Journal of http://jag.sagepub.com/ Applied Gerontology
In-Hospital Mortality and Unintentional Falls Among Older Adults in the United States Darcy K. McMaughan Moudouni and Charles D. Phillips Journal of Applied Gerontology 2013 32: 923 originally published online 7 June 2012 DOI: 10.1177/0733464812445615 The online version of this article can be found at: http://jag.sagepub.com/content/32/8/923
Published by: http://www.sagepublications.com
On behalf of:
Southern Gerontological Society
Additional services and information for Journal of Applied Gerontology can be found at: Email Alerts: http://jag.sagepub.com/cgi/alerts Subscriptions: http://jag.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
>> Version of Record - Nov 13, 2013 OnlineFirst Version of Record - Jun 7, 2012 What is This?
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
445615 uni and PhillipsJournal of Applied Gerontology
JAG32810.1177/0733464812445615Moudo
Brief Report
In-Hospital Mortality and Unintentional Falls Among Older Adults in the United States
Journal of Applied Gerontology 32(8) 923–935 © The Author(s) 2012 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0733464812445615 jag.sagepub.com
Darcy K. McMaughan Moudouni1 and Charles D. Phillips2
Abstract Purpose of the Study: To estimate the odds of death associated with documented unintentional falls and acute care hospitalization among older adults in the United States. Design and Method: Data were abstracted from the 2005 Nationwide Inpatient Sample (NIS) and odds of death were modeled using logistic regression. Results: The age 65 and older fall rate per 1,000 discharges was 53.0 while the mortality rate for those who fell was 33.2. Older-old (odds ration [OR] = 2.93; confidence interval [CI] = [2.50, 3.43]), men (OR = 1.64, CI = [1.54, 1.75]), and non-White (OR = 1.09; CI = [1.01, 1.19]) had higher odds of death compared to younger-old, women, and Whites. Additional comorbidity (OR = 3.41, CI = [3.05, 3.82]), dehydration (OR = 1.14; CI = [1.05, 1.25]) and intracranial fractures (OR = 4.46; CI = [4.02, 4.95]) resulted in greater odds of death. Implications: Among older adults who experienced a fall and hospitalization, odds of mortality appear influenced by factors beyond injury severity related to falling. Additional research is necessary to delineate the mechanisms behind these phenomena to inform the public about falls-prevention programs. Manuscript received: May 13, 2011; final revision received: February 24, 2012; accepted: March 16, 2012. 1
exas A&M Health Science Center, School of Rural Public Health, Department of Health T Policy and Management, Program in Aging, Disability, and Long-Term Care Policy 2 School of Rural Public Health, College Station, TX, USA Corresponding Author: Darcy K. McMaughan Moudouni, Texas A&M Health Science Center School of Rural Public Health, SRPH Administration Building, College Station, TX 77843-1266, USA. Email: [email protected]
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
924
Journal of Applied Gerontology 32(8)
Keywords fall mortality, HCUP, NIS, Older adults, hospitalization
Introduction One in three older adults fall at least once each year (Hausdorff, Rios, & Edelberg, 2001; Tinetti, Doucette, Claus, & Marottoli, 1995). Oftentimes falls predicate a general decline in health—to the point that older adults who experience a fall associate this event with an increased probability of losing their ability to live independently (Stoeckel & Porell, 2009). Twenty to thirty percent of these falls result in injuries requiring medical attention (Alexander, Rivara, & Wolf, 1992)— consequently, on an annual basis as many as 1.8 million older adults who fell and were treated in emergency rooms, 433,000 are hospitalized and 13,700 die (Stevens, Ryan, & Kresnow, 2006). Reflecting the high rate of fall injuries and deaths, fall-related medical expenses for the year 2000 were estimated at US$19 billion, with fall-related deaths accounting for US$200 million in expenditures (Stevens, Corso, Finkelstein, & Miller, 2006). The 1990s saw fall-mortality rates steadily increase (Kannus et al., 1999; Kannus, Parkkari, Niemi, & Palvanen, 2005; Paulozzi, Ballesteros, & Stevens, 2006; Stevens, Ryan, et al., 2006). Falls now constitute the leading cause of unintentional death and injury among older adults in the United States (National Center for Injury Prevention and Control, 2006). The rise of fall deaths among older adults necessitates understanding those factors associated with falling and dying. Using nationally representative hospital discharge data from 2005, this study examines correlates of mortality and documented unintentional falls in inpatients aged 65 years and older in an effort to understand who is falling and dying.
Design and Method Data These data focus on in-hospital death associated with documented unintentional falls requiring medical care, regardless of where the fall occurred and regardless of the primary diagnosis (Tinetti et al., 1994). All data used in our analyses were derived from the 2005 Nationwide Inpatient Sample (NIS), Healthcare Cost and Utilization Project (HCUP NIS, 2005), Agency for Healthcare Research and Quality. The NIS contained 39,200,000 discharge abstracts from 1,054 community hospitals in 37 states and used a stratified cluster sample design. The NIS first stratified all community hospitals in the United States by region, urban or rural location, teaching status, ownership, and bedsize. From this stratified population,
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
925
a random sample of hospitals (about 20% of the total community hospitals in the United States) was selected for inclusion in the database. All discharges from each hospital are included in the NIS. The NIS’s high rates of external injury coding completeness and accuracy (Coben, Steiner, Barrett, Merrill, & Adamson, 2006) makes it both useful and appropriate for injury analysis.
Measurement The study sample included discharges with an unintentional fall and an age greater than 65 years (N = 140,870). An underlying diagnosis of unintentional fall was based on a primary external cause of injury code (E code) of accidental fall from the ICD-9 (E880-E888; Paulozzi et al., 2006; Stevens, Ryan, et al., 2006; Thomas, Stevens, Sarmiento, & Wald, 2008). E codes are three-to-fourdigit codes used by medical records personnel to classify an external cause of injury based on mechanism (e.g., poisoning, fall, drowning, motor vehicle accident), intent (e.g., homicide, suicide, accidental or unintentional), and location (e.g., home, work; Abellera, Annest, Conn, & Kohn, 2005). The fall sample used in this research consisted of any discharge with a primary E code indicating an unintentional fall, regardless of the primary diagnosis. Dependent variable. The outcome of interest was in-hospital death, denoted by a hospital discharge status of “died.” Independent variables. Demographic covariates included age, sex, race/ethnicity, admission source, rurality, average household income for the patient’s zipcode, and comorbidities. Age, gender, and race/ethnicity have been recognized as important predictors of falls and fall-mortality (Kannus et al., 1999; Paulozzi et al., 2006; Steinman, 2008; Stevens, Corso, et al., 2006). Age was categorized into four categories (65-74, 75-84, 85-94, 95 and above) for age group comparisons. Race was coded as White, non-White, and missing. Nine of the states included in the NIS did not report race, resulting in a relatively large amount of missing data on this variable. Admission source indicated whether the patient arrived at the hospital from an emergency room or another source. An adaptation of the Charlson Index was included to indicate a cumulative increase in likelihood of 1-year mortality due to the severity of the effect of comorbidities. This variable served as our covariate for illness severity or health status (Charlson, Pompei, Ales, & MacKenzie, 1987; Stagg, 2006). This adaptation uses a sum of weighted comorbidities from the 17 illnesses included in the original Charlson Index to produce truncated outcome scores (Deyo, Cherkin, & Ciol, 1992) using algorithms developed for administrative data (Quan et al., 2005). The Charlson Index is considered a valid and reliable tool for predicting mortality in health care administrative data (de Groot, Beckerman, Lankhorst, & Bouter, 2003).
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
926
Journal of Applied Gerontology 32(8)
Fractures of the hip (CCS 226), upper limb (CCS 229), lower limb (CCS 230), intracranial fractures (CCS 233), and other fractures (CCS 231) were selected based on frequency of the diagnosis (top 5%) in the first four diagnosis categories in the discharge record. Additional diagnoses generally associated with older ages (urinary tract infection and dehydration) were included, based on their prevalence and their exclusion from the Charlson Index. Where applicable,0 ICD-9 codes were collapsed into main codes or combined to form Clinical Classification diagnoses (CCS) using HCUP software, as coding at the ICD-9 group level is more accurate than coding using higher levels of specificity (Langley, Stephenson, Thorpe, & Davie, 2006).
Analytic Strategy Comorbid conditions and patient demographics significantly associated with death were modeled using logistic regression analysis, with death as the outcome. A bivariate logistic model of the relationship between age and death constituted the initial analysis. Addition covariates were systematically added to the logistic model by first inspecting the relationship between each individual variable and the dependent variable—variables exhibiting a statistically significant association with the outcome of interest (death) were included in the final model. Analyses were performed using STATA 11. Confidence intervals were calculated using the bootstrap techniques available in the SVY:LOGISTIC command. This approach provides appropriate variance estimates in the face of stratification and a lack of independence among observations (e.g., discharges from the same hospital). No specific hospital characteristics were used as covariates in the model; this research focused on individual characteristics.
Results In this nationally representative sample, both mortality and fall rates were higher among older adults. The weight-adjusted overall mortality rate for all discharges was 20.9 deaths per 1,000. Focusing just on discharges of individuals aged 65 or older, the overall weight-adjusted mortality rate was 44.0 deaths per 1,000 discharges. The overall weight-adjusted fall rate was 26.1 per 1,000 discharges. For those 65 or older, the adjusted fall rate was 53.0 per 1,000 discharges. Among those discharges with an E code indicating an accidental fall, the overall adjusted mortality rate was 26.6, while the adjusted mortality rate among those who were aged 65 years or older and fell was 33.2. Table 1 describes the sample. The majority of the falls sample was older-old patients who were largely female, White, and admitted from the emergency room.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
927
Moudouni and Phillips
Table 1. Descriptive Statistics: Hospital Discharges of Older Persons With a Fall.a
Sex*** Female Race/ethnicity* White Black Hispanic Asian/Pacific Islander Other Missing Age at admission*** 65-74 75-84 85-94 95** Admission source*** ER Another hospital Another facility (including long-term care) Court/routine/other Rurality (patient’s count of residence)* Urban (large metro) Rural (small metro, micro, noncore) Median household income for patient’s ZIP* US$1-US$36,999 US$37,000-US$45,999 US$46,000-US$60,000 US$61,000 or more Charlson Index***b 1 2 3 Fractures Hip fracture*** Lower limb fracture*** Upper limb fracture*** Intracranial fracture*** Other fracture*** Other diagnoses Urinary tract infection*** Dehydration***
Total N =140,870
Death n= 4,689
% 70.83
% 54.9
69.01 2.59 3.36 1.08 1.34 22.61
71.04 2.58 2.73 1.34 1.30 21.01
20.14 41.52 34.42 3.92
15.08 38.32 40.18 6.42
77.26 3.10 2.00 17.64
77.29 5.10 2.79 14.82
52.35 47.65
51.03 48.97
23.30 24.92 25.70 26.08
22.05 25.76 25.66 26.53
39.87 32.62 27.50
21.50 31.50 47.45
32.87 8.40 8.79 6.65 13.91
29.69 3.71 2.99 23.65 9.87
12.06 15.68
7.70 16.83
Note: ER = emergency room. a Total sample and discharge status of death (Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2005). b Charlson Index is a measure of disease burden based on the Quan et al. (2005) algorithm for 1-year mortality risk. *p < .05. **p < .01. ***p < .001.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
928
Journal of Applied Gerontology 32(8)
Most of the discharges were for persons 75 to 84 (42 %) or 85 to 94 years of age (34%). Due to the age of the sample members (above 64), the vast majority of the discharges used Medicare as the primary payer (93%). Out of 140,870 events among those aged 65 or older and diagnosed with a fall, 4,989 deaths occurred during hospitalization (about 3%). The likelihood of death increased with age. Also, among those older persons hospitalized and who had a diagnosis of a fall, the death rate was higher for men (5.1%) than women (2.5%). Table 2 presents the falls model that included all the covariates. Odds of death among individuals aged 75 to 84 were 1.28 times higher than among those aged 65 to 74 (OR = 1.28, 95% CI = [1.15, 1.44]). Being 85 to 94 years of age was associated with an 80% (OR = 1.81, 95% CI = [1.61, 2.04]) increase in the likelihood of death from a fall, compared to those aged 65 to 74. The oldest-old displayed the greatest odds of death; patients aged 95 and older had a 3-time greater risk of death (OR = 2.93, 95% CI = [2.50, 3.43]) than those aged 65 to 74. Older men were more likely to die than older women (OR = 1.64, 95% CI = [1.54, 1.75]) and non-White older inpatients were more likely to die than their White counterparts (OR = 1.09, 95% CI = [1.01, 1.19]). Patients with an unintentional fall and a Moderate Charlson Index (score of 2) were almost 2 times as likely to die during hospitalization (OR = 1.82, 95% CI = [1.65, 2.00]) compared to those with a Charlson Index of 1. This risk increased with greater comorbidity severity, as individuals with a Charlson Index score of 3 were more than 3 times more likely to die during hospitalization (OR = 3.41, 95% CI = [3.05, 3.82]) compared to individuals with a Charlson Index of 1. Likewise, patients who were dehydrated were at a 1.14-time greater risk of death than patients who had no documented dehydration (OR = 1.14, 95% CI = [1.05, 1.25]). Sustaining an intracranial injury dramatically increased the odds of dying (OR = 4.46, 95% CI = [4.02, 4.95]). All other fractures decreased the likelihood of death (lower limb fracture, OR = 0.57, 95% CI = [0.48, 0.68]; upper limb fracture, OR = 0.42, 95% CI = [0.35, 0.50]; and other fractures, OR = 0.78, 95% CI = [0.70, 0.88]) compared to not having the fracture, as did presence of a urinary tract infection (OR = 0.66, 95% CI = [0.59, 0.73]) compared to not having a urinary tract infection. Having a hip fracture decreased the likelihood of death compared to not having a hip fracture but was not significant in the final model (OR = 0.99, 95% CI = [0.91, 1.09]).
Discussion This study explored the relationship between unintentional falls that occurred in any location, hospitalization, and death among older adults hospitalized for any
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
929
Table 2. Adjusted Relationships: Predictors of Fall Mortality.a Age at admission (65-74 referent) 75-84 85-94 95** Sex (female referent) Male Race/ethnicity (White referent) Non-White Admission source (ER referent) Non-ER source Rurality (urban referent) Rural Median household income/ZIP (US$1-US$36,999 referent) US$37,000-US$45,999 US$46,000-US$60,999 US$61,000 and higher Charlson Index (1 referent)b 2 3 Fractures Hip fracture Lower limb fracture Upper limb fracture Intracranial fracture Other fracture Other diagnoses Dehydration Urinary tract infection
1.28 (1.15-1.44)*** 1.81 (1.61-2.04)*** 2.93 (2.50-3.43)*** 1.64 (1.54-1.75)*** 1.09 (1.01-1.19)* 0.99 (0.87-1.13) 1.08 (0.98-1.19) 1.09 (0.99-1.19) 1.04 (0.95-1.14) 1.06 (0.96-1.17) 1.82 (1.65-2.00)*** 3.41 (3.05-3.82)*** 0.99 (0.91-1.09) 0.57 (0.48-0.68)*** 0.42 (0.35-0.50)*** 4.46 (4.02-4.95)*** 0.78 (0.70-0.88)*** 1.14 (1.05-1.25)*** 0.66 (0.59-0.73)***
Note: ER = emergency room. N = 140,870. a Final model odds ratio (95% CI) (Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2005). b Charlson Index is a measure of disease burden based on the Quan et al. (2005) algorithm for 1-year mortality risk. *p < .05. **p < .01. ***p < .001.
reason. Several socioeconomic factors beyond injury or health status (as defined by the Charlson Index) predicted higher odds of death, including older ages, being male, and being a member of an ethnic/racial minority. On its own, older age at
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
930
Journal of Applied Gerontology 32(8)
admission was significantly associated with a higher risk of death. This relationship remained unchanged after controlling for other sociodemographic and health-status indicators. These greater odds of death in older ages may result from population increases in longevity and associated frailty in the older years, combined with an increase in prescription drug use (including polypharmacy; Cumming et al., 1991; Hanlon et al., 2009). In the study sample, men were more likely to die than women, a finding previously reported (Kannus et al., 1999; Kannus et al., 2005; Stevens, Ryan, et al., 2006) and perhaps due to men engaging in potentially riskier behavior or women being more likely to engage in risk avoidance (Byrnes, Miller, & Schafer, 1999; Courtenay, 2000; Turner & McClure, 2003). The majority of the common fall fractures (upper and lower limb, fractures of the hip, and “other fractures”) were less likely to result in mortality than falls involving head injuries. Intracranial fractures drastically raised the odds of death. Individuals with a diagnosis of intracranial fracture had odds of dying 446% higher than those without a skull fracture. Fall-induced severe head injuries among the older persons are increasing (Kannus et al., 1999). Traumatic brain injuries (TBIs) explain approximately half of fatal falls and 8% of nonfatal fall hospitalizations in older Americans (and a substantial proportion of the fall-mortality costs; Stevens, Corso, et al., 2006; Thomas et al., 2008). This study is restricted by the limited range of information provided by hospital administrative data (Koehler et al., 2006). E codes are limited in that they rely on the willingness and ability of health professionals using the codes. As of 2005, only 26 states plus the District of Columbia require collection of injury E codes (Abellera et al., 2005). However, previous works indicate that E codes underestimate the proportion of injuries due to falls through the failure to record E codes, rather than misapplication of codes (Barrett, Steiner, & Coben, 2005; Carlson, Nugent, Grill, & Sayer, 2010) or systematic misclassification (Smith & Langley, 1998). Due to regulatory requirements concerning death certificates issued for injury deaths, E coding is generally more complete for mortality data than morbidity data (Annest et al., 2008). We used the HCUP, which reports a relatively high level of injury E code completeness compared to other national data sets: an average of 87.2% completeness for injury records from the states providing hospital discharge data. More than half of those states (19 out of 33) submitted E codes on at least 90% of their discharge data (Barrett et al., 2005). In addition, E codes for accidental falls are frequently used to examine costs and risk factors associated with falling (Roudsari, Ebel, Corso, Molinari, & Koepsell, 2005). These codes are considered a reliable source of information on injury intent and mechanism in hospital discharge data, especially for unintentional falls (Langley et al., 2006; LeMier, Cummings, & West, 2001). Still, the possibility of bias in our results due to the failure to appropriately use E codes must be acknowledged.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
931
Our findings can potentially be used for the refinement of fall-prevention programs and risk adjustment for public reporting. As highlighted in this study, men potentially have higher odds of fall-related mortality, yet older women tend to seek out fall-prevention programs more than older men (Calhoun et al., 2011). Programs that target groups at a high risk for death generally do so after individuals have been hospitalized (Moore et al., 2010). Broader outreach efforts that identify and include high-risk groups may be achieved through fall-prevention program refinement via social marketing aimed at tailoring existing programs to the specific characteristics of the intended recipients, as in the case of the N’Balance program (Clark et al., 2011) and the A Matter of Balance program (Healy et al., 2008). Similarly, selecting programs based on known efficacy with target populations (such as the Balance/Volunteer Lay Leader model fall-prevention program for rural older adults) can enhance dissemination (Smith et al., 2010). However, to this date, there are no especially effective interventions that target populations at higher risk of death from a fall and no real recognition in programs of the differential risk of death for different individuals or different types of falls (Gillespie et al., 2009). Understanding differences in patients’ risk of death that are unrelated to the quality of hospital care they received is essential as the American health care system moves forward toward accountability through consumer empowerment and financial incentives based on performance. Certain patient characteristics, not under the control of the hospital, may make death more likely—this includes both sociodemographic characteristics and the comorbidities that individuals bring with them to the acute care setting. Risk adjustment of important outcomes, like mortality, without adjusting for such factors may introduce significant bias into the process of rating hospital performance. Using sociodemographic characteristics associated with an increased risk of death for the older adults who fell (e.g., gender, age) could be useful in U.S. Department of Health and Human Services (HHS) programs focusing on processes and outcomes of inpatient care, such as Hospital Compare and ValueBased Purchasing (which include death from hip fractures among the ratings for patient safety).
Acknowledgment The authors received departmental support for this project.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
932
Journal of Applied Gerontology 32(8)
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References Abellera, J., Annest, J. L., Conn, J. M., & Kohn, M. (2005). How states are collecting and using cause of injury data: 2004 update to the 1997 report (A report by the council of state and territorial epidemiologists; data committee injury control and emergency health services section, American public health association; and state and territorial injury prevention directors association). Atlanta, GA: The Council of State and Territorial Epidemiologists. Alexander, B. H., Rivara, F. P., & Wolf, M. E. (1992). The cost and frequency of hospitalization for fall-related injuries in older adults. American Journal of Public Health, 82, 1020-1023. Annest, J. L., Fingerhut, L. A., Gallagher, S. S., Grossman, D. C., Hedegaard, H., Johnson, . . . Trent, R. B. (2008, March 28). Strategies to improve external causeof-injury coding in state-based hospital discharge and emergency department data systems. Morbidity and Mortality Weekly Report, 57(RR01), 1-15. Barrett, M., Steiner, C., & Coben, J. (2005). Healthcare cost and utilization project (HCUP) e code evaluation report 2004 (HCUP methods series report No. 200406). Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from http://www.hcup-us.ahrq.gov/reports/methods.jsp Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A meta-analysis. Psychological Bulletin, 125, 367-383. Calhoun, R., Meischke, H., Hammerback, K., Bohl, A., Poe, P., Williams, B., & Phelan, E. A. (2011). Older adults’ perceptions of clinical fall prevention programs: A qualitative study. Journal of Aging Research, Article ID 867341, 7. doi:10.4061/2011/867341 Carlson, K. F., Nugent, S. M., Grill, J., & Sayer, N. A. (2010). Accuracy of external cause-of-injury coding in VA polytrauma patient discharge records. Journal of Rehabilitation Research & Development, 47, 689-698. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Disease, 40, 373-383. Clark, L., Thoreson, S., Goss, C. W., Marquez Zimmer, L., Marosits, M., & DiGuiseppi, C. (2011, April). Understanding fall meaning and context in marketing balance classes to older adults. Journal of Applied Gerontology. Retrieved from http://jag.sagepub.com/content/early/2011/03/31/0733464811 399896
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
933
Coben, J. H., Steiner, C. A., Barrett, M., Merrill, C. T., & Adamson, D. (2006). Completeness of cause of injury coding in healthcare administrative databases in the United States, 2001. Injury Prevention, 12(3), 199-201. Courtenay, W. H. (2000). Constructions of masculinity and their influence on men’s well-being: A theory of gender and health. Social Science and Medicine, 50(1), 1385-1401. Cumming, R. G., Miller, J. P., Kelsey, J. L., Davis, P., Arfken, C. L., Birge, S. J., & Peck W. A. (1991). Medications and multiple falls in elderly people: The St. Louis OASIS study. Age and Ageing, 20, 455-461. Deyo, R. A., Cherkin, D. C., & Ciol, M. A. (1992). Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. Journal of Clinical Epidemiology, 45, 613-619. de Groot, V., Beckerman, H., Lankhorst, G. J., & Bouter, L. M. (2003). How to measure comorbidity: A critical review of available methods. Journal of Clinical Epidemiology, 56(3), 221-229. Gillespie, L. D., Robertson, M. C., Gillespie, W. J., Lamb, S. E., Gates, S., Cumming, R. G., & Rowe, B. H. (2009). Interventions for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews, 2, Article CD007146. doi:10.1002/14651858.CD007146.pub2 Hanlon, J. T., Boudreau, R. M., Roumani, Y. F., Newman, A. B., Ruby, C. M., Wright, R. M., . . . Studenski, S. A. (2009). Number and dosage of central nervous system medications on recurrent falls in community elders: The health, aging and body composition study. Journals of Gerontology: Series A, 64A, 492-498. Hausdorff, J. M., Rios, D. A., & Edelberg, H. K. (2001). Gait variability and fall risk in community-living older adults: A 1-year prospective study. Archives of Physical Medicine and Rehabilitation, 82, 1050-1056. Healthcare Cost and Utilization Project Nationwide Inpatient Sample. (2005). Healthcare cost and utilization project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from www.hcup-us.ahrq.gov/nisoverview.jsp Healy, T. C., Peng, C., Haynes, M. S., McMahon, E. M., Botler, J. L., & Gross, L. (2008). The feasibility and effectiveness of translating a matter of balance into a volunteer lay leader model. Journal of Applied Gerontology, 27(1), 34-51. Retrieved from http://jag.sagepub.com/content/27/1/34 Kannus, P., Parkkari, J., Koskinen, S., Niemi, S., Palvanen, M., Jarvinen, M., & Vuori, I. (1999). Fall-induced injuries and deaths among older adults. Journal of the American Medical Association, 281, 1895-1899. Kannus, P., Parkkari, J., Niemi, S., & Palvanen, M. (2005). Fall-induced deaths among elderly people. American Journal of Public Health, 95, 422-424. Koehler, S. A., Weiss, H. B., Shakir, A., Shaeffer, S., Ladham, S., Rozin, L., & Wecht, C. H. (2006). Accurately assessing elderly fall deaths using hospital discharge and vital
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
934
Journal of Applied Gerontology 32(8)
statistics data. American Journal of Forensic Medicine and Pathology, 27(1), 30-35. Langley, J., Stephenson, S., Thorpe, C., & Davie, G. (2006). Accuracy of injury coding under ICD-9 for New Zealand public hospital discharges. Injury Prevention, 12(1), 58-61. LeMier, M., Cummings, P., & West, T. A. (2001). Accuracy of external cause of injury codes reported in Washington State hospital discharge records. Injury Prevention, 7, 334-338. Moore, M., Williams, B., Ragsdale, S., LoGerfo, J. P., Goss, J. R., Schreuder, A. B., & Phelan, E. A. (2010). Translating a multifactorial fall prevention intervention into practice: A controlled evaluation of a fall prevention clinic. Journal of the American Geriatrics Society, 58, 357-363. National Center for Injury Prevention and Control. (2006). Web-based injury statistics query and reporting system. Retrieved from http:/www.cdc.gov/ncipc/wisqars Paulozzi, L. J., Ballesteros, M. F., & Stevens, J. A. (2006). Recent trends in mortality from unintentional injury in the United States. Journal of Safety Research, 37, 277-283. Quan, H., Sundararajan, V., Halfon, P., Fong, A., Burnand, B., Luthi, J. C., . . . Ghali, W. A. (2005). Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Medical Care, 43, 1130-1139. Roudsari, B. S., Ebel, B. E., Corso, P. S., Molinari, N.-A. M., & Koepsell, T. D. (2005). The acute medical care costs of fall-related injuries among the U.S. older adults. Injury, 36, 1316-1322. Smith, G., & Langley, J. (1998). Drowning surveillance: How well do e-codes identify submersion fatalities? Injury Prevention, 4, 135-139. Smith, M. L., Ahn, S., Sharkey, J. R., Horel, S., Mier, N., & Ory, M. G. (2010, August). Successful falls prevention programming for older adults in Texas: Rural-urban variations. Journal of Applied Gerontology. Retrieved from http://jag.sagepub. com/content/early/2010/10/09/0733464810378407 Stagg, V. (2006). CHARLSON: Stata module to calculate Charlson Index of Comorbidity (Statistical software components S456719). Chestnut Hill, MA: Boston College Department of Economics. Steinman, B. A. (2008, May). Self-reported vision, upper/lower limb disability, and fall risk in older adults. Journal of Applied Gerontology. Retrieved from http:// jag.sagepub.com/content/27/4/406 Stevens, J. A., Corso, P. S., Finkelstein, E. A., & Miller, T. R. (2006). The costs of fatal and non-fatal falls among older adults. Injury Prevention, 12(5), 290-295. Stevens, J. A., Ryan, G., & Kresnow, M. S. (2006). Fatalities and injuries from falls among older adults—United States, 1993-2003 and 2001-2005. Morbidity & Mortality Weekly Report, 55, 1221-1224.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
935
Stoeckel, K. J., & Porell, F. (2009). Do older adults anticipate relocating? The relationship between housing relocation expectations and falls. Journal of Applied Gerontology. Retrieved from http://jag.sagepub.com/content/29/2/231 Thomas, K. E., Stevens, J. A., Sarmiento, K., & Wald, M. M. (2008). Fall-related traumatic brain injury deaths and hospitalizations among older adults—United States, 2005. Journal of Safety Research, 39, 269-272. Tinetti, M. E., Baker, D. I., McAvay, G., Claus, E. B., Garett, P., Gottschalk, M., . . . Horwitz, R. I. (1994). A multifactorial intervention to reduce the risk of falling among elderly people living in the community. New England Journal of Medicine, 331, 821-827. Tinetti, M. E., Doucette, J., Claus, E., & Marottoli, R. (1995). Risk factors for serious injury during falls by older persons in the community. Journal of the American Geriatrics Society, 43, 1214-1221. Turner, C., & McClure, R. (2003). Age and gender differences in risk-taking behavior as an explanation for high incidence of motor vehicle crashes as a driver in young males. Injury Control and Safety Promotion, 10(3), 123-130.
Author Biographies Darcy K. McMaughan Moudouni, PhD, is an assistant professor and researcher for the Program in Aging, Disability and Long-Term Care Policy at the Texas A&M Health Science Center School of Rural Public Health. Charles D. Phillips, PhD, MPH, is a regent’s professor and head of the Program in Aging, Disability and Long-Term Care Policy at the Texas A&M Health Science Center School of Rural Public Health. Both authors are in the Department of Health Policy and Management.
In-Hospital Mortality and Unintentional Falls Among Older Adults in the United States Darcy K. McMaughan Moudouni and Charles D. Phillips Journal of Applied Gerontology 2013 32: 923 originally published online 7 June 2012 DOI: 10.1177/0733464812445615 The online version of this article can be found at: http://jag.sagepub.com/content/32/8/923
Published by: http://www.sagepublications.com
On behalf of:
Southern Gerontological Society
Additional services and information for Journal of Applied Gerontology can be found at: Email Alerts: http://jag.sagepub.com/cgi/alerts Subscriptions: http://jag.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
>> Version of Record - Nov 13, 2013 OnlineFirst Version of Record - Jun 7, 2012 What is This?
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
445615 uni and PhillipsJournal of Applied Gerontology
JAG32810.1177/0733464812445615Moudo
Brief Report
In-Hospital Mortality and Unintentional Falls Among Older Adults in the United States
Journal of Applied Gerontology 32(8) 923–935 © The Author(s) 2012 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0733464812445615 jag.sagepub.com
Darcy K. McMaughan Moudouni1 and Charles D. Phillips2
Abstract Purpose of the Study: To estimate the odds of death associated with documented unintentional falls and acute care hospitalization among older adults in the United States. Design and Method: Data were abstracted from the 2005 Nationwide Inpatient Sample (NIS) and odds of death were modeled using logistic regression. Results: The age 65 and older fall rate per 1,000 discharges was 53.0 while the mortality rate for those who fell was 33.2. Older-old (odds ration [OR] = 2.93; confidence interval [CI] = [2.50, 3.43]), men (OR = 1.64, CI = [1.54, 1.75]), and non-White (OR = 1.09; CI = [1.01, 1.19]) had higher odds of death compared to younger-old, women, and Whites. Additional comorbidity (OR = 3.41, CI = [3.05, 3.82]), dehydration (OR = 1.14; CI = [1.05, 1.25]) and intracranial fractures (OR = 4.46; CI = [4.02, 4.95]) resulted in greater odds of death. Implications: Among older adults who experienced a fall and hospitalization, odds of mortality appear influenced by factors beyond injury severity related to falling. Additional research is necessary to delineate the mechanisms behind these phenomena to inform the public about falls-prevention programs. Manuscript received: May 13, 2011; final revision received: February 24, 2012; accepted: March 16, 2012. 1
exas A&M Health Science Center, School of Rural Public Health, Department of Health T Policy and Management, Program in Aging, Disability, and Long-Term Care Policy 2 School of Rural Public Health, College Station, TX, USA Corresponding Author: Darcy K. McMaughan Moudouni, Texas A&M Health Science Center School of Rural Public Health, SRPH Administration Building, College Station, TX 77843-1266, USA. Email: [email protected]
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
924
Journal of Applied Gerontology 32(8)
Keywords fall mortality, HCUP, NIS, Older adults, hospitalization
Introduction One in three older adults fall at least once each year (Hausdorff, Rios, & Edelberg, 2001; Tinetti, Doucette, Claus, & Marottoli, 1995). Oftentimes falls predicate a general decline in health—to the point that older adults who experience a fall associate this event with an increased probability of losing their ability to live independently (Stoeckel & Porell, 2009). Twenty to thirty percent of these falls result in injuries requiring medical attention (Alexander, Rivara, & Wolf, 1992)— consequently, on an annual basis as many as 1.8 million older adults who fell and were treated in emergency rooms, 433,000 are hospitalized and 13,700 die (Stevens, Ryan, & Kresnow, 2006). Reflecting the high rate of fall injuries and deaths, fall-related medical expenses for the year 2000 were estimated at US$19 billion, with fall-related deaths accounting for US$200 million in expenditures (Stevens, Corso, Finkelstein, & Miller, 2006). The 1990s saw fall-mortality rates steadily increase (Kannus et al., 1999; Kannus, Parkkari, Niemi, & Palvanen, 2005; Paulozzi, Ballesteros, & Stevens, 2006; Stevens, Ryan, et al., 2006). Falls now constitute the leading cause of unintentional death and injury among older adults in the United States (National Center for Injury Prevention and Control, 2006). The rise of fall deaths among older adults necessitates understanding those factors associated with falling and dying. Using nationally representative hospital discharge data from 2005, this study examines correlates of mortality and documented unintentional falls in inpatients aged 65 years and older in an effort to understand who is falling and dying.
Design and Method Data These data focus on in-hospital death associated with documented unintentional falls requiring medical care, regardless of where the fall occurred and regardless of the primary diagnosis (Tinetti et al., 1994). All data used in our analyses were derived from the 2005 Nationwide Inpatient Sample (NIS), Healthcare Cost and Utilization Project (HCUP NIS, 2005), Agency for Healthcare Research and Quality. The NIS contained 39,200,000 discharge abstracts from 1,054 community hospitals in 37 states and used a stratified cluster sample design. The NIS first stratified all community hospitals in the United States by region, urban or rural location, teaching status, ownership, and bedsize. From this stratified population,
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
925
a random sample of hospitals (about 20% of the total community hospitals in the United States) was selected for inclusion in the database. All discharges from each hospital are included in the NIS. The NIS’s high rates of external injury coding completeness and accuracy (Coben, Steiner, Barrett, Merrill, & Adamson, 2006) makes it both useful and appropriate for injury analysis.
Measurement The study sample included discharges with an unintentional fall and an age greater than 65 years (N = 140,870). An underlying diagnosis of unintentional fall was based on a primary external cause of injury code (E code) of accidental fall from the ICD-9 (E880-E888; Paulozzi et al., 2006; Stevens, Ryan, et al., 2006; Thomas, Stevens, Sarmiento, & Wald, 2008). E codes are three-to-fourdigit codes used by medical records personnel to classify an external cause of injury based on mechanism (e.g., poisoning, fall, drowning, motor vehicle accident), intent (e.g., homicide, suicide, accidental or unintentional), and location (e.g., home, work; Abellera, Annest, Conn, & Kohn, 2005). The fall sample used in this research consisted of any discharge with a primary E code indicating an unintentional fall, regardless of the primary diagnosis. Dependent variable. The outcome of interest was in-hospital death, denoted by a hospital discharge status of “died.” Independent variables. Demographic covariates included age, sex, race/ethnicity, admission source, rurality, average household income for the patient’s zipcode, and comorbidities. Age, gender, and race/ethnicity have been recognized as important predictors of falls and fall-mortality (Kannus et al., 1999; Paulozzi et al., 2006; Steinman, 2008; Stevens, Corso, et al., 2006). Age was categorized into four categories (65-74, 75-84, 85-94, 95 and above) for age group comparisons. Race was coded as White, non-White, and missing. Nine of the states included in the NIS did not report race, resulting in a relatively large amount of missing data on this variable. Admission source indicated whether the patient arrived at the hospital from an emergency room or another source. An adaptation of the Charlson Index was included to indicate a cumulative increase in likelihood of 1-year mortality due to the severity of the effect of comorbidities. This variable served as our covariate for illness severity or health status (Charlson, Pompei, Ales, & MacKenzie, 1987; Stagg, 2006). This adaptation uses a sum of weighted comorbidities from the 17 illnesses included in the original Charlson Index to produce truncated outcome scores (Deyo, Cherkin, & Ciol, 1992) using algorithms developed for administrative data (Quan et al., 2005). The Charlson Index is considered a valid and reliable tool for predicting mortality in health care administrative data (de Groot, Beckerman, Lankhorst, & Bouter, 2003).
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
926
Journal of Applied Gerontology 32(8)
Fractures of the hip (CCS 226), upper limb (CCS 229), lower limb (CCS 230), intracranial fractures (CCS 233), and other fractures (CCS 231) were selected based on frequency of the diagnosis (top 5%) in the first four diagnosis categories in the discharge record. Additional diagnoses generally associated with older ages (urinary tract infection and dehydration) were included, based on their prevalence and their exclusion from the Charlson Index. Where applicable,0 ICD-9 codes were collapsed into main codes or combined to form Clinical Classification diagnoses (CCS) using HCUP software, as coding at the ICD-9 group level is more accurate than coding using higher levels of specificity (Langley, Stephenson, Thorpe, & Davie, 2006).
Analytic Strategy Comorbid conditions and patient demographics significantly associated with death were modeled using logistic regression analysis, with death as the outcome. A bivariate logistic model of the relationship between age and death constituted the initial analysis. Addition covariates were systematically added to the logistic model by first inspecting the relationship between each individual variable and the dependent variable—variables exhibiting a statistically significant association with the outcome of interest (death) were included in the final model. Analyses were performed using STATA 11. Confidence intervals were calculated using the bootstrap techniques available in the SVY:LOGISTIC command. This approach provides appropriate variance estimates in the face of stratification and a lack of independence among observations (e.g., discharges from the same hospital). No specific hospital characteristics were used as covariates in the model; this research focused on individual characteristics.
Results In this nationally representative sample, both mortality and fall rates were higher among older adults. The weight-adjusted overall mortality rate for all discharges was 20.9 deaths per 1,000. Focusing just on discharges of individuals aged 65 or older, the overall weight-adjusted mortality rate was 44.0 deaths per 1,000 discharges. The overall weight-adjusted fall rate was 26.1 per 1,000 discharges. For those 65 or older, the adjusted fall rate was 53.0 per 1,000 discharges. Among those discharges with an E code indicating an accidental fall, the overall adjusted mortality rate was 26.6, while the adjusted mortality rate among those who were aged 65 years or older and fell was 33.2. Table 1 describes the sample. The majority of the falls sample was older-old patients who were largely female, White, and admitted from the emergency room.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
927
Moudouni and Phillips
Table 1. Descriptive Statistics: Hospital Discharges of Older Persons With a Fall.a
Sex*** Female Race/ethnicity* White Black Hispanic Asian/Pacific Islander Other Missing Age at admission*** 65-74 75-84 85-94 95** Admission source*** ER Another hospital Another facility (including long-term care) Court/routine/other Rurality (patient’s count of residence)* Urban (large metro) Rural (small metro, micro, noncore) Median household income for patient’s ZIP* US$1-US$36,999 US$37,000-US$45,999 US$46,000-US$60,000 US$61,000 or more Charlson Index***b 1 2 3 Fractures Hip fracture*** Lower limb fracture*** Upper limb fracture*** Intracranial fracture*** Other fracture*** Other diagnoses Urinary tract infection*** Dehydration***
Total N =140,870
Death n= 4,689
% 70.83
% 54.9
69.01 2.59 3.36 1.08 1.34 22.61
71.04 2.58 2.73 1.34 1.30 21.01
20.14 41.52 34.42 3.92
15.08 38.32 40.18 6.42
77.26 3.10 2.00 17.64
77.29 5.10 2.79 14.82
52.35 47.65
51.03 48.97
23.30 24.92 25.70 26.08
22.05 25.76 25.66 26.53
39.87 32.62 27.50
21.50 31.50 47.45
32.87 8.40 8.79 6.65 13.91
29.69 3.71 2.99 23.65 9.87
12.06 15.68
7.70 16.83
Note: ER = emergency room. a Total sample and discharge status of death (Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2005). b Charlson Index is a measure of disease burden based on the Quan et al. (2005) algorithm for 1-year mortality risk. *p < .05. **p < .01. ***p < .001.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
928
Journal of Applied Gerontology 32(8)
Most of the discharges were for persons 75 to 84 (42 %) or 85 to 94 years of age (34%). Due to the age of the sample members (above 64), the vast majority of the discharges used Medicare as the primary payer (93%). Out of 140,870 events among those aged 65 or older and diagnosed with a fall, 4,989 deaths occurred during hospitalization (about 3%). The likelihood of death increased with age. Also, among those older persons hospitalized and who had a diagnosis of a fall, the death rate was higher for men (5.1%) than women (2.5%). Table 2 presents the falls model that included all the covariates. Odds of death among individuals aged 75 to 84 were 1.28 times higher than among those aged 65 to 74 (OR = 1.28, 95% CI = [1.15, 1.44]). Being 85 to 94 years of age was associated with an 80% (OR = 1.81, 95% CI = [1.61, 2.04]) increase in the likelihood of death from a fall, compared to those aged 65 to 74. The oldest-old displayed the greatest odds of death; patients aged 95 and older had a 3-time greater risk of death (OR = 2.93, 95% CI = [2.50, 3.43]) than those aged 65 to 74. Older men were more likely to die than older women (OR = 1.64, 95% CI = [1.54, 1.75]) and non-White older inpatients were more likely to die than their White counterparts (OR = 1.09, 95% CI = [1.01, 1.19]). Patients with an unintentional fall and a Moderate Charlson Index (score of 2) were almost 2 times as likely to die during hospitalization (OR = 1.82, 95% CI = [1.65, 2.00]) compared to those with a Charlson Index of 1. This risk increased with greater comorbidity severity, as individuals with a Charlson Index score of 3 were more than 3 times more likely to die during hospitalization (OR = 3.41, 95% CI = [3.05, 3.82]) compared to individuals with a Charlson Index of 1. Likewise, patients who were dehydrated were at a 1.14-time greater risk of death than patients who had no documented dehydration (OR = 1.14, 95% CI = [1.05, 1.25]). Sustaining an intracranial injury dramatically increased the odds of dying (OR = 4.46, 95% CI = [4.02, 4.95]). All other fractures decreased the likelihood of death (lower limb fracture, OR = 0.57, 95% CI = [0.48, 0.68]; upper limb fracture, OR = 0.42, 95% CI = [0.35, 0.50]; and other fractures, OR = 0.78, 95% CI = [0.70, 0.88]) compared to not having the fracture, as did presence of a urinary tract infection (OR = 0.66, 95% CI = [0.59, 0.73]) compared to not having a urinary tract infection. Having a hip fracture decreased the likelihood of death compared to not having a hip fracture but was not significant in the final model (OR = 0.99, 95% CI = [0.91, 1.09]).
Discussion This study explored the relationship between unintentional falls that occurred in any location, hospitalization, and death among older adults hospitalized for any
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
929
Table 2. Adjusted Relationships: Predictors of Fall Mortality.a Age at admission (65-74 referent) 75-84 85-94 95** Sex (female referent) Male Race/ethnicity (White referent) Non-White Admission source (ER referent) Non-ER source Rurality (urban referent) Rural Median household income/ZIP (US$1-US$36,999 referent) US$37,000-US$45,999 US$46,000-US$60,999 US$61,000 and higher Charlson Index (1 referent)b 2 3 Fractures Hip fracture Lower limb fracture Upper limb fracture Intracranial fracture Other fracture Other diagnoses Dehydration Urinary tract infection
1.28 (1.15-1.44)*** 1.81 (1.61-2.04)*** 2.93 (2.50-3.43)*** 1.64 (1.54-1.75)*** 1.09 (1.01-1.19)* 0.99 (0.87-1.13) 1.08 (0.98-1.19) 1.09 (0.99-1.19) 1.04 (0.95-1.14) 1.06 (0.96-1.17) 1.82 (1.65-2.00)*** 3.41 (3.05-3.82)*** 0.99 (0.91-1.09) 0.57 (0.48-0.68)*** 0.42 (0.35-0.50)*** 4.46 (4.02-4.95)*** 0.78 (0.70-0.88)*** 1.14 (1.05-1.25)*** 0.66 (0.59-0.73)***
Note: ER = emergency room. N = 140,870. a Final model odds ratio (95% CI) (Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2005). b Charlson Index is a measure of disease burden based on the Quan et al. (2005) algorithm for 1-year mortality risk. *p < .05. **p < .01. ***p < .001.
reason. Several socioeconomic factors beyond injury or health status (as defined by the Charlson Index) predicted higher odds of death, including older ages, being male, and being a member of an ethnic/racial minority. On its own, older age at
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
930
Journal of Applied Gerontology 32(8)
admission was significantly associated with a higher risk of death. This relationship remained unchanged after controlling for other sociodemographic and health-status indicators. These greater odds of death in older ages may result from population increases in longevity and associated frailty in the older years, combined with an increase in prescription drug use (including polypharmacy; Cumming et al., 1991; Hanlon et al., 2009). In the study sample, men were more likely to die than women, a finding previously reported (Kannus et al., 1999; Kannus et al., 2005; Stevens, Ryan, et al., 2006) and perhaps due to men engaging in potentially riskier behavior or women being more likely to engage in risk avoidance (Byrnes, Miller, & Schafer, 1999; Courtenay, 2000; Turner & McClure, 2003). The majority of the common fall fractures (upper and lower limb, fractures of the hip, and “other fractures”) were less likely to result in mortality than falls involving head injuries. Intracranial fractures drastically raised the odds of death. Individuals with a diagnosis of intracranial fracture had odds of dying 446% higher than those without a skull fracture. Fall-induced severe head injuries among the older persons are increasing (Kannus et al., 1999). Traumatic brain injuries (TBIs) explain approximately half of fatal falls and 8% of nonfatal fall hospitalizations in older Americans (and a substantial proportion of the fall-mortality costs; Stevens, Corso, et al., 2006; Thomas et al., 2008). This study is restricted by the limited range of information provided by hospital administrative data (Koehler et al., 2006). E codes are limited in that they rely on the willingness and ability of health professionals using the codes. As of 2005, only 26 states plus the District of Columbia require collection of injury E codes (Abellera et al., 2005). However, previous works indicate that E codes underestimate the proportion of injuries due to falls through the failure to record E codes, rather than misapplication of codes (Barrett, Steiner, & Coben, 2005; Carlson, Nugent, Grill, & Sayer, 2010) or systematic misclassification (Smith & Langley, 1998). Due to regulatory requirements concerning death certificates issued for injury deaths, E coding is generally more complete for mortality data than morbidity data (Annest et al., 2008). We used the HCUP, which reports a relatively high level of injury E code completeness compared to other national data sets: an average of 87.2% completeness for injury records from the states providing hospital discharge data. More than half of those states (19 out of 33) submitted E codes on at least 90% of their discharge data (Barrett et al., 2005). In addition, E codes for accidental falls are frequently used to examine costs and risk factors associated with falling (Roudsari, Ebel, Corso, Molinari, & Koepsell, 2005). These codes are considered a reliable source of information on injury intent and mechanism in hospital discharge data, especially for unintentional falls (Langley et al., 2006; LeMier, Cummings, & West, 2001). Still, the possibility of bias in our results due to the failure to appropriately use E codes must be acknowledged.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
931
Our findings can potentially be used for the refinement of fall-prevention programs and risk adjustment for public reporting. As highlighted in this study, men potentially have higher odds of fall-related mortality, yet older women tend to seek out fall-prevention programs more than older men (Calhoun et al., 2011). Programs that target groups at a high risk for death generally do so after individuals have been hospitalized (Moore et al., 2010). Broader outreach efforts that identify and include high-risk groups may be achieved through fall-prevention program refinement via social marketing aimed at tailoring existing programs to the specific characteristics of the intended recipients, as in the case of the N’Balance program (Clark et al., 2011) and the A Matter of Balance program (Healy et al., 2008). Similarly, selecting programs based on known efficacy with target populations (such as the Balance/Volunteer Lay Leader model fall-prevention program for rural older adults) can enhance dissemination (Smith et al., 2010). However, to this date, there are no especially effective interventions that target populations at higher risk of death from a fall and no real recognition in programs of the differential risk of death for different individuals or different types of falls (Gillespie et al., 2009). Understanding differences in patients’ risk of death that are unrelated to the quality of hospital care they received is essential as the American health care system moves forward toward accountability through consumer empowerment and financial incentives based on performance. Certain patient characteristics, not under the control of the hospital, may make death more likely—this includes both sociodemographic characteristics and the comorbidities that individuals bring with them to the acute care setting. Risk adjustment of important outcomes, like mortality, without adjusting for such factors may introduce significant bias into the process of rating hospital performance. Using sociodemographic characteristics associated with an increased risk of death for the older adults who fell (e.g., gender, age) could be useful in U.S. Department of Health and Human Services (HHS) programs focusing on processes and outcomes of inpatient care, such as Hospital Compare and ValueBased Purchasing (which include death from hip fractures among the ratings for patient safety).
Acknowledgment The authors received departmental support for this project.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
932
Journal of Applied Gerontology 32(8)
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References Abellera, J., Annest, J. L., Conn, J. M., & Kohn, M. (2005). How states are collecting and using cause of injury data: 2004 update to the 1997 report (A report by the council of state and territorial epidemiologists; data committee injury control and emergency health services section, American public health association; and state and territorial injury prevention directors association). Atlanta, GA: The Council of State and Territorial Epidemiologists. Alexander, B. H., Rivara, F. P., & Wolf, M. E. (1992). The cost and frequency of hospitalization for fall-related injuries in older adults. American Journal of Public Health, 82, 1020-1023. Annest, J. L., Fingerhut, L. A., Gallagher, S. S., Grossman, D. C., Hedegaard, H., Johnson, . . . Trent, R. B. (2008, March 28). Strategies to improve external causeof-injury coding in state-based hospital discharge and emergency department data systems. Morbidity and Mortality Weekly Report, 57(RR01), 1-15. Barrett, M., Steiner, C., & Coben, J. (2005). Healthcare cost and utilization project (HCUP) e code evaluation report 2004 (HCUP methods series report No. 200406). Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from http://www.hcup-us.ahrq.gov/reports/methods.jsp Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A meta-analysis. Psychological Bulletin, 125, 367-383. Calhoun, R., Meischke, H., Hammerback, K., Bohl, A., Poe, P., Williams, B., & Phelan, E. A. (2011). Older adults’ perceptions of clinical fall prevention programs: A qualitative study. Journal of Aging Research, Article ID 867341, 7. doi:10.4061/2011/867341 Carlson, K. F., Nugent, S. M., Grill, J., & Sayer, N. A. (2010). Accuracy of external cause-of-injury coding in VA polytrauma patient discharge records. Journal of Rehabilitation Research & Development, 47, 689-698. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Disease, 40, 373-383. Clark, L., Thoreson, S., Goss, C. W., Marquez Zimmer, L., Marosits, M., & DiGuiseppi, C. (2011, April). Understanding fall meaning and context in marketing balance classes to older adults. Journal of Applied Gerontology. Retrieved from http://jag.sagepub.com/content/early/2011/03/31/0733464811 399896
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
933
Coben, J. H., Steiner, C. A., Barrett, M., Merrill, C. T., & Adamson, D. (2006). Completeness of cause of injury coding in healthcare administrative databases in the United States, 2001. Injury Prevention, 12(3), 199-201. Courtenay, W. H. (2000). Constructions of masculinity and their influence on men’s well-being: A theory of gender and health. Social Science and Medicine, 50(1), 1385-1401. Cumming, R. G., Miller, J. P., Kelsey, J. L., Davis, P., Arfken, C. L., Birge, S. J., & Peck W. A. (1991). Medications and multiple falls in elderly people: The St. Louis OASIS study. Age and Ageing, 20, 455-461. Deyo, R. A., Cherkin, D. C., & Ciol, M. A. (1992). Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. Journal of Clinical Epidemiology, 45, 613-619. de Groot, V., Beckerman, H., Lankhorst, G. J., & Bouter, L. M. (2003). How to measure comorbidity: A critical review of available methods. Journal of Clinical Epidemiology, 56(3), 221-229. Gillespie, L. D., Robertson, M. C., Gillespie, W. J., Lamb, S. E., Gates, S., Cumming, R. G., & Rowe, B. H. (2009). Interventions for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews, 2, Article CD007146. doi:10.1002/14651858.CD007146.pub2 Hanlon, J. T., Boudreau, R. M., Roumani, Y. F., Newman, A. B., Ruby, C. M., Wright, R. M., . . . Studenski, S. A. (2009). Number and dosage of central nervous system medications on recurrent falls in community elders: The health, aging and body composition study. Journals of Gerontology: Series A, 64A, 492-498. Hausdorff, J. M., Rios, D. A., & Edelberg, H. K. (2001). Gait variability and fall risk in community-living older adults: A 1-year prospective study. Archives of Physical Medicine and Rehabilitation, 82, 1050-1056. Healthcare Cost and Utilization Project Nationwide Inpatient Sample. (2005). Healthcare cost and utilization project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from www.hcup-us.ahrq.gov/nisoverview.jsp Healy, T. C., Peng, C., Haynes, M. S., McMahon, E. M., Botler, J. L., & Gross, L. (2008). The feasibility and effectiveness of translating a matter of balance into a volunteer lay leader model. Journal of Applied Gerontology, 27(1), 34-51. Retrieved from http://jag.sagepub.com/content/27/1/34 Kannus, P., Parkkari, J., Koskinen, S., Niemi, S., Palvanen, M., Jarvinen, M., & Vuori, I. (1999). Fall-induced injuries and deaths among older adults. Journal of the American Medical Association, 281, 1895-1899. Kannus, P., Parkkari, J., Niemi, S., & Palvanen, M. (2005). Fall-induced deaths among elderly people. American Journal of Public Health, 95, 422-424. Koehler, S. A., Weiss, H. B., Shakir, A., Shaeffer, S., Ladham, S., Rozin, L., & Wecht, C. H. (2006). Accurately assessing elderly fall deaths using hospital discharge and vital
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
934
Journal of Applied Gerontology 32(8)
statistics data. American Journal of Forensic Medicine and Pathology, 27(1), 30-35. Langley, J., Stephenson, S., Thorpe, C., & Davie, G. (2006). Accuracy of injury coding under ICD-9 for New Zealand public hospital discharges. Injury Prevention, 12(1), 58-61. LeMier, M., Cummings, P., & West, T. A. (2001). Accuracy of external cause of injury codes reported in Washington State hospital discharge records. Injury Prevention, 7, 334-338. Moore, M., Williams, B., Ragsdale, S., LoGerfo, J. P., Goss, J. R., Schreuder, A. B., & Phelan, E. A. (2010). Translating a multifactorial fall prevention intervention into practice: A controlled evaluation of a fall prevention clinic. Journal of the American Geriatrics Society, 58, 357-363. National Center for Injury Prevention and Control. (2006). Web-based injury statistics query and reporting system. Retrieved from http:/www.cdc.gov/ncipc/wisqars Paulozzi, L. J., Ballesteros, M. F., & Stevens, J. A. (2006). Recent trends in mortality from unintentional injury in the United States. Journal of Safety Research, 37, 277-283. Quan, H., Sundararajan, V., Halfon, P., Fong, A., Burnand, B., Luthi, J. C., . . . Ghali, W. A. (2005). Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Medical Care, 43, 1130-1139. Roudsari, B. S., Ebel, B. E., Corso, P. S., Molinari, N.-A. M., & Koepsell, T. D. (2005). The acute medical care costs of fall-related injuries among the U.S. older adults. Injury, 36, 1316-1322. Smith, G., & Langley, J. (1998). Drowning surveillance: How well do e-codes identify submersion fatalities? Injury Prevention, 4, 135-139. Smith, M. L., Ahn, S., Sharkey, J. R., Horel, S., Mier, N., & Ory, M. G. (2010, August). Successful falls prevention programming for older adults in Texas: Rural-urban variations. Journal of Applied Gerontology. Retrieved from http://jag.sagepub. com/content/early/2010/10/09/0733464810378407 Stagg, V. (2006). CHARLSON: Stata module to calculate Charlson Index of Comorbidity (Statistical software components S456719). Chestnut Hill, MA: Boston College Department of Economics. Steinman, B. A. (2008, May). Self-reported vision, upper/lower limb disability, and fall risk in older adults. Journal of Applied Gerontology. Retrieved from http:// jag.sagepub.com/content/27/4/406 Stevens, J. A., Corso, P. S., Finkelstein, E. A., & Miller, T. R. (2006). The costs of fatal and non-fatal falls among older adults. Injury Prevention, 12(5), 290-295. Stevens, J. A., Ryan, G., & Kresnow, M. S. (2006). Fatalities and injuries from falls among older adults—United States, 1993-2003 and 2001-2005. Morbidity & Mortality Weekly Report, 55, 1221-1224.
Downloaded from jag.sagepub.com at UNIV OF BRAZIL on April 30, 2014
Moudouni and Phillips
935
Stoeckel, K. J., & Porell, F. (2009). Do older adults anticipate relocating? The relationship between housing relocation expectations and falls. Journal of Applied Gerontology. Retrieved from http://jag.sagepub.com/content/29/2/231 Thomas, K. E., Stevens, J. A., Sarmiento, K., & Wald, M. M. (2008). Fall-related traumatic brain injury deaths and hospitalizations among older adults—United States, 2005. Journal of Safety Research, 39, 269-272. Tinetti, M. E., Baker, D. I., McAvay, G., Claus, E. B., Garett, P., Gottschalk, M., . . . Horwitz, R. I. (1994). A multifactorial intervention to reduce the risk of falling among elderly people living in the community. New England Journal of Medicine, 331, 821-827. Tinetti, M. E., Doucette, J., Claus, E., & Marottoli, R. (1995). Risk factors for serious injury during falls by older persons in the community. Journal of the American Geriatrics Society, 43, 1214-1221. Turner, C., & McClure, R. (2003). Age and gender differences in risk-taking behavior as an explanation for high incidence of motor vehicle crashes as a driver in young males. Injury Control and Safety Promotion, 10(3), 123-130.
Author Biographies Darcy K. McMaughan Moudouni, PhD, is an assistant professor and researcher for the Program in Aging, Disability and Long-Term Care Policy at the Texas A&M Health Science Center School of Rural Public Health. Charles D. Phillips, PhD, MPH, is a regent’s professor and head of the Program in Aging, Disability and Long-Term Care Policy at the Texas A&M Health Science Center School of Rural Public Health. Both authors are in the Department of Health Policy and Management.
