Summary
An Overview of the 3rd International Conference on Measuring the Burden of Injury - Movement Towards Consensus
Stephen Luchter
Ellen J.
MacKenzie
The third international conference on Measuring the Burden of Injury was held in Baltimore, Maryland on May 15th and 16th, 2000. The prior two conferences were held at Fremantle, Western Australia in 1996, and at Noordwijkerhout, The Netherlands in 1998. Nominally associated with the World Injury Conference movement, the Burden of Injury conferences have focused on research concerning methods of measuring injury outcomes and their application.
The driving force behind holding three international conferences on “burden of injury,” with another conference planned for 2002, is a need felt within the international injury science community to work towards some level of agreement on a number of issues concerning injury outcome measures. Many of the instruments being used to measure injury outcome were not developed with injury in mind, so there are concerns about their applicability. Also, since there are a number of reasons for measuring injury outcome in addition to determining current health status, for example, as a basis for cost/effectiveness analysis of interventions, there are concerns about how well the measures work in that arena.
The conference had as its theme, “Towards Consensus.” Conference participants received a pre-conference report incorporating the results of a literature survey of existing instruments and their application to injury and a copy of the glossary prepared by ECOSA. At the conference, invited speakers presented overviews of major topics or described particular instruments. Summaries of these presentations comprise the body of this report. The results of recent research were presented in posters. Poster abstracts are incorporated in an appendix.
This article is intended to summarize the main lessons learned. These fall into two primary categories; improved communication and insights into the applicability of individual instruments for measuring different aspects of the burden of injury.
Improving Communication
John Graham’s presentation on cost effectiveness/cost benefit provides a thought- provoking look back to the philosophical roots from which many of the differences in approach to measuring the burden of injury have evolved. An understanding of these differences should greatly improve communication.
Although the philosophical basis for most measures of injury burden is utilitarianism, usually characterized as “the greatest good for the greatest number,” how one operationalizes this varies greatly depending on which aspect of the utilitarian philosophy one espouses. “Liberalism” refers to both libertarians, who focus on rights of citizens that government may not violate, and egalitarians, who are concerned with rights of citizens that government is obliged to supply. Communitarians believe that public policy should evolve out of community-defined values (whether ethically or religiously defined). These philosophical bases do not have much interest in economic efficiency, where “economic efficiency” as applied to injury control means that if those citizens who benefit from an injury-control intervention had to bear its entire cost, they would consider it worthwhile. If those who benefit would not be willing to pay for its cost, the intervention is “inefficient.”
There are two competing strands of utilitarianism, subjective and objective. Subjective utilitarians believe that the principle of “consumer sovereignty” should govern value choices. This principal is generally construed as “beneficiaries would be willing to pay.” In applying this principal to injury control, the subjectivist believes that it is the personal perspective of the consumer or patient, assuming they have been fully informed of the issues, that should determine what is a “benefit” or “cost” and how much money should be expended to prevent or treat a particular trauma-induced impairment. Hence, strict subjective utilitarians believe that cost-benefit analysis is the appropriate tool for application to injury control interventions. The objectivist utilitarians, however, do not completely trust consumer or patient perceptions in the field of health and thus prefer cost-effectiveness analysis.
Another area where there has often been misunderstanding was discussed by Gale Whiteneck. He stressed the fact that practitioners and researchers have a very different approach to the problem of measuring the burden of injury. As a direct care-giver in a rehabilitation setting, Dr. Whiteneck’s concern is the patient’s progress towards full functioning. On the other hand, as he sees it, researchers in general are not focused on a particular patient but rather on how to measure the efficacy of a particular treatment, or the societal impact of a particular injury.
The contribution of these two presentations is that they sharply define the basis for differing approaches to measuring the burden of injury. As a result, communication should be enhanced from this point forward.
Health Status and Quality of Life Measures
An overview of health status and quality of life measures was presented by Ellen MacKenzie. The two well established approaches to quantifying health status and quality of life stem from two different traditions. The psychometric approach focuses on gathering data on the relationship among a number of variables and using statistical techniques to develop weightings for the different factors. The preference-based approach depends on determining how people value different states and develops weightings based on the application of utility theory. Four indices in the former category were discussed in detail at the conference; the Sickness Impact Profile (SIP), the Short Form 36 (SF-36), the Children’s Health Questionnaire (CHQ) and the Functional Independence Measure (FIM). Three indices based on preference weights were discussed; the Quality of Well Being (QWB), the Functional Capacity Index (FCI) and the Euroqol.
Charles Mock described the well established tool for developing a health profile, the Sickness Impact Profile (SIP). This instrument consists of 136 statements that the respondent answers with a yes or no. These statements reflect activities of daily living grouped into categories: physical dimension, psychosocial dimension, sleep and rest, eating, work, home management, and recreation. Responses are scaled into scores from 0-100, where scores of 0-3 generally represent little or no disability, 4-9 mild disability, 10-19 moderate disability, and > 20 severe disability. The SIP has been demonstrated to be reliable and internally consistent in a number of settings, including trauma, and it appears to be sensitive to less severe outcomes. In some cases other measures appeared to provide a more accurate measure of injury outcome. A major drawback to its use is the time to administer, which typically is between 20 and 30 minutes.
Currently, the most commonly used generic measure of health status is the Medical Outcome Study Short-Form 36 health survey (SF-36), a self-administered questionnaire consisting of 36 items that requires 5-10 minutes to complete. This health status measure was described by Branko Kopjar. The 36 items generate a profile of scores across eight dimensions: Physical Functioning, Social Functioning, Role Limitations (physical problems), Role Limitation (emotional problems), Mental Health, Vitality, Pain, General Health and health change. Studies have shown that the SF-36 has high internal consistency, good criterion validity, replicability and reliability. There have been several studies that applied SF-36 to patients with acute injuries.
One population for which the general health status measures are not applicable is children. Shanti Ameratunga described the CHQ, which is being developed to fill that void. Two instruments are available, one intended for completion by parents of children aged 5 years and older (98, 50 and 28 item versions available); and another for completion by children aged 10 years and older (87 item version with a shorter version under development). The raw scale scores of the CHQ are converted to a 0-100 scale, with a higher score indicative of more favorable health and well-being. This index appears to be a promising generic tool for measuring the burden of injury in children, however, no studies reporting its utility in this context have been published to date. There is also a concern that the CHQ may not provide information in the area of cognitive function (e.g., in the context of school and learning).
Another special population for which general health status measures may not be the most appropriate is the population undergoing rehabilitation. John Corrigan described the Functional Independence Measure (FIM), widely applied to this population. The FIM measures independent performance in self-care, sphincter control, transfers, locomotion, communication, and social cognition. FIM item ratings range from 1 to 7. A FIM item score of 7 is categorized as “complete independence,” while a score of 1 is “total assistance” (performs less than 25% of task). Experienced users can complete the scale in 5 to 10 minutes for subjects they know; a structured phone interview requires 15 to 20 minutes for experienced users evaluating subjects of whom they have no prior knowledge. The effectiveness of the FIM for monitoring short- and long-term outcomes of injury is more equivocal than its utility during intervention. This may be due simply to severity of the condition or to ceiling effects.
The presentation by Maria Segui-Gomez focused on the issues unique to preference based measures. These measures combine mortality-, morbidity-, and quality of life related issues in ways that health status measures based on the psychometric tradition cannot. There are numerous details that must be considered in developing a preference-based measure. At a minimum, the ideal health-state classification upon which the quality weights are derived should reflect: (a) the domains that are important to the problem under consideration, and (b) quality weights that are population-based, economic theory-based (particularly if we are dealing with policy decisions), interval-scaled, and measured or transformed onto an interval scale where the reference point death, has a score of 0.0 and the reference point optimal health has a score of 1.0. The preferences should, at minimum, incorporate the effects of morbidity on productivity and leisure, and this is particularly true if the measure is to be used in an economic evaluation exercise.
Troy Holbrook described one preference index that has been applied to traumatic injury, the Quality of Well Being scale, (QWB). This scale includes a 27 item symptom-scale and three scales of function: Mobility, Physical Activity, and Social Activity. Each symptom and step on these scales has its own associated preference weight. The overall score is calculated by adding (algebraically) the decrements to full function. The QWB is available in both interviewer-administered and self-administered formats for all ages. The instrument can also be administered by proxy for all age groups when necessary.
Preliminary results of an application of the Functional Capacity Index (FCI) were presented by Rod McClure. The FCI maps Abbreviated Injury Scale (AIS 1990) unique identifiers into total body scores that reflect expected levels of reduced functional capacity experienced by a person twelve months after sustaining the specified injury. This total body score is a number between 0 (no limitations) and 1 (death) and is comprised of weighted summed scores representing the injury’s effect on each of 10 attributes: Eating, Excretory Function, Sexual Function, Ambulation, Hand/Arm Function, Bending/ Lifting, Vision, Hearing, Speech, and Cognitive Function. The aim of the study was to test the validity of the FCI by comparing the predicted total body functional capacity with that empirically observed in a sample of people twelve months after their injury. Function was measured at twelve months for 619 patients seen in a Queensland trauma center. Mean time of questionnaire completion for those who were fully recovered was 2 minutes and for the residual disability group was 17 minutes. Preliminary results suggest that the Index has face validity and has the potential for widespread use as a measure of injury outcome.
The European Quality of Life Scale (EuroQol), a generic instrument for describing and valuing health-related quality of life was described by Sari Ponzer. The current version, the EQ-5D, includes measures of mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension is divided into three degrees of severity: no problem, some problems, or major problems. The EuroQol appears to have potential as an instrument for measuring outcome after injuries in populations with pronounced morbidity. It is simple and easy to use, however there appear to be ceiling effects and the lack of a cognitive dimension has been questioned.
Cost Measures
Although methods for measuring the economic costs resulting from injury are more highly developed than many of the health status and quality of life measures, a number of issues remain. Presentations covered methods for calculating costs of multiple injuries, a summary of the costs of injuries in the Netherlands, a review of current thinking on “willingness to pay” methods, and a look at some emerging issues in estimating societal costs.
Multiple injuries are the norm in many unintentional injury situations. Calculating the costs to society of these situations has been difficult. Delia Hendrie presented two models being developed for estimating the costs of multiple injuries, an additive/non additive model and a generalized linear model. Preliminary results using Australian data show that the generalized linear model explained thirty-six percent of the variation in the total cost of injuries. Although encouraging, these results indicate that further work is needed.
A model for calculating the costs of injury within the European Union was described by Saakje Mulder. The model incorporates the following cost elements: follow-up and aftercare by the treating physician; emergency transport; Emergency Department treatment; other outpatient care; daytime nursing; clinical nursing; clinical/therapeutic intervention; rehabilitation; nursing home care; outpatient physiotherapy; home care; and medication. Using data from the Netherlands, the model showed that almost fifty percent of the total costs are attributable to hospital and nursing home care, thirty percent to care in outpatient departments (half of which was for emergency assistance) and twenty percent to extramural and other care. The direct costs of injuries accounted for 3.4 percent of the national health care budget.
The current thinking about the application of “willingness to pay” in the UK was presented by David Ball. Dr. Ball argues that in addition to known difficulties in making estimates of the value of a statistical life, there may be a useful analogue from basic physics; that there is only a probability function rather than a single value for quantities such as value of statistical life. On this basis he recommends a range be used rather than a point value, and that it not be to four significant figures.
Ted Miller gave a presentation covering a variety of issues that should be considered when estimating the economic costs resulting from injury. Among the more salient points were the need to use life tables, employment/unemployment rates and earnings that more accurately reflect the demographics of the population being analyzed, and the need to incorporate appropriate costs for family, other caregiver and employer costs.
Data Issues
Two presentations were included in the agenda that covered data issues of particular interest to injury researchers. One covered the status of the International Collaborative Effort on Injury Statistics and the other covered the plans for further evolution of the Abbreviated Injury Scale.
There is widespread agreement that comparability of data is a critical aspect of learning from each other. Lois Fingerhut described the efforts underway and some of the products of that effort. One product is the International Classification of External Cause of Injury, (ICECI), which provides a much richer set of descriptors than the external cause codes within the International Classification of Diseases. Several approaches to the development of a minimum data set were also presented.
A topic of interest to those engaged in trauma related work, the future of the Abbreviated Injury Scale, was presented by Tom Gennerelli. The next step in the evolution of this index will focus on comparability and compatibility with other existing or proposed injury scoring systems including the Organ Injury Scaling (American Association for the Surgery of Trauma), the Orthopedic Trauma Association (OTA) Fracture Classification system and the ICD10- CM. In addition, the plan is to expand the AIS pre-dot code to more precisely include location (aspect, side), bilateral injuries and common injury combinations. Finally, there are plans to continue to refine the injury descriptors so that they are capable of describing threat to life as well as impairment and functional capacity.
Next Steps
Although not part of the formal program, there was considerable discussion among the participants that a meeting once every two years is not sufficient. The program committee is considering the formation of a working group, open to all who are interested. The working group would continue to develop issues such as the determination of the applicability of measures for the burden of injury and the consideration of broader adoption of the ECOSA glossary.