Identifying patients at risk of not achieving meaningful gains in long-term postsurgical patient-reported outcome measures (PROMs) is important for improving patient monitoring and facilitating ...presurgical decision support. Machine learning may help automatically select and weigh many predictors to create models that maximize predictive power. However, these techniques are underused among studies of total joint arthroplasty (TJA) patients, particularly those exploring changes in postsurgical PROMs. QUESTION/PURPOSES: (1) To evaluate whether machine learning algorithms, applied to hospital registry data, could predict patients who would not achieve a minimally clinically important difference (MCID) in four PROMs 2 years after TJA; (2) to explore how predictive ability changes as more information is included in modeling; and (3) to identify which variables drive the predictive power of these models.
Data from a single, high-volume institution's TJA registry were used for this study. We identified 7239 hip and 6480 knee TJAs between 2007 and 2012, which, for at least one PROM, patients had completed both baseline and 2-year followup surveys (among 19,187 TJAs in our registry and 43,313 total TJAs). In all, 12,203 registry TJAs had valid SF-36 physical component scores (PCS) and mental component scores (MCS) at baseline and 2 years; 7085 and 6205 had valid Hip and Knee Disability and Osteoarthritis Outcome Scores for joint replacement (HOOS JR and KOOS JR scores), respectively. Supervised machine learning refers to a class of algorithms that links a mapping of inputs to an output based on many input-output examples. We trained three of the most popular such algorithms (logistic least absolute shrinkage and selection operator (LASSO), random forest, and linear support vector machine) to predict 2-year postsurgical MCIDs. We incrementally considered predictors available at four time points: (1) before the decision to have surgery, (2) before surgery, (3) before discharge, and (4) immediately after discharge. We evaluated the performance of each model using area under the receiver operating characteristic (AUROC) statistics on a validation sample composed of a random 20% subsample of TJAs excluded from modeling. We also considered abbreviated models that only used baseline PROMs and procedure as predictors (to isolate their predictive power). We further directly evaluated which variables were ranked by each model as most predictive of 2-year MCIDs.
The three machine learning algorithms performed in the poor-to-good range for predicting 2-year MCIDs, with AUROCs ranging from 0.60 to 0.89. They performed virtually identically for a given PROM and time point. AUROCs for the logistic LASSO models for predicting SF-36 PCS 2-year MCIDs at the four time points were: 0.69, 0.78, 0.78, and 0.78, respectively; for SF-36 MCS 2-year MCIDs, AUROCs were: 0.63, 0.89, 0.89, and 0.88; for HOOS JR 2-year MCIDs: 0.67, 0.78, 0.77, and 0.77; for KOOS JR 2-year MCIDs: 0.61, 0.75, 0.75, and 0.75. Before-surgery models performed in the fair-to-good range and consistently ranked the associated baseline PROM as among the most important predictors. Abbreviated LASSO models performed worse than the full before-surgery models, though they retained much of the predictive power of the full before-surgery models.
Machine learning has the potential to improve clinical decision-making and patient care by helping to prioritize resources for postsurgical monitoring and informing presurgical discussions of likely outcomes of TJA. Applied to presurgical registry data, such models can predict, with fair-to-good ability, 2-year postsurgical MCIDs. Although we report all parameters of our best-performing models, they cannot simply be applied off-the-shelf without proper testing. Our analyses indicate that machine learning holds much promise for predicting orthopaedic outcomes. LEVEL OF EVIDENCE: Level III, diagnostic study.
Background
Medicare is rapidly moving toward using patient-reported outcome measures (PROMs) for outcomes assessment and justification of orthopaedic and other procedures. Numerous measures have been ...developed to study knee osteoarthritis (OA); however, many of these surveys are long, disruptive to clinic flow, and result in incomplete data capture and/or low followup rates. The Knee injury and Osteoarthritis Outcome (KOOS) physical function short-form (KOOS-PS), while shorter, ignores pain, which is a primary concern of patients with advanced knee OA.
Questions/purposes
Our objective was to derive and validate a short-form survey focused on the patient with end-stage knee OA undergoing TKA.
Methods
Using our hospital’s knee replacement registry, we retrospectively identified 2291 patients with knee OA who underwent primary unilateral TKA and had completed preoperative and 2-year postoperative PROMs. We assessed 30 items from the 42-item KOOS that were quantitatively most difficult for patients to perform before TKA and qualitatively most relevant to patients with end-stage knee OA. Rasch analysis identified the KOOS, JR, a seven-item instrument, representing a single dimension, which we define as “knee health” because it reflects aspects of pain, symptom severity, and activities of daily living (ADL) including movements or activities that are directly relevant and difficult for patients with advanced knee OA. We assessed the internal consistency, external validity (versus KOOS and WOMAC domains), responsiveness, and floor and ceiling effects of the KOOS, JR. External validation was performed using calculated KOOS, JR scores in collaboration with a nationally representative joint replacement registry, the Function and Outcomes Research for Comparative Effectiveness in Total Joint Replacement (FORCE-TJR).
Results
Internal consistency for the KOOS, JR was high (Person Separation Index, 0.84; and 0.85 FORCE), external validity against other validated knee surveys was excellent (Spearman correlation coefficient, ρ 0.54–0.91), particularly for the KOOS pain (ρ 0.89 95% CI, 0.88–0.91 Hospital for Special Surgery HSS; and 0.91 95% CI, 0.90–0.93 FORCE) and KOOS ADL (ρ 0.87 95% CI, 0.85–0.88 HSS; and 0.84 95% CI, 0.81–0.87 FORCE). The KOOS, JR responsiveness (standardized response means, 1.79 95% CI, 1.70–1.88 HSS; and 1.70 95% CI, 1.54–1.86 FORCE) was high and floor 0.4–1.2%) and ceiling (18.8–21.8%) effects were favorable.
Conclusions
The new short knee PROM, the KOOS, JR, provides a single score representing “knee health” as it combines pain, symptoms, and functional limitations in a single score. This short-form PROM is patient-relevant and efficient.
Level of Evidence
Level III, diagnostic study.
Recent investigations have confirmed the impact of spinopelvic mobility upon functional acetabular position. However, it is critical to remember the impact of femoral component position as it ...contributes to combined version of the hip joint. It has been demonstrated that implant position can affect articular wear, the potential for prosthetic impingement as well as a cause of joint instability. Ensuring an impingement-free range of motion at the time of hip replacement is critical to long-term survivorship. During hip reconstruction, there is significant latitude in acetabular component positioning but femoral component version is often dictated by native femoral anatomy and in some instances cannot be adjusted. Understanding the role that femoral version plays in prosthetic hip joint kinematics should provide surgeons with a better understanding of this complex relationship.
We sought is to determine the mechanism of failure among primary total knee arthroplasties (TKAs) performed at a single high-volume institution by asking the following research questions: (1) What ...are the most common failure modes for modern TKA designs? and (2) What are the preoperative risk factors for failure following primary TKA?
From May 2007 to December 2012, 18,065 primary TKAs performed on 16,083 patients at a single institution were recorded in a prospective total joint arthroplasty registry with a minimum of 5-year follow-up. We retrospectively reviewed patient charts to determine a cause of failure for primary TKAs. A cox proportional hazard model was used to determine the risk of revision surgery following primary TKA.
The most common reasons for failure within 2 years after TKA were infection and stiffness. The multivariable regression identified the following preoperative risk factors for TKA failure: history of drug abuse (hazard ratio HR 4.68; P = 0.03), deformity/mechanical preoperative diagnosis (HR 3.52; P < .01), having a constrained condylar knee implant over posterior-stabilized implant (HR 1.99; P < .01), post-traumatic/trauma preoperative diagnosis (HR 1.78; P = .03), and younger age (HR 0.61; P < .01)
These findings add to the growing data that primary TKAs are no longer failing from polyethylene wear-related issues. This study identified preoperative risk factors for failure of primary TKAs, which may be useful information for developing strategies to improve outcomes following TKA.
Background:
The quality and state of satisfaction reporting after total knee replacement (TKR) is variable.
Questions/Purposes:
The purposes of this systematic review were (1) to examine the ...available literature on patient satisfaction after TKR, (2) to evaluate the quality of available evidence, and (3) to identify predictors of patient satisfaction after TKR.
Methods:
A systematic review of the MEDLINE database was performed. The initial search yielded 1219 studies. The inclusion criteria were English language, clinical outcome study with primary outcome related to TKR for osteoarthritis, and patient-reported satisfaction included as an outcome measure. Studies were assessed for demographics, methodology for reporting satisfaction, and factors influencing satisfaction.
Results:
Two hundred eight studies, including 95, 560 patients who had undergone TKR, met all inclusion and exclusion criteria; 112 (53.8%) of these studies were published in the past 3 years. Satisfaction was most commonly measured using an ordinal scale. Twenty-seven studies (13%) used a validated satisfaction survey. Eighty-three percent of studies reported more than 80% satisfaction. The most commonly reported predictor of satisfaction was post-operative patient-reported functional outcome. Pre-operative anxiety/depression was the most common pre-operative predictor of dissatisfaction.
Conclusion:
There are numerous studies reporting patient satisfaction after TKR, and publication on the topic has been increasing over the past decade. However, the majority of studies represent lower levels of evidence and use heterogeneous methods for measuring satisfaction, and few studies use validated satisfaction instruments. In general, the majority of studies report satisfaction rates ranging from 80 to 100%, with post-operative functional outcome and relief of pain being paramount determinants for achieving satisfaction.
Background
Patient-reported outcome measures (PROMs) are increasingly in demand for outcomes evaluation by hospitals, administrators, and policymakers. However, assessing total hip arthroplasty (THA) ...through such instruments is challenging because most existing measures of hip health are lengthy and/or proprietary.
Questions/purposes
The objective of this study was to derive a patient-relevant short-form survey based on the Hip disability and Osteoarthritis Outcome Score (HOOS), focusing specifically on outcomes after THA.
Methods
We retrospectively evaluated patients with hip osteoarthritis who underwent primary unilateral THA and who had completed preoperative and 2-year postoperative PROMs using our hospital’s hip replacement registry. The 2-year followup in this population was 81% (4308 of 5351 patients). Of these, 2371 completed every item on the HOOS before surgery and at 2 years, making them eligible for the formal item reduction analysis. Through semistructured interviews with 30 patients, we identified items in the HOOS deemed qualitatively most important to patients with hip osteoarthritis. The original HOOS has 40 items, the four quality-of-life items were excluded a priori, five were excluded for being redundant, and one was excluded based on patient-relevance surveys. The remaining 30 items were evaluated using Rasch modeling to yield a final six-item HOOS, Joint Replacement (HOOS, JR), representing a single construct of “hip health.” We calculated HOOS, JR scores for the Hospital for Special Surgery (HSS) cohort and validated this new score for internal consistency, external validity (versus HOOS and WOMAC domains), responsiveness to THA, and floor and ceiling effects. Additional external validation was performed using calculated HOOS, JR scores in collaboration with the Function and Outcomes Research for Comparative Effectiveness in Total Joint Replacement (FORCE-TJR) nationally representative joint replacement registry (n = 910).
Results
The resulting six-item PROM (HOOS, JR) retained items only from the pain and activities of daily living domains. It showed high internal consistency (Person Separation Index, 0.86 HSS; 0.87 FORCE), moderate to excellent external validity against other hip surveys (Spearman’s correlation coefficient, 0.60–0.94), very high responsiveness (standardized response means, 2.03 95% CI, 1.84–2.22 FORCE; and 2.38 95% CI, 2.27–2.49 HSS), and favorable floor (0.6%–1.9%) and ceiling (37%–46%) effects. External validity was highest for the HOOS pain (Spearman’s correlation coefficient, 0.87 95% CI, 0.86–0.89 HSS; and 0.87 95% CI, 0.84–0.90 FORCE) and HOOS activities of daily living (Spearman’s correlation coefficient, 0.94 95% CI, 0.93–0.95 HSS; and 0.94 95% CI, 0.93–0.96 FORCE) domains in the HSS validation cohort and the FORCE-TJR cohort.
Conclusions
The HOOS, JR provides a valid, reliable, and responsive measure of hip health for patients undergoing THA. This short-form PROM is patient relevant and efficient.
Level of Evidence
Level III, diagnostic study.
As our recognition of the complexity of the hip-spine relationship is based on stability, there are several intraoperative strategies that surgeons may consider. First, patient positioning on the ...operative table plays an important role in reducing pelvic motion and tilt. Then, consider determining the femoral version before acetabular preparation and component insertion. The concept of a combined version of both components is an excellent guide to position. However, femoral version is often dictated by native femoral torsion, and therefore, it is the acetabular version that must be adjusted accordingly. Finally, based upon preoperative planning and intraoperative verification, the use of enabling technologies such as navigation, robotics, and other smart tools appears to play a major role of ever increasing importance in accurate component placement.
Background
Adverse local tissue reactions (ALTR) around metal-on-metal (MOM) hip arthroplasties are increasingly being recognized as a cause of failure. These reactions may be associated with ...intraoperative tissue damage and complication rates as high as 50% after revision. Although MRI can identify ALTR in MOM hips, it is unclear whether the MRI findings predict those at revision surgery.
Questions/purposes
We therefore (1) identified which MRI characteristics correlated with histologically confirmed ALTR (using the aseptic lymphocytic vasculitis-associated lesions ALVAL score) and intraoperative tissue damage and (2) developed a predictive model using modified MRI to detect ALVAL and quantify intraoperative tissue damage.
Methods
We retrospectively reviewed 68 patients with failed MOM hip arthroplasties who underwent preoperative MRI and subsequent revision surgery. Images were analyzed to determine synovial volume, osteolysis, and synovial thickness. The ALVAL score was used to grade tissue samples, thus identifying a subset of patients with ALTR. Intraoperative tissue damage was graded using a four-point scale. Random forest analysis determined the sensitivity and specificity of MRI characteristics in detecting ALVAL (score ≥ 5) and intraoperative tissue damage.
Results
Maximal synovial thicknesses and synovial volumes as determined on MRI correlated with the ALVAL score and were higher in cases of severe intraoperative tissue damage. Our MRI predictive model showed sensitivity and specificity of 94% and 87%, respectively, for detecting ALVAL and 90% and 86%, respectively, for quantifying intraoperative tissue damage.
Conclusions
MRI is sensitive and specific in detecting ALVAL and tissue damage in patients with MOM hip implants. MRI can be used as a screening tool to guide surgeons toward timely revision surgery.
Level of Evidence
Level III, diagnostic study. See Guidelines for Authors for a complete description of levels of evidence.
Sitting radiographs have been used as a pre-operative tool to plan patient-specific total hip arthroplasty (THA) component position that would improve hip stability. Previous work has demonstrated ...that spinal mobility may impact functional acetabular position when seated. We sought to determine whether patients who dislocate following THA have different sitting spinopelvic alignment or acetabular component orientation compared to patients who did not dislocate.
A consecutive series of 1000 patients underwent post-operative low-dose biplanar spine-to-ankle lateral radiographs in standing and sitting positions 1 year following THA. Twelve patients (1% of all patients) experienced hip dislocation. Patients were categorized as having normal lumbar spines (without radiographic arthrosis) or as having lumbar multi-level degenerative disc disease. Measurements of spinopelvic alignment parameters (including sacral slope, lumbar lordosis, and proximal femur angles) and acetabular component orientation in sitting position (functional inclination and functional anteversion) were performed.
Patients who dislocated had significantly less spine flexion, less change in pelvic tilt, and more hip flexion from standing to sitting positions compared to patients with normal spines. In sitting position, dislocators had acetabular components with less functional inclination and less functional anteversion.
This study demonstrates that patients with fixed spinopelvic alignment from standing to sitting position are at higher risk of hip dislocation. Imaging patients from standing to sitting position using this technique can provide valuable information on whether a patient has fixed spinopelvic alignment with postural changes and is therefore at higher risk of dislocation.
Abstract Pelvic tilt (PT) affects the functional anteversion and inclination of acetabular components in total hip arthroplasty (THA). One-hundred and thirty-eight consecutive patients who underwent ...unilateral primary THA were reviewed. Most cases had some degree of pre-operative PT, with 17% having greater than 10° of PT on standing pre-operative radiographs. There was no significant change in PT following THA. A computer model of a hemispheric acetabular component implanted in a range of anatomic positions in a pelvis with varying PT was created to determine the effects of PT on functional anteversion and inclination. Based on the study results, tilt-adjustment of the acetabular component position based on standing pre-operative imaging will likely improve functional component position in most patients undergoing THA.
PDF
