Non-fatal injuries among Chinese aged 65 years and older: findings from the Fourth National Health Services Survey
- 1Department of Epidemiology and Health Statistics, School of Public Health, Central South University, Changsha, Hunan, China
- 2Center for Health Statistics and Information, the Ministry of Health of China
- Correspondence to Dr Susan P. Baker, Center for Injury Research and Policy, Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, 624 North Broadway, Baltimore, MD 21205, USA;
Contributors GH and SPB participated in the design and data analysis and have seen and approved the final version.
- Accepted 8 March 2010
- Published Online First 30 June 2010
Objective To understand the epidemiology of non-fatal injuries among adults aged 65 years and older in China.
Design Cross-sectional survey (the Fourth National Health Services Survey of China).
Participants Urban and rural residents aged 65 years and older from 56 400 households in China.
Main Outcome Measures The incidence rate was calculated as the number of persons injured in the previous 12 months divided by the population×1000.
Results The incidence rate of non-fatal injuries among elderly individuals in the previous 12 months was 37.5 per 1000 population. Home, street, working environment, and public buildings were the most common places of occurrence, accounting for more than 90% of injuries. Falls were the leading cause of non-fatal injuries. After adjusting for other factors, Han people were 39% more likely to be injured than non-Han people, and the divorced and the widowed were found to have, respectively, 4.6 and 2.2 times the risk of injury compared with single persons, p<0.05. Education, per capita household income and urbanisation did not significantly affect the injury risk when confounding factors were controlled for.
Conclusion Almost 4% of adults aged 65 years and over sustain injuries each year in China. Falls should be a priority of injury prevention for elderly people, efficient home injury prevention programmes need to be developed, and the divorced and widowed should be targeted as groups at high risk of injury.
According to the Chinese Injury Prevention Report by the Ministry of Health of China, approximately 200 million injuries occur annually, resulting in 700 thousand deaths, one million disabilities, 14 million hospitalisations and 60 million emergency department visits in China.1 Zhou et al2 report that injuries account for an annual loss of 12.6 million potentially productive years of life, a loss greater than for any disease group; the estimated annual economic cost of injury is almost four times the total public health services budget of China. Despite its importance, injury prevention has been little addressed and is not yet regarded as a priority by the government.3–5
Older adults have the highest mortality rate from injuries.6 In 2007, adults aged 65 years and over accounted for 9% of the total population.7 However, we know little about the epidemiological characteristics of injuries among elderly adults. The few existing injury studies report incidence rates of the entire population, ranging from 59 to 130 per 1000 population based on local samples.8–11 None of the published studies based on local samples reported the epidemiological characteristics of non-fatal injuries among elderly people. Moreover, none of the existing studies focus on the influence of socioeconomic status (SES) on injuries, although increasing attention has been given to the effects of SES as a determinant of health.12
The National Health Services Survey of China was initiated in 1993 by the Ministry of Health of China and is now conducted every 5 years, to help the central government understand the need for and supply of health services. Details on injuries were excluded from the first three National Health Services Surveys (1993, 1998 and 2003) and were first included in the fourth survey in 2008.
The objective of the present study was to provide the epidemiological details of non-fatal injuries among Chinese aged 65 years and older, using the Fourth National Health Services Survey, so as to offer a sound basis for governmental decisions related to injury control in elderly people and provide new evidence to understand the impact of SES on non-fatal injuries.
The data came from the Fourth National Health Services Survey, which was organised by the Ministry of Health of China and was completed through face-to-face interviews between June 15 and July 10 2008. A three-stage stratified random sampling procedure was used to select households for interview. The entire nation, except for Hong Kong and Macau, was clustered by the government administrative system (ie, city, county, town and village). A total of 94 cities or counties, 470 towns and 940 villages was randomly selected. The targeted cities were further divided into neighbourhood communities (the smallest administrative unit in urban areas). Sixty households were selected from each community or village at random, yielding 56 400 households. Each household had the same probability of being selected at random. All members of selected households were invited to participate in the survey.13 Analyses of sample representativeness show good agreement between sample counties (90 selected counties) and the whole country in economic status, age composition, birth rate, death rate, education level and unemployment rate.13 The percentage of persons aged 65 years and over was 4.9% in both the sample counties and the whole country.
The interview was conducted by trained medical doctors. Before the survey, interviewers received specific training and passed the examination for field interview. During the survey, interviewers visited each household up to three times. Interviewers explained the purposes and confidentiality of the survey, and then invited family members to participate. Respondents could refuse to participate in the survey. Their participation in the survey was regarded as giving oral consent.13
All adults were asked to answer the questions for themselves. For adults who were not at home at interview time, other adults in their family were asked to answer the questions on their behalf. A check of each questionnaire was undertaken by a district survey manager at the end of each interview day. A second interview was required for those interviewees whose questionnaires included important missing values, contradictions between relevant questions, or unrealistic answers for certain questions.13
The survey defined injury as ‘received medical treatment or caused activity restriction for 1 day or more in the previous 12 months'. The Fourth National Health Services Survey questionnaire included five questions related to injuries: (1) Were you injured in the past 12 months? (2) How many times were you injured in the past 12 months? (Then, anyone who had been injured was asked questions 3–5). (3) What was the cause of the most recent injury? (4) Where did it occur? and (5) How severe was it?
The 13 most common causes of non-fatal injuries were listed as the options for question 3: transportation, fall, falling object, cutting/piercing, explosion, animal bite, drowning, suffocation, electric current, fire/burn, unintentional poisoning, assault, self-harm and other. The options for answering question 4 were: home, road, workplace, school, public building and other place. The severity of injury was classified into four categories, consisting of disability, 10 or more days' hospitalisation, 2–9 days' hospitalisation and 1 day's hospitalisation or received medical treatment.
Seven available sociodemographic variables in the survey questionnaire were chosen to portray the epidemiological characteristics of non-fatal injuries among adults aged 65 years and older: (1) sex; (2) race; (3) age; (4) marital status; (5) education; (6) per capita household income and (7) urbanisation.
Because Han people account for approximately 90% of the total population of China, we combined all the minorities into a single category, the non-Han. Persons aged 65 years and older were divided into three age groups, 65–74, 75–84 and 85 years and over. The marital variable contained five categories: unmarried, married, divorced, widowed and other. Education was divided into primary education and lower (≤5 years), secondary education (6–11 years) and college and over (≥12 years). The 56 400 households included in the survey were divided equally into four classes based on per capita household income: lowest (<2500 Chinese yuan), low (2500–3999 Chinese yuan), high (4000–7199 Chinese yuan) and highest (≥7200 Chinese yuan).
The survey adopted a seven-class classification of urbanisation based on cluster analysis of three socioeconomic indicators, one education indicator, four population indicators and two health indicators.13 The seven-class classification includes large cities (the most urbanised), middle cities, small cities, first-class rural areas, second-class rural areas, third-class rural areas and fourth-class rural areas (the least urbanised). The first three categories constitute urban areas, and the rest comprise rural areas.13
The incidence rate was calculated as the number of persons injured in the previous 12 months divided by the population×1000. Because the number of injuries that a person experiences in a time unit is believed to meet the Poisson distribution, Poisson regression is recommended to identify factors that influence injury occurrence.14 When observations fall into groups or clusters, such as observations on individuals nested in families, hospitals, or firms, it may be unreasonable to assume that observations in the same cluster are independent, and multilevel models are suggested in this case.15 Because the persons from the same community/village might cluster due to the common living environment, we fitted a two-level Poisson regression to examine the significance of sociodemographic variables. The incidence rate ratio (IRR) was used to quantify the impact of influencing factors on the occurrence of injuries; p<0.05 was selected as the statistically significant level.
The Fourth National Health Services Survey included 19 505 adults aged 65 years and older (table 1). Men accounted for 48% of the sample. Han people comprised the majority of study subjects (89%). The age groups of 65–74, 75–84 and 85 years and over constituted 65%, 30% and 5% of our sample, respectively, and 76% of adults in the sample received only primary education (5 years and less). The proportion of urban interviewees was 37%.
The incidence rate of injuries among elderly people in the previous 12 months was 37.5/1000 population, with a 95% CI of 34.9 to 40.2 (table 1). The incidence rate among elderly men (35.2/1000) was lower than among elderly women (39.6/1000); the unadjusted male:female IRR was 0.86 (p<0.05). Compared with the minorities, Han people had a higher incidence rate (38.5/1000 vs 30.4/1000 for the non-Han) as will be discussed later, with an unadjusted IRR of 1.47 (p<0.05). Adults aged 85 years and over and 75–84 years had a higher risk of injuries than adults aged 65–74 years; their unadjusted IRR were 1.2 and 1.4, respectively (p<0.05). The unadjusted risks of injury among the divorced and widowed were 4.6 and 2.4, respectively, times the risk in single adults (p<0.05).
Of those who were injured, urban women were twice as likely as urban men to have been injured three or more times in the previous year (12% vs 6%) (table 2). More than 70% of injuries occurred at home or in the street. Public buildings were the third most common place of injury for rural residents, accounting for more than 10% of locations. A higher percentage of all injuries resulted in disability in urban residents than in rural residents (8% vs 4%), whereas the percentage hospitalised for 10 or more days in urban areas was lower than in rural areas (11% vs 19%).
Falls were the most common injury for both urban and rural elderly people, with an incidence rate higher than 20/1000, which explained more than two-thirds of all injuries (table 3). For urban residents, transportation, animal bites, falling objects, cut/pierced and poisoning were also common injuries; the rates for these causes in urban men were higher than for urban women, although the differences were not significant. Among rural residents, transportation and animal bites were the second and third most common injuries.
After adjusting for other variables, race and marital status were significant (p<0.05) (table 4). Han people were 39% more likely to sustain an injury than the non-Han. The divorced and the widowed had 4.6 and 2.2 times the risk of injury, respectively, as elderly people who were single. Education, per capita household income and urbanisation did not significantly affect the injury risk when confounding factors were controlled for.
The Fourth National Health Services Survey of China in 2008 has, for the first time, estimated the number of non-fatal injuries in the previous 12 months. Based on a nationwide representative sample, approximately 6.6 million injuries occurred in 4.6 million persons, a rate of 37.5 persons per 1000 population. This result is far lower than the estimation from the Chinese injury prevention report, of approximately 19 million injuries occurrences annually.1 The 12-month period for recalling the injuries may partly explain the lower incidence rate in this survey, because a long period of recall has been found to reduce the estimates for all injuries and especially for less severe injuries.16 Another possibility is that the definition of injury in the survey was more restrictive, defining injury as resulting in medical treatment or activity restriction for 1 day or more in the previous 12 months. Other studies defined an injury as meeting one of three criteria: (1) diagnosed by medical institutions; (2) needed treatment or medical care; and (3) resulted in missed work/school for at least half a day).7–10
The strength of our analysis is that our findings offer injury details not available elsewhere. The identification of the most common locations and causes of recent injury is helpful for the development of interventions and priorities. Although psychological symptoms,17 18 lack of related knowledge19 and living alone19 have been linked to an increased risk of injury among elderly people in China, these are insufficient for identifying the most common contributing factors and the most promising preventive measures. In addition to the recent multifaceted interventions in community settings that were reported to be useful in preventing falls among older persons in China,20 research is needed to identify the protective and risk factors for falls.
Other investigators have found that low SES was associated with a higher risk of non-fatal injuries among working-age adults.21 22 In contrast, a study of adults of all ages found no association between SES (household income, education and urbanisation) and non-fatal injuries.23 Similarly, we found no association for elderly people. Weaker socioeconomic inequalities among elderly people were interpreted by Kelly and Miles-Doan22 as one possible explanation for the different pattern in elderly people. Because of insufficient sample size, we did not examine the association between SES and any specific cause of injury.
The divorced and the widowed were found to be at substantially higher risk of injury than older adults who are single. The absence of a care giver might be an important risk factor, especially for those who cannot care for themselves. The results indicate that divorced and widowed elderly people should be targeted as high-risk groups for injury, and prevention programmes need to be developed as soon as possible in China. Surprisingly, non-Han individuals were found to have a lower risk of injury than Han people. Because no ethnicity-related studies have been published so far, we ascribe the ethnic difference to two possibilities: (1) sample representativeness and (2) unknown protective factors in the various minorities. These two assumptions are worthy of examination in future studies.
Our findings are mainly limited by the quality of the data. Because it was the first time for China to include injury questions in its national health services survey, the designers were not experienced in injury surveys. Several defects were found to exist in the questionnaire of the Fourth National Health Services Survey, such as the 12-month period for recalling the injuries. A second limitation is that the study did not include information on fatal injury. The reasons for the exclusion of fatal injuries include: fatal injury data can be obtained through the Chinese Vital Registration data; and the National Health Service Survey cannot provide stable estimates of injury mortality rates; national injury mortality rates in 2008 were urban 31/100 000 and rural 53/100 000,24 and the total sample for the Fourth National Health Services Survey included only approximately 200 000 residents.
Further studies are needed to identify key factors contributing to the most common injuries and to develop efficient prevention programmes for elderly people.
What is already known on this topic
Injuries are a serious threat to the health of adults aged 65 years and older.
The epidemiological characteristics of non-fatal injuries among elderly people in China have not been reported.
None of the existing studies have examined the influence of SES on non-fatal injuries among elderly people in China.
What this study adds
The incidence rate of non-fatal injuries among elderly people in the previoius 12 months was 37.5 per 1000 population
Home, street, working environment and public buildings were the most common places of occurrence.
Falls were the leading cause of non-fatal injuries for elderly people.
Divorced and widowed elderly people had, respectively, 4.6 and 2.2 times the risk of injury compared with single persons.
Education, per capita household income and urbanisation did not significantly affect the injury risk when confounding factors were controlled for.
Funding This study was supported by the Center for Health Statistics and Information of Ministry of Health of China, the China Medical Board of New York (CMB) and the Center for Injury Research and Prevention, Centers for Disease Control and Prevention (grant CCR302486).
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.