Risk Factors for Accidental Injuries Among the Elderly from the Survey on Ageing and Independence
Unintentional injuries have the potential to harm the well-being of the elderly. Results from the National Population Health Survey (NPHS) revealed that an estimated 10% of Canadians over the age of 55 experienced unintentional injuries that were serious enough to limit their normal daily activities. Additionally, for individuals over the age of 65 unintentional injuries have continued to remain one of the most frequent causes of hospitalization and death with mortality rates from unintentional injuries claiming 96 per 100,000 population.
Raina et al. report that the most frequent injuries among seniors are the result of falling, motor vehicle crashes, pedestrian-related injuries and self-inflicted injuries, each with there own myriad of risk factors. Unintentional injuries adversely affect indirect (i.e., health) and direct costs (health care costs, out-of-pocket expenses) to society and the outcomes for elderly are typically more severe than for any other age groups. However, with the exception of falling among seniors, research on unintentional injuries among seniors has been relatively sparse. This lack of evidence has had the effect of hindering reduction of injuries/prevention of injuries among seniors. Given that unintentional injuries do not all require medical care, self-report surveys provide a means of examining unintentional injuries among seniors that may not be captured by hospitalization and mortality data.
Survey on Ageing and Independence (SAI)
Statistics Canada conducted the SAI in September 1991 to examine factors that contribute to the independence and quality of life of Canadian seniors. Specifically, the SAI collected information on retirement status, type of employment, work characteristics and preparatory activities for retirement in the Canadian population. Details of socio-demographics, social support resources, health status, physical activities, activity limitations, and dwelling characteristics were also collected.
The SAI collected data on a representative sample of approximately 20,000 individuals between the ages of 45 to 102. For each household contacted, one person over the age of 45 was interviewed via the telephone; however, 10% of the sessions were conducted through face-to-face interviews in the participant's home. The sample excluded individuals living in remote areas of Canada, residents of institutions, individuals living on Indian Reserves and full-time members of the Canadian Armed Forces.
Purpose
This paper will report on the prevalence and risk factors for unintentional injuries for individuals over the age of 65 using SAI data.
Methods
The independent variables examined here were : (1) socio-demographic factors; (2) health practices; (3) social relationships; (4) frailty and (5) exposure to risk of injury. The dependent variable consisted of accident-related injuries outside of the home. The questions posed to participants were: "In the past twelve months, were you injured in an accident away from your home (not including automobile accidents), that altered your routine for at least one day?" and, further, "Thinking about the most recent accident, what injuries did you have ... (i.e., cuts? bruises? dislocations? fractures? sprains/strains? tenderness?)". The responses were subsequently coded into a binary variable: experienced no accident-related injury (0) or experienced an accident-related injury (1).
Data Analysis
Logistic regression models were used to estimate the probability of an injury. Only the independent variables found to be significant at the bivariate level were further analyzed in multivariate models. The final logistic regression model was used to estimate the adjusted odds ratios for the main and interactive effects for the variables investigated. All of the data used within this analysis have been weighted in order to correct for the SAI design effect; however, the same weights were recalibrated to reflect the actual sample size.
Results
Univariate Results
The majority of the 10,059 participants were between the age ranges of 65 to 69 (37%) and 70 to 74 (28%), were female (57%) and had an elementary or some secondary education (65%). Approximately 91% of the sample reported avoiding alcohol consumption, while 79% reported that they were non-smokers. Approximately 94% claimed to have regular rest and sleep patterns. In the overall sample, 60% were married, 92% of the sample received support from close family members and/or friends, while 8% did not receive support from family or friends. (Table 1) Approximately 67% of the overall sample felt that their activity was not limited by health. Similarly, 64% reported that they perceived themselves to be in good health.
With respect to activity level, the largest proportion of the subjects (40%) claimed to be as active as individuals of the same age. Approximately 69% reported that they provided assistance to others, while an estimated 12% of the seniors stated that their homes were in need of minor or major repairs. (Table 2) About 3% of the overall sample reported experiencing an injury. With respect to these injuries 2.4% (n=241) received treatment from a health care professional, 0.8% (n= 82) self-treated, and 0.1% (n=7) received treatment from a health care professional and self-treated the injury.
Multivariate Model
Gender, marital status, education, alcohol consumption, rest and sleep patterns, activity limited by health, providing assistance to others, activity levels compared to other of the same age, and home maintenance status were the significant independent variables in the final logistic regression model. An interaction between gender and home maintenance was also present.
Individuals that were not married had a 1.66 times greater risk of injury. Further, individuals that avoided alcohol (O.R.=1.72) and assisted others (O.R.=1.64) were also at an increased risk of injury, as compared to those that consumed alcohol and did not assist other individuals respectively. Having a high school diploma or some post-secondary education appeared to have a protective effect against injuries external to the home, while a college diploma or university education was not significant.
Having regular rest and sleep patterns, and not having an activity limitation were also found to have a protective effect against experiencing an accident-related injury external to the home.
The interaction between gender and home maintenance status showed that females were at a greater risk of injury, as compared to men. Risk increased for females as home maintenance progressed from no repairs (O.R.= 1.19), to minor repairs (O.R. = 1.84), to major repairs (O.R. = 3.42). Men that had major repairs appeared to be protected against injury risk, as compared to men who had homes without repairs to be completed. In fact, risk of injury was lowest for men with homes in need of major repairs (O.R. = 0.20).
Discussion and Concluding Remarks
National data dealing with risk factors for injuries among the elderly is relatively uncommon. The present study revealed that 3.3% of Canadian seniors experienced accidents away from their home that altered their routine for at least one day, while the NPHS found that 10% of older adults encountered accidental injuries. However, the NPHS examined all types of injuries and injury locations, whereas this SAI analysis did not include motor vehicle accidents, and only examined accidents away from the homes, and not in the homes of the participants.
The most striking trend with these data is that risk factors identified for accidental injuries are consistent with previously identified risk factors for falls. For instance, being female and poor health, as measured by inadequate rest, smoking, and activity limitations for health reasons have been implicated as risk factors for falls in previous research. It would seem plausible that the majority of injuries reported in the SAI are attributed to falling given that falls account for most injuries that occur in individuals over the age of 65, and the injuries sustained (i.e., cuts, bruises, dislocations, fractures, sprains/strains, tenderness) are consistent with injuries related to falling. However, the questions posed by the SAI did not allow for differentiation of the cause of injury, and as such, it was impossible to determine whether the injuries were definitely the result of a fall.
Individuals that assisted others were more likely to experience an injury. The possible explanation for this finding is twofold: individuals that assist others increase their opportunities to injure themselves (e.g., repairing others homes, lifting physicallydependent individuals), or individuals that did not assist others are unable to do so and thus do not have the same level of exposure to situations in which they could injure themselves. Given that this finding is relatively unsupported in the literature, further examination of its contribution to injury risk is warranted.
An interaction term between gender and home maintenance was significant within the final model. Specifically, females tended to have an increased risk of injury as repair status increased, particularly for the major repairs category. However, being male and having a home in need of repair, appeared to have a protective effect against injuries. It is possible that for injuries outside of the home (outcome variable in question) other factors related to home repair may actually place seniors (particularly women in the major repairs category) at risk. For example, differential mortality may partially account for the differences between males and females. Specifically, frail or unhealthy males may have been "selected out" (died), thus leaving a greater proportion of healthy males within the population, that were not susceptible to experiencing a fall (see Hirdes & Forbes (1993) for an additional discussion on differential mortality). An alternative possibility for the gender differences is the use of imprecise measures, and the exclusion of some other factor(s) that lead to incomplete coverage of all mechanisms linking gender and fall-related injuries. For instance, do women have certain medical conditions or use particular medications that make them more susceptible to falling? and have all diseases and/or medications been measured? or are the numbers too small to be statistically significant within the analyses? The lack of information pertaining to disease diagnoses and medications does not permit this type of analysis to be completed with the SAI data. Future work is warranted in this relatively understudied area in order to better understand the relationship between gender and home maintenance status, with the inclusion of broader questions pertaining to health.
What these findings contribute to the literature is a move beyond univariate analysis of data as in Raina et al. (1999), and provision of more detailed and predictive information pertaining to risk factors for accidental injuries. Identification of such risk factors provides health care professionals with insight into factors that, if modified, have the potential to decrease accidental injuries and improve/maintain quality of life. It is important to note that the cross-sectional nature of the SAI does not allow for the event to be temporally related to the individual characteristics of the situation, and thus cause-and-effect relationships cannot be drawn.
The SAI also did not differentiate between specific causes of the injuries (i.e., fall, motor vehicle accident), but rather the severity of the injury incurred. Further, given that the data were collected in 1991 the prevalence estimates within the SAI may be different today; regardless, the risk factor results would remain relatively stable. Despite these limitations, the SAI collected information (i.e., information pertaining to home repairs) that has not been collected in other national surveys, and as such, provides insight into risk factors that place individuals at risk, and thus require further investigation. Therefore, conducting comprehensive national longitudinal studies that include broad multidisciplinary frameworks addressed in the NPHS or the SAI are needed.