Development and Assessment of Machine Learning Models for Individualized Risk Assessment of Mastectomy Skin Flap Necrosis.

OBJECTIVE: To develop, validate, and evaluate ML algorithms for predicting MSFN. BACKGROUND: MSFN is a devastating complication that causes significant distress to patients and physicians by prolonging recovery time, compromising surgical outcomes, and delaying adjuvant therapy. METHODS: We conducted comprehensive review of all consecutive patients who underwent mastectomy and immediate implant-based reconstruction from January 2018 to December 2019. Nine supervised ML algorithms were developed to predict MSFN. Patient data were partitioned into training (80%) and testing (20%) sets. RESULTS: We identified 694 mastectomies with immediate implant-based reconstruction in 481 patients. The patients had a mean age of 50 ± 11.5 years, years, a mean body mass index of 26.7 ± 4.8 kg/m 2 , and a median follow-up time of 16.1 (range, 11.9-23.2) months. MSFN developed in 6% (n = 40) of patients. The random forest model demonstrated the best discriminatory performance (area under curve, 0.70), achieved a mean accuracy of 89% (95% confidence interval, 83-94), and identified 10 predictors of MSFN. Decision curve analysis demonstrated that ML models have a superior net benefit regardless of the probability threshold. Higher body mass index, older age, hypertension, subpectoral device placement, nipple-sparing mastectomy, axillary nodal dissection, and no acellular dermal matrix use were all independently associated with a higher risk of MSFN. CONCLUSIONS: ML algorithms trained on readily available perioperative clinical data can accurately predict the occurrence of MSFN and aid in individualized patient counseling, preoperative optimization, and surgical planning to reduce the risk of this devastating complication.

Is the Centralization of Complex Surgical Procedures an Unintended Spillover Effect of Global Capitation? - Insights from the Maryland Global Budget Revenue Program.

OBJECTIVE: To determine if global budget revenue (GBR) models incent the centralization of complex surgical care. SUMMARY BACKGROUND: In 2014, Maryland initiated a statewide GBR model. While prior research has shown improvements in cost and outcomes for surgical care post-GBR implementation, the mechanism remains unclear. METHODS: Utilizing state inpatient databases, we compared the proportion of adults undergoing elective complex surgeries (gastrectomy, pneumonectomy/lobectomy, proctectomies, and hip/knee revision) at high-concentration hospitals (HCHs) in Maryland and control states. Annual concentration, per procedure, was defined as hospital volume divided by state volume. HCHs were defined as hospitals with a concentration at least at the 75 th percentile in 2010. We estimated the difference-in-differences (DiD) of the probability of patients undergoing surgery at HCHs before and after GBR implementation. FINDINGS: Our sample included 122,882 surgeries. Following GBR implementation, all procedures were increasingly performed at HCHs in Maryland. States satisfied the parallel trends assumption for the centralization of gastrectomy and pneumonectomy/lobectomy. Post-GBR, patients were more likely to undergo gastrectomy (DiD: 5.5 p.p., 95% CI [2.2, 8.8]) and pneumonectomy/lobectomy (DiD: 12.4 p.p., 95% CI [10.0, 14.8]) at an HCH in Maryland compared with control states. For our hip/knee revision analyses, we assumed persistent counterfactuals and noted a positive DiD post-GBR implementation (DiD: 4.8 p.p., 95% CI [1.3, 8.2]). No conclusion could be drawn for proctectomy due to different pre-GBR trends. CONCLUSIONS: GBR implementation is associated with increased centralization for certain complex surgeries. Future research is needed to explore the impact of centralization on patient experience and access.

Determinants of telemedicine adoption among financially distressed patients with cancer during the COVID-19 pandemic: insights from a nationwide study.

PURPOSE: Telemedicine use during the COVID-19 pandemic among financially distressed patients with cancer, with respect to the determinants of adoption and patterns of utilization, has yet to be delineated. We sought to systematically characterize telemedicine utilization in financially distressed patients with cancer during the COVID-19 pandemic. METHODS: We conducted a cross-sectional analysis of nationwide survey data assessing telemedicine use in patients with cancer during the COVID-19 pandemic collected by Patient Advocate Foundation (PAF) in December 2020. Patients were characterized as financially distressed by self-reporting limited financial resources to manage out-of-pocket costs, psychological distress, and/or adaptive coping behaviors. Primary study outcome was telemedicine utilization during the pandemic. Secondary outcomes were telemedicine utilization volume and modality preferences. Multivariable and Poisson regression analyses were used to identify factors associated with telemedicine use. RESULTS: A convenience sample of 627 patients with cancer responded to the PAF survey. Telemedicine adoption during the pandemic was reported by 67% of patients, with most (63%) preferring video visits. Younger age (19-35 age compared to ≥ 75 age) (OR, 6.07; 95% CI, 1.47-25.1) and more comorbidities (≥ 3 comorbidities compared to cancer only) (OR, 1.79; 95% CI, 1.13-2.65) were factors associated with telemedicine adoption. Younger age (19-35 years) (incidence rate ratios [IRR], 1.78; 95% CI, 24-115%) and higher comorbidities (≥ 3) (IRR; 1.36; 95% CI, 20-55%) were factors associated with higher utilization volume. As area deprivation index increased by 10 units, the number of visits decreased by 3% (IRR 1.03, 95% CI, 1.03-1.05). CONCLUSIONS: The rapid adoption of telemedicine may exacerbate existing inequities, particularly among vulnerable financially distressed patients with cancer. Policy-level interventions are needed for the equitable and efficient provision of this service.

Assessment of Costs and Care Quality Associated With Major Surgical Procedures After Implementation of Maryland's Capitated Budget Model.

Importance: In 2014, Maryland implemented the all-payer model, a distinct hospital funding policy that applied caps on annual hospital expenditures and mandated reductions in avoidable complications. Expansion of this model to other states is currently being considered; therefore, it is important to evaluate whether Maryland's all-payer model is achieving the desired goals among surgical patients, who are an at-risk population for most potentially preventable complications. Objective: To examine the association between the implementation of Maryland's all-payer model and the incidence of avoidable complications and resource use among adult surgical patients. Design, Setting, and Participants: This comparative effectiveness study used hospital discharge records from the Healthcare Cost and Utilization Project state inpatient databases to conduct a difference-in-differences analysis comparing the incidence of avoidable complications and the intensity of health resource use before and after implementation of the all-payer model in Maryland. The analytical sample included 2 983 411 adult patients who received coronary artery bypass grafting (CABG), carotid endarterectomy (CEA), spinal fusion, hip or knee arthroplasty, hysterectomy, or cesarean delivery between January 1, 2008, and December 31, 2016, at acute care hospitals in Maryland (intervention state) and New York, New Jersey, and Rhode Island (control states). Data analysis was conducted from July 2019 to July 2021. Exposures: All-payer model. Main Outcomes and Measures: Complications (infectious, cardiovascular, respiratory, kidney, coagulation, and wound) and health resource use (ie, hospital charges). Results: Of 2 983 411 total patients in the analytical sample, 525 262 patients were from Maryland and 2 458 149 were from control states. Across Maryland and the control states, there were statistically significant but not clinically relevant differences in the preintervention period with regard to patient age (mean [SD], 49.7 [19.0] years vs 48.9 [19.3] years, respectively; P < .001), sex (22.7% male vs 21.4% male; P < .001), and race (0.3% vs 0.4% American Indian, 2.8% vs 4.5% Asian or Pacific Islander, 25.9% vs 12.7% Black, 4.7% vs 11.9% Hispanic, and 63.5% vs 63.4% White; P < .001). After implementation of the all-payer model in Maryland, significantly lower rates of avoidable complications were found among patients who underwent CABG (-11.3%; 95% CI, -13.8% to -8.7%; P < .001), CEA (-1.6%; 95% CI, -2.9% to -0.3%; P = .02), hip arthroplasty (-0.8%; 95% CI, -1.0% to -0.5%; P < .001), knee arthroplasty (-0.4%; 95% CI, -0.7% to -0.1%; P = .01), and cesarean delivery (-1.0%; 95% CI, -1.3% to -0.7%; P < .001). In addition, there were significantly lower increases in index hospital costs in Maryland among patients who underwent CABG (-$6236; 95% CI, -$7320 to -$5151; P < .001), CEA (-$730; 95% CI, -$1367 to -$94; P = .03), spinal fusion (-$3253; 95% CI, -$3879 to -$2627; P < .001), hip arthroplasty (-$328; 95% CI, -$634 to -$21; P = .04), knee arthroplasty (-$415; 95% CI, -$643 to -$187; P < .001), cesarean delivery (-$300; 95% CI, -$380 to -$220; P < .001), and hysterectomy (-$745; 95% CI, -$974 to -$517; P < .001). Significant changes in patient mix consistent with a younger population (eg, a shift toward private/commercial insurance for orthopedic procedures, such as spinal fusion [4.3%; 95% CI, 3.4%-5.2%; P < .001] and knee arthroplasty [1.6%; 95% CI, 1.0%-2.3%; P < .001]) and a lower comorbidity burden across surgical procedures (eg, CABG: -0.7% [95% CI, -0.1% to -0.5%; P < .001]; hip arthroplasty: -3.0% [95% CI, -3.6% to -2.3%; P < .001]) were also observed. Conclusions and Relevance: In this study, patients who underwent common surgical procedures had significantly fewer avoidable complications and lower hospital costs, as measured against the rate of increase throughout the study, after implementation of the all-payer model in Maryland. These findings may be associated with changes in the patient mix.

Understanding the Relationship between Breast Reconstruction Subtype and Risk of Financial Toxicity: A Single-Institution Pilot Study.

BACKGROUND: The impact of breast reconstruction on financial toxicity remains poorly understood despite growing awareness. The authors sought to illustrate the relationship between breast reconstruction subtypes and the risk of financial toxicity. METHODS: The authors conducted a single-institution cross-sectional survey of all female breast cancer patients undergoing any form of breast reconstruction between January of 2018 and June of 2019. Financial toxicity was measured by means of the validated Comprehensive Score for Financial Toxicity instrument. Demographics, clinical course, and coping strategies were abstracted from a purpose-built survey and electronic medical records. Multivariable linear regression was performed to identify associations with financial toxicity. RESULTS: The authors' analytical sample was 350 patients. One hundred eighty-four (52.6 percent) underwent oncoplastic reconstruction, 126 (36 percent) underwent implant-based reconstruction, and 40 (11.4 percent) underwent autologous reconstruction. Oncoplastic reconstruction recipients were older, had a higher body mass index, and were more likely to have supplemental insurance and receive adjuvant hormonal therapy. No significant differences in the risk of financial toxicity were uncovered across breast reconstruction subtypes (p = 0.53). Protective factors against financial toxicity were use of supplemental insurance (p = 0.0003) and escalating annual household income greater than $40,000 (p < 0.0001). Receipt of radiation therapy was positively associated with worsening financial toxicity (-2.69; 95 CI percent, -5.22 to -0.15). Financial coping strategies were prevalent across breast reconstruction subtypes. CONCLUSIONS: Breast reconstruction subtype does not differentially impact the risk of financial toxicity. Increasing income and supplemental insurance were found to be protective, whereas receipt of radiation therapy was positively associated with financial toxicity. Prospective, multicenter studies are needed to identify the main drivers of out-of-pocket costs and financial toxicity in breast cancer care.

Development of Machine Learning Algorithms for the Prediction of Financial Toxicity in Localized Breast Cancer Following Surgical Treatment.

PURPOSE: Financial burden caused by cancer treatment is associated with material loss, distress, and poorer outcomes. Financial resources exist to support patients but identification of need is difficult. We sought to develop and test a tool to accurately predict an individual's risk of financial toxicity based on clinical, demographic, and patient-reported data prior to initiation of breast cancer treatment. PATIENTS AND METHODS: We surveyed 611 patients undergoing breast cancer therapy at MD Anderson Cancer Center. We collected data using the validated COmprehensive Score for financial Toxicity (COST) patient-reported outcome measure alongside other financial indicators (credit score, income, and insurance status). We also collected clinical and perioperative data. We trained and tested an ensemble of machine learning (ML) algorithms (neural network, regularized linear model, support vector machines, and a classification tree) to predict financial toxicity. Data were randomly partitioned into training and test samples (2:1 ratio). Predictive performance was assessed using area-under-the-receiver-operating-characteristics-curve (AUROC), accuracy, sensitivity, and specificity. RESULTS: In our test sample (N = 203), 48 of 203 women (23.6%) reported significant financial burden. The algorithm ensemble performed well to predict financial burden with an AUROC of 0.85, accuracy of 0.82, sensitivity of 0.85, and specificity of 0.81. Key clinical predictors of financial burden from the linear model were neoadjuvant therapy (ßregularized, .11) and autologous, rather than implant-based, reconstruction (ßregularized, .06). Notably, radiation and clinical tumor stage had no effect on financial burden. CONCLUSION: ML models accurately predicted financial toxicity related to breast cancer treatment. These predictions may inform decision making and care planning to avoid financial distress during cancer treatment or enable targeted financial support. Further research is warranted to validate this tool and assess applicability for other types of cancer.

Self-Reported Risk Factors for Financial Distress and Attitudes Regarding Cost Discussions in Cancer Care: A Single-Institution Cross-Sectional Pilot Study of Breast Reconstruction Recipients.

BACKGROUND: High treatment costs associated with breast cancer are a substantial burden to patients and society. Despite mounting awareness, patient perspectives about the value of cost discussions in breast reconstruction and risk factors for financial distress are unknown. METHODS: The authors performed a single-institution, cross-sectional survey of all women who underwent breast reconstruction following mastectomy or lumpectomy for breast cancer or risk reduction. Questions were derived from previously published survey items, and the authors leveraged regression analysis to identify patient-level risk factors for major financial distress. RESULTS: A total of 647 of 2293 patients returned the survey questionnaires (28.2 percent response rate). From the 647 respondents, 399 (62 percent) underwent breast reconstruction, and of these, 140 (35 percent) reported that total treatment expenses were higher than expected. One hundred twenty-nine breast reconstruction patients (32 percent) paid over $5000 in out-of-pocket costs. Two hundred eighty-four (71 percent) felt that surgeons should explain the estimated out-of-pocket costs when choosing a type of breast reconstruction and 205 (51 percent) believed that a financial consultation should be scheduled with every new cancer diagnosis. However, only 52 patients (13 percent) reported having had cost discussions with the treatment team. The incidence of major financial distress was n = 70 (18 percent), and following regression analysis, higher credit score and annual income were associated with a 66 percent and 69 percent risk reduction, respectively. CONCLUSIONS: Recipients of breast reconstruction demonstrate unanticipated and unplanned financial strain related to out-of-pocket expenses and believe that cost-consciousness should impact treatment decisions. Lower income and credit score are associated with financial distress. Cost discussions may optimize decision-making in preference sensitive conditions.

Exploring provider- and practice-level drivers of cost-consciousness in breast cancer reconstruction-secondary analysis of a survey of the American Society of Plastic Surgeons.

BACKGROUND: The role of physicians in dampening health care costs is a renewed focus of policy-makers. We examined provider- and practice-level factors affecting four domains of cost-consciousness among plastic surgeons performing breast reconstruction. METHODS: Secondary analysis was performed on the survey responses of 329 surgeons who routinely performed breast reconstruction. Using a 5-point Likert scale, we queried four domains of cost-consciousness: out-of-pocket cost awareness, cost discussions, cognizance of patients' financial burden, and attitudes regarding cost discussions. Multivariable linear regression was performed to identify provider- and practice-level factors affecting these domains according to composite scores. RESULTS: Overall cost-consciousness scores (CS) were moderate and ranged from 2.14 to 4.30. There were no significant differences across practice settings. Male gender (p = 0.048), Hispanic ethnicity (p = 0.021), and increasing clinical experience (p = 0.015) were associated with higher out-of-pocket cost awareness. Increasing surgeon experience was also associated with having cost discussions (p = 0.039). No provider- or practice-level factors were associated with cognizance of patients' financial burden. Salaried physicians displayed a more positive attitude toward out-of-pocket cost discussions (p = 0.049). On pairwise testing, the out-of-pocket cost awareness was significantly different between Hispanic surgeons and white surgeons (4.30 vs. 3.55), and between surgeons with more than 20 years' experience and with less than 5 years' experience (3.87 vs. 3.37). CONCLUSIONS: Surgeon gender, ethnicity, and experience and practice compensation type inform various domains of cost-consciousness in breast reconstruction. Structural and behavioral interventions could possibly increase physicians' cost-consciousness.

Interview Scores Correlate with Fellow Microsurgical Skill and Performance.

BACKGROUND: The interview process for surgical trainees aims to select those individuals who will perform best during training and have the greatest potential as future surgeons. The objective of this study was to evaluate the relationship between criteria assessed at interview, technical skills, and performance, for the first time, to optimize the selection process for a Microsurgery fellowship. METHODS: Twenty microsurgery fellows in three consecutive annual cohorts at a single academic center were prospectively evaluated. At interview, subjects were scored for multiple standardized domains. At the start and at end of the fellowship, microsurgical technical skill was assessed both in the laboratory and operating room (OR) using a validated assessment tool. At the end of the fellowship, there was a final evaluation of performance. RESULTS: At the start, microsurgical skill significantly correlated with almost all domains evaluated at interview, most closely with prior plastic surgery training experience. At the end of the fellowship, skill level improved in all trainees, with the greatest improvement made by the lowest ranked and skilled trainees. The highest ranked trainees, however, made the greatest improvement in speed. CONCLUSIONS: The results of this study, for the first time, validate the current interview process to correctly select the highest performing and most skilled candidates and support the effectiveness of a 1-year microsurgical fellowship in improving microsurgical skill in all trainees, irrespective of their initial ability. The importance of valuing the relative quality of prior training and experience at selection is also highlighted.

Factors Correlating With Microsurgical Performance: A Clinical and Experimental Study.

BACKGROUND: Microsurgery is one the most complex surgical skills to master. The factors correlating with microsurgical performance, however, are poorly understood. Understanding these factors will aid in the training and assessment of microsurgeons. METHODS: A total of 29 microsurgery fellows enrolled in a dedicated 1-year microsurgery fellowship were included in the study. For the clinical evaluations, microsurgical anastomosis performance was evaluated during multiple procedures in the operating room at the start, midpoint, and end of the fellowship by all departmental faculty using a validated microsurgical assessment tool. For the laboratory evaluations, blinded video recordings of each fellow performing an arterial femoral anastomosis in a live rat model at the start and end of the fellowship were evaluated using 3 validated microsurgical global ratings scale tools. Correlations between performance and the factors assessed by the tools were evaluated. RESULTS: In the clinical study there were a total of 474 anastomosis evaluations; clinical performance correlated best with speed, instrument handling, and motion. In the laboratory study 58 evaluations were conducted, and performance tracked most closely with instrument handling, flow of operation, and operative steps, as well as correlating significantly inversely with time taken. The most common errors committed were unequal stitch bites, wrong grasp/damage tissue, and loose knot. CONCLUSIONS: Speed, both subjective and objective, instrument handling, operative flow, and motion, were relevant to performance of a microsurgical anastomosis. A prospective trial is now necessary to determine whether these factors should be considered in definitions of competency in microsurgery training pathways.

Robotic microsurgery: validating an assessment tool and plotting the learning curve.

BACKGROUND: The surgical robot has emerged as a potentially useful tool in microsurgery. The purpose of this study was to develop a validated assessment instrument and assess the learning curve for robotic microsurgery. The authors hypothesized that subjects would demonstrate improvement across multiple domains of performance with repetition of robotic microsurgical tasks. METHODS: In part 1, a novel assessment instrument called the Structured Assessment of Robotic Microsurgical Skills was tested. Four blinded expert evaluators graded six robotic microsurgery videos and interrater reliability was determined. In part 2, a cohort of 10 participants at various skill levels performed five robotic microvascular anastomoses. All 50 sessions were subjected to blind evaluation using the assessment instrument. Primary outcome measures included changes in operative time over the five sessions, and changes in assessment scores for all skill areas. RESULTS: Interrater reliability for the Structured Assessment of Robotic Microsurgical Skills instrument was excellent for each skill area, demonstrated by Cronbach alpha scores greater than 0.9 in each category across evaluators. All skill areas improved significantly for all participants, and operative time decreased for all participants over the course of the study. The results showed an initial steep ascent in technical skill acquisition followed by more gradual improvement, and a steady decrease in operative time to as short as 9 minutes. CONCLUSIONS: The Structured Assessment of Robotic Microsurgery Skills is a valid instrument for assessing microsurgical skill. Subjects at all levels of training, ranging from minimal microsurgical experience to expert microsurgeons, gained proficiency over the course of five robotic sessions.

Tracking the learning curve in microsurgical skill acquisition.

BACKGROUND: Despite advances in surgical training, microsurgery is still based on an apprenticeship model. To evaluate skill acquisition and apply targeted feedback to improve their training model, the authors applied the Structured Assessment of Microsurgery Skills to the training of microsurgical fellows. They hypothesized that subjects would demonstrate measurable improvement in performance throughout the study period and consistently across evaluators. METHODS: Seven fellows were evaluated during 118 microsurgical cases by 16 evaluators over three 1-month evaluation periods in 1 year (2010 to 2011). Evaluators used the Structured Assessment of Microsurgery Skills questionnaire, which measures dexterity, visuospatial ability, operative flow, and judgment. To validate the data, microsurgical anastomoses in rodents performed by the fellows in a laboratory at the beginning and end of the study period were evaluated by five blinded plastic surgeons using the same questionnaire. Primary outcomes were change in scores between evaluation periods and interevaluator reliability. RESULTS: Between the first two evaluation periods, all skill areas and overall performance improved significantly. Between the second two periods, most skill areas improved, but only a few improved significantly. Operative errors decreased significantly between the first and subsequent periods (81 versus 36; p < 0.05). In the laboratory study, all skills were significantly (p < 0.05) or marginally (0.05 ≤ p < 0.10) improved between time points. The overall interevaluator reliability of the questionnaire was acceptable (α = 0.67). CONCLUSIONS: The Structured Assessment of Microsurgery Skills questionnaire is a valid instrument for assessing microsurgical skill, providing individualized feedback with acceptable interevaluator reliability. Use of the questionnaire is anticipated to enhance microsurgical training.