performance. In the scenarios examined, complementary codesharing reduces alliance-wide revenues by up to 1 %. Losses increase when a carrier experiences high local demand or a high degree of codeshare demand, and disseminate over the whole network. Virtual codeshares also cause losses of 0.3 % to 1.5 % depending on the discount level offered by the marketing carrier and on the demand structure. Finally, the authors formulate a set of managerial implications based on these findings.     

Model-Based Decision Support in Manufacturing and Service Networks

In this paper, we sketch some of the challenges that should be addressed in future research efforts for model-based decision support in manufacturing and service networks. This includes integration issues, taking into account the autonomy of the decision-making entities in face of information asymmetry, the modeling of preferences of the decision makers, efficiently determining robust solutions, i.e. solutions that are insensitive with respect to changes in the problem data, and a reduction of the time needed for model building and usage. The problem solution cycle includes problem analysis, the design of appropriate algorithms and their performance assessment. We are interested in a prototypical integration of the proposed methods within application systems, which can be followed up with field tests of the extended application systems. We argue that the described research agenda requires the interdisciplinary collaboration of business and information systems engineering researchers with colleagues from management science, computer science, and operations research. In addition, we present some exemplifying, illustrative examples of relevant research results.