RESUMO
The scheduling of operating room (OR) slots requires the accurate prediction of surgery duration. We evaluated the performance of existing Moving Average (MA) based estimates with novel machine learning (ML)-based models of surgery durations across two sites in the US and Singapore. We used the Duke Protected Analytics Computing Environment (PACE) to facilitate data-sharing and big data analytics across the US and Singapore. Data from all colorectal surgery patients between 1 January 2012 and 31 December 2017 in Singapore and, 1 January 2015 to 31 December 2019 in the US were used, and 7585 cases and 3597 single and multiple procedure cases from Singapore and US were included. The ML models were based on categorical gradient boosting (CatBoost) models trained on common data fields shared by both institutions. The procedure codes were based on the Table of Surgical Procedure (TOSP) (Singapore) and the Current Procedural Terminology (CPT) codes (US). The two types of codes were mapped by surgical experts. The CPT codes were then transformed into the relative value unit (RVU). The ML models outperformed the baseline MA models. The MA, scheduled durations and procedure codes were found to have higher loadings as compared to surgeon factors. We further demonstrated the use of the Duke PACE in facilitating data-sharing and big data analytics.
RESUMO
OBJECTIVE: To develop a 2-stage discrete events simulation (DES) based framework for the evaluation of elective surgery cancellation strategies and resumption scenarios across multiple operational outcomes. MATERIALS AND METHODS: Study data was derived from the data warehouse and domain knowledge on the operational process of the largest tertiary hospital in Singapore. 34,025 unique cases over 43 operating rooms (ORs) and 18 surgical disciplines performed from 1 January 2019 to 31 May 2020 were extracted for the study. A clustering approach was used in stage 1 of the modelling framework to develop the groups of surgeries that followed distinctive postponement patterns. These clusters were then used as inputs for stage 2 where the DES model was used to evaluate alternative phased resumption strategies considering the outcomes of OR utilization, waiting times to surgeries and the time to clear the backlogs. RESULTS: The tool enabled us to understand the elective postponement patterns during the COVID-19 partial lockdown period, and evaluate the best phased resumption strategy. Differences in the performance measures were evaluated based on 95% confidence intervals. The results indicate that two of the gradual phased resumption strategies provided lower peak OR and bed utilizations but required a longer time to return to BAU levels. Minimum peak bed demands could also be reduced by approximately 14 beds daily with the gradual resumption strategy, whilst the maximum peak bed demands by approximately 8.2 beds. Peak OR utilization could be reduced to 92% for gradual resumption as compared to a minimum peak of 94.2% with the full resumption strategy. CONCLUSIONS: The 2-stage modelling framework coupled with a user-friendly visualization interface were key enablers for understanding the elective surgery postponement patterns during a partial lockdown phase. The DES model enabled the identification and evaluation of optimal phased resumption policies across multiple important operational outcome measures. LAY ABSTRACT: During the height of the COVID-19 pandemic, most healthcare systems suspended their non-urgent elective surgery services. This strategy was undertaken as a means to expand surge capacity, through the preservation of structural resources (such as operating theaters, ICU beds, and ventilators), consumables (such as personal protective equipment and medications), and critical healthcare manpower. As a result, some patients had less-essential surgeries postponed due to the pandemic. As the first wave of the pandemic waned, there was an urgent need to quickly develop optimal strategies for the resumption of these surgeries. We developed a 2-stage discrete events simulation (DES) framework based on 34,025 unique cases over 43 operating rooms (ORs) and 18 surgical disciplines performed from 1 January 2019 to 31 May 2020 captured in the Singapore General Hospital (SGH) enterprise data warehouse. The outcomes evaluated were OR utilization, waiting times to surgeries and time to clear the backlogs. A user-friendly visualization interface was developed to enable decision makers to determine the most promising surgery resumption strategy across these outcomes. Hospitals globally can make use of the modelling framework to adapt to their own surgical systems to evaluate strategies for postponement and resumption of elective surgeries.
RESUMO
BACKGROUND: Type 2 diabetes mellitus (T2DM) and its related complications represent a growing economic burden for many countries and health systems. Diabetes complications can be prevented through better disease control, but there is a large gap between the recommended treatment and the treatment that patients actually receive. The treatment of T2DM can be challenging because of different comprehensive therapeutic targets and individual variability of the patients, leading to the need for precise, personalized treatment. OBJECTIVE: The aim of this study was to develop treatment recommendation models for T2DM based on deep reinforcement learning. A retrospective analysis was then performed to evaluate the reliability and effectiveness of the models. METHODS: The data used in our study were collected from the Singapore Health Services Diabetes Registry, encompassing 189,520 patients with T2DM, including 6,407,958 outpatient visits from 2013 to 2018. The treatment recommendation model was built based on 80% of the dataset and its effectiveness was evaluated with the remaining 20% of data. Three treatment recommendation models were developed for antiglycemic, antihypertensive, and lipid-lowering treatments by combining a knowledge-driven model and a data-driven model. The knowledge-driven model, based on clinical guidelines and expert experiences, was first applied to select the candidate medications. The data-driven model, based on deep reinforcement learning, was used to rank the candidates according to the expected clinical outcomes. To evaluate the models, short-term outcomes were compared between the model-concordant treatments and the model-nonconcordant treatments with confounder adjustment by stratification, propensity score weighting, and multivariate regression. For long-term outcomes, model-concordant rates were included as independent variables to evaluate if the combined antiglycemic, antihypertensive, and lipid-lowering treatments had a positive impact on reduction of long-term complication occurrence or death at the patient level via multivariate logistic regression. RESULTS: The test data consisted of 36,993 patients for evaluating the effectiveness of the three treatment recommendation models. In 43.3% of patient visits, the antiglycemic medications recommended by the model were concordant with the actual prescriptions of the physicians. The concordant rates for antihypertensive medications and lipid-lowering medications were 51.3% and 58.9%, respectively. The evaluation results also showed that model-concordant treatments were associated with better glycemic control (odds ratio [OR] 1.73, 95% CI 1.69-1.76), blood pressure control (OR 1.26, 95% CI, 1.23-1.29), and blood lipids control (OR 1.28, 95% CI 1.22-1.35). We also found that patients with more model-concordant treatments were associated with a lower risk of diabetes complications (including 3 macrovascular and 2 microvascular complications) and death, suggesting that the models have the potential of achieving better outcomes in the long term. CONCLUSIONS: Comprehensive management by combining knowledge-driven and data-driven models has good potential to help physicians improve the clinical outcomes of patients with T2DM; achieving good control on blood glucose, blood pressure, and blood lipids; and reducing the risk of diabetes complications in the long term.
Assuntos
Diabetes Mellitus Tipo 2 , Glicemia , Diabetes Mellitus Tipo 2/tratamento farmacológico , Humanos , Reprodutibilidade dos Testes , Estudos Retrospectivos , Resultado do TratamentoRESUMO
AIM: We propose a nurse scheduling framework based on a set of performance measures that are aligned with multiple outcome measures. A case study for the emergency department is presented. METHODS: A total of 142,564 emergency department attendances over 1 year were included in this study. Operational requirements, constraints and historical workload data were translated into a mixed-integer sequential goal programming model, which considers the following outcome measures: (1) nurse-patient ratios; (2) number of favourable/unfavourable shifts; and (3) dispersion of rest days. Computational studies compared the performance of the mixed-integer sequential goal programming results with manually generated historical nurse schedules. RESULTS: The maximum nurse-patient ratio deviation against the target was approximately 10% compared to 47% generated by the historical rosters (a 10% deviation translates to approximately two nurses). An on-line decision support system, which integrates shift preferences, staff databases and a workload forecasting module, was also developed. CONCLUSION: A decision support system based on the mixed-integer sequential goal programming modelling framework was proposed. The application of the model in a case study for an emergency department demonstrated improvements over existing manual scheduling methods. IMPLICATIONS FOR NURSING MANAGEMENT: This study demonstrates a mathematical, programming-based decision support system, which allows for managerial priorities and nurse preferences to be jointly considered in the automatic generation of nurse rosters.