Arch Orthop Trauma Surg DOI 10.1007/s00402-014-1959-y
Orthopaedic Surgery
Risk factors for complications and in‑hospital mortality following hip fractures: a study using the National Trauma Data Bank Philip J. Belmont Jr. · E’Stephan J. Garcia · David Romano · Julia O. Bader · Kenneth J. Nelson · Andrew J. Schoenfeld
Received: 29 October 2012 © Springer-Verlag Berlin Heidelberg (outside the USA) 2014
Abstract Study design Retrospective review of prospectively collected data. Objective To describe the impact of patient demographics, injury-specific factors, and medical co-morbidities on outcomes after hip fracture using the National Sample Program (NSP) of the National Trauma Data Bank (NTDB). Methods The 2008 NSP-NTDB was queried to identify patients sustaining hip fractures. Patient demographics, comorbidities, injury-specific factors, and outcomes (including mortality and complications) were recorded and a national estimate model was developed. Unadjusted differences for risk factors were evaluated using t test/Wald Chi square analyses. Weighted logistic regression and sensitivity analyses were performed to control for all factors in the model. Results The weighted sample contained 44,419 incidents of hip fracture. The average age was 72.7. Sixty-two percent of the population was female and 80 % was white. The mortality rate was 4.5 % and 12.5 % sustained at least one
complication. Seventeen percent of patients who sustained at least one complication died. Dialysis, presenting in shock, cardiac disease, male sex, and ISS were significant predictors of mortality, while dialysis, obesity, cardiac disease, diabetes, and a procedure delay of ≥2 days influenced complications. The major potential modifiable risk factor appears to be time to procedure, which had a significant impact on complications. Conclusions This is the first study to postulate predictors of morbidity and mortality following hip fracture in a US national model. While many co-morbidities appear to be influential in predicting outcome, some of the more significant factors include the presence of shock, dialysis, obesity, and time to surgery. Level of evidence Prognostic study, Level II. Keywords Hip fracture · Mortality · Complications
Introduction Investigation Performed at William Beaumont Army Medical Center, El Paso, TX P. J. Belmont Jr. · E. J. Garcia · D. Romano · K. J. Nelson Department of Orthopaedic Surgery, William Beaumont Army Medical Center, Texas Tech University Health Sciences Center, 5005 N. Piedras Street, El Paso, TX 79920, USA J. O. Bader Statistical Consulting Laboratory, University of Texas at El Paso, El Paso, TX, USA A. J. Schoenfeld (*) Department of Orthopaedic Surgery, University of Michigan, 2800 Plymouth Road, Building 10, Rm G016, Ann Arbor, MI 48109, USA e-mail: [email protected]
Hip fractures are one of the most important causes of longterm disability and a significant public health issue [1]. Hip fracture rates vary widely [2, 3], yet the mean number of hip fractures in the United States has been estimated between 414 and 957 per 100,000 [4]. Within the United States, there are 340,000 hip fracture-related hospitalizations annually and this is expected to increase as a result of demographic aging [5–7]. Hip fractures carry high associated mortality and complication rates despite advances in regional anesthesia and surgical techniques [8]. An increased risk of death after hip fracture is well documented with a one-year mortality rate of 12–33 % [3, 9–13] and patients often fail to attain prefracture levels of functionality [14–16].
13
While the goal of surgical treatment of hip fractures is to achieve a painless and stable lower extremity and return patients to their prefracture functionality, the occurrence of perioperative complications can exert adverse effects that may result in further disability and even death [13, 17, 18]. Moreover, complications after surgery have been shown to increase healthcare costs and prolong the recovery period [19]. When examining the literature concerning mortality and morbidity after hip fracture, the effect of patient variables such as age, sex and medical co-morbidities is rarely analyzed in randomized trials, while patient factors such as fracture type and pattern are underreported in observational studies [20]. This investigation utilized the 2008 National Sample Program (NSP) of the National Trauma Data Bank (NTDB) to identify demographic risks for post-operative morbidity and mortality for patients who sustained a hip fracture. We hypothesized that the methodology of the NSP-NTDB would allow the development of a predictive model capable of describing the influence of demographic factors, injuryspecific predictors, medical co-morbidities and time to surgery on morbidity and mortality following hip fracture. To the best of our knowledge, this study is among the first to engage in predictive modeling using a sample representative of the US population.
Materials and methods The 2008 version of the NSP of the NTDB was obtained through application to the American College of Surgeons (ACS). The methodology behind the NTDB and its NSP has been extensively described in several publications [21–25] and is also accessible in the program’s user handbook [26]. The NTDB-NSP is a stratified sample of 100 national sites selected by the ACS on the basis of a mathematical model indicating that a weighted sample derived from centerspecific data would be representative of the national demographic [26]. Weighting takes into account region of origin, trauma designation, and number of Emergency Room visits [21, 26]. The 2008 NSP was developed using information submitted by 90 national medical centers. National estimates are calculated using the NSP program and rely on original patient-based data submitted by the institution as well as adjustments that take into account the probability of hospital selection [26]. The NTDB and NSP have been utilized extensively and have been found to be a reliable means of developing prognostic models for complications, mortality, and other factors related to trauma [21–24]. A search was performed of the NSP using the International Classification of Disease (ICD-9) codes contained in the 820.XX series to identify patients who sustained hip fractures. Based on ICD-9 classification, fractures were categorized as femoral neck (820.00, 01, 02, 03, 09, 10, 11, 12,
13
Arch Orthop Trauma Surg
13, 19, 80, 90), pertrochanteric/intertrochanteric (820.20, 21, 30, 31) and subtrochanteric (820.22, 32). For all individuals determined as having sustained a hip fracture, demographic details were censored, including age, race and gender. Injury-specific factors were also assessed, such as mechanism of injury, presence of shock at presentation (defined by an initial systolic blood pressure
Orthopaedic Surgery
Risk factors for complications and in‑hospital mortality following hip fractures: a study using the National Trauma Data Bank Philip J. Belmont Jr. · E’Stephan J. Garcia · David Romano · Julia O. Bader · Kenneth J. Nelson · Andrew J. Schoenfeld
Received: 29 October 2012 © Springer-Verlag Berlin Heidelberg (outside the USA) 2014
Abstract Study design Retrospective review of prospectively collected data. Objective To describe the impact of patient demographics, injury-specific factors, and medical co-morbidities on outcomes after hip fracture using the National Sample Program (NSP) of the National Trauma Data Bank (NTDB). Methods The 2008 NSP-NTDB was queried to identify patients sustaining hip fractures. Patient demographics, comorbidities, injury-specific factors, and outcomes (including mortality and complications) were recorded and a national estimate model was developed. Unadjusted differences for risk factors were evaluated using t test/Wald Chi square analyses. Weighted logistic regression and sensitivity analyses were performed to control for all factors in the model. Results The weighted sample contained 44,419 incidents of hip fracture. The average age was 72.7. Sixty-two percent of the population was female and 80 % was white. The mortality rate was 4.5 % and 12.5 % sustained at least one
complication. Seventeen percent of patients who sustained at least one complication died. Dialysis, presenting in shock, cardiac disease, male sex, and ISS were significant predictors of mortality, while dialysis, obesity, cardiac disease, diabetes, and a procedure delay of ≥2 days influenced complications. The major potential modifiable risk factor appears to be time to procedure, which had a significant impact on complications. Conclusions This is the first study to postulate predictors of morbidity and mortality following hip fracture in a US national model. While many co-morbidities appear to be influential in predicting outcome, some of the more significant factors include the presence of shock, dialysis, obesity, and time to surgery. Level of evidence Prognostic study, Level II. Keywords Hip fracture · Mortality · Complications
Introduction Investigation Performed at William Beaumont Army Medical Center, El Paso, TX P. J. Belmont Jr. · E. J. Garcia · D. Romano · K. J. Nelson Department of Orthopaedic Surgery, William Beaumont Army Medical Center, Texas Tech University Health Sciences Center, 5005 N. Piedras Street, El Paso, TX 79920, USA J. O. Bader Statistical Consulting Laboratory, University of Texas at El Paso, El Paso, TX, USA A. J. Schoenfeld (*) Department of Orthopaedic Surgery, University of Michigan, 2800 Plymouth Road, Building 10, Rm G016, Ann Arbor, MI 48109, USA e-mail: [email protected]
Hip fractures are one of the most important causes of longterm disability and a significant public health issue [1]. Hip fracture rates vary widely [2, 3], yet the mean number of hip fractures in the United States has been estimated between 414 and 957 per 100,000 [4]. Within the United States, there are 340,000 hip fracture-related hospitalizations annually and this is expected to increase as a result of demographic aging [5–7]. Hip fractures carry high associated mortality and complication rates despite advances in regional anesthesia and surgical techniques [8]. An increased risk of death after hip fracture is well documented with a one-year mortality rate of 12–33 % [3, 9–13] and patients often fail to attain prefracture levels of functionality [14–16].
13
While the goal of surgical treatment of hip fractures is to achieve a painless and stable lower extremity and return patients to their prefracture functionality, the occurrence of perioperative complications can exert adverse effects that may result in further disability and even death [13, 17, 18]. Moreover, complications after surgery have been shown to increase healthcare costs and prolong the recovery period [19]. When examining the literature concerning mortality and morbidity after hip fracture, the effect of patient variables such as age, sex and medical co-morbidities is rarely analyzed in randomized trials, while patient factors such as fracture type and pattern are underreported in observational studies [20]. This investigation utilized the 2008 National Sample Program (NSP) of the National Trauma Data Bank (NTDB) to identify demographic risks for post-operative morbidity and mortality for patients who sustained a hip fracture. We hypothesized that the methodology of the NSP-NTDB would allow the development of a predictive model capable of describing the influence of demographic factors, injuryspecific predictors, medical co-morbidities and time to surgery on morbidity and mortality following hip fracture. To the best of our knowledge, this study is among the first to engage in predictive modeling using a sample representative of the US population.
Materials and methods The 2008 version of the NSP of the NTDB was obtained through application to the American College of Surgeons (ACS). The methodology behind the NTDB and its NSP has been extensively described in several publications [21–25] and is also accessible in the program’s user handbook [26]. The NTDB-NSP is a stratified sample of 100 national sites selected by the ACS on the basis of a mathematical model indicating that a weighted sample derived from centerspecific data would be representative of the national demographic [26]. Weighting takes into account region of origin, trauma designation, and number of Emergency Room visits [21, 26]. The 2008 NSP was developed using information submitted by 90 national medical centers. National estimates are calculated using the NSP program and rely on original patient-based data submitted by the institution as well as adjustments that take into account the probability of hospital selection [26]. The NTDB and NSP have been utilized extensively and have been found to be a reliable means of developing prognostic models for complications, mortality, and other factors related to trauma [21–24]. A search was performed of the NSP using the International Classification of Disease (ICD-9) codes contained in the 820.XX series to identify patients who sustained hip fractures. Based on ICD-9 classification, fractures were categorized as femoral neck (820.00, 01, 02, 03, 09, 10, 11, 12,
13
Arch Orthop Trauma Surg
13, 19, 80, 90), pertrochanteric/intertrochanteric (820.20, 21, 30, 31) and subtrochanteric (820.22, 32). For all individuals determined as having sustained a hip fracture, demographic details were censored, including age, race and gender. Injury-specific factors were also assessed, such as mechanism of injury, presence of shock at presentation (defined by an initial systolic blood pressure
