The American Journal of Managed Care March 2015
Trends in Mortality Following Hip Fracture in Older Women
Objectives: To examine contemporary trends in mortality fol-lowing hip fracture among older postmenopausal women in an integrated healthcare delivery system.
Study Design: Retrospective cohort study of 13,550 women aged ≥65 years with hip fracture during 2000 to 2010. Methods: Demographic factors, comorbidity index score, fracture history, early rehospitalization, and all-cause mortality within 1 year following hip fracture were examined using health plan data-bases and records. Temporal trends, risk factors, and the associa-tion of race/ethnicity and mortality within 1 year post fracture were examined using multivariable logistic regression.
Results: Among 13,550 women with hip fracture, 84.6% were aged ≥75 years: 83.6% were white, 2.8% black, 5.6% Hispanic, 4.5% Asian, and 3.5% of other/unknown race. Following hip frac-ture, 2.4% died during the index hospitalization, while 12.3% were rehospitalized within 30 days of discharge. Infection, pneumonia, and cardiovascular conditions were the most common nonor-thopedic indications for readmission. Mortality rates at 6 months (17%) and 1 year (22.8%) following hip fracture were high and in-creased with age. Greater comorbidity and early rehospitalization were associated with increased mortality risk, while Asian and Hispanic race/ethnicity were associated with lower mortality risk (vs white). Temporal trends demonstrated a small but significant reduction in mortality risk during 2004 to 2010. Conclusions: While hip fracture morbidity and mortality remain high, temporal trends suggest recent declines in mortality risk, with risk of death following hip fracture lower for Asian and His-panic women. Future studies should examine potential benefits of targeted interventions within integrated healthcare settings and factors contributing to observed racial/ethnic differences in post fracture survival.
Am J Manag Care. 2015;21(3):e206-e214
While hip fracture mortality remains high, significant declines in mortality risk and important racial/ethnic differences were seen following hip fracture in older women receiving care within a large integrated healthcare delivery system.
- Older age, increased comorbidity, and early rehospitalization were associated with an increased risk of death within 1 year following hip fracture.
- Racial/ethnic differences in post fracture survival were also seen; compared with women of white race, Asian and Hispanic women had lower mortality risk during the first year.
- Temporal trends demonstrated small but significant mortality risk reduction, where the potential benefits of targeted healthcare interventions should be further examined.
Among US women 65 years and older, the annual incidence of hip fractures is estimated at 793 per 100,000 women, based on Medicare claims data from 2005.1 While hip fractures constitute only 14% of all osteoporotic fractures,2 they result in substantial morbidity and mortality,1-6 with short-term mortality risk 2- to 8-fold higher than that of age-matched controls.4,7,8 One-year mortality rates are in the range of 22%,1 with excess mortality risk highest during the first 3 to 6 months following hip fracture.4,7 Given the projected increase of the aging US population over the next 2 decades9—particularly among those over 65 years old and those of nonwhite race/ethnicity,9,10 in whom less is known regarding contemporary fracture outcomes—hip fractures remain a significant public health concern.
Numerous risk factors contribute to hip fracture mortality in women, including age, fracture type, functional status, comorbidities, post fracture care, and rehabilitation.1,7,11-13 Examination of 20-year trends since 1985 demonstrated national declines in 1-year mortality following hip fracture, largely due to mortality reduction prior to 1998.1 These findings may be potentially attributable in part to advances in surgical and post operative medical care.1 Survival rates following hip fracture also vary by race/ethnicity, with historical mortality rates slightly higher for blacks compared with whites.14 Few studies have examined differences in mortality following hip fracture among US women of other race/ ethnicities.
After experiencing hip fracture, only a subset of women will regain their pre-fracture quality of life, while many suffer permanent deficits in activities of daily living6,15; it is estimated that up to 17% of remaining post fracture life may be spent in a nursing facility.6 Hospital readmission rates are high following hip fracture and have been mostly attributable to nonorthopedic/nonsurgical conditions.3,16 In one study, infections (21%), followed by cardiac conditions (12%), were the leading causes of readmission within the first 6 months, while surgical complications accounted for only 11% of readmissions.16 Rehospitalization following hip fracture has also been shown to be an independent predictor of adverse morbidity and mortality outcome.3,5,16
In the contemporary era, limited data pertaining to hip fracture morbidity and mortality exist in large community-based practice settings serving demographically diverse populations. This study examines early rehospitalization and mortality outcome following hip fracture during 2000 to 2010 within a northern California integrated healthcare delivery system, with a specific focus on racial/ethnic differences and temporal trends in mortality risk.
Kaiser Permanente Northern California (KPNC) is a large, integrated healthcare delivery system providing care to over 3 million members. There are more than 150 medical offices and 20 hospitals providing care, with centralized databases of all network and nonnetwork hospitalization discharge diagnoses, ambulatory visit diagnoses, and pharmacy records. The Kaiser Foundation Research Institute’s Institutional Review Board approved the study and a waiver of informed consent was obtained due to the nature of the study.
Using health plan databases, we identified all female health plan members 65 years or older with a principal hospital discharge diagnosis of a proximal femur fracture between January 1, 2000, and December 31, 2010. A proximal femur (hip) fracture was defined by a fracture of the femoral neck (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 820.00-820.03, 820.09, 820.8) or pertrochanteric region (ICD-9-CM 820.20, 820.21), excluding open frac-tures and those associated with major trauma (ICD-9-CM E800-E848). The final cohort was established by ascertaining the first qualifying hip fracture per woman. Women without continuous health plan membership in the year prior to hip fracture were excluded from these analyses in order to define a cohort with ongoing membership and to allow for ascertainment of baseline co-variates. Temporal trends and race/ethnic differences in femur fracture among KPNC women have been separately reported.17
Age and self-reported race/ethnicity (classified as non-Hispanic white, black, Hispanic, Asian, or other/unknown) were obtained from administrative databases. A clinical comorbidity index score was calculated based on diagnosis and procedure codes obtained from all hospitalization, emergency, and ambulatory visits in the prior year (including the index fracture admission) using a modification of the Charlson Comorbidity Index established by Deyo and colleagues.18 Fracture history (fractures occurring after age 40 and prior to hip fracture) was obtained by identifying prior outpatient and hospitalization diagnoses of fractures involving the spine, trunk, upper and lower extremities (ICD-9-CM 805, 807-815, 817-825, 827-829), excluding open fractures, fractures associated with spinal cord injury, fractures of the head/fingers/toes, and hospi-talized fractures associated with high-energy trauma (ICD-9-CM E800-E848).
Early rehospitalization was defined as rehospitalization within 1 to 30 days following discharge from the initial hip fracture hospitalization (same- or next-day readmissions or transfers with a principal diagnosis of hip fracture were considered part of the initial hospitalization, affecting only 0.8% of the cohort). The principal diagnosis assigned to the rehospitalization event was categorized as follows: 1) cardiovascular condition: hypertension, ischemic heart disease, cardiomyopathy, heart failure, cardiogenic shock, conduction disorder, dysrhythmia, or tachycardia (ICD-9-CM 401-405, 410-414, 425-428, 785.0, 785.51); 2) venous embolism/ thrombosis, pulmonary embolism (ICD-9-CM 415.1, 453); 3) cerebrovascular disease: hemorrhage, occlusion, ischemia, late effects (ICD-9-CM 430-438); 4) pneumonia, pneumonitis, influenza, respiratory insufficiency or failure (ICD-9-CM 480-488, 507, 518.81-84); 5) chronic obstructive pulmonary disease/asthma/bronchitis (ICD-9-CM 466.0, 490-493, 496, 519.11); 6) urinary tract infection (ICD-9-CM 599.0); 7) cognitive impairment or altered mental status (ICD-9-CM 290-298, 331, 348.3, 780.0-1, 780.97); 8) acute or chronic kidney disease/failure (ICD-9-CM 584-586); 9) gastritis, gastrointestinal ulcers/hemorrhage (ICD-9-CM 531-535, 526.12-13, 578.9); 10) glucose-, electrolyte-, or fluid-related disorder (ICD-9-CM 249-251, 275.2-4, 276); 11) Clostridium difficile in-fection (ICD-9-CM 8.45); 12) skin infection, ulcer, gangrene, or wound infection (ICD-9-CM 681-682, 686, 707, 785.4, 998.5); 13) bacteremia/septicemia, sepsis, septic shock (ICD-9-CM 038, 790.7, 785.52, 995.91-92); 14) malignant neoplasm (ICD-9-CM 140-208); 15) nonpelvic, nonhip/femur fracture (ICD-9-CM 800-807, 809-819, 822-829); 16) pelvic- or femur-related fracture/aftercare (ICD-9-CM 733.14-15, 808, 820, 821, V54.13, V54.14, V43.64, V54.13, V54.23, V54.8-9); and 17) surgical implant or hardware-related issue (ICD-9-CM 996.4-996.6, 996.70, 996.77-996.79, V43.64).
Differences between subgroups (eg, by age and race/ ethnicity) were compared using the χ2 test. Multivariable logistic regression was used to evaluate the association of race/ethnicity, calendar year, and early rehospitalization with mortality outcome, adjusting for confounders associated with both the predictor and outcome of interest. We did not specifically examine for potential interactions. Calendar year was represented as a categorical variable (using the year 2000 or 2004 as reference) and as a continuous variable when evaluating for a linear trend. A 2-sided P value of <.05 was used as the criterion for statistical significance. All analyses were performed using SAS 9.3 (SAS Institute, Cary, North Carolina).
We initially identified 14,059 female health plan members 65 years and older who suffered a low-trauma hip fracture in the femoral neck or pertrochanter during 2000 to 2010. After excluding 509 women without continuous health plan membership in the year prior to fracture, the final cohort included 13,550 women, with a mean age of 82.5 ± 7.4 years and 84.6% aged 75 years or over. The cohort was predominantly white (83.6%), with 2.8% black, 5.6% Hispanic, 4.5% Asian, and 3.5% of other/unknown race. During the index hospitalization, 2.4% died, includ-ing a slightly higher proportion among women 85 years and older (3.3%) compared with women aged 65 to 74 and 75 to 84 years (both 1.8%; P <.001).