Machine learning decision support model for radical cystectomy discharge planning
PURPOSE/OBJECTIVE:Timely and appropriate discharge placement for patients who have undergone radical cystectomy (RC) remains challenging. Our objective was to improve the discharge planning process by creating a machine learning model that helps to predict the need for non-home hospital discharge to a higher level of care. MATERIALS AND METHODS/METHODS:Patients undergoing elective radical cystectomy for bladder cancer from 2014-2019 were identified in the ACS-NSQIP database. A gradient boosted decision tree was trained on selected predischarge variables to predict discharge location, dichotomized into home and non-home. We used threshold-moving to calibrate model predictions and evaluated model performance on a testing set using receiver operating characteristic and precision recall curves. Model performance was further examined in subgroups of interest. RESULTS AND CONCLUSIONS/CONCLUSIONS:A total of 11,881 patients met inclusion criteria with a mean age of 68.6 years. 10.6% of patients undergoing RC had non-home discharges. Our model predicting non-home discharge achieved an area under the receiver operating characteristic curve of 0.80 and an average precision of 0.33. After threshold-moving, our model had a recall of 0.757 and a precision of 0.211. Top variables by importance were septic shock occurrence, ventilator-use greater than 48 hours, organ space surgical site infection and unplanned intubation. Our model shows strong performance in identifying patients who required non-home discharge to higher levels of care, outperforming commonly used clinical indices and prior work. Modern machine learning techniques may be applied to support more timely and appropriate clinical decision making.
Reply by Authors
Demographic and Practice Trends of Rural Urologists in the U.S.: Implications for Workforce Policy
Introduction:There is a shortage in the number of urologists needed to satisfy the needs of an aging U.S. population. The urologist shortage may have a pronounced impact on aging rural communities. Our objective was to describe the demographic trends and scope of practice of rural urologists using data from the American Urological Association Census.Methods:We conducted a retrospective analysis of American Urological Association Census survey data over a 5-year period (2016-2020), including all U.S.-based practicing urologists. Metropolitan (urban) and nonmetropolitan (rural) practice classifications were based on rural-urban commuting area codes for the primary practice location zip code. We conducted descriptive statistics of demographics, practice characteristics and specific rural-focused survey items.Results:In 2020, rural urologists were older (60.9 years, 95% CI 58.5-63.3 vs 54.6 years, 95% CI 54.0-55.1) and were in practice longer (25.4 years, 95% CI 23.2-27.5 vs 21.2 years, 95% CI 20.8-21.5) than urban counterparts. Since 2016, mean age and years in practice increased for rural urologists but remained stable for urban urologists, suggesting an influx of younger urologists to urban areas. Compared with urban urologists, rural urologists had significantly less fellowship training and more frequently worked in solo practice, multispecialty groups and private hospitals.Conclusions:The urological workforce shortage will particularly impact rural communities and their access to urological care. We hope our findings will inform and empower policymakers to develop targeted interventions to expand the rural urologist workforce.
Publisher Correction: Impact of type of minimally invasive approach on open conversions across ten common procedures in different specialties
Demographic and Practice Trends of Rural Urologists in the U.S.: Implications for Workforce Policy
INTRODUCTION/BACKGROUND:There is a shortage in the number of urologists needed to satisfy the needs of an aging U.S. POPULATION/METHODS:The urologist shortage may have a pronounced impact on aging rural communities. Our objective was to describe the demographic trends and scope of practice of rural urologists using data from the American Urological Association Census. METHODS:We conducted a retrospective analysis of American Urological Association Census survey data over a 5-year period (2016-2020), including all U.S.-based practicing urologists. Metropolitan (urban) and nonmetropolitan (rural) practice classifications were based on rural-urban commuting area codes for the primary practice location zip code. We conducted descriptive statistics of demographics, practice characteristics and specific rural-focused survey items. RESULTS:In 2020, rural urologists were older (60.9 years, 95% CI 58.5-63.3 vs 54.6 years, 95% CI 54.0-55.1) and were in practice longer (25.4 years, 95% CI 23.2-27.5 vs 21.2 years, 95% CI 20.8-21.5) than urban counterparts. Since 2016, mean age and years in practice increased for rural urologists but remained stable for urban urologists, suggesting an influx of younger urologists to urban areas. Compared with urban urologists, rural urologists had significantly less fellowship training and more frequently worked in solo practice, multispecialty groups and private hospitals. CONCLUSIONS:The urological workforce shortage will particularly impact rural communities and their access to urological care. We hope our findings will inform and empower policymakers to develop targeted interventions to expand the rural urologist workforce.
Reply by Authors
Impact of type of minimally invasive approach on open conversions across ten common procedures in different specialties
BACKGROUND:Conversion rates during minimally invasive surgery are generally examined in the limited scope of a particular procedure. However, for a hospital or payor, the cumulative impact of conversions during commonly performed procedures could have a much larger negative effect than what is appreciated by individual surgeons. The aim of this study is to assess open conversion rates during minimally invasive surgery (MIS) across common procedures using laparoscopic/thoracoscopic (LAP/VATS) and robotic-assisted (RAS) approaches. STUDY DESIGN/METHODS:Retrospective cohort study using the Premier Database on patients who underwent common operations (hysterectomy, lobectomy, right colectomy, benign sigmoidectomy, low anterior resection, inguinal and ventral hernia repair, and partial nephrectomy) between January 2013 and September 2015. ICD-9 and CPT codes were used to define procedures, modality, and conversion. Propensity scores were calculated using patient, hospital, and surgeon characteristics. Propensity-score matched analysis was used to compare conversions between LAP/VATS and RAS for each procedure. RESULTS:A total of 278,520 patients had MIS approaches of the ten operations. Conversion occurred in 5% of patients and was associated with a 1.77Â day incremental increase in length of stay and $3441 incremental increase in cost. RAS was associated with a 58.5% lower rate of conversion to open surgery compared to LAP/VATS. CONCLUSION/CONCLUSIONS:At a health system or payer level, conversion to open is detrimental not just for the patient and surgeon but also puts a significant strain on hospital resources. Use of RAS was associated with less than half of the conversion rate observed for LAP/VATS.
Does histologic subtype impact overall survival in observed T1a kidney cancers compared with competing risks? Implications for biopsy as a risk stratification tool
OBJECTIVES/OBJECTIVE:We sought to assess if adding a biopsy proven histologic subtype to a model that predicts overall survival that includes variables representing competing risks in observed, biopsy proven, T1a renal cell carcinomas, enhances the model's performance. METHODS:The National Cancer Database was assessed (years 2004-2015) for patients with observed T1a renal cell carcinoma who had undergone renal mass biopsy. Kaplan-Meier curves were utilized to estimate overall survival stratified by histologic subtype. We utilized C-index from a Cox proportional hazards model to evaluate the impact of adding histologic subtypes to a model to predict overall survival for each stage. RESULTS:Of 132â€‰958 T1a renal masses identified, 1614 had biopsy proven histology and were managed non-operatively. Of those, 61% were clear cell, 33% papillary, and 6% chromophobe. Adjusted Kaplan-Meier curves demonstrated a difference in overall survival between histologic subtypes (Pâ€‰=â€‰0.010) with greater median overall survival for patients with chromophobe (85.1â€‰months, hazard rate 0.45, Pâ€‰=â€‰0.005) compared to clear cell (64.8â€‰months, reference group). Adding histology to a model with competing risks alone did not substantially improve model performance (C-index 0.65 vs 0.64 respectively). CONCLUSIONS:Incorporation of histologic subtype into a risk stratification model to determine prognostic overall survival did not improve modeling of overall survival compared with variables representing competing risks in patients with T1a renal cell carcinoma managed with observation. These results suggest that performing renal mass biopsy in order to obtain tumor histology may have limited utility. Future studies should further investigate the overall utility of renal mass biopsy for observed T1a kidney cancers.
Safety of repeat blue light cystoscopy with hexaminolevulinate (HAL) in the management of bladder cancer: Results from a phase III, comparative, multi-center study
PURPOSE/OBJECTIVE:The therapeutic benefit of intravesical instillation of hexaminolevulinate (HAL) at the time of transurethral resection of bladder tumor (TURBT) has been demonstrated in multiple studies. The purpose of this study was to prospectively assess the safety of repeated administration of HAL from a phase III pre-trial planned analysis. MATERIALS AND METHODS/METHODS:All patients evaluated in the study received at least 1 dose of HAL at the time of office cystoscopy, and a subset of these patients (nâ€¯=â€¯103, 33.2%) received a second dose a few weeks later at the time of TURBT. Adverse events (AEs) were recorded, and the safety of repeat use of HAL was determined by comparing the proportion of patients with AEs considered causally related to HAL in the surveillance examination compared to the OR examination. Association between categorical variables was tested using Fisher's Exact Test, and a P < 0.05 was considered statistically significant. RESULTS:HAL-related AEs were experienced by 6 patients (2.2%) during surveillance cystoscopy and 3 patients (3.4%) following TURBT (Pâ€¯=â€¯0.76); 181 patients (59.5%) had prior exposure to HAL before enrolling in the study with no difference in the number of AEs when comparing prior exposure to HAL to no prior exposure (Pâ€¯=â€¯0.76). Of the patients who previously received intravesical therapy, 8 (2.9%) had at least 1 AE during surveillance compared to 3 (9.7%) who had no prior intravesical therapy (Pâ€¯=â€¯0.09). CONCLUSIONS:Repeat use of HAL is safe even when administered within a few weeks of receiving a dose of intravesical therapy.
A Prospective Pilot Study Investigating Performance of 18F-Fluciclovine PET Imaging for Detection of Prostate Cancer 2 Years Following Primary Partial Gland Cryoablation
Purpose/UNASSIGNED:The goal of partial gland ablation (PGA) is to eradicate focal lesions of clinically significant prostate cancer (csPCa) with minimal adverse impact on functional outcomes. The primary objective of this study is to characterize the performance of 18F-Fluciclovine PET imaging for detection of prostate cancer following PGA. Materials and Methods/UNASSIGNED:Subjects 2Â years following primary partial gland cryoablation (PPGCA) were invited to participate in an IRB-approved study providing they met the following inclusion criteria: a single reported mpMRI region of interest (ROI) concordant with biopsy Gleason Grade Group (GGG)â€‰<â€‰4, no gross extra-prostatic extension on mpMRI, and no GGGâ€‰>â€‰1 or GGG 1 with a core lengthâ€‰>â€‰6Â mm on contralateral systematic biopsy. 18F-Fluciclovine PET MRI imaging of the prostate was performed followed by in and out-of-field biopsies. Results/UNASSIGNED:Twenty-seven men who met eligibility criteria participated in the prospective study. In-field and out-of-field csPCa recurrence rate was 7.4% and 22.2%, respectively. The sensitivity and positive predictive value of mpMRI and PET imaging did not reach performance to reliably inform who should undergo prostate biopsy. Conclusion/UNASSIGNED:At 2Â years following PPGCA, the rate of in-field csPCa was exceedingly low indicating a limited role for imaging to inform in-field biopsy decisions. The csPCa detection rate of out-of-field recurrence was 22% which provides an opportunity for imaging to inform out-of-field biopsy decisions. Based on our findings, 18F-Fluciclovine PET MRI cannot be used to inform who should undergo out-of-field prostate biopsy at 2Â years following PPGCA.