ABSTRACT
BACKGROUND: Adjunctive screw fixation has been shown to be reliable in achieving acetabular component stability in revision total hip arthroplasty (THA). The purpose of this study was to assess the effect of inferior screw placement on acetabular component failure following revision THA. We hypothesized that inferior screw fixation would decrease acetabular failure rates. METHODS: We reviewed 250 patients who had Paprosky Type II or III defects who underwent acetabular revision between 2001 and 2021 across three institutions. Demographic factors, the number of screws, location of screw placement (superior versus inferior), use of augments and/or cup-cage constructs, Paprosky classification, and presence of discontinuity were documented. Multivariate regression was performed to identify the independent effect of inferior screw fixation on the primary outcome of aseptic rerevision of the acetabular component. RESULTS: At a mean follow-up of 53.4 months (range, 12 to 261), 16 patients (6.4%) required re-revision for acetabular loosening. There were 140 patients (56.0%) who had inferior screw fixation, all of whom did not have neurovascular complications during screw placement. Patients who had inferior screws had a lower rate of acetabular rerevision than those who only had superior screw fixation (2.1 versus 11.8%, P = .0030). Multivariate regression demonstrates that inferior screw fixation decreased the likelihood of rerevision for acetabular loosening when compared to superior screw fixation alone (odds ratio: 0.1, confidence interval: 0.03 to 0.5; P = .0071). No other risk factors were identified. CONCLUSIONS: Inferior screw fixation is a safe and reliable technique to reduce acetabular component failure following revision THA in cases of severe acetabular bone loss.
ABSTRACT
BACKGROUND: Advanced imaging modalities are expensive, and access to advanced imaging services may vary by socioeconomic factors, creating the potential for unwarranted variations in care. Ankle sprains are a common injury for which variations in MRI use can occur, both via underuse of indicated MRIs (appropriate use) and overuse of nonindicated MRIs (inappropriate use). High-value, equitable healthcare would decrease inappropriate use and increase appropriate use of MRI for this common injury. It is unknown whether socioeconomic factors are associated with underuse of indicated MRIs and overuse of nonindicated MRIs for ankle sprains. QUESTIONS/PURPOSES: Using ankle sprains as a paradigm injury, given their high population incidence, we asked: (1) Does MRI use for ankle sprains vary by insurance type? (2) After controlling for relevant confounding variables, did patients who received an MRI have higher odds of undergoing ankle surgery? METHODS: Between 2011 and 2019, a total of 6,710,223 patients were entered into the PearlDiver Mariner Patient Records Database with a diagnosis of ankle sprain. We considered patients with continuous enrollment in the database for at least 1 year before and 2 years after the diagnosis as potentially eligible. Based on that, 68% (4,567,106) were eligible; a further 20% (1,372,478) were excluded because of age younger than 18 years, age at least 65 years with Medicaid insurance, or age < 65 years with Medicare insurance. Another 0.1% (9169) had incomplete data, leaving 47% (3,185,459) for analysis here. Patients with Medicaid insurance differed from patients with Medicare Advantage or private insurance with respect to age, gender, region, and comorbidity burden. The primary outcome was ankle MRI occurring within 12 months after diagnosis. The use of ankle surgery after MRI in each cohort was measured as a secondary outcome. We used multivariable logistic regression models to evaluate the association between insurance type and MRI use while adjusting for age, gender, region, and comorbidity burden. Separate multivariable regression models were created to evaluate the association between receiving an MRI and subsequent ankle surgery for each insurance type, adjusting for age, gender, region, and comorbidity burden. Within 12 months of an ankle sprain diagnosis, 1% (3522 of 339,457) of patients with Medicaid, 2% (44,793 of 2,627,288) of patients with private insurance, and 1% (1660 of 218,714) of patients with Medicare Advantage received an MRI. RESULTS: After controlling for age, gender, region, and comorbidity burden, patients with Medicaid had lower odds of receiving an MRI within 12 months after ankle sprain diagnosis than patients with private insurance (odds ratio 0.60 [95% confidence interval 0.57 to 0.62]; p < 0.001). Patients with Medicaid who received an MRI had higher adjusted odds of undergoing subsequent ankle surgery (OR 23 [95% CI 21 to 26]; p < 0.001) than patients with private insurance (OR 12.7 [95% CI 12 to 13]; p < 0.001). CONCLUSION: Although absolute MRI use was generally low, there was substantial relative variation by insurance type. Given the high incidence of ankle sprains in the general population, these relative differences can translate to tens of thousands of MRIs. Further studies are needed to evaluate the reasons for decreased appropriate MRI use in patients with Medicaid and overuse of MRI in patients with private insurance. The establishment of clinical practice guidelines by orthopaedic professional societies and more stringent gatekeeping for MRI use by health insurers could reduce unwarranted variations in MRI use. LEVEL OF EVIDENCE: Level III, prognostic study.
ABSTRACT
OBJECTIVES: We aimed to develop deep learning models using longitudinal chest X-rays (CXRs) and clinical data to predict in-hospital mortality of COVID-19 patients in the intensive care unit (ICU). METHODS: Six hundred fifty-four patients (212 deceased, 442 alive, 5645 total CXRs) were identified across two institutions. Imaging and clinical data from one institution were used to train five longitudinal transformer-based networks applying five-fold cross-validation. The models were tested on data from the other institution, and pairwise comparisons were used to determine the best-performing models. RESULTS: A higher proportion of deceased patients had elevated white blood cell count, decreased absolute lymphocyte count, elevated creatine concentration, and incidence of cardiovascular and chronic kidney disease. A model based on pre-ICU CXRs achieved an AUC of 0.632 and an accuracy of 0.593, and a model based on ICU CXRs achieved an AUC of 0.697 and an accuracy of 0.657. A model based on all longitudinal CXRs (both pre-ICU and ICU) achieved an AUC of 0.702 and an accuracy of 0.694. A model based on clinical data alone achieved an AUC of 0.653 and an accuracy of 0.657. The addition of longitudinal imaging to clinical data in a combined model significantly improved performance, reaching an AUC of 0.727 (p = 0.039) and an accuracy of 0.732. CONCLUSIONS: The addition of longitudinal CXRs to clinical data significantly improves mortality prediction with deep learning for COVID-19 patients in the ICU. KEY POINTS: ⢠Deep learning was used to predict mortality in COVID-19 ICU patients. ⢠Serial radiographs and clinical data were used. ⢠The models could inform clinical decision-making and resource allocation.