Time series modelling and forecasting of emergency department overcrowding

Towards reliable forecasting of healthcare capacity needs: A scoping review and evidence mapping

BACKGROUND

The COVID-19 pandemic has highlighted the critical importance of robust healthcare capacity planning and preparedness for emerging crises. However, healthcare systems must also adapt to more gradual temporal changes in disease prevalence and demographic composition over time. To support proactive healthcare planning, statistical capacity forecasting models can provide valuable information to healthcare planners. This systematic literature review and evidence mapping aims to identify and describe studies that have used statistical forecasting models to estimate healthcare capacity needs within hospital settings.

METHOD

Studies were identified in the databases MEDLINE and Embase and screened for relevance before items were defined and extracted within the following categories: forecast methodology, measure of capacity, forecast horizon, healthcare setting, target diagnosis, validation methods, and implementation.

RESULTS

84 studies were selected, all focusing on various capacity outcomes, including number of hospital beds/ patients, staffing, and length of stay. The selected studies employed different analytical models grouped in six items; discrete event simulation (N = 13, 15 %), generalized linear models (N = 21, 25 %), rate multiplication (N = 15, 18 %), compartmental models (N = 14, 17 %), time series analysis (N = 22, 26 %), and machine learning not otherwise categorizable (N = 12, 14 %). The review further provides insights into disease areas with infectious diseases (N = 24, 29 %) and cancer (N = 12, 14 %) being predominant, though several studies forecasted healthcare capacity needs in general (N = 24, 29 %). Only about half of the models were validated using either temporal validation (N = 39, 46 %), cross-validation (N = 2, 2 %) or/and geographical validation (N = 4, 5 %).

CONCLUSION

The forecasting models' applicability can serve as a resource for healthcare stakeholders involved in designing future healthcare capacity estimation. The lack of routine performance validation of the used algorithms is concerning. There is very little information on implementation and follow-up validation of capacity planning models.

Feasibility of forecasting future critical care bed availability using bed management data

OBJECTIVES

This project aims to determine the feasibility of predicting future critical care bed availability using data-driven computational forecast modelling and routinely collected hospital bed management data.

METHODS

In this proof-of-concept, single-centre data informatics feasibility study, regression-based and classification data science techniques were applied retrospectively to prospectively collect routine hospital-wide bed management data to forecast critical care bed capacity. The availability of at least one critical care bed was forecasted using a forecast horizon of 1, 7 and 14 days in advance.

RESULTS

We demonstrated for the first time the feasibility of forecasting critical care bed capacity without requiring detailed patient-level data using only routinely collected hospital bed management data and interpretable models. Predictive performance for bed availability 1 day in the future was better than 14 days (mean absolute error 1.33 vs 1.61 and area under the curve 0.78 vs 0.73, respectively). By analysing feature importance, we demonstrated that the models relied mainly on critical care and temporal data rather than data from other wards in the hospital.

DISCUSSION

Our data-driven forecasting tool only required hospital bed management data to forecast critical care bed availability. This novel approach means no patient-sensitive data are required in the modelling and warrants further work to refine this approach in future bed availability forecast in other hospital wards.

CONCLUSIONS

Data-driven critical care bed availability prediction was possible. Further investigations into its utility in multicentre critical care settings or in other clinical settings are warranted.

Estimating emergency department crowding with stochastic population models

Environments such as shopping malls, airports, or hospital emergency-departments often experience crowding, with many people simultaneously requesting service. Crowding highly fluctuates, with sudden overcrowding "spikes". Past research has either focused on average behavior, used context-specific models with a large number of parameters, or machine-learning models that are hard to interpret. Here we show that a stochastic population model, previously applied to a broad range of natural phenomena, can aptly describe hospital emergency-department crowding. We test the model using data from five-year minute-by-minute emergency-department records. The model provides reliable forecasting of the crowding distribution. Overcrowding is highly sensitive to the patient arrival-flux and length-of-stay: a 10% increase in arrivals triples the probability of overcrowding events. Expediting patient exit-rate to shorten the typical length-of-stay by just 20 minutes (8.5%) cuts the probability of severe overcrowding events by 50%. Such forecasting is critical in prevention and mitigation of breakdown events. Our results demonstrate that despite its high volatility, crowding follows a dynamic behavior common to many systems in nature.

Evaluating the Patient Boarding during Omicron Surge in Hong Kong: Time Series Analysis

The fifth wave of COVID-19 outbreaks in Hong Kong (HK) from January to March 2022 has the highest confirmed cases and deaths compared with previous waves. Severe hospital boarding (to inpatient wards) was noted in various Emergency Departments (EDs). Our objective is to identify factors associated with hospital boarding during Omicron surge in HK. We conducted a retrospective cohort study including all ED visits and inpatient (IP) ward admissions from January 1st to March 31st, 2022. Vector Autoregression model evaluated the effects of a single variable on the targeted hospital boarding variables. Admissions from elderly homes with 6 lag days held the highest positive value of statistical significance (t-stat = 2.827, P < .05) caused prolonged admission waiting time, while medical patients with 4 lag days had the highest statistical significance (t-stat = 2.530, P < .05) caused an increased number of boarding patients. Within one week after impulses, medical occupancy's influence on the waiting time varied from 0.289 on the 1st day to -0.315 on the 7th day. While occupancy of medical wards always positively affected blocked number of patients, and its response was maximized at 0.309 on the 2nd day. Number of confirmed COVID-19 cases was not the sole significant contributor, while occupancy of medical wards was still a critical factor associated with patient boarding. Increasing ward capacity and controlling occupancy were suggested during the outbreak. Moreover, streamlining elderly patients in ED could be an approach to relieve pressure on the healthcare system.

Artificial intelligence and discrete-event simulation for capacity management of intensive care units during the Covid-19 pandemic: A case study

The Covid-19 pandemic has pushed the Intensive Care Units (ICUs) into significant operational disruptions. The rapid evolution of this disease, the bed capacity constraints, the wide variety of patient profiles, and the imbalances within health supply chains still represent a challenge for policymakers. This paper aims to use Artificial Intelligence (AI) and Discrete-Event Simulation (DES) to support ICU bed capacity management during Covid-19. The proposed approach was validated in a Spanish hospital chain where we initially identified the predictors of ICU admission in Covid-19 patients. Second, we applied Random Forest (RF) to predict ICU admission likelihood using patient data collected in the Emergency Department (ED). Finally, we included the RF outcomes in a DES model to assist decision-makers in evaluating new ICU bed configurations responding to the patient transfer expected from downstream services. The results evidenced that the median bed waiting time declined between 32.42 and 48.03 min after intervention.

XAIRE: An ensemble-based methodology for determining the relative importance of variables in regression tasks. Application to a hospital emergency department

Nowadays it is increasingly important in many applications to understand how different factors influence a variable of interest in a predictive modeling process. This task becomes particularly important in the context of Explainable Artificial Intelligence. Knowing the relative impact of each variable on the output allows us to acquire more information about the problem and about the output provided by a model. This paper proposes a new methodology, XAIRE, that determines the relative importance of input variables in a prediction environment, considering multiple prediction models in order to increase generality and avoid bias inherent in a particular learning algorithm. Concretely, we present an ensemble-based methodology that promotes the aggregation of results from several prediction methods to obtain a relative importance ranking. Also, statistical tests are considered in the methodology in order to reveal significant differences between the relative importance of the predictor variables. As a case study, XAIRE is applied to the arrival of patients in a Hospital Emergency Department, which has resulted in one of the largest sets of different predictor variables in the literature. Results show the extracted knowledge related to the relative importance of the predictors involved in the case study.

Machine learning based forecast for the prediction of inpatient bed demand

Background

Overcrowding is a serious problem that impacts the ability to provide optimal level of care in a timely manner. High patient volume is known to increase the boarding time at the emergency department (ED), as well as at post-anesthesia care unit (PACU). Furthermore, the same high volume increases inpatient bed transfer times, which causes delays in elective surgeries, increases the probability of near misses, patient safety incidents, and adverse events.

Objective

The purpose of this study is to develop a Machine Learning (ML) based strategy to predict weekly forecasts of the inpatient bed demand in order to assist the resource planning for the ED and PACU, resulting in a more efficient utilization.

Methods

The data utilized included all adult inpatient encounters at Geisinger Medical Center (GMC) for the last 5 years. The variables considered were class of inpatient encounter, observation, or surgical overnight recovery (SORU) at the time of their discharge. The ML based strategy is built using the K-means clustering method and the Support Vector Machine Regression technique (K-SVR).

Results

The performance obtained by the K-SVR strategy in the retrospective cohort amounts to a mean absolute percentage error (MAPE) that ranges between 0.49 and 4.10% based on the test period. Additionally, results present a reduced variability, which translates into more stable forecasting results.

Conclusions

The results from this study demonstrate the capacity of ML techniques to forecast inpatient bed demand, particularly using K-SVR. It is expected that the implementation of this model in the workflow of bed capacity management will create efficiencies, which will translate in a more reliable, inexpensive and timely care for patients.

Towards accurate prediction of patient length of stay at emergency department: a GAN-driven deep learning framework

Recently, the hospital systems face a high influx of patients generated by several events, such as seasonal flows or health crises related to epidemics (e.g., COVID'19). Despite the extent of the care demands, hospital establishments, particularly emergency departments (EDs), must admit patients for medical treatments. However, the high patient influx often increases patients' length of stay (LOS) and leads to overcrowding problems within the EDs. To mitigate this issue, hospital managers need to predict the patient's LOS, which is an essential indicator for assessing ED overcrowding and the use of the medical resources (allocation, planning, utilization rates). Thus, accurately predicting LOS is necessary to improve ED management. This paper proposes a deep learning-driven approach for predicting the patient LOS in ED using a generative adversarial network (GAN) model. The GAN-driven approach flexibly learns relevant information from linear and nonlinear processes without prior assumptions on data distribution and significantly enhances the prediction accuracy. Furthermore, we classified the predicted patients' LOS according to time spent at the pediatric emergency department (PED) to further help decision-making and prevent overcrowding. The experiments were conducted on actual data obtained from the PED in Lille regional hospital center, France. The GAN model results were compared with other deep learning models, including deep belief networks, convolutional neural network, stacked auto-encoder, and four machine learning models, namely support vector regression, random forests, adaboost, and decision tree. Results testify that deep learning models are suitable for predicting patient LOS and highlight GAN's superior performance than the other models.

Effectiveness of cascading time series models based on meteorological factors in improving health risk prediction

Meteorological factors, which are periodic and regular in a long run, have an unignorable impact on human health. Accurate health risk prediction based on meteorological factors is essential for optimal allocation of resource in healthcare units. However, due to the non-stationary and non-linear nature of the original hospitalization sequence, traditional methods are less robust in predicting it. This study aims to investigate hospital admission prediction models using time series pre-processing algorithms and deep learning approach based on meteorological factors. Using the electronic medical record data from Panyu Central Hospital and meteorological data of Panyu district from 2003 to 2019, 46,089 eligible patients with lower respiratory tract infections (LRTIs) and four meteorological factors were identified to build and evaluate the prediction models. A novel hybrid model, Cascade GAM-CEEMDAN-LSTM Model (CGCLM), was established in combination with generalized additive model (GAM), complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), and long-short term memory (LSTM) networks for predicting daily admissions of patients with LRTIs. The experimental results show that CGCLM multistep method proposed in this paper outperforms single LSTM model in the prediction of health risk time series at different time window sizes. Moreover, our results also indicate that CGCLM has the best prediction performance when the time window is set to 61 days (RMSE = 1.12, MAE = 0.87, R² = 0.93). Adequate extraction of exposure-response relationships between meteorological factors and diseases and suitable handling of sequence pre-processing have an important role in time series prediction. This hybrid climate-based model for predicting LRTIs disease can also be extended to time series prediction of other epidemic disease.

The Dynamics of Patient Visits to a Public Hospital Pediatric Emergency Department: A Time-Series Model

BACKGROUND

The overcrowding of emergency departments (EDs) is an increasingly relevant public health problem. The main aims of this study were to identify and analyze temporal periodicities of a self-referred pediatric ED (PED), correlate them with meteorological and calendar variables and build a robust forecasting model.

METHODS

An 8-year administrative data set (2010-2017) of the daily number of admissions to the PED of a public hospital in Lisbon, Portugal, was used (n = 670,379). A time-series model of the daily number of visits was built, including temporal periodicities, the Portuguese school calendar, and a meteorological comfort index (humidex).

RESULTS

Several temporal cycles were identified: 1 year (peak in January/February related to respiratory infections in younger children and infants), 6 months (peaks in May and October with an increase in the admissions of older children and adolescents with trauma, gastrointestinal infections and atopic symptoms), 4 months (related to annual school vacations), 1 week (lower admission values on Saturday), and half a week (low from Friday to Monday morning). School calendar and humidex were significantly correlated with daily admissions. The model yielded a mean absolute percentage error of 10.7% ± 1.10% when cross-validation was performed for the full data set.

CONCLUSION

Although PED visits are multifactorial, they may be predicted and explained by a relatively small number of variables. Such a model may be easily reproduced in different settings and represents a relevant tool to improve quality in EDs through correctly adapting human resources to ED demand.

Dynamics of the Third Wave of COVID-19 from the Perspective of the Emergency Department in a Large Regional Hospital—Single Center Observational Study

Background: The outbreak of the Coronavirus Disease 2019 (COVID-19) pandemic has caused many significant social and economic changes. The consecutive waves of the epidemic in various countries have had dissimilar courses depending on the methods used to combat it. The aim of this study was to determine the dynamics of the third wave of COVID-19 from the perspective of emergency departments (ED). Methods: This was a retrospective review of medical records from ED. The authors have identified the most frequent symptoms. Prognostic factors have been chosen—prognostic scales, length of stay (LOS)—and a number of resources required have been calculated. Results: As the time passed, there were fewer patients and they presented mild symptoms. A statistically significant difference was observed in the median of blood oxygenation measurement (p = 0.00009), CRP level (p = 0.0016), and admission rate. Patients admitted to the hospital required more resources at ED. LOS was shorter in patients discharged home (p < 0.0001). Conclusions: The blood oxygen saturation (SPO2) and CPR levels can be helpful in decision-making regarding medical treatment. The fast-track for patients in good clinical condition may shorten the duration of stay in ED, and reduce the number of required resources.

Prediction of Obstetric Patient Flow and Horizontal Allocation of Medical Resources Based on Time Series Analysis

Objective: Given the ever-changing flow of obstetric patients in the hospital, how the government and hospital management plan and allocate medical resources has become an important problem that needs to be urgently solved. In this study a prediction method for calculating the monthly and daily flow of patients based on time series is proposed to provide decision support for government and hospital management.

Methods: The historical patient flow data from the Department of Obstetrics and Gynecology of the First Hospital of Jilin University, China, from January 1, 2018, to February 29, 2020, were used as the training set. Seven models such as XGBoost, SVM, RF, and NNAR were used to predict the daily patient flow in the next 14 days. The HoltWinters model is then used to predict the monthly flow of patients over the next year.

Results: The results of this analysis and prediction model showed that the obstetric inpatient flow was not a purely random process, and that patient flow was not only accompanied by the random patient flow but also showed a trend change and seasonal change rule. ACF,PACF,Ljung_box, and residual histogram were then used to verify the accuracy of the prediction model, and the results show that the Holtwiners model was optimal. R2, MAPE, and other indicators were used to measure the accuracy of the 14 day prediction model, and the results showed that HoltWinters and STL prediction models achieved high accuracy.

Conclusion: In this paper, the time series model was used to analyze the trend and seasonal changes of obstetric patient flow and predict the patient flow in the next 14 days and 12 months. On this basis, combined with the trend and seasonal changes of obstetric patient flow, a more reasonable and fair horizontal allocation scheme of medical resources is proposed, combined with the prediction of patient flow.

Forecasting emergency department admissions

The emergency department of a hospital plays an extremely important role in the healthcare of patients. To maintain a high quality service, clinical professionals need information on how patient flow will evolve in the immediate future. With accurate emergency department forecasts it is possible to better manage available human resources by allocating clinical staff before peak periods, thus preventing service congestion, or releasing clinical staff at less busy times. This paper describes a solution developed for the presentation of hourly, four-hour, eight-hour and daily number of admissions to a hospital’s emergency department. A 10-year history (2009–2018) of the number of emergency admissions in a Portuguese hospital was used. To create the models several methods were tested, including exponential smoothing, SARIMA, autoregressive and recurrent neural network, XGBoost and ensemble learning. The models that generated the most accurate hourly time predictions were the recurrent neural network with one-layer (sMAPE = 23.26%) and with three layers (sMAPE = 23.12%) and XGBoost (sMAPE = 23.70%). In terms of efficiency, the XGBoost method has by far outperformed all others. The success of the recurrent neuronal network and XGBoost machine learning methods applied to the prediction of the number of emergency department admissions has been demonstrated here, with an accuracy that surpasses the models found in the literature.

Implementing a material planning and control method for special nutrition in a Brazilian public hospital

This study aims to (i) propose a demand forecast model for special nutrition materials in the context of health services, and (ii) comparatively evaluate three inventory management and control systems (periodic review, continuous review and mixed) for special nutrition materials. For that, we carried out a case study in a Brazilian public teaching hospital where data and information collection were conducted over a span of 22 months (from January 2018 and were consolidated until October 2019). A six-step approach was followed to propose the demand forecasting models and, later, evaluate the inventory control systems for special nutrition materials. Results indicate that if the organization implements the proposed inventory management method, there could be savings of up to 33% in the stock values managed by the healthcare organization. This research shows the planning and control of special nutrition materials in an integrated manner. Demand forecasting methods have been combined with inventory management to promote systemic improvements to healthcare organization.

Monitoring patient flow in a hospital emergency department: ARMA-based nonparametric GLRT scheme

Overcrowding in emergency departments (EDs) is a primary concern for hospital administration. They aim to efficiently manage patient demands and reducing stress in the ED. Detection of abnormal ED demands (patient flows) in hospital systems aids ED managers to obtain appropriate decisions by optimally allocating the available resources following patient attendance. This paper presents a monitoring strategy that provides an early alert in an ED when an abnormally high patient influx occurs. Anomaly detection using this strategy involves the amalgamation of autoregressive-moving-average (ARMA) time series models with the generalized likelihood ratio (GLR) chart. A nonparametric procedure based on kernel density estimation is employed to determine the detection threshold of the ARMA-GLR chart. The developed ARMA-based GLR has been validated through practical data from the ED at Lille Hospital, France. Then, the ARMA-based GLR method's performance was compared to that of other commonly used charts, including a Shewhart chart and an exponentially weighted moving average chart; it proved more accurate.

How New Mexico Leveraged a COVID-19 Case Forecasting Model to Preemptively Address the Health Care Needs of the State: Quantitative Analysis

BACKGROUND

Prior to the COVID-19 pandemic, US hospitals relied on static projections of future trends for long-term planning and were only beginning to consider forecasting methods for short-term planning of staffing and other resources. With the overwhelming burden imposed by COVID-19 on the health care system, an emergent need exists to accurately forecast hospitalization needs within an actionable timeframe.

OBJECTIVE

Our goal was to leverage an existing COVID-19 case and death forecasting tool to generate the expected number of concurrent hospitalizations, occupied intensive care unit (ICU) beds, and in-use ventilators 1 day to 4 weeks in the future for New Mexico and each of its five health regions.

METHODS

We developed a probabilistic model that took as input the number of new COVID-19 cases for New Mexico from Los Alamos National Laboratory's COVID-19 Forecasts Using Fast Evaluations and Estimation tool, and we used the model to estimate the number of new daily hospital admissions 4 weeks into the future based on current statewide hospitalization rates. The model estimated the number of new admissions that would require an ICU bed or use of a ventilator and then projected the individual lengths of hospital stays based on the resource need. By tracking the lengths of stay through time, we captured the projected simultaneous need for inpatient beds, ICU beds, and ventilators. We used a postprocessing method to adjust the forecasts based on the differences between prior forecasts and the subsequent observed data. Thus, we ensured that our forecasts could reflect a dynamically changing situation on the ground.

RESULTS

Forecasts made between September 1 and December 9, 2020, showed variable accuracy across time, health care resource needs, and forecast horizon. Forecasts made in October, when new COVID-19 cases were steadily increasing, had an average accuracy error of 20.0%, while the error in forecasts made in September, a month with low COVID-19 activity, was 39.7%. Across health care use categories, state-level forecasts were more accurate than those at the regional level. Although the accuracy declined as the forecast was projected further into the future, the stated uncertainty of the prediction improved. Forecasts were within 5% of their stated uncertainty at the 50% and 90% prediction intervals at the 3- to 4-week forecast horizon for state-level inpatient and ICU needs. However, uncertainty intervals were too narrow for forecasts of state-level ventilator need and all regional health care resource needs.

CONCLUSIONS

Real-time forecasting of the burden imposed by a spreading infectious disease is a crucial component of decision support during a public health emergency. Our proposed methodology demonstrated utility in providing near-term forecasts, particularly at the state level. This tool can aid other stakeholders as they face COVID-19 population impacts now and in the future.

Performance evaluation of Emergency Department patient arrivals forecasting models by including meteorological and calendar information: A comparative study

The volume of daily patient arrivals at Emergency Departments (EDs) is unpredictable and is a significant reason of ED crowding in hospitals worldwide. Timely forecast of patients arriving at ED can help the hospital management in early planning and avoiding of overcrowding. Many different ED patient arrivals forecasting models have previously been proposed by using time series analysis methods. Even though the time series methods such as Linear and Logistic Regression, Autoregressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), Exponential Smoothing (ES), and Artificial Neural Network (ANN) have been explored extensively for the ED forecasting model development, the few significant limitations of these methods associated in the analysis of time series data make the models inadequate in many practical situations. Therefore, in this paper, Machine Learning (ML)-based Random Forest (RF) regressor, and Deep Neural Network (DNN)-based Long Short-Term Memory (LSTM) and Convolutional Neural network (CNN) methods, which have not been explored to the same extent as the other time series techniques, are implemented by incorporating meteorological and calendar parameters for the development of forecasting models. The performances of the developed three models in forecasting ED patient arrivals are evaluated. Among the three models, CNN outperformed for short-term (3 days in advance) patient arrivals prediction with Mean Absolute Percentage Error (MAPE) of 9.24% and LSTM performed better for moderate-term (7 days in advance) patient arrivals prediction with MAPE of 8.91% using weather forecast information. Whereas, LSTM model outperformed with MAPE of 8.04% compared to 9.53% by CNN and 10.10% by RF model for current day prediction of patient arrivals using 3 days past weather information. Thus, for short-term ED patient arrivals forecasting, DNN-based model performed better compared to RF regressor ML-based model.

Mixed time series approaches for forecasting the daily number of hospital blood collections

PURPOSE

Provide new methods to predict the number of hospital blood collections.

METHODS

The registered outpatients and blood collection patients in a large hospital in China in the period from March 2018 to April 2019 were enrolled in the study. Firstly, we analyzed the time series characteristics of the daily blood collection patients and their correlation with the number of daily outpatients. Then, we used the time series ARIMA and linear regression methods to build the periodic trend model of the blood collections number prediction and the regression prediction model with the number of registered outpatients as an independent variable. Finally, we built a combined prediction model considering mixed time series to predict the number of blood collections in the hospital.

RESULTS

The combined prediction model has a higher accuracy and can better explore the characteristics of the number of blood collections compared with other models. It can also give some suggestions for a reasonable blood collection management.

CONCLUSION

The combined prediction model of mixed time series can reflect the change in the blood collections number due to the influence of internal and external factors and can realize the blood collection prediction with a higher accuracy providing a new method for the prediction of the blood collections number.

Forecasting emergency department overcrowding: A deep learning framework

As the demand for medical cares has considerably expanded, the issue of managing patient flow in hospitals and especially in emergency departments (EDs) is certainly a key issue to be carefully mitigated. This can lead to overcrowding and the degradation of the quality of the provided medical services. Thus, the accurate modeling and forecasting of ED visits are critical for efficiently managing the overcrowding problems and enable appropriate optimization of the available resources. This paper proposed an effective method to forecast daily and hourly visits at an ED using Variational AutoEncoder (VAE) algorithm. Indeed, the VAE model as a deep learning-based model has gained special attention in features extraction and modeling due to its distribution-free assumptions and superior nonlinear approximation. Two types of forecasting were conducted: one- and multi-step-ahead forecasting. To the best of our knowledge, this is the first time that the VAE is investigated to improve forecasting of patient arrivals time-series data. Data sets from the pediatric emergency department at Lille regional hospital center, France, are employed to evaluate the forecasting performance of the introduced method. The VAE model was evaluated and compared with seven methods namely Recurrent Neural Network (RNN), Long short-term memory (LSTM), Bidirectional LSTM (BiLSTM), Convolutional LSTM Network (ConvLSTM), restricted Boltzmann machine (RBM), Gated recurrent units (GRUs), and convolutional neural network (CNN). The results clearly show the promising performance of these deep learning models in forecasting ED visits and emphasize the better performance of the VAE in comparison to the other models.

Medical service demand forecasting using a hybrid model based on ARIMA and self-adaptive filtering method

BACKGROUND

Accurate forecasting of medical service demand is beneficial for the reasonable healthcare resource planning and allocation. The daily outpatient volume is characterized by randomness, periodicity and trend, and the time series methods, like ARIMA are often used for short-term outpatient visits forecasting. Therefore, to further enlarge the prediction horizon and improve the prediction accuracy, a hybrid prediction model integrating ARIMA and self-adaptive filtering method is proposed.

METHODS

The ARIMA model is first used to identify the features like cyclicity and trend of the time series data and to estimate the model parameters. The parameters are then adjusted by the steepest descent algorithm in the adaptive filtering method to reduce the prediction error. The hybrid model is validated and compared with traditional ARIMA by several test sets from the Time Series Data Library (TSDL), a weekly emergency department (ED) visit case from literature study, and the real cases of prenatal examinations and B-ultrasounds in a maternal and child health care center (MCHCC) in Ningbo.

RESULTS

For TSDL cases the prediction accuracy of the hybrid prediction is improved by 80-99% compared with the ARIMA model. For the weekly ED visit case, the forecasting results of the hybrid model are better than those of both traditional ARIMA and ANN model, and similar to the ANN combined data decomposition model mentioned in the literature. For the actual data of MCHCC in Ningbo, the MAPE predicted by the ARIMA model in the two departments was 18.53 and 27.69%, respectively, and the hybrid models were 2.79 and 1.25%, respectively.

CONCLUSIONS

The hybrid prediction model outperforms the traditional ARIMA model in both accurate predicting result with smaller average relative error and the applicability for short-term and medium-term prediction.

Excess Patient Visits for Cough and Pulmonary Disease at a Large US Health System in the Months Prior to the COVID-19 Pandemic: Time-Series Analysis

BACKGROUND

Accurately assessing the regional activity of diseases such as COVID-19 is important in guiding public health interventions. Leveraging electronic health records (EHRs) to monitor outpatient clinical encounters may lead to the identification of emerging outbreaks.

OBJECTIVE

The aim of this study is to investigate whether excess visits where the word "cough" was present in the EHR reason for visit, and hospitalizations with acute respiratory failure were more frequent from December 2019 to February 2020 compared with the preceding 5 years.

METHODS

A retrospective observational cohort was identified from a large US health system with 3 hospitals, over 180 clinics, and 2.5 million patient encounters annually. Data from patient encounters from July 1, 2014, to February 29, 2020, were included. Seasonal autoregressive integrated moving average (SARIMA) time-series models were used to evaluate if the observed winter 2019/2020 rates were higher than the forecast 95% prediction intervals. The estimated excess number of visits and hospitalizations in winter 2019/2020 were calculated compared to previous seasons.

RESULTS

The percentage of patients presenting with an EHR reason for visit containing the word "cough" to clinics exceeded the 95% prediction interval the week of December 22, 2019, and was consistently above the 95% prediction interval all 10 weeks through the end of February 2020. Similar trends were noted for emergency department visits and hospitalizations starting December 22, 2019, where observed data exceeded the 95% prediction interval in 6 and 7 of the 10 weeks, respectively. The estimated excess over the 3-month 2019/2020 winter season, obtained by either subtracting the maximum or subtracting the average of the five previous seasons from the current season, was 1.6 or 2.0 excess visits for cough per 1000 outpatient visits, 11.0 or 19.2 excess visits for cough per 1000 emergency department visits, and 21.4 or 39.1 excess visits per 1000 hospitalizations with acute respiratory failure, respectively. The total numbers of excess cases above the 95% predicted forecast interval were 168 cases in the outpatient clinics, 56 cases for the emergency department, and 18 hospitalized with acute respiratory failure.

CONCLUSIONS

A significantly higher number of patients with respiratory complaints and diseases starting in late December 2019 and continuing through February 2020 suggests community spread of SARS-CoV-2 prior to established clinical awareness and testing capabilities. This provides a case example of how health system analytics combined with EHR data can provide powerful and agile tools for identifying when future trends in patient populations are outside of the expected ranges.

Prediction of Daily Blood Sampling Room Visits Based on ARIMA and SES Model

This paper is aimed at establishing a combined prediction model to predict the demand for medical care in terms of daily visits in an outpatient blood sampling room, which provides a basis for rational arrangement of human resources and planning. On the basis of analyzing the comprehensive characteristics of the randomness, periodicity, trend, and day-of-the-week effects of the daily number of blood collections in the hospital, we firstly established an autoregressive integrated moving average model (ARIMA) model to capture the periodicity, volatility, and trend, and secondly, we constructed a simple exponential smoothing (SES) model considering the day-of-the-week effect. Finally, a combined prediction model of the residual correction is established based on the prediction results of the two models. The models are applied to data from 60 weeks of daily visits in the outpatient blood sampling room of a large hospital in Chengdu, for forecasting the daily number of blood collections about 1 week ahead. The result shows that the MAPE of the combined model is the smallest overall, of which the improvement during the weekend is obvious, indicating that the prediction error of extreme value is significantly reduced. The ARIMA model can extract the seasonal and nonseasonal components of the time series, and the SES model can capture the overall trend and the influence of regular changes in the time series, while the combined prediction model, taking into account the comprehensive characteristics of the time series data, has better fitting prediction accuracy than a single model. The new model can well realize the short-to-medium-term prediction of the daily number of blood collections one week in advance.

Assessing healthcare service quality using routinely collected data: Linking information systems in emergency care

Emergency departments need to continuously calculate quality indicators in order to perform structural improvements, improvements in the daily routine, and ad-hoc improvements in everyday life. However, many different actors across multiple disciplines collaborate to provide emergency care. Hence, patient-related data is stored in several information systems, which in turn makes the calculation of quality indicators more difficult. To address this issue, we aim to link and use routinely collected data of the different actors within the emergency care continuum. In order to assess the feasibility of linking and using routinely collected data for quality indicators and whether this approach adds value to the assessment of emergency care quality, we conducted a single case study in a German academic teaching hospital. We analyzed the available data of the existing information systems in the emergency continuum and linked and pre-processed the data. Based on this, we then calculated four quality indicators (Left Without Been Seen, Unplanned Reattendance, Diagnostic Efficiency, and Overload Closure). Lessons learned from the calculation and results of the discussions with staff members that had multiple years of work experience in the emergency department provide a better understanding of the quality of the emergency department, the related challenges during the calculation, and the added value of linking routinely collected data.

Emergency patient flow forecasting in the radiology department

The accurate forecast of radiology emergency patient flow is of great importance to optimize appointment scheduling decisions. This study used a multi-model approach to forecast daily radiology emergency patient flow with consideration of different patient sources. We constructed six linear and nonlinear models by considering the lag effects and corresponding time factors. The autoregressive integrated moving average and least absolute shrinkage and selection operator (Lasso) were selected from the category of linear models, whereas linear-and-radial support vector regression models, random forests and adaptive boosting were chosen from the category of nonlinear models. The models were applied to 4-year daily emergency visits data in the radiology department of West China Hospital in Chengdu, China. The mean absolute percentage error of six models ranged from 8.56 to 9.36 percent for emergency department patients, whereas it varied from 10.90 to 14.39 percent for ward patients. The best-performing model for total radiology visits was Lasso, which yielded a mean absolute percentage error of 7.06 percent. The arrival patterns of emergency department and total radiology emergency patient flows could be modeled by linear processes. By contrast, the nonlinear model performed best for ward patient flow. These findings will benefit hospital managers in managing efficient patient flow, thus improving service quality and increasing patient satisfaction.

Predicting Outpatient Appointment Demand Using Machine Learning and Traditional Methods

Traditional methods have long been used for clinical demand forecasting. Machine learning methods represent the next evolution in forecasting, but model choice and optimization remain challenging for achieving optimal results. To determine the best method to predict demand for outpatient appointments comparing machine learning and traditional methods, this retrospective study analyzed "appointment requests" at a major outpatient department in a destination medical center. Two separate locations (A and B) were assessed with 20 traditional, hybrid (traditional + machine learning) and machine learning methods to determine the best forecasting outcome (lowest Forecast Standard Error, FSE). Data characteristics from both datasets were examined. 20 forecasting models were then assessed and compared for the best result. Location A's data displayed a cyclical and non-trending pattern while Location B's displayed a cyclical and trending pattern. Both Location A and B yielded the feature engineered XGBoost model (machine learning) with the lowest out-of-sample FSE. It is important to carefully analyze and understand the underlying data set pattern and then test a variety of traditional, machine learning, and hybrid prediction methods to achieve optimal predictive results. Additionally, the use of feature engineering or hybrid methods can augment the usefulness of machine learning methods.

An exhaustive review and analysis on applications of statistical forecasting in hospital emergency departments

Emergency departments (EDs) provide medical treatment for a broad spectrum of illnesses and injuries to patients who arrive at all hours of the day. The quality and efficient delivery of health care in EDs are associated with a number of factors, such as patient overall length of stay (LOS) and admission, prompt ambulance diversion, quick and accurate triage, nurse and physician assessment, diagnostic and laboratory services, consultations and treatment. One of the most important ways to plan the healthcare delivery efficiently is to make forecasts of ED processes. The aim this study is thus to provide an exhaustive review for ED stakeholders interested in applying forecasting methods to their ED processes. A categorisation, analysis and interpretation of 102 papers is performed for review. This exhaustive review provides an insight for researchers and practitioners about forecasting in EDs in terms of showing current state and potential areas for future attempts.

Theoretical bounds and approximation of the probability mass function of future hospital bed demand

Failing to match the supply of resources to the demand for resources in a hospital can cause non-clinical transfers, diversions, safety risks, and expensive under-utilized resource capacity. Forecasting bed demand helps achieve appropriate safety standards and cost management by proactively adjusting staffing levels and patient flow protocols. This paper defines the theoretical bounds on optimal bed demand prediction accuracy and develops a flexible statistical model to approximate the probability mass function of future bed demand. A case study validates the model using blinded data from a mid-sized Massachusetts community hospital. This approach expands upon similar work by forecasting multiple days in advance instead of a single day, providing a probability mass function of demand instead of a point estimate, using the exact surgery schedule instead of assuming a cyclic schedule, and using patient-level duration-varying length-of-stay distributions instead of assuming patient homogeneity and exponential length of stay distributions. The primary results of this work are an accurate and lengthy forecast, which provides managers better information and more time to optimize short-term staffing adaptations to stochastic bed demand, and a derivation of the minimum mean absolute error of an ideal forecast.

Modelling and forecasting daily surgical case volume using time series analysis

Hospitals and outpatient surgery centres are often plagued by a recurring staff management question: "How can we plan our nursing schedule weeks in advance, not knowing how many and when patients will require surgery?" Demand for surgery is driven by patient needs, physician constraints, and weekly or seasonal fluctuations. With all of these factors embedded into historical surgical volume, we use time series analysis methods to forecast daily surgical case volumes, which can be extremely valuable for estimating workload and labour expenses. Seasonal Autoregressive Integrated Moving Average (SARIMA) modelling is used to develop a statistical prediction model that provides short-term forecasts of daily surgical demand. We used data from a Level 1 Trauma Centre to build and evaluate the model. Our results suggest that the proposed SARIMA model can be useful for estimating surgical case volumes 2-4 weeks prior to the day of surgery, which can support robust and reliable staff schedules.

Applying advanced analytics to guide emergency department operational decisions: A proof-of-concept study examining the effects of boarding

BACKGROUND

Emergency Department (ED) leaders are increasingly confronted with large amounts of data with the potential to inform and guide operational decisions. Routine use of advanced analytic methods may provide additional insights.

OBJECTIVES

To examine the practical application of available advanced analytic methods to guide operational decision making around patient boarding.

METHODS

Retrospective analysis of the effect of boarding on ED operational metrics from a single site between 1/2015 and 1/2017. Times series were visualized through decompositional techniques accounting for seasonal trends, to determine the effect of boarding on ED performance metrics and to determine the impact of boarding "shocks" to the system on operational metrics over several days.

RESULTS

There were 226,461 visits with the mean (IQR) number of visits per day was 273 (258-291). Decomposition of the boarding count time series illustrated an upward trend in the last 2-3 quarters as well as clear seasonal components. All performance metrics were significantly impacted (p<0.05) by boarding count, except for overall Press Ganey scores (p<0.65). For every additional increase in boarder count, overall length-of-stay (LOS) increased by 1.55min (0.68, 1.50). Smaller effects were seen for waiting room LOS and treat and release LOS. The impulse responses indicate that the boarding shocks are characterized by changes in the performance metrics within the first day that fade out after 4-5days.

CONCLUSION

In this study regarding the use of advanced analytics in daily ED operations, time series analysis provided multiple useful insights into boarding and its impact on performance metrics.

Hospital daily outpatient visits forecasting using a combinatorial model based on ARIMA and SES models

Background

Accurate forecasting of hospital outpatient visits is beneficial for the reasonable planning and allocation of healthcare resource to meet the medical demands. In terms of the multiple attributes of daily outpatient visits, such as randomness, cyclicity and trend, time series methods, ARIMA, can be a good choice for outpatient visits forecasting. On the other hand, the hospital outpatient visits are also affected by the doctors’ scheduling and the effects are not pure random. Thinking about the impure specialty, this paper presents a new forecasting model that takes cyclicity and the day of the week effect into consideration.

Methods

We formulate a seasonal ARIMA (SARIMA) model on a daily time series and then a single exponential smoothing (SES) model on the day of the week time series, and finally establish a combinatorial model by modifying them. The models are applied to 1 year of daily visits data of urban outpatients in two internal medicine departments of a large hospital in Chengdu, for forecasting the daily outpatient visits about 1 week ahead.

Results

The proposed model is applied to forecast the cross-sectional data for 7 consecutive days of daily outpatient visits over an 8-weeks period based on 43 weeks of observation data during 1 year. The results show that the two single traditional models and the combinatorial model are simplicity of implementation and low computational intensiveness, whilst being appropriate for short-term forecast horizons. Furthermore, the combinatorial model can capture the comprehensive features of the time series data better.

Conclusions

Combinatorial model can achieve better prediction performance than the single model, with lower residuals variance and small mean of residual errors which needs to be optimized deeply on the next research step.

Early Detection of Peak Demand Days of Chronic Respiratory Diseases Emergency Department Visits Using Artificial Neural Networks

Chronic respiratory diseases, mainly asthma and chronic obstructive pulmonary disease (COPD), affect the lives of people by limiting their activities in various aspects. Overcrowding of hospital emergency departments (EDs) due to respiratory diseases in certain weather and environmental pollution conditions results in the degradation of quality of medical care, and even limits its availability. A useful tool for ED managers would be to forecast peak demand days so that they can take steps to improve the availability of medical care. In this paper, we developed an artificial neural network based classifier using multilayer perceptron with back propagation algorithm that predicts peak event (peak demand days) of patients with respiratory diseases, mainly asthma and COPD visiting EDs in Dallas County of Texas in the United States. The precision and recall for peak event class were 77.1% and 78.0%, respectively, and those for nonpeak events were 83.9% and 83.2%, respectively. The overall accuracy of the system is 81.0%.

Time series analysis of temporal trends in the pertussis incidence in Mainland China from 2005 to 2016

Short-term forecast of pertussis incidence is helpful for advanced warning and planning resource needs for future epidemics. By utilizing the Auto-Regressive Integrated Moving Average (ARIMA) model and Exponential Smoothing (ETS) model as alterative models with R software, this paper analyzed data from Chinese Center for Disease Control and Prevention (China CDC) between January 2005 and June 2016. The ARIMA (0,1,0)(1,1,1)₁₂ model (AICc = 1342.2 BIC = 1350.3) was selected as the best performing ARIMA model and the ETS (M,N,M) model (AICc = 1678.6, BIC = 1715.4) was selected as the best performing ETS model, and the ETS (M,N,M) model with the minimum RMSE was finally selected for in-sample-simulation and out-of-sample forecasting. Descriptive statistics showed that the reported number of pertussis cases by China CDC increased by 66.20% from 2005 (4058 cases) to 2015 (6744 cases). According to Hodrick-Prescott filter, there was an apparent cyclicity and seasonality in the pertussis reports. In out of sample forecasting, the model forecasted a relatively high incidence cases in 2016, which predicates an increasing risk of ongoing pertussis resurgence in the near future. In this regard, the ETS model would be a useful tool in simulating and forecasting the incidence of pertussis, and helping decision makers to take efficient decisions based on the advanced warning of disease incidence.

Agents endowed with uncertainty management behaviors to solve a multiskill healthcare task scheduling

Health organizations are complex to manage due to their dynamic processes and distributed hospital organization. It is therefore necessary for healthcare institutions to focus on this issue to deal with patients' requirements. We aim in this paper to develop and implement a management decision support system (DSS) that can help physicians to better manage their organization and anticipate the feature of overcrowding. Our objective is to optimize the Pediatric Emergency Department (PED) functioning characterized by stochastic arrivals of patients leading to its services overload. Human resources allocation presents additional complexity related to their different levels of skills and uncertain availability dates. So, we propose a new approach for multi-healthcare task scheduling based on a dynamic multi-agent system. Decisions about assignment and scheduling are the result of a cooperation and negotiation between agents with different behaviors. We therefore define the actors involved in the agents' coalition to manage uncertainties related to the scheduling problem and we detail their behaviors. Agents have the same goal, which is to enhance care quality and minimize long waiting times while respecting degrees of emergency. Different visits to the PED services and regular meetings with the medical staff allowed us to model the PED architecture and identify the characteristics and different roles of the healthcare providers and the diverse aspects of the PED activities. Our approach is integrated in a DSS for the management of the Regional University Hospital Center (RUHC) of Lille (France). Our survey is included in the French National Research Agency (ANR) project HOST (Hôpital: Optimisation, Simulation et évitement des Tensions (ANR-11-TecSan-010: http://host.ec-lille.fr/wp-content/themes/twentyeleven/docsANR/R0/HOST-WP0.pdf)).

Forecasting Daily Patient Outflow From a Ward Having No Real-Time Clinical Data

Background: Modeling patient flow is crucial in understanding resource demand and prioritization. We study patient outflow from an open ward in an Australian hospital, where currently bed allocation is carried out by a manager relying on past experiences and looking at demand. Automatic methods that provide a reasonable estimate of total next-day discharges can aid in efficient bed management. The challenges in building such methods lie in dealing with large amounts of discharge noise introduced by the nonlinear nature of hospital procedures, and the nonavailability of real-time clinical information in wards.

Forecasting the Emergency Department Patients Flow

Emergency department (ED) have become the patient's main point of entrance in modern hospitals causing it frequent overcrowding, thus hospital managers are increasingly paying attention to the ED in order to provide better quality service for patients. One of the key elements for a good management strategy is demand forecasting. In this case, forecasting patients flow, which will help decision makers to optimize human (doctors, nurses…) and material(beds, boxs…) resources allocation. The main interest of this research is forecasting daily attendance at an emergency department. The study was conducted on the Emergency Department of Troyes city hospital center, France, in which we propose a new practical ED patients classification that consolidate the CCMU and GEMSA categories into one category and innovative time-series based models to forecast long and short term daily attendance. The models we developed for this case study shows very good performances (up to 91,24 % for the annual Total flow forecast) and robustness to epidemic periods.

Time series analysis of influenza incidence in Chinese provinces from 2004 to 2011

Influenza as a severe infectious disease has caused catastrophes throughout human history, and every pandemic of influenza has produced a great social burden. We compiled monthly data of influenza incidence from all provinces and autonomous regions in mainland China from January 2004 to December 2011, comprehensively evaluated and classified these data, and then randomly selected 4 provinces with higher incidence (Hebei, Gansu, Guizhou, and Hunan), 2 provinces with median incidence (Tianjin and Henan), 1 province with lower incidence (Shandong), using time series analysis to construct an ARIMA model, which is based on the monthly incidence from 2004 to 2011 as the training set. We exerted the X-12-ARIMA procedure for modeling due to the seasonality these data implied. Autocorrelation function (ACF), partial autocorrelation function (PACF), and automatic model selection were to determine the order of the model parameters. The optimal model was decided by a nonseasonal and seasonal moving average test. Finally, we applied this model to predict the monthly incidence of influenza in 2012 as the test set, and the simulated incidence was compared with the observed incidence to evaluate the model's validity by the criterion of both percentage variability in regression analyses (R) and root mean square error (RMSE). It is conceivable that SARIMA (0,1,1)(0,1,1)12 could simultaneously forecast the influenza incidence of the Hebei Province, Guizhou Province, Henan Province, and Shandong Province; SARIMA (1,0,0)(0,1,1)12 could forecast the influenza incidence in Gansu Province; SARIMA (3,1,1)(0,1,1)12 could forecast the influenza incidence in Tianjin City; and SARIMA (0,1,1)(0,0,1)12 could forecast the influenza incidence in Hunan Province. Time series analysis is a good tool for prediction of disease incidence.

Using Statistical Forecasting to Optimize Staff Scheduling in Healthcare Organizations

Modern-day business environment of healthcare organizations demands the maximization of operational effectiveness and quality with optimal cost. Therefore, healthcare executives are often required to make difficult decisions based on subjective experience and judgement. An example of such a decision is scheduling of resources to fulfil demand for service. The effective use of statistical forecasting can lead to better personnel scheduling decisions based on estimates of patient arrival rates, resulting in improvement in quality of service as well as reduction of cost.

The purpose of this article is to demonstrate the typical steps involved in applying forecasting techniques in patient care: This demonstration involves use of statistical techniques like Analysis of Variance (ANOVA) to identify factors driving demand, and Auto Regressive Integrated Moving Average (ARIMA) to develop a forecasting model for optimal staff scheduling in healthcare organizations based on patient arrival rates. The models are developed and subsequently tested on a set of real data gathered from a regional hospital located in the US. Statistically significant difference in average patient count was found among different days of the week. The findings of the research suggests that resources like cleaning personnel can be better utilized by allocating different proportions of resources to different parts of the week, based on the understanding of different patient load over these time periods.

Use of an agent-based simulation model to evaluate a mobile-based system for supporting emergency evacuation decision making

Recently, mass casualty incidents (MCIs) have been occurring frequently and have gained international attention. There is an urgent need for scientifically proven and effective emergency responses to MCIs, particularly as the severity of incidents is continuously increasing. The emergency response to MCIs is a multi-dimensional and multi-participant dynamic process that changes in real-time. The evacuation decisions that assign casualties to different hospitals in a region are very important and impact both the results of emergency treatment and the efficiency of medical resource utilization. Previously, decisions related to casualty evacuation were made by an incident commander with emergency experience and in accordance with macro emergency guidelines. There are few decision-supporting tools available to reduce the difficulty and psychological pressure associated with the evacuation decisions an incident commander must make. In this study, we have designed a mobile-based system to collect medical and temporal data produced during an emergency response to an MCI. Using this information, our system's decision-making model can provide personal evacuation suggestions that improve the overall outcome of an emergency response. The effectiveness of our system in reducing overall mortality has been validated by an agent-based simulation model established to simulate an emergency response to an MCI.