Residential status and risk of hip fracture

Age and Ageing 1999; 28: 135–139

䉷 1999, British Geriatrics Society

Residential status and risk of hip fracture ROBYN NORTON, A. JOHN C AMPBELL1, IAN R. REID2, MEG BUTLER, ROCHELLE CURRIE, ELIZABETH ROBINSON3, HARLEY GRAY2,4 Injury Prevention Research Centre, Department of Community Health, University of Auckland, Private Bag 92019, Auckland, New Zealand 1 Faculty of Medicine, University of Otago Medical School, PO Box 913, Dunedin, New Zealand 2 Departments of Medicine and 3Community Health, University of Auckland, Auckland, New Zealand 4 Surgical Services, Middlemore Hospital, Auckland, New Zealand Correspondence should be addressed to: R. Norton. Fax: (+61) 2 992 66830. Email: [email protected]

Abstract Objective: to examine the association between residential status and risk of hip fracture in older people. Design: population-based case–control study. Setting: Auckland, New Zealand. Subjects: a random sample of all individuals ⱖ 60 years, hospitalized with a fracture of the proximal femur between July 1991 and February 1994. Controls were age and gender frequency-matched to the cases, randomly selected from a random sample of general practitioners. Main outcome measures: radiographically-confirmed fracture of the proximal femur. Fractures sustained as a result of major trauma, such as in a motor vehicle crash, and those associated with pre-existing pathological conditions were excluded. Results: individuals living in institutions were almost four times more likely to sustain a hip fracture [age- and gender-adjusted odds ratio (OR) = 3.8; 95% confidence interval (CI): 3.0–4.8] than those living in private homes. After adjustment for potential confounding factors, the risk of hip fracture associated with living in an institution remained significantly increased (P < 0.0001), although the magnitude of the risk was somewhat diminished (OR = 2.2; 95% CI: 1.5–3.5). Conclusions: living in an institution is associated with an increased risk of hip fracture in older people. Specific factors that place these individuals at increased risk need to be identified, in order to develop intervention strategies. Keywords: hip fracture, institutions, residential status, risk factors

Introduction Hip fractures are an important cause of morbidity and disability in older people [1], and their incidence is expected to increase, particularly as the size of the older population increases [2]. A recent review of interventions for the prevention of falls and fall-related injuries identified some strategies that are effective in preventing falls and/or minimizing their impact once they occur [3]. Consequently, preventative efforts based on such strategies may counter increases in the anticipated incidence of hip fractures world-wide. However, these efforts might be further enhanced if high-risk groups and the factors associated with increased risk in these groups were identified.

Those living in institutional care are at increased risk of hip fracture compared with those living in private homes [4–6]. The increased risks in such individuals may be due to factors independently associated with both living in an institution and hip fracture [5]. However, factors associated with institutional care may independently increase the risks of hip fracture. To date, few case–control or cohort studies have included both individuals living in institutions and those living in private homes, and thus they have been unable to examine the risks of hip fracture by residential status, while controlling for potential confounding factors. We report here the results of a case–control study of hip fracture in Auckland, New Zealand which examined the association between residential status and hip fracture.

135

R. Norton et al.

Subjects and methods Subjects Auckland has a population of about one million people, of whom 15% are aged 60 years or older. Between July 1991 and February 1994, all older individuals (ⱖ60 years) hospitalized with a radiographically-confirmed fracture of the proximal femur were identified prospectively through the ward registers of Auckland and Middlemore hospitals. Almost all individuals with a fracture of the proximal femur in the Auckland region are admitted directly to one of these public hospitals [7]. Demographic information was sought on each person with a hip fracture. Information was not obtained for those who: (i) normally lived outside Auckland, (ii) had sustained a fracture as a result of major trauma, such as in a motor vehicle crash or (iii) were known on admission to have a fracture that was the result of some pre-existing pathological condition, such as cancer. A random sample of patients for whom demographic information was obtained was then systematically selected from ward registers, using a pre-arranged sampling procedure, for inclusion in the Auckland Hip Fracture Study. These individuals were contacted post-operatively, before discharge and invited to participate in the study. Next-of-kin or care-givers (proxy respondents) were invited to participate in the study if potential respondents were identified as being cognitively impaired, i.e. if next-of-kin or care-givers indicated that the patient was incapable of understanding and completing the questionnaire or if the individual incorrectly answered at least three out of 10 items on the mental status questionnaire [8]. Proxy respondents were also sought for those patients who died before interview, were too ill or for some other reason were unable to participate personally. Controls, age and gender frequency-matched to the cases, were randomly selected from the practices of a random sample of all general practitioners, from the same geographic region from which the cases were identified. An age- and gender-profile of patients aged 60 years or older was developed from information obtained from those general practitioners who agreed to participate. Controls were then randomly selected (using computer-generated random numbers) from these practices, with a sampling probability in proportion to the number of eligible patients in each practice. The general practitioners then initially contacted these patients and, if they agreed, the patients were invited to participate in the study by the Auckland Hip Fracture Study staff. Proxy respondents were invited to participate if required. Data collection After providing signed agreement to participate, cases

136

and controls completed an interviewer-administered questionnaire and had body measurements taken. Information was sought on potential risk factors for hip fracture, including residential status—defined in terms of living in a private homes versus living in an institution (rest home, private hospital or retirement village). Those living independently in retirement villages were also categorized as living in private homes. Statistical analysis Analyses were conducted using logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) [9]. Age- and gender-adjusted ORs were calculated separately for residential status and for potential confounding factors. The latter included variables that are associated both with the risk of living in an institution and with the risk of hip fracture: cognitive impairment, marital status, number of previous fractures since the age of 40 years, number of falls in the last year, pre-existing serious illness (Parkinson’s disease, stroke, heart failure, or epilepsy), level of functional impairment (defined using Katz activities of daily living scale [10], with a score of A representing independence and increasing levels of dependence represented by subsequent letters of the alphabet), weight and number of months in current residence. Multivariate ORs associated with living in an institution were then calculated, adjusting for the potential confounding factors. Investigation of the effects of adding other risk factors sequentially was undertaken. Variables were added in the order suggested by a best subsets selection [9], while ensuring that dummy categorical variables were not split. Ethical issues This study was undertaken with the approval of the Northern Regional Health Authority ethics committee.

Results Over two and a half years, 1832 hip fractures were sustained by 1774 older people in the Auckland region, equivalent to an annual incidence of 459.4 individuals per 100 000 population aged 60 years or older [7]. Information was sought on 936 (51.1%) of the 1832 hip fracture patients and agreement to participate in the study was obtained for 911 (97.3%). Of the 118 general practitioners invited to participate in the study, 84 (71.2%) agreed to do so. A total of 910 (81.4%) of the 1118 individuals who were selected as controls agreed to participate. The gender- and age-profile of cases and controls was similar (Table 1), although the control population comprised a somewhat higher proportion of women

Residential status and hip fracture Table 1. Gender and age distribution of cases and controls. Values are numbers (percentages) Cases (n = 911)

Controls (n = 910)

.......................................................................................................

Gender Female Male

699 (76.7) 212 (23.3)

724 (79.6) 186 (20.4)

Age group (years) 60–69 70–79 80–89 90þ

55 (6.0) 254 (27.9) 453 (49.7) 149 (16.4)

85 (9.3) 296 (32.5) 421 (46.3) 108 (11.9)

and tended to be slightly younger (mean age = 80.5 years) than the case population (mean age = 82 years). In all later analyses, age and gender adjustments were undertaken. Individuals living in institutions were more likely to have sustained a hip fracture than those living in private homes (age- and gender-adjusted OR = 3.8; 95% CI: 3.0–4.8; Table 2). The risk of hip fracture was

also increased in individuals who were cognitively impaired, were not currently or had never married, had previous fractures, had fallen twice or more in the previous 12 months, had a history of serious or chronic illness, required assistance in activities of daily living, had low body weights and had been in their current residence for less than 6 months (all P < 0.001). After adjustment for each of these potential confounding factors, the risk of hip fracture associated with living in an institution was still significantly increased (OR = 2.2; 95% CI: 1.5–3.5; P < 0.0001), although the magnitude of the risk was smaller (Table 3).

Discussion We found that living in an institution was associated with an increased risk of hip fracture in older people. In part, this increased risk was explained by factors that both increase an individual’s risk of living in an institution and their risk of sustaining a hip fracture. However, a twofold increased risk of hip fracture persisted after controlling for these potentially

Table 2. Residential status and other potential confounding factors in individuals with hip fractures and in controls No. (and %) of subjects ..................................................................................................................................................

Cases

Controls

Age- and gender-adjusted OR (95% CI)

P-value

..................................................................................................................................................................................................

Residential status Private home Institution

565 (62.0) 346 (38.0)

782 (85.9) 128 (14.1)

1 3.8 (3.0–4.8)

–