Abstract
Introduction
Hip fractures remain one of the most prevalent and deadly conditions afflicting those 65 years and older. For other health conditions (e.g. myocardial infarction), hospital location is associated with poorer health outcomes. To our knowledge, no study has investigated the relationship between hip fracture morality rate in the United States between urban and rural hospital settings.
Methods
A retrospective cohort study was conducted to examine differences in in-hospital mortality between groups treated in rural, urban-teaching, and urban-non-teaching hospitals, as well as public and private hospitals. Mortality rates were also compared for variances between surgical treatment, sex, insurance, patient location, race, and income. Discharge data was collected for 256,240 inpatient stays from the 2012 National Inpatient Sample (NIS), Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality.
Results
Odds of mortality were 14.6% greater in rural hospital hip fracture patients compared to Urban/Non-Teaching centers (p < 0.05).
Conclusions
Results from this study lend support to necessitate further research investigating prospective barriers to care of those in rural settings. This may point to limitations in resources and trained medical and surgical specialists in rural hospitals and need for continued research to mitigate such findings.
Keywords: Hip fracture, Mortality, Rural, Teaching hospitals, Social determinants
Level of evidence
Level III, Retrospective Cohort Study, Prognostic.
Introduction
An aging population constitutes a key demographic factor to healthcare in coming years. There are 76 million Americans born between 1946 and 1964, deemed the “Baby Boomers.“. By 2029 – at which time all Boomers will be over the age of 65 – this generation will comprise 20% of the US population.1 Improved management of chronic conditions, with a concomitant increase in obesity rates, is expected to increase the prevalence of arthritis, the need for lower extremity surgeries, and, notably, the occurrence of hip fractures.2 The incidence of hip fractures exponentially rises after the age of 70. This has been linked in past literature to the increased propensity for falls in the elderly. Thirty-percent of those over the age of 65 will experience one fall per year, which increases to 50% in those over 80 years old3 significantly increasing risk for hip fracture.
By 2050 the worldwide incidence for hip fractures is projected to rise to 6.26 million cases per year, with North America encompassing 70% of them.4 This poses profound socioeconomic and medical consequences. On average, $40,000 in direct medical costs are incurred in the first year following hip fracture, compounded by an additional $50,000 in indirect costs in subsequent years.5 However, the greatest costs are lives themselves. One-year mortality rate due to hip fracture is as high as 36% in elderly populations.6 Haentjens et al. conducted a MEDLINE and EMBASE meta-analysis review of 39 adult cohorts and reported adults 80 years and older have a 5- to 8-fold increased risk for all-cause mortality during the first three months after hip fracture compared to those of younger populations.7 Extraordinary mortality following hip fracture in the elderly may be linked to secondary complications including pulmonary embolism, infection, heart failure, surgical trauma, and complications secondary to surgery.8 In spite of this, there is some evidence that hospital type may strongly influence patient outcomes and mortality rates following hip fracture.
Hospital type can be characterized by size, geographic location, personnel employed (i.e. teaching versus non-teaching), and funding (e.g. for-profit versus non-profit). One study investigated hospital type (e.g. government funded, non-profit, minor teaching, and major teaching) and its influence on four-week and one-year survival following stroke, coronary artery disease, congestive heart failure, and hip fracture. It revealed all disease prognoses were worst at for-profit non-teaching hospitals, however differences were most notable in those with hip fractures.9 Other studies observed that prognosis at teaching hospitals following hip fracture is better in both Canadian10 and United States2 hospitals. However, to our knowledge, no study has investigated the relationship between hip fracture morality rate or hospital load in the United States between urban and rural settings.
With an imminent upsurge in hip fractures on the horizon, it is essential to extrapolate similar correlations for health disparities following hip fracture amongst hospital settings in an effort to truncate accompanying morbidity and mortality. Thus, our aim is to determine differences in hip fracture mortality amongst hospital types (size, location and teaching vs. non-teaching, public vs. private). Specifically, we compared rural, urban-teaching, and urban-non-teaching hospitals for differences in patient outcomes. We also investigated mortality reduction by various surgical intervention common to hip fractures (e.g. open reduction internal fixation versus total hip arthroplasty).
Materials and methods
A retrospective cohort study was conducted to examine differences in in-hospital mortality between groups treated in rural, urban-teaching, and urban-non-teaching hospitals, as well as public and private hospitals. Mortality rates were also compared for variances between surgical treatment, sex, insurance, patient location, race, and income. Discharge data was collected from the 2012 National Inpatient Sample (NIS), Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality.11 The 2012 HCUP NIS database includes 7,296,968 discharge cases. The first diagnosis code generally corresponds to the primary diagnosis for hospital stay. As such, discharge records included up to 25 ICD-9 Diagnosis codes. Discharges included patients 65 years and older with a diagnosis of fractures of the femoral neck and all subclassifications including transcervical fracture closed and open, pertrochanteric fracture of femur closed and open, and closed and open fracture of unspecified part of the neck of the femur (ICD-9: 820, 820. x, and 820. xx). The use of NIS data required a weighting procedure to make reliable national estimates. Collectively, the final sample size was 256,240. All results reflect the weighting technique.12
Dependent variables
The dependent variable is a measure of in-hospital mortality, coded 0 = No, and 1 = Yes.
Independent variables
Hospital Location/Teaching Status is coded 1 = Rural, 2 = Urban-Non-Teaching, and 3 = Urban-Teaching. Hospital Control/Ownership is coded 1 = Public, 2 = Private/Non-Profit, and 3 = Private/Investor-Owned.
The HCUP NIS includes 15 Healthcare Common Procedure Coding System (HCPCS) codes in separate fields. For comparing among procedure type, five procedures were included reflecting those that required emergent repair of the hip: 1 = Code 7855: Internal fixation of bone without fracture reduction, femur; 2 = Code 7915: Closed reduction of fracture with internal fixation, femur; 3 = Code 7935: Open reduction of fracture with internal fixation, femur; 4 = Code 8151: Total hip replacement; 5 = Code 8152: Partial hip replacement.
A sixth category, titled “All Other Procedures”, was utilized as a reference category, made up of the following procedure codes: 0070: Revision of hip replacement, both acetabular and femoral components, 0071: Revision of hip replacement, acetabular component, 0072: Revision of hip replacement, femoral component, 0073: Revision of hip replacement, acetabular liner and/or femoral head only, 0074: Hip bearing surface, metal-on-polyethylene, 0075: Hip bearing surface, metal-on-metal, 0076: Hip bearing surface, ceramic-on-ceramic, 0077: Hip bearing surface, ceramic-on-polyethylene, 0085: Resurfacing hip, total, acetabulum and femoral head, 0086: Resurfacing hip, partial, femoral head, 0087: Resurfacing hip, partial, acetabulum, 7975: Closed reduction of dislocation of hip, 7985: Open reduction of dislocation of hip, 8005: Arthrotomy for removal of prosthesis without replacement, hip, 8015: Other arthrotomy, hip, 8025: Arthroscopy, hip, 8035: Biopsy of joint structure, hip, 8045: Division of joint capsule, ligament, or cartilage, hip, 8075: Synovectomy, hip, 8085: Other local excision or destruction of lesion of joint, hip, 8095: Other excision of joint, hip, 8121: Arthrodesis of hip, 8140: Repair of hip, not elsewhere classified, 8153: Revision of hip replacement, not otherwise specified, 8312: Adductor tenotomy of hip, and 8418: Disarticulation of hip.
Control variables for the study were as follows: Age in years at admission, sex coded 0 = Female, 1 = Male, Number of Chronic Conditions on admission, Insurance Status coded as 1 = Private Insurance/HMO, 2 = Medicaid, 3 = Medicare, Patient Location is coded as 1 = “Central” counties of metro areas of≥1 million population, 2 = “Fringe” counties of metro areas of≥1 million population, 3 = Counties in metro areas of 250,000–999,999 population, 4 = Counties in metro areas of 50,000–249,999 population, 5 = Micropolitan counties, and 6 = Not metropolitan or micropolitan counties, Race coded 1 = Caucasian, 2 = African American, 3 = Hispanic, 4 = Asian/Pacific Islander, 5 = Native American, and Income categorized in the NIS as median household income national quartile for patient Zip Code. 1 = 1st Quartile (lowest income), 2 = 2nd Quartile, 3 = 3rd Quartile, and 4 = 4th Quartile (highest income).
Statistical analysis
SPSS ® 24.013 was used for all data analyses. Table 1 provides a summary of discharges included in the sample. Chi-square analyses were performed to compare mortality rates between groups based on hospital location/teaching status, hospital control/ownership, procedure type, sex, insurance status, patient location, race, and income (Table 2). Binary logistic regression was used to determine influence of hospital type on mortality while controlling for several socio-demographic controls (Table 3).
Table 1.
Characteristics of Hip Fracture Patient Discharges, n = 256,240.
| DEPENDENT VARIABLES | |
|---|---|
| Died during hospitalization, # (%) | 6395 (2.5) |
| INDEPENDENT VARIABLES | |
| Age, mean (SD) | 82.20 (7.12) |
| No. of Chronic Conditions, mean (SD) | 5.85 (2.98) |
| Procedure Type | |
| All other procedures, # (%) | 33,235 (13.0) |
| Partial hip replacement, # (%) | 76,605 (29.9) |
| Open reduction of fracture w/internal fixation, # (%) | 66,005 (25.8) |
| Closed reduction of fracture w/internal fixation, # (%) | 54,295 (21.2) |
| Internal fixation of bone without fracture reduction, # (%) | 17,790 (6.9) |
| Total hip replacement | 8310 (3.2) |
| Female sex, # (%) | 186,110 (72.6) |
| Primary Payer | |
| Medicare, # (%) | 236,110 (92.1) |
| Medicaid, # (%) | 2140 (0.8) |
| Private/HMO, # (%) | 13,500 (5.3) |
| Patient Location | |
| Central counties of metro areas≥1 million population, # (%) | 61,120 (23.9) |
| Fringe counties of metro areas of≥1 million population, # (%) | 60,415 (23.6) |
| Counties in metro areas of 250,000 to 999,999 population, # (%) | 52,725 (20.6) |
| Counties in metro areas of 50,000 to 249,999, # (%) | 25,945 (10.1) |
| Micropolitan counties, # (%) | 32,325 (12.6) |
| Not metropolitan or micropolitan, # (%) | 23,265 (9.1) |
| Race | |
| Caucasian, # (%) | 211,430 (82.5) |
| African American, # (%) | 8775 (3.4) |
| Hispanic, # (%) | 11,725 (4.6) |
| Asian/Pacific Islander, # (%) | 4175 (1.6) |
| Native American, # (%) | 1195 (0.5) |
| Median Household Income for Zip Code | |
| Lowest Income Quartile, # (%) | 67,220 (26.2) |
| 2nd Income Quartile, # (%) | 65,235 (25.5) |
| 3rd Income Quartile, # (%) | 60,945 (23.8) |
| Highest Income Quartile, # (%) | 58,280 (22.7) |
| Hospital Location/Teaching Status | |
| Rural, # (%) | 38,635 (15.1) |
| Urban/Non-teaching, # (%) | 114,910 (44.8) |
| Urban/Teaching, # (%) | 102,695 (40.1) |
| Hospital Control/Ownership | |
| Public, # (%) | 26,590 (10.4) |
| Private/Non-profit, # (%) | 191,925 (74.9) |
| Private/Investor owned, # (%) | 37,725 (14.7) |
Table 2.
Comparison of percent mortality observed in various groups.
| Hospital Location/Teaching Status | Rural | Urban/Non-teaching | Urban/Teaching | Pearson Chi-square, df, p | |||
|---|---|---|---|---|---|---|---|
| % Mortality | 3.0* | 2.3 | 2.5 | χ = 50.9, df = 2, p < 0.001 | |||
| * Mortality rate for Rural Hospital Location group is statistically different from other groups based on Hospital Location at p < 0.05 | |||||||
| Hospital Control/Ownership | Public | Private/Non-profit | Private/Investor Owned | Pearson Chi-square, df, p | |||
| % Mortality | 2.5 | 2.3* | 2.5* | χ = 9.46, df = 2, p = 0.009 | |||
| * Mortality rates for Private/Non-profit and Private/Investor Owned significantly differ at p < 0.05, but each does not differ from the Public control hospital group | |||||||
| Procedure Type | All other procedures | Partial hip replacement | Open reduction of fracture w/internal fixation | Closed reduction of fracture w/internal fixation | Internal fixation of bone without fracture reduction | Total hip replacement | Pearson Chi-square, df, p |
| % Mortality | 6.2* | 2.2 | 1.9 | 1.8 | 1.4 | 1.9 | χ = 2,235, df = 5, p < 0.001 |
| * Mortality rate for All Other Procedures groups is statistically different from all other groups based on Procedure Type at p < 0.05, | |||||||
| Sex | Male | Female | Pearson Chi-square, df, p | ||||
| % Mortality | 3.9* | 2.0 | χ = 767, df = 1, p < 0.001 | ||||
| Primary Payer | Medicare | Medicaid | Private/HMO | Pearson Chi-square, df, p | |||
| % Mortality | 2.5 | 1.2* | 2.2 | χ = 19.5, df = 25, p < 0.001 | |||
| * Mortality rate for Medicaid group is statistically different from all other Primary Payer groups at p < 0.05 | |||||||
| Patient Location | Central counties of metro areas≥1 million population | Fringe counties of metro areas of≥1 million population | Counties in metro areas of 250,000 to 999,999 population | Counties in metro areas of 50,000 to 249,999 | Micropolitan counties | Not metropolitan or micropolitan | Pearson Chi-square, df, p |
| % Mortality | 2.1* | 2.2* | 2.8 | 2.6 | 2.9 | 2.9 | χ = 122, df = 5, p < 0.001 |
| * Mortality rates for two indicated groups based on Patient Location are not statistically different, but each is statistically different from the other four groups based on Patient Location at p < 0.05 | |||||||
| Race | Caucasian | African American | Hispanic | Asian/Pacific Islander | Native American | Pearson Chi-square, df, p | |
| % Mortality | 2.5 | 2.7 | 1.9* | 1.7* | 3.3 | χ = 32.0, df = 4, p < 0.001 | |
| * Mortality rates for two indicated groups based on Race are not statistically different, but each is statistically different from the other three groups based on Race at p < 0.05 | |||||||
| Median Household Income by Zip Code | Lowest Income Quartile | 2nd Income Quartile | 3rd Income Quartile | Highest Income Quartile | Pearson Chi-square, df, p | ||
| % Mortality | 2.5 | 2.8 | 2.7 | 2.0* | χ = 79.9, df = 3, p < 0.001 | ||
| * Mortality rate for Highest Income group is significantly different from three other income groups at p < 0.05 | |||||||
Table 3.
Binary logistic regression predicting in-hospital mortality.
| Independent Variables | Regression Coefficient β (S.E) | Odds ratio (95% CI) |
|---|---|---|
| Hospital Location/Teaching Status (Reference: Rural) | ||
| Urban/Non-teaching | −0.158 (0.058)* | 0.854 (0.762–0.958) |
| Urban/Teaching | −0.006 (0.058) | 0.995 (0.888–1.114) |
| Hospital Control/Ownership (Reference: Public) | ||
| Private/Non-profit | 0.041 (0.047) | 1.014 (0.950–1.141) |
| Private/Investor owned | 0.058 (0.058) | 1.059 (0.946–1.186) |
| Age | 0.059 (0.002)*** | 1.06 |
| No. of Chronic Conditions | 0.142 (0.004)*** | 1.15 |
| Procedure Type (Reference: All other procedures) | ||
| Partial hip replacement | −1.139 (0.037)*** | 0.320 (0.298–0.344) |
| Open reduction of fracture w/internal fixation | −1.172 (0.039)*** | 0.310 (0.287–0.335) |
| Closed reduction of fracture w/internal fixation | −1.314 (0.043)*** | 0.269 (0.247–0.293) |
| Internal fixation of bone without fracture reduction | −1.501 (0.072)*** | 0.223 (0.194–0.257) |
| Total hip replacement | −0.0928 (0.086)*** | 0.395 (0.334–0.468) |
| Male (Reference: Female) | 0.651 (0.028)*** | 1.917 (1.814–2.025) |
| Primary Payer (Reference: Medicare) | ||
| Medicaid | −0.600 (0.229)** | 0.549 (0.350–0.861) |
| Private/HMO | −0.010 (0.066) | 0.990 (0.870–1.127) |
| Patient Location (Reference: Not metro/micropolitan) | ||
| Central counties of metro areas≥1 million population | −0.501 (0.065)*** | 0.606 (0.533–0.688) |
| Fringe counties of metro areas of≥1 million population | −0.248 (0.066)*** | 0.780 (0.686–0.887) |
| Counties in metro areas of 250,000 to 999,999 population | −0.090 (0.062) | 0.914 (0.809–1.033) |
| Counties in metro areas of 50,000 to 249,999 | −0.056 (0.068) | 0.946 (0.827–1.081) |
| Micropolitan counties | −0.069 (0.058) | 0.933 (0.832–1.046) |
| Race (Reference: Caucasian) | ||
| African American | 0.163 (0.071)* | 1.177 (1.024–1.354) |
| Hispanic | −0.209 (0.071)** | 0.811 (0.705–0.933) |
| Asian/Pacific Islander | −0.351 (0.0124)* | 0.704 (0.552–0.898) |
| Native American | 0.245 (0.175) | 1.277 (0.906–1.800) |
| Median Household Income (Reference: 4th Quartile) | ||
| Lowest Income Quartile | 0.291 (0.047)*** | 1.338 (1.219–1.468) |
| 2nd Income Quartile | 0.363 (0.045)*** | 1.438 (1.317–1.569) |
| 3rd Income Quartile | 0.387 (0.043)*** | 1.473 (1.355–1.600) |
*p < 0.05, **p < 0.01, ***p < 0.001.
Results
Table 1 summarizes patient characteristics for those included in this study. The cohort is predominately female (72.6%), from metropolitan areas with greater than 1 million population (47.5%), had Medicare (92.1%), and were Caucasian (82.5%). The majority of patients sought care at Urban non-teaching hospitals (44.8%) and Private non-profit hospitals (74.9%). The most common procedure fracture repair was partial hip replacement (hemiarthroplasty), constituting 29.9% of all surgeries. The mortality rate for this cohort was 2.5%.
Chi-square analysis was used to determine differences in mortality rates amongst groups based on hospital type (ownership and location/teaching status), procedure type, patient sex, primary insurance payer, patient location, race, and income (Table 2). Regarding Hospital Location/Teaching status, mortality rate following hip fracture observed in Rural hospitals was 3.0%, statistically greater (p < 0.05) than the percent mortality observed in Urban/Non-Teaching (2.3%) and Urban/Teaching (2.5%) hospitals. The mortality rate at Private/Non-Profit hospitals was 2.3%, significantly lower (p < 0.05) than the 2.5% mortality observed at Private/Investor-Owned hospitals. The mortality rate for All Other Procedures (6.2%) to treat hip fractures was statistically greater (p < 0.05) than rates for partial hip replacement (2.2%), open reduction of fracture with internal fixation (ORIF) (1.9%), closed reduction of fracture with internal fixation (CRIF) (1.8%), internal fixation of bone without fracture reduction (1.4%), and total hip replacement (THA) (1.9%) (chi-square = 2235, p < 0.001).
Mortality in central counties of metro areas with at least 1 million residents (2.1%), and in fringe counties of metro areas with at least 1 million residents (2.2%) was lower than in smaller geographic regions (p < 0.05). Mortality in Hispanics (1.9%) and Asian/Pacific Islanders (1.7%) was lower than the mortality rate for Caucasian, African American, and Native American populations (p < 0.05). Using zip code for household income stratification revealed those in the highest (4th) income quartile experienced a significantly lower percent mortality (2%) than those of lower income quartiles (p < 0.05).
Finally, binary logistic regression (Table 3) revealed that Hospital Control/Ownership did not influence patient mortality. We observed no difference in the odds of mortality between Private/Non-Profit, or Private/Investor-Owned hospitals, compared to Public hospitals. However, we found that patients at Urban/Non-Teaching hospitals had 14.6% lower odds of death (OR = 0.854, 95% CI: 0.762–0.958) compared to patients at Rural hospitals. Regarding procedure type, the odds of mortality were lower for all procedures ranging from 40% lower for THA (OR = 0.395, 95% CI: 0.334–0.468) to 78% lower for internal fixation of bone without fracture reduction (OR = 0.223, 95%: 0.194–0.257) compared to All Other Procedures.
With respect to control variables, for every one unit (year) increase in age, patients experience about 6% increased odds of mortality (b = 0.059, OR = 1.06, p < 0.001), while for every one unit increase in the number of chronic conditions, the odds of mortality increase by 15% (b = 0.142, OR = 1.15, p < 0.001). Males experience 92% increased odds of mortality compared to females (OR = 1.917, 95% CI:1.814–2.025). African Americans experienced 17% increased odds of mortality compared to Caucasian patients (OR = 1.177, 95% CI: 1.024–1.354), while Hispanics (OR = 0.811, 95% CI: 0.705–0.933) and Asian/Pacific Islanders (OR = 0.704, 95% CI: 0.552–0.898) experienced 19% and 30% lower odds of mortality, respectively. Finally, regarding income, mortality increased as income quartile dropped. Odds of mortality for the lowest (1st) income quartile were 34% (OR = 1.338, 95% CI: 1.219–1.468) greater compared to the highest (4th) income quartile.
Discussion
Clinical significance
Aging is a natural, irreversible process, demonstrating profound physiological and pathological manifestations. These changes bring an increased risk for falls. The mortality rate of falls increases in both sexes and all racial ethnic groups in those over age 75, accounting for 70% of accidental deaths.14 Brauer et al. reported that from 1986 to 2005 hip fractures accounted for 20% of Medicare hospital claims of those 65 years and older, with nearly 1 in 100 Americans experiencing hip fracture in their lifetime.6 Cooper et al. documented that one year following hip fracture, 40% of patients are still unable to walk independently, 60% have difficulty with essential activities of daily living, 80% are restricted in other activities (e.g. cooking, driving), and 27% enter nursing homes for the first time.15 With a surging senior citizen population, these issues become even more pressing, further necessitating the identification and mitigation of post-fracture variables leading to high morbidity and mortality in the elderly.
Hospital location/type
Differences in mortality rates between Urban/Teaching and Rural hospital can be traced to the tendency of complex cases in rural sites to be transferred to teaching settings, as seen with a study investigating patients following cervical spine surgery.16 In a study of heart failure patients, teaching hospitals had more experienced physicians and nurses, standardized algorithms for care, and better access to resources for patient education and follow-up.17 In addition, rural hospitals had reduced staffing (particularly on weekends) leading to incomplete handover between caregivers at shift changes and diminished physician cross-coverage, as well as limited support services and reduced community services (e.g. home care) for patients following discharge.17
This study provides evidence that patients presenting with hip fractures may also demonstrate lower mortality rates at both Teaching and Non-Teaching Urban hospitals. Compared to Rural hospitals, there was a 14.6% reduction in odds of mortality at Urban/Non-Teaching hospitals. Urban settings offer closer proximity to care and greater clinician to population ratios.10 As hip fractures are an acute medical emergency, optimal outcomes depend on quick diagnosis and appropriate treatment. Rapid repair of the fracture requires access to a trauma center and specialists. However, access to either of these may be limited in rural settings as 34.8 million Americans do not have access to a trauma center within a 1-h drive of their home.18 While we cannot definitively deduce the reason for increased differences in mortality in patients treated at rural versus urban non-teaching and teaching settings, access to resources, physician cross-over, incomplete handover at shift change, and limited services following discharge are probable factors.
Hospital ownership
Literature has demonstrated health outcome differences between hospitals based on ownership type (i.e. Private versus Public). Some studies have attributed these differences to variance in patient selection, arguing that Public hospitals tend to treat populations that are older, have lower socioeconomic status, lead riskier lifestyles, and have higher levels of comorbidities than those attending private hospitals.19,20 Although we observed no differences between Public and Private hospitals overall, there were significant differences (p < 0.05) in mortality between Private Non-Profit (2.5%) and Private Investor (2.3%) funded hospitals. This parallels past literature. A meta-analysis by Devereaux et al., that studied 38 million patients with various conditions from 26,000 Canadian hospitals, demonstrated a significantly increased risk for death at Private/Investor hospitals compared to Private/Non-Profit (RR = 1.095, CI = 1.050–1.141, p < 0.0001).21
Schlesinger et al. demonstrated investor hospitals have higher prices, higher incidence of nursing home adverse events, barriers to indigent patients, and higher mortality rates in both hospitals and dialysis facilities, compared to private non-profits.22 Differences in hip fracture mortality may be due to the Private/Non-Profit hospitals’ slower to tendency for change in services when demands fluctuate. Schlesinger et al. also discussed how Private/Non-Profits classically utilize services not already on the markets. This leads to innovation of programs not yet available at Private/Investor hospitals (e.g. Healthcare Maintenance Organizations during the 1930s and Hospice in the 1980s).22 This study yields data supporting variances in hip fracture mortality may be witnessed by private hospital ownership.
Procedures
We observed lower rates of hip fracture mortality (ranging from 1.4% to 2.2% by Internal Fixation of Bone Without Fracture Reduction, Closed Reduction of Fracture with Internal Fixation, Open Reduction of Fracture with Internal Fixation, Partial Hip Replacement, and Total Hip Replacement compared to the mortality rate for combined measure of All Other Procedure types (6.3%). The increased odds of mortality for “All Other Procedures” suggests that these are used for more severe fractures require increasingly invasive and less often practiced procedures. Lower volumes of higher complexity cases may potentiate morbidity and mortality of approaches listed in “All Other Procedures”. Additionally, increased odds of mortality could be due to various confounding variables: multiple injuries upon presentation, complicated fracture mechanism, patient comorbidities, surgeon case volume, and iatrogenic issues. Overall, surgical modality for hip fracture should be selected based on a number of elements, including but not limited to patient specific factors, age at presentation, and fracture type.
Limitations
One significant limitation to this study is the use of ICD coding for case identification. Lost in the use of ICD codes is insight into complexity of the case. In addition, variances demonstrated by hospital ownership may, in part, be affected by financial factors. Finally, we aggregated a number of procedures into our combined “All Other Procedures” variable that might otherwise be quite different from each other. Future studies could attempt to study each of these procedures separately to determine differences among these procedures as well as between these and the five most common procedures we studied. Further investigation of mortality differences among individual procedures may reveal other mortality differences and provide evidence to better address these differences.
Conclusion
Of 256,240 patients studied in this investigation, we observed 14.6% increased odds of mortality for hip fracture patients treated in Rural hospitals compared to Urban/Non-Teaching hospitals (p < 0.05). We also observed decreased odds of mortality for the five most common procedures compared to all other procedure types. These results may provide additional evidence that suggests barriers to care in rural settings. Future research should focus on the types of resources that are lacking in rural settings to determine better policy initiatives and resource allocation to rural hospitals. Ultimately, better policy and resource allocations may address the hip fracture mortality disparity between urban and rural hospitals.
CRediT authorship contribution statement
Brendan J. Farley: Writing - original draft, Writing - review & editing, Visualization, Supervision. Brian M. Shear: Writing - original draft, Writing - review & editing, Visualization, Supervision. Vivian Lu: Formal analysis, Writing - review & editing, Investigation. Kyla Walworth: Formal analysis, Investigation. Kevin Gray: Formal analysis, Investigation. Matt Kirsch: Formal analysis, Investigation. John M. Clements: Conceptualization, Methodology, Formal analysis, Validation, Investigation, Software, Resources, Data curation, Writing - review & editing, Visualization, Supervision, Project administration.
Decalaration of competing interest
The authors report no conflicts of interest and no funding
Acknowledgements
None.
References
- 1.Fry R. US Dept of Health and Human Services National Center for Health Statistics; 2016. Millenials Overtake Baby Boomers as America's Largest Population. s.L. [Google Scholar]
- 2.The Effect of Hospital Type and Surgical Delay on Mortality after Surgery for Hip Fractures . Weller I, Wai K, Jaglal S, Kreder J. s.l. JBJS, 2005, Vol. vol. 87. [DOI] [PubMed]
- 3.Cvecka J., Tirpakova V., Sedliak M. Physical activity in elderly. Eur J Transl Myol. 2015;25:249–252. doi: 10.4081/ejtm.2015.5280. 4, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kannus P., Parkkari J., Sievanen H. Epidemiology of hip fractures. Bone. 1996;18:57–63. doi: 10.1016/8756-3282(95)00381-9. 1, s.l. [DOI] [PubMed] [Google Scholar]
- 5.Burge R., Dawson-Hughes B., Solomon D. Incidence and economic burden of osteoporosis related fractures in the United States, 2005-2025. J Bone Miner Res. 2007;22:465–475. doi: 10.1359/jbmr.061113. s.l. [DOI] [PubMed] [Google Scholar]
- 6.Brauer C., Coca-Perraillon M., Cutler D. Incidence and mortality of hip fractures in the United States. J Am Med Assoc. 2009;302:1573–1579. doi: 10.1001/jama.2009.1462. 14, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Haentjens P., Magaziner J., Colon-Emeric C. Meta-analysis: excess mortality after hip fracture among older women and men. Annala of Internal Medicine. 2010 doi: 10.1059/0003-4819-152-6-201003160-00008. s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Panula J., Pihlakamaki H., Mattila V. Mortality and cause of death in hip fracture patients aged 65 or older - a population-based study. BMC Msk Disorders. 2011;12 doi: 10.1186/1471-2474-12-105. 105, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Taylor D., Whellan D., Sloan F. Effects of admission to a teaching hospital on the cost and quality of care for medicare beneficiaries. N Engl J Med. 1999;340:293–299. doi: 10.1056/NEJM199901283400408. s.l. [DOI] [PubMed] [Google Scholar]
- 10.The Effect of Hospital Type and Surgical Delay on Mortality after Surgery for Hip Fracture. Weller I, Wai E, Jaglal S, Kreder H. 3, s.l. : JBJS, 2005, Vols. 87-B. [DOI] [PubMed]
- HCUP . Agency for Healthcare Research and Quality; Rockville, MD: 2012. Healthcare cost and utilization Project. HCUP National Inpatient Sample (NIS) [Google Scholar]
- 12.Houchens R., Ross D., Elixhauser A. US Agency for Healthcare Research and Quality; Rockville, MD: 2015. Final Report on Calculating National Inpatient Sample (NIS) Variances for Data Years 2012 and Later. UCUP Methods Series Report #2015-09ONLINE. [Google Scholar]
- 13.Corp, IBM . IBM Corp; Armonk, NY: 2016. IBM SPSS Statistics for Windows. [Google Scholar]
- 14.Fuller G. Falls in the elderly. Am Fam Physician. 2000;1:2159–2168. 61-7, s.l. [PubMed] [Google Scholar]
- 15.Cooper C. The crippling consequences of fractures and their impact on quality of life. Am J Med. 1997;103:12–19. doi: 10.1016/s0002-9343(97)90022-x. 2-1, s.l. [DOI] [PubMed] [Google Scholar]
- 16.Outcomes Of Cervical Spine Surgery in Teaching and Non-teaching Hospitals. Fineberg S, Oglesby M, Patel A, et al 13, s.l: Spine (Phila Pa 1976), 2013, Vol. vol. 38, pp. 1089-1096. [DOI] [PubMed]
- 17.McAlister F., Anita G., Sumit M. Postdischarge outcomes in heart failure are better for teaching hospitals and weekday discharges. Circulation. 2013;6 doi: 10.1161/CIRCHEARTFAILURE.113.000336. 5, s.l. [DOI] [PubMed] [Google Scholar]
- 18.Hsia R., Shen Y. Possible geographical barriers to trauma centers access for vulnerable patients in the United States: an analysis of urban and rural communities. Arch Surg. 2011;146:46–52. doi: 10.1001/archsurg.2010.299. 1, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Holom G., Hagen T. Quality differences between private for-profit, private non-profit and public hospitals in Norway: a retrospective national register-based study of acute readmission rates following total hip and knee arthroplasties. BMJ Open. 2017;7 doi: 10.1136/bmjopen-2016-015771. 8, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tynkkynen L., Vrangbaek K. Comparing public and private providers: a scoping review of hospital services in Europe. BMC Health Serv Res. 2018;18:141. doi: 10.1186/s12913-018-2953-9. 1, s.l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Devereaux P., Peter T., Lacchetti C. A systematic review and meta-analysis of studies comparing mortality rates of private for-profit and private not-for-profit hospitals. CMAJ (Can Med Assoc J) 2002;166:1399–1406. 11, s.l. [PMC free article] [PubMed] [Google Scholar]
- 22.Schlesinger M., Bradford H. How nonprofits matter in American medicine, and what to do about it. Health Aff. 2006;25 doi: 10.1377/hlthaff.25.w287. 1, s.l. [DOI] [PubMed] [Google Scholar]