Abstract
Objectives
Delayed fixation and frailty are established risk factors for morbidity and mortality in patients undergoing fixation for intertrochanteric (IT) hip fractures. These variables, however, have been considered independently. Is frailty an independent risk factor for delayed fixation in patients with IT fractures?
Methods
A retrospective review of prospectively collected cross-sectional data was performed using the National Inpatient Sample (NIS) database from 2015Q4 to 2021. Diagnostic and procedural ICD-10 codes were used to identify 16338 patients at least 60 years old who underwent a surgical intervention for an IT fracture. Frailty scores were calculated using the 11-point Modified Frailty Index (11-mFI) and the 19-point Charlson Comorbidity Index (CCI). Delayed fixation was defined as surgeries delayed more than 48 hours from admission and occurred in 866 (5.3%) of encounters. Multivariable logistic regression was used to assess associations between 11-mFI or CCI and delayed fixation.
Results
The probability of delayed fixation increased 41% for each integer increase in 11-mFI score (aOR 1.41, 95% CI: 1.33–1.51) and 17% for each integer increase in CCI score (aOR 1.17, 95% CI: 1.14–1.21). The median cost of admission was $15,209.17 (IQR $11,960.82–$19,995.16) for patients whose surgery was within 48 hours and $20,592.03 (IQR $15,688.88–$29,375.30) for patients with delayed fixation.
Conclusions
Two frailty scores, 11-mFI and CCI, were independent predictors of delayed fixation for adult patients with IT fractures. Frailty indices can be useful tools in risk assessment, shared decision-making, and multidisciplinary coordination of care.
Introduction
Hip fractures are a significant cause of morbidity and mortality, with recent estimates of a 15–36% mortality rate at one-year post-fracture.1-4 The management of these patients has focused on minimizing surgical delay, defined as surgery performed more than 48 hours after admission. 5 A meta-analysis of 31424 patients of at least 60 years old with hip fractures reported a 20% reduced mortality risk within 12 months in patients operated on within 48 hours. 6 Moreover, IT fractures are estimated to account for 42.5% of all hip fractures and contribute $2.63 billion annual direct medical costs to the United States (U.S.) healthcare system and, therefore, warrant focused investigations into treatment delays. 7
Frailty, defined as progressively diminished physiologic reserve, has been associated with increased risk of falls8-10 and adverse surgical outcomes.11,12 Several studies have evaluated frailty indices as predictors of poor outcomes in hip fracture patients.13-16 A large hip fracture registry study reported the 11-mFI independently predicted functional dependence at discharge. 17 De Haan et al. prospectively validated the original and adjusted CCI in hip fracture surgery and reported moderate discrimination (AUCs ∼0.67–0.72) for 30-day and 1-year mortality. 18 While these studies suggest the 11-mFI and CCI relate to postoperative functional outcomes and mortality, none have evaluated these indices as predictors of surgical delay in IT fractures.6,19,20 This study aims to examine the ability of the 11-mFI and the CCI in predicting risk of delayed surgical fixation IT fractures using a nationally representative database.
Materials and Methods
Database Characteristics
Patient hospital encounters were identified using the 2015 fourth quarter (Q4) through 2019 data from the Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS). The NIS is the largest all-payor inpatient database in the U.S. and approximates 20% of all hospital discharges, representing 7 million hospital stays annually. This timeline was chosen to capture the years that NIS began using the International Classification of Diseases, 10th revision, Clinical Modification (ICD-10-CM) codes. All facets of the HCUP Nationwide Database Data Use Agreement were maintained. This study was reviewed and approved by the New York Medical College Institutional Review Board, study #23970.
Patient Selection
A total of 16 338 initial hospital encounters were identified for patients at least 60 years of age21,22 with ICD-10-CM codes for IT fracture and ICD-10-PCS codes for insertion of intramedullary internal fixation devices (Appendix A). Patient characteristics included age, sex, race, income quartile, and primary payor.8,9 The mean patient age was 80.5 years (SD 9 years). Patients were 68.4% female and 84.3% White. The largest payor was Medicare, 87.6%. Comorbidities assessed included obesity as a binary variable. Use of long-term anticoagulation therapy and receipt of echocardiogram during the admission were captured as clinically relevant risk factors for morbidity.23,24 In total, 1.5% of patients required an echocardiogram and 17.4% were on long-term anticoagulation therapy. Of patients who experienced surgical delay, 2.9% required an echocardiogram and 29.0% were on long-term anticoagulation.
Income quartile was based on median household income by ZIP code. The most common income quartile was the second (26th–50th percentile), 26.4%. The institutional characteristics included hospital region, teaching and rural status, and bed size. Almost half of the cases were treated in large bed-size hospitals (45.5%) in the South (41.8%). Sixty-three percent of patients were treated at teaching hospitals in an urban setting. Median cost of admission was $15,209.17 (IQR $11,960.82–$19,995.16) for patients whose surgery was within 48 hours and $20,592.03 (IQR $15,688.88–$29,375.30) for patients with delayed fixation.
Frailty
The 11-Item Modified Frailty Index (11-mFI) is Unweighted, With a Single Point for Each Variable, Whereas the Modified Charlson Comorbidity Index (mCCI) is a Weighted Score With Higher Points for Medical Comorbidities Such as Cancer and is Composed of Slightly Different Variables
11-mFI: 11-item Modified Frailty Index mCCI: modified Charlson Comorbidity Index.
HTN: Hypertension.
CHF: Congestive heart failure.
COPD: Chronic pulmonary obstructive disease.
PNA: Pneumonia.
TIA: Transient ischemic attack.
Adult Patients Over 60 years of Age With Intertrochanteric Hip Fractures Identified in the National Inpatient Sample 2015Q4–2019 Were Stratified by Surgery Within (n = 15 472) or More Than 48 hours (n = 866) From Admission. Patient and Hospital Characteristics and Frailty Scores for the Two Groups Were Compared
*NIS and HCUP guidelines require that number counts are not published for cells containing less than 10 patients count.
11-mFI: 11-item Modified Frailty Index.
mCCI: modified Charlson Comorbidity Index.
AIDS: Acquired Immunodeficiency Syndrome.
Outcomes
The primary outcome was delay to surgery defined as greater than 48 hours from admission to fixation. As NIS lacks patient identifiers or outcomes beyond the acute care episode, only in-hospital outcomes were captured.
Statistical Analysis
Patient characteristics were summarized by delayed fixation status using counts (percentages) and means (standard deviations) and compared groups with Pearson Chi-square tests (categorical) and one-way ANOVA (continuous) to describe the study population. Multivariable logistic regression models were fit with frailty (11-mFI or CCI) as the primary predictor and delayed fixation as the dependent variable. Covariates were selected based on clinical relevance and included age, sex, race, echocardiogram requirement, and anticoagulation status. Additional significant covariates were selected via forward stepwise regression. The association between frailty and delayed fixation was evaluated in two separate models (11-mFI model and CCI model). All statistical tests were two-sided with a significance threshold of P < 0.05. Analyses were performed using Python (available from https://www.python.org/).
Source of Funding
This work was supported by the New York Medical College Medical Student Summer Research Fellowship Grant.
Results
Patient and Hospital Characteristics
Of the 16 338 patient encounters identified, 15 472 (94.7%) of patients did not experience surgical delay and 866 (5.3%) of patients experienced a delay greater than 48 hours. Demographic, hospital, and clinical characteristics as well as frailty scores of these two patient groups were recorded (Table 2). In bivariate analysis, sex, race, solid cancer, dementia, diabetes, hypertension, chronic pulmonary disease, peripheral vascular disease, echocardiogram, long-term anticoagulation therapy, hospital bed-size, hospital region, 11-mFI, and CCI were significantly associated with surgical delay. The final multivariable models included race, diabetes, hypertension, echocardiogram, long-term anticoagulation, hospital bed-size, hospital region, 11-mFI, and CCI predicting surgical delay. Sex and age were included as established characteristics that affect surgical delay and frailty. 27
Multivariable Analysis of Delayed Fixation
Adjusted Odds Ratios and 95% Confidence Intervals for Surgical Delay More Than 48 hours Were Calculated Using Multivariate Logistic Regression Models for 11-Item Modified Frailty Index (11-mFI) and Modified Charlson Comorbidity Index (mCCI) Frailty Scores for Adult Patients Over 60 Years of Age With Intertrochanteric Hip Fractures
11-mFI: 11-item Modified Frailty Index; mCCI: modified Charlson Comorbidity Index.
CI: Confidence Interval.
CCI: The probability of delayed fixation increased 17% for each integer increase in CCI score (aOR 1.17, 95% CI: 1.14–1.21). Identifying as Black or Asian/Pacific Islander, requiring an echocardiogram, and use of long-term anticoagulant therapy increased the odds of delayed fixation. Admission to medium bed-sized hospitals and hospitals in the Midwest or West decreased the odds of delayed fixation. Age, sex, hypertension, and diabetes had no significant effect on delayed fixation risk (Table 3).
Discussion
The objective of this study was to identify patient- and system-level factors independently associated with delayed hip fracture surgical fixation and, critically, to compare the predictive utility of two validated risk assessment tools: the 11-factor Modified Frailty Index (11-mFI) and the Charlson Comorbidity Index (CCI). The most salient finding is the significant and independent association of both frailty and comorbidity burden with an increased probability of delayed fixation across both multivariable models. Specifically, the risk of delayed fixation increased by 47% for every integer increase in the 11-item modified frailty score and a 17% increase observed per unit increase in the 19-item CCI score. This disparity in measures reflects the difference in instrument scales; nonetheless, these findings reinforce the clinical relevance of comprehensive frailty assessment and suggest that the broader systemic vulnerability captured by the 11-mFI may be a more critical determinant of delayed time-to-surgery than a purely comorbidity-based score.
Predictive indices have become increasingly valued in clinical shared decision making, pre-operative counseling, and risk stratification. In patients with hip fractures, the CCI was predictive of postoperative complications, 28 30-day and 1-year mortality,18,29 and 10-year survivorship. 30 The 11-mFI score was predictive of in-hospital complications and mortality, 31 1- and 2-year mortality rates, 32 and functional dependence at discharge. 17 One study found that surgical delays increased complications only in higher-risk fracture patients. 33 This study builds on this foundation by identifying significant relationships between both 11-mFI and CCI scores and increased risk of surgical delay.
While this study was not specifically designed to examine costs, patients with surgical delay had higher median costs associated with their admission. The modified frailty index has been examined as a potential risk adjustment tool in determining future hip and femur fracture payment bundles by identifying patients who may incur higher costs of treatment. 34 Similarly, this study suggests the CCI and 11-mFI may be appropriate alternatives which can determine a patient’s potential complexity and associated cost of care.
Beyond frailty indices, this study’s models identified several predictors of delayed fixation: identifying as Black or Asian/Pacific Islander, requiring a pre-operative echocardiogram, and long-term anticoagulant therapy. Conversely, treatment at medium bed-size hospitals and hospitals located in the Midwest or West regions demonstrated a protective association against delayed fixation which corroborated prior studies.35,36 These findings highlight that the timing of surgery is not solely determined by inherent patient health but is also influenced by institutional resource allocation and management pathways. For example, a study of patients with hip fractures reported elimination of racial differences in postoperative outcomes with universal health insurance and standardized protocols in an integrated managed health system. 37
The observed protective effect of hypertension and diabetes in the 11-mFI model—a contrast to the null finding in the CCI model—further suggests a nuanced interaction between well-managed chronic disease and overall physiological reserve. It is possible this finding may reflect a loss of statistical power as the CCI includes 3 categories for diabetes severity, whereas 11-mFI uses a binary method; thus, patients with less severe disease may have decreased the overall risk effect.38,39 As a national database study, the sample is subject to selection bias. The impact of diabetes on delayed fixation is controversial. Arpan et al (2025) found that uncontrolled diabetes was a significant contributor in hip fracture surgical delay. Yet, Joseph et al (2023) found that diabetes was not associated, whereas cardiology comorbidities and consults were associated with increased risk of surgical delay. 40 Reduced surgical delay risk for patients with diabetes and hypertension may reflect risk management pathways and chronic outpatient management. 41
The associations between echocardiogram requirement and long-term anticoagulation therapy with increased risk of surgical delay were consistent with prior studies.23,24 Neuraxial analgesia is preferred in this patient population and is often prohibited by the use of long-term anticoagulation medications. Perioperative management of patients receiving long-term anticoagulation therapy is a difficult issue where the risk of intraoperative blood loss must be weighed against thromboembolism. NIS procedural variables were suspected to underreport these measures given the low percentages of long-term anticoagulation use and echocardiogram in this sample.
A strength of this study is that several known risk factors for delayed surgery persist when considering frailty, which is itself a risk factor. These results have important implications for perioperative risk assessment and multidisciplinary coordination of care. Patients who present with IT fractures can be assessed quickly with either the 11-mFI and CCI frailty indices for risk of surgical delay.
Limitations
This study has several limitations. First, frailty is a multifactorial physiological state with varied clinical phenotypes. The 11-mFI and CCI indices primarily rely on comorbidities and thus may be limited by their concrete natures. For example, the scores do not reflect the following clinical example: a patient who maintained the same diagnoses but has progressive diminished function such as requiring walking aids or exercise endurance. Thus, there is a risk of confounding by variables not captured by the frailty indices. Patient-reported outcomes offer a potential solution to this limitation.
Second, although multivariable regression was performed, there is an additional risk of residual confounding due to the constraints of the NIS dataset. Granular details that could explain delays to surgery, such as lab and imaging results, or operating room availability, are not captured in the NIS dataset. Whether a patient was transferred to another facility is unavailable and this event often increases the duration until surgery. ZIP codes are not an ideal measure of socioeconomic status but were used as an income measurement in this study due to limitations of the NIS dataset.
Third, collecting cross-sectional in-hospital data, the NIS is vulnerable to selection bias. Relevant details such as surgeon preference, institutional guidelines, and anesthesia type are not contained within the dataset.
Fourth, as with any large database study, the results of the present study may have been influenced by coding discrepancies, data misclassifications, and missing data. Lastly, the NIS is limited to single episodes of acute care and lacks longitudinal outcomes.
Conclusion
As the population ages and the prevalence of IT fractures increases, prediction models of surgical risk and complications are paramount. This study found the 11-mFI and CCI frailty indices to be significant predictors of surgical delay risk. These frailty instruments can be useful and accessible tools for risk assessment and multidisciplinary care.
Supplemental Material
Supplemental Material - Two Frailty Scores Predict Delayed Fracture Fixation in Patients With Intertrochanteric Hip Fracture
Supplemental Material for Two Frailty Scores Predict Delayed Fracture Fixation in Patients With Intertrochanteric Hip Fracture by Jenny Yang, Maxwell Ruffner, Mary Katharyn Fatehi, BS, Victor Koltenyuk, Wanda Horn, Jay Ayar, Elizabeth Drugge, and Anna R. Cooper, in Montefiore Einstein Journal of Musculoskeletal Medicine and Surgery
Footnotes
Ethical Considerations
This study received ethical approval from the New York Medical College IRB (approval #23970) on December 12, 2024. This is an IRB-approved retrospective study, all patient information was de-identified, and patient consent was not required. Patient data will not be shared with third parties.
Consent to Participate
The Ethics Committee of the New York Medical College waived the need for ethics approval and patient consent for the collection, analysis, and publication of the retrospectively obtained and anonymized data for this non-interventional study.
Author Contributions
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the New York Medical College Medical Student Summer Research Fellowship Grant (no grant number).
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Disclosures
The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
