Development and preliminary testing of Health Equity Across the AI Lifecycle (HEAAL): A framework for healthcare delivery organizations to mitigate the risk of AI solutions worsening health inequities.

The use of data-driven technologies such as Artificial Intelligence (AI) and Machine Learning (ML) is growing in healthcare. However, the proliferation of healthcare AI tools has outpaced regulatory frameworks, accountability measures, and governance standards to ensure safe, effective, and equitable use. To address these gaps and tackle a common challenge faced by healthcare delivery organizations, a case-based workshop was organized, and a framework was developed to evaluate the potential impact of implementing an AI solution on health equity. The Health Equity Across the AI Lifecycle (HEAAL) is co-designed with extensive engagement of clinical, operational, technical, and regulatory leaders across healthcare delivery organizations and ecosystem partners in the US. It assesses 5 equity assessment domains-accountability, fairness, fitness for purpose, reliability and validity, and transparency-across the span of eight key decision points in the AI adoption lifecycle. It is a process-oriented framework containing 37 step-by-step procedures for evaluating an existing AI solution and 34 procedures for evaluating a new AI solution in total. Within each procedure, it identifies relevant key stakeholders and data sources used to conduct the procedure. HEAAL guides how healthcare delivery organizations may mitigate the potential risk of AI solutions worsening health inequities. It also informs how much resources and support are required to assess the potential impact of AI solutions on health inequities.

Implementation of a validated post-operative opioid nomogram into clinical gynecologic surgery practice: A quality improvement initiative.

Objectives: The Gynecologic Oncology Postoperative Opioid use Predictive (GO-POP) calculator is a validated tool to provide evidence-based guidance on post-operative opioid prescribing. The objective of this study was to evaluate the impact of the implementation of GO-POP within an academic Gynecologic Oncology division. Methods: Two cohorts of patients (pre-implementation and post-implementation) who underwent surgery were compared with reference to GO-POP calculator implementation. All patients were included in the post-implementation group, regardless of GO-POP calculator use. An additional expanded-implementation cohort was used to compare pain control between GO-POP users and non-GO-POP users prospectively. Wilcoxon rank sum tests or ANOVA for continuous variables and Chi-square or Fisher's exact tests were used to categorical variables. Results: The median number of pills prescribed post-operatively decreased from 15 pills (Q1: 10, Q3: 20) to 10 pills (Q1: 8, Q3: 14.8) after implementation (p < 0.001). In the expanded-implementation cohort (293 patients), 41% patients were prescribed opioids using the GO-POP calculator. An overall median of 10 pills were prescribed with no difference by GO-POP calculator use (p = 0.26). Within the expanded-implementation cohort, refill requests (5% vs 9.2%; p = 0.26), clinician visits (0.8% vs 0.6%, p = 1), ED or urgent care visits (0% vs 2.3%, p = 0.15) and readmissions (0% vs 1.7%, p = 0.27) for pain did not differ between those prescribed opioids with and without the GO-POP calculator. Conclusions: A 33% reduction in post-operative opioid pills prescribed was seen following implementation of the GO-POP calculator into the Gynecologic Oncology division without increasing post-operative pain metrics or encounters for refill requests.

Accelerating health system innovation: principles and practices from the Duke Institute for Health Innovation.

The Duke Institute for Health Innovation (DIHI) was launched in 2013. Frontline staff members submit proposals for innovation projects that align with strategic priorities set by organizational leadership. Funded projects receive operational and technical support from institute staff members and a transdisciplinary network of collaborators to develop and implement solutions as part of routine clinical care, ranging from machine learning algorithms to mobile applications. DIHI's operations are shaped by four guiding principles: build to show value, build to integrate, build to scale, and build responsibly. Between 2013 and 2021, more than 600 project proposals have been submitted to DIHI. More than 85 innovation projects, both through the application process and other strategic partnerships, have been supported and implemented. DIHI's funding has incubated 12 companies, engaged more than 300 faculty members, staff members, and students, and contributed to more than 50 peer-reviewed publications. DIHI's practices can serve as a model for other health systems to systematically source, develop, implement, and scale innovations.

Implementation of a telehealth videoconference to improve hospital-to-skilled nursing care transitions: Preliminary data.

BACKGROUND: Transition-related patient safety errors are high among patients discharged from hospitals to skilled nursing facilities (SNFs), and interventions are needed to improve communication between hospitals and SNF providers. Our objective was to describe the implementation of a pilot telehealth videoconference program modeled after Extension for Community Health Outcomes-Care Transitions and examine patient safety errors and readmissions. METHODS: A multidisciplinary telehealth videoconference program was implemented at two academic hospitals for patients discharged to participating SNFs. Process measures, patient safety errors, and hospital readmissions were evaluated retrospectively for patients discussed at weekly conferences between July 2019-January 2020. Results were mapped to the constructs of the Reach, Effectiveness, Adoption, Implementation, Maintenance (RE-AIM) model. Descriptive statistics were reported for the conference process measures, patient and index hospitalization characteristics, and patient safety errors. The primary clinical outcome was all-cause 30-day readmissions. An intention-to-treat (ITT) analysis was conducted using logistic regression models fit to compare the probability of 30-day hospital readmission in patients discharged to participating SNFs across 7 months prior to after telehealth project implementation. RESULTS: There were 263 patients (67% of eligible patients) discussed during 26 telehealth videoconferences. Mean discussion time per patient was 7.7 min and median prep time per patient was 24.2 min for the hospital pharmacist and 10.3 min for the hospital clinician. A total of 327 patient safety errors were uncovered, mostly related to communication (54%) and medications (43%). Differences in slopes (program period vs. pre-implementation) of the probability of readmission across the two time periods were not statistically significant (OR 0.95, [95% CI 0.75, 1.19]). CONCLUSIONS: A pilot care innovations telehealth videoconference between hospital-based and SNF provider teams was successfully implemented within a large health system and enhanced care transitions by optimizing error-prone transitions. Future work is needed to understand process flow within nursing homes and its impact on clinical outcomes.

Development and Validation of Machine Learning Models to Predict Admission From Emergency Department to Inpatient and Intensive Care Units.

STUDY OBJECTIVE: This study aimed to develop and validate 2 machine learning models that use historical and current-visit patient data from electronic health records to predict the probability of patient admission to either an inpatient unit or ICU at each hour (up to 24 hours) of an emergency department (ED) encounter. The secondary goal was to provide a framework for the operational implementation of these machine learning models. METHODS: Data were curated from 468,167 adult patient encounters in 3 EDs (1 academic and 2 community-based EDs) of a large academic health system from August 1, 2015, to October 31, 2018. The models were validated using encounter data from January 1, 2019, to December 31, 2019. An operational user dashboard was developed, and the models were run on real-time encounter data. RESULTS: For the intermediate admission model, the area under the receiver operating characteristic curve was 0.873 and the area under the precision-recall curve was 0.636. For the ICU admission model, the area under the receiver operating characteristic curve was 0.951 and the area under the precision-recall curve was 0.461. The models had similar performance in both the academic- and community-based settings as well as across the 2019 and real-time encounter data. CONCLUSION: Machine learning models were developed to accurately make predictions regarding the probability of inpatient or ICU admission throughout the entire duration of a patient's encounter in ED and not just at the time of triage. These models remained accurate for a patient cohort beyond the time period of the initial training data and were integrated to run on live electronic health record data, with similar performance.

Solving the Value Equation: Assessing Surgeon Performance Using Risk-Adjusted Quality-Cost Diagrams and Surgical Outcomes.

Quality-cost diagrams have been used previously to assess interventions and their cost-effectiveness. This study explores the use of risk-adjusted quality-cost diagrams to compare the value provided by surgeons by presenting cost and outcomes simultaneously. Colectomy cases from a single institution captured in the National Surgical Quality Improvement Program database were linked to hospital cost-accounting data to determine costs per encounter. Risk adjustment models were developed and observed average cost and complication rates per surgeon were compared to expected cost and complication rates using the diagrams. Surgeons were surveyed to determine if the diagrams could provide information that would result in practice adjustment. Of 55 surgeons surveyed on the utility of the diagrams, 92% of respondents believed the diagrams were useful. The diagrams seemed intuitive to interpret, and making risk-adjusted comparisons accounted for patient differences in the evaluation.

The Readmission Event after Vascular Surgery: Causes and Costs.

BACKGROUND: The study evaluates the readmission diagnoses after vascular surgical interventions and the associated hospital costs. METHODS: Patients readmitted after undergoing carotid artery stenting (CAS), carotid endarterectomy (CEA), infrarenal endovascular abdominal aortic aneurysm repair (EVAR), open abdominal aortic aneurysm repair (OAAA), suprainguinal revascularization (SUPRA), or infrainguinal revascularization (INFRA) between January 1, 2008 and October 20, 2013 at a single academic institution were retrospectively identified. Demographic, preoperative, and postoperative event variables were obtained by chart review. The diagnoses and the costs of the readmission event were obtained by chart review and from hospital financial data. Readmission indications were grouped as unrelated or planned readmissions, procedure-specific complications, wound complications, cardiac causes, and other. Univariate analyses of categorical variables were performed with χ2 or Fisher exact test where appropriate. Continuous variables were analyzed using the Wilcoxon rank-sum test. RESULTS: A total of 1,170 patient records were identified. Thirty-day readmission occurred in 112 patients (9.6%). The readmission rate was significantly different between groups: 4.5% in CAS (n = 8/177), 8.5% in CEA (21/246), 5.8% in EVAR (18/312), 11.4% in OAAA (4/35), 15.6% in INFRA (33/212), 13.5% in SUPRA (24/178), and 40% in combined SUPRA and INFRA (4/10) (P < 0.0001). Readmissions were unrelated or planned in 19.6% of patients. Wound complications were the most common readmission diagnoses (36.6%, 41/112).There was a difference in the distribution of readmission indications among procedure groups, with wound complications being predominant in INFRA and SUPRA groups (60.6% and 58.3%, respectively), and cardiac events predominantly in EVAR patients (42%) (P < 0.001). In univariable analysis of predictors of readmission, significant preoperative factors were chronic obstructive pulmonary disease, renal insufficiency, and lower hematocrit. Significant postoperative predictors included any postoperative complication, number of complications, increased length of stay, wound complications, postoperative infections, blood transfusion, and reoperation. The median hospital cost for readmission for wound complications was 29,723 USD (interquartile range 23,841-36,878), and for cardiac complications was 39,784 USD (26,305-46,918). The median cost of readmission for bypass graft occlusion was 33,366 USD (20,530-43,170). The median length of stay also differed depending on the readmission diagnosis and was highest for bypass graft occlusion (8.5 days). CONCLUSIONS: Readmissions after vascular procedures are associated with high cost and hospital bed utilization. Wound complications continue to be the dominant readmission etiology. The characterization of these costs and risk factors in this study can allow for resource allocation to minimize preventable related readmissions. A significant proportion of readmissions after vascular interventions are planned or unrelated, which should be taken into consideration in metric benchmarking and performance comparisons.

Time is Money: Hospital Costs Associated With Video-Assisted Thoracoscopic Surgery Lobectomies.

BACKGROUND: Proposed changes in health care will place an increasing burden on surgeons to care for patients more efficiently to minimize cost. We reviewed costs surrounding video-assisted thoracoscopic surgery (VATS) lobectomies to see where changes could be made to ensure maximum value. METHODS: We queried The Society of Thoracic Surgeons database for all VATS lobectomies performed for lung cancer from January 2011 to December 2013. Clinical data were linked with hospital financial data to determine hospital expenditures for each patient. RESULTS: In all, 263 VATS lobectomies were included. Mean operating room time was 236 minutes, and median length of stay was 4 days. Mean hospital cost was $19,769. The majority of cost (58%) was attributed to operating room and floor costs (length of stay), and the majority of operating room costs were secondary to room rate and staplers. A total of 77 complications, as defined by STS, occurred in the cohort; 41 patients had only one complication, 11 patients had two complications, and 6 patients had three or more complications. The occurrence of one complication was associated with a net loss of $496 whereas two complications in a patient led to a $3,882 net loss. Overall, complications were independently correlated with significant cost increases. CONCLUSIONS: Our study shows that the most significant costs associated with VATS lobectomies relate to operating room time, stapler use, floor charges, and cost associated with complications. Cost-reducing strategies will need to concentrate on optimizing operating room times and reducing length of stay while simultaneously minimizing complications.

Tanning beds: A call to action for further educational and legislative efforts.

BACKGORUND AND OBJECTIVES: Melanoma is steadily increasing over the past decade. Recent studies confirmed a link between tanning bed use and melanoma. We sought to determine the prevalence and frequency of tanning bed among young patients with melanoma. METHODS: We retrospectively analyzed tanning bed use among young melanoma patients compared to controls selected from the hospital medical records. A telephone survey investigated family history of melanoma or skin cancer, hair color, eye color, skin type, tanning bed use, and patient awareness of dangers of tanning bed use. RESULTS: A total of 601 melanoma cases were identified; 265 (44%) completed the telephone survey as did 195 (31%) controls. Of these 460 subjects, 260 were female. Females were 3.0 times more likely to have used a tanning bed. Melanoma patients had natural light color hair, blue-green eyes, and 2.0 times more likely to use a tanning bed than controls. Among the tanning bed users, 90% were aware of danger of tanning bed. CONCLUSIONS: Our study found that tanning beds were more likely to be used by young women, the majority of whom are aware of the associated risks. Eliminating the use of tanning beds should be considered to decrease the incidence of melanoma.

Association between hospital finances, payer mix, and complications after hyperthermic intraperitoneal chemotherapy: deficiencies in the current healthcare reimbursement system and future implications.

BACKGROUND: Despite increasing implementation of cytoreductive surgery (CRS) and hyperthermic intraperitoneal chemotherapy (HIPEC), there are little data on its financial implications. We analyzed hospital cost and reimbursement data within the context of insurance provider type and postoperative complications. METHODS: Clinicopathologic variables, hospital costs, and reimbursement for all patients undergoing CRS/HIPEC at a single institution from 2009 to 2013 were analyzed. RESULTS: A total of 64 patients underwent CRS/HIPEC. Median PCI score was 19, and average operative time was 550 min. Tumor histology included appendiceal (n = 40; 62 %), colorectal (n = 16; 25 %), goblet cell (n = 5; 8 %), and mesothelioma (n = 3; 5 %). Median length-of-stay was 13 days. Complications occurred in 42 patients (66 %), including 13 (20 %) with major (Clavien grade III-IV) complications. Payer mix included 42 private insurance and 22 Medicare/Medicaid. Financial data was available for 56 patients: average total hospital cost was $49,248 and reimbursement was $63,771, for a hospital profit of $14,523/patient. Despite similar costs between Medicare/Medicaid and private-insurance patients, Medicare/Medicaid reimbursed much less ($30,713 vs $80,747; p < 0.001), resulting in a net loss of $17,342 per patient. For private-insured patients, major complications were associated with increased cost and increased reimbursement, resulting in a net profit of $36,285, compared with a net loss of $54,274 in Medicare/Medicaid patients. CONCLUSIONS: CRS/HIPEC is profitable in privately insured patients, even for those with major complications, but loses money in patients with Medicare/Medicaid. Under a future bundled-reimbursement system, complications will be negatively associated with profit. With these impending changes, hospitals must place emphasis on value, recalculate the reimbursement necessary for financial viability, and focus on decreasing costs and minimizing complications.

The Association Between Hospital Finances and Complications After Complex Abdominal Surgery: Deficiencies in the Current Health Care Reimbursement System and Implications for the Future.

OBJECTIVE: To determine the relationship between complications after 3 common general surgery procedures and per-episode hospital finances. BACKGROUND: With impending changes in health care reimbursement, maximizing the value of care delivered is paramount. Data on the relative clinical and financial impact of postoperative complications are necessary for directing surgical quality improvement efforts. METHODS: We reviewed the medical records of patients enrolled in the American College of Surgeons' National Surgical Quality Improvement Program who underwent pancreaticoduodenectomy, hepatectomy, and colectomy at a single academic institution between September 2009 and August 2012. Clinical outcomes data were subsequently linked with hospital billing data to determine hospital finances associated with each episode. We describe the association between postoperative complications, hospital length of stay, and different financial metrics. Multivariable linear regression modeling tested linear association between postoperative outcomes and cost data. RESULTS: There was a positive association between the number of surgical complications, payments, length of stay, total charges, total costs, and contribution margin for the three procedures. Multivariable models indicated that complications were independently associated with total cost among the selected procedures. Payments increased with complications, offsetting increased costs. CONCLUSIONS: In the current fee-for-service environment, the financial incentives are misaligned with quality improvement efforts. As we move to a value-driven method of reimbursement, administrators and health care providers alike will need to focus on improving the quality of patient care while remaining conscious of the cost of care delivered. Reducing complications effectively improves value.

Cost comparison analysis of open versus laparoscopic distal pancreatectomy.

BACKGROUND: In comparison with open distal pancreatectomy (ODP), laparoscopic distal pancreatectomy (LDP) is associated with fewer complications and shorter hospital stays, but comparative cost data for the two approaches are limited. METHODS: Records of all distal pancreatectomies carried out from January 2009 to June 2013 were reviewed and stratified according to operative complexity. Patient factors and outcomes were recorded. Total variable costs (TVCs) were tabulated for each patient, and stratified by category [e.g. 'floor', 'operating room' (OR), 'radiology']. Costs for index admissions and 30-day readmissions were compared between LDP and ODP groups. RESULTS: Of 153 procedures, 115 (70 LDP, 45 ODP) were selected for analysis. The TVC of the index admission was US$3420 less per patient in the LDP group (US$10 480 versus US$13 900; P = 0.06). Although OR costs were significantly greater in the LDP cohort (US$5756 versus US$4900; P = 0.02), the shorter average hospitalization in the LDP group (5.2 days versus 7.7 days; P = 0.01) resulted in a lower overall cost. The total cost of index hospitalization combined with readmission was significantly lower in the LDP cohort (US$11 106 versus US$14 803; P = 0.05). CONCLUSIONS: In appropriately selected patients, LDP is more cost-effective than ODP. The increased OR cost associated with LDP is offset by the shorter hospitalization. These data clarify targets for further cost reductions.