Habermann, Elizabeth

ABSTRACT Background Evaluating fairness in artificial intelligence (AI) is essential before clinical implementation to prevent unintended harm across diverse patient populations. This study assesses performance differences of an image-based surgical site infection (SSI) detection model across age, gender, geography, race, and skin color. Methods Model development/validation study using retrospective, multicenter data across surgical subspecialties from adults who submitted an image of their incision via the electronic health record within 30 days of their procedure between 2019-2024. Area under receiver operating curve (AUROC) was compared by age (<55 vs. ≥55 years old), gender (male vs. female), geography (Rochester, Minnesota vs. Jacksonville, Florida vs. Scottsdale, Arizona vs. regional Minnesota/Wisconsin areas), race (White/Caucasian vs. non-White/Caucasian), and skin color (Fitpatrick I-III vs. IV-VI and Monk A-E vs. F-J). Results Training data contained 13,702 images from 4,274 patients between 2019-2022 (50.1% ≥55 years old; 63.2% female; 91.6% White/Caucasian; 58.5% Rochester, 17.3% Florida, 10.5% Arizona; 96.3% light skin tone), of which 1,348 images (9.8%) had an SSI. No differences in SSI detection existed for age, gender, race, or skin color (AUROC 0.83-0.87, p>0.09), however geographical differences were significant: lowest AUROC 0.83 (95% CI 0.77-0.88) vs. highest AUROC 0.87 (95% CI 0.84-0.89, p=0.01). Temporal validation using 4,242 images from 1,353 patients between 2022-2023 showed intact fairness across subgroups (ΔAUROC ≤0.06) except geography (ΔAUROC ≤0.16). Conclusions The model behaved fairly across age, gender, race, and skin color, however geographical differences persisted. These findings have practical implications underscoring the importance of evaluating bias in clinical AI before deployment.

ABSTRACT Background ACS‐NSQIP began capturing operative approach in 2022. With increased adoption of minimally invasive (MIS) techniques for malignancy, a contemporary review of benchmark outcomes for open, laparoscopic and robotic gastrectomy is warranted. Methods We performed a retrospective cohort study of adults in ACS‐NSQIP (2022–2024) undergoing partial or total gastrectomy (CPT 43631‐4 and 43620‐2) for cancer. Outcomes included operative time, length of stay (LOS), unplanned conversion‐to‐open, and 30‐day morbidity, mortality, and readmission. Multivariable logistic regression evaluated patient and operative factors associated with morbidity. Results Among 1,459 patients (median age 66, 60.4% male), 66.8% underwent partial gastrectomy and 33.2% total gastrectomy. Approaches were 61.2% open, 19.5% laparoscopic, and 19.3% robotic. Thirty‐day mortality (1.8%), major complications (18.2%), and readmissions (10.3%) did not differ by approach. Unplanned conversion‐to‐open occurred in 14.0% of MIS. Robotic had longer median operative time of +1.2 and +1.7 h than laparoscopic and open and median −1 and −2 days shorter LOS respectively ( p < 0.001). Hypoalbuminemia (OR 3.10), CKD ≥ 3 (OR 1.75), and anemia (OR 1.48) were independently associated with major complications ( p < 0.05). Conclusion Operative approach influences operative time and LOS but not short‐term morbidity and mortality. The newly available ACS‐NSQIP operative approach variable enables approach‐specific benchmarks for cancer‐related gastrectomy.

Abstract Background Hospital‐at‐home programs (HaH) in the United States have evolved to include a virtual‐hybrid delivery model, where all physician encounters are virtual and partnered with a home care team. Objective To examine whether a virtual hybrid HaH program enabled by technology has similar clinical outcomes to traditional brick‐and‐mortar (B&M) hospital care. Methods We conducted a pragmatic trial at three hospitals, randomizing 1150 acutely ill patients requiring hospital care between July 10, 2023, and October 31, 2023 one‐to‐one into two groups: intervention (HaH) and control (B&M). The primary analysis was an intention‐to‐treat non‐inferiority analysis of the primary outcome, which was a composite of 30‐day all‐cause mortality and unplanned readmissions. Secondary outcomes included 30‐day readmission, all‐cause mortality, and patient experience. Results The mean age was 67.8 (standard deviation [SD] 16.3) years, and 52.2% were female. The primary outcome occurred in 99 (17.3%) HaH patients and 114 (19.8%) B&M patients (odds ratio [OR] 0.85, 95% confidence interval [CI] 0.63–1.14, p = .28), meeting the non‐inferiority criterion. Thirty‐day unplanned readmission occurred in 84 (14.7%) HaH patients and 101 (17.5%) B&M patients (OR 0.81, 95% CI 0.59–1.11, p = .19). Thirty‐day all‐cause mortality occurred in 25 (4.4%) HaH patients and 19 (3.3%) B&M patients (OR 1.34, 95% CI 0.73–2.46, p = .35). No HaH patients died while receiving their hospital care at home. HaH program was associated with a higher likelihood of patients reporting feeling extremely comfortable or very comfortable (84.4% HaH; 60.9% B&M, p = .001). Conclusion A hybrid HaH model is a safe and comfortable alternative to traditional B&M hospital care.

Background: Traditional Medicare (TM) claims data are widely used by researchers and registries to report survivorship following total knee arthroplasty (TKA). The purpose of the present study was to investigate whether the mass exodus of patients from TM to Medicare Advantage (MA) has compromised the fidelity of TM data. Methods: We identified 11,717 Medicare-eligible patients (15,282 knees) who had undergone primary TKA from 2000 to 2020 at a single institution. Insurance type was analyzed, and 84% of TKAs were covered by TM. The rates of survivorship free from revision or reoperation were calculated for patients with TM coverage. The same survivorship end points were recalculated after censoring of patients who transitioned to MA after primary TKA, thereby modeling the impact of losing patients from the TM dataset. Differences in survivorship were compared. The mean duration of follow-up was 10 years. Results: From 2000 to 2020, there was a decrease in TM insurance (from 94% to 68%) and a corresponding increase in MA insurance (from 0% to 19%) among patients undergoing TKA. Following TKA, 25% of patients with TM coverage switched to MA. For patients with TM at the time of surgery, the 15-year rates of survivorship free from any reoperation or revision were 90% and 96%, respectively. When patients were censored upon transition from TM to MA, the 15-year rates of survivorship free from any reoperation (92% versus 90%; hazard ratio [HR] = 1.2; p = 0.001) or any revision (97% versus 96%; HR = 1.3; p = 0.002) were significantly higher. Conclusions: One in 4 patients left TM for MA after primary TKA, effectively making them lost to follow-up within TM datasets. The mass exodus of patients out of TM resulted in falsely elevated estimates of survivorship free from reoperation and from revision, with increasing divergence in survivorship over time, when MA data were excluded. As MA continues to grow, efforts to incorporate these data will become increasingly important. Level of Evidence: Therapeutic Level III . See Instructions for Authors for a complete description of levels of evidence.

ABSTRACT Background Surgical resection for pancreas ductal adenocarcinoma (PDAC) remains the mainstay of treatment. Regardless of enhanced survival rates, disparities in patient outcomes and experiences driven by the social determinants of health (SDH) persist. The area deprivation index (ADI) comprises 17 indicators related to education, employment, housing‐quality, and poverty. We sought to examine the impact of ADI on surgical outcomes of PDAC patients. Methods Patients who underwent pancreatoduodenectomy or distal pancreatectomy for PDAC between January 2011 and December 2022 were identified. ADI was calculated using patient zip codes and categorized into quartiles, with the highest indicating the most marginalized. The primary outcome was loss to follow‐up after surgical resection. To account for confounders, a competing risk multivariable regression analysis was used. Results A total of 1001 patients had a mean age of 66.6 (±9.64), with 46.3% ( n = 463) of patients being female. A majority, 94.6% ( n = 947), identified as white, and 64.1% ( n = 641) had at least a college degree. The median length of follow‐up after surgery was 1.8 years (interquartile range: 0.9–3.5). Multivariable analysis adjusting for competing risk of death, showed that patients who are least marginalized are 1.57 times more likely to have their follow‐up than those most marginalized (hazard ratio: 1.57, 95% confidence interval: 1.08–2.29, p = 0.017). Conclusion SDH impact many aspects of patient's care including a higher risk of loss to follow‐up for marginalized patients after surgery for PDAC. Future efforts should seek to identify and lower barriers faced by marginalized patients with system‐level changes to ensure equitable access.

Individual or pooled commercial claims data sources such as the IBM MarketScan and PearlDiver provide information from health-care encounters by individuals enrolled in participating health insurance plans. These data sources contain deidentified data on demographic characteristics, enrollment start and end dates, inpatient and outpatient procedures and medical diagnoses with associated service dates and settings, and dispensed medications. Although there are concerns that long-term follow-up is limited because of interruptions in the continuity of coverage and reliance on billing data may overrecord or underrecord diagnoses and confounders, these data sources are nevertheless valuable for orthopaedic studies focusing on short-term complications, costs, and utilization.

Abstract: Direct data abstraction from a patient’s chart by experienced medical professional data abstractors has been the historical gold standard for quality and accuracy in clinical medical research. The limiting challenge to population-wide studies for quality and public health purposes is the high personnel costs associated with very large-scale efforts of this type. Two historically related programs that are at least partially able to successfully circumvent this problem and provide high-quality data relating to surgical procedures and the early postoperative period are reviewed in this article. Both utilize similar data abstraction efforts by specially trained and qualified medical abstractors of a sample subset of the total procedures performed at participating hospitals. The Veterans Affairs Surgical Quality Improvement Program (VASQIP), detailed by Nicholas J. Giori, MD, PhD, in the first section of this article, makes use of trained abstractors and has undergone recent additions and updates, including the development of an associated total hip registry for the VA system. The data elements and data integrity provided by both of these programs establish important benchmarks for other “big data” efforts, which often attempt to use alternative less-expensive methods of data collection in order to achieve more widespread or even nationwide data collection. In the second section, Elizabeth B. Habermann, PhD, MPH, provides a detailed review of the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP), the data elements collected, and examples of the range of quality improvement and outcomes studies in orthopaedic surgery that it has made possible, along with information on data that have not been collected and the resulting limitations. The ACS NSQIP was actually modeled after the very similar earlier effort started by the United States Department of Veterans Affairs (VA).

BACKGROUND: The development of major low anterior resection syndrome (LARS) after low anterior resection is severely detrimental to quality of life, yet awareness of it by clinicians and patients and the frequency of treatment of LARS is unclear. STUDY DESIGN: Patients who underwent low anterior resection for sigmoid or rectal cancer at a tertiary center between 2007 and 2017 (n = 798) were surveyed in 2019 to assess LARS symptoms and report medications or treatment received for LARS. LARS scores were calculated (score range 0–42) and normalized to published data on LARS prevalence in the general population in Europe, stratified by age (<50 or ≥50) and sex. RESULTS: Of the 594 patients (74%) who returned the survey, 255 (43%) were identified as having major LARS (LARS score ≥30). This prevalence was significantly higher than published normative data from Denmark and Amsterdam when stratified by age greater than or less than 50 and sex. Patients with major LARS infrequently reported current use of first-line therapies (antidiarrheal medications 32%, fiber supplements 16%, and both 13%). Only 3% reported receiving second-line therapy of transanal irrigations and/or pelvic floor rehabilitation, and only 1% had undergone third-line therapy of sacral nerve stimulator implantation. CONCLUSION: Major LARS is common yet seemingly underrecognized by clinicians because less than half of patients are on first-line therapy and practically none are on second- and third-line therapies. Long-term follow-up of patients after low anterior resection, improved preoperative and postoperative education, and continued symptom assessment is necessary to improve treatment of major LARS.