1
|
Kiganda C, Akcayol MA. Forecasting the Spread of COVID-19 Using Deep Learning and Big Data Analytics Methods. SN COMPUTER SCIENCE 2023; 4:374. [PMID: 37193218 PMCID: PMC10155670 DOI: 10.1007/s42979-023-01801-5] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 06/02/2022] [Accepted: 03/22/2023] [Indexed: 05/18/2023]
Abstract
To contain the spread of the COVID-19 pandemic, there is a need for cutting-edge approaches that make use of existing technology capabilities. Forecasting its spread in a single or multiple countries ahead of time is a common strategy in most research. There is, however, a need for all-inclusive studies that capitalize on the entire regions on the African continent. This study closes this gap by conducting a wide-ranging investigation and analysis to forecast COVID-19 cases and identify the most critical countries in terms of the COVID-19 pandemic in all five major African regions. The proposed approach leveraged both statistical and deep learning models that included the autoregressive integrated moving average (ARIMA) model with a seasonal perspective, the long-term memory (LSTM), and Prophet models. In this approach, the forecasting problem was considered as a univariate time series problem using confirmed cumulative COVID-19 cases. The model performance was evaluated using seven performance metrics that included the mean-squared error, root mean-square error, mean absolute percentage error, symmetric mean absolute percentage error, peak signal-to-noise ratio, normalized root mean-square error, and the R2 score. The best-performing model was selected and used to make future predictions for the next 61 days. In this study, the long short-term memory model performed the best. Mali, Angola, Egypt, Somalia, and Gabon from the Western, Southern, Northern, Eastern, and Central African regions, with an expected increase of 22.77%, 18.97%, 11.83%, 10.72%, and 2.81%, respectively, were the most vulnerable countries with the highest expected increase in the number of cumulative positive cases.
Collapse
Affiliation(s)
- Cylas Kiganda
- Computer Science Department, Institute of Informatics, Gazi University, Ankara, Turkey
| | - Muhammet Ali Akcayol
- Computer Science Department, Institute of Informatics, Gazi University, Ankara, Turkey
| |
Collapse
|
2
|
Lu X, Qiu H. Explainable prediction of daily hospitalizations for cerebrovascular disease using stacked ensemble learning. BMC Med Inform Decis Mak 2023; 23:59. [PMID: 37024922 PMCID: PMC10080841 DOI: 10.1186/s12911-023-02159-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/03/2022] [Accepted: 03/23/2023] [Indexed: 04/08/2023] Open
Abstract
BACKGROUND With the prevalence of cerebrovascular disease (CD) and the increasing strain on healthcare resources, forecasting the healthcare demands of cerebrovascular patients has significant implications for optimizing medical resources. METHODS In this study, a stacking ensemble model comprised of four base learners (ridge regression, random forest, gradient boosting decision tree, and artificial neural network) and a meta learner (elastic net) was proposed for predicting the daily number of hospital admissions (HAs) for CD using the historical HAs data, air quality data, and meteorological data in Chengdu, China from 2015 to 2018. To solve the label imbalance problem, a re-weighting method based on label distribution smoothing was integrated into the meta learner. We trained the model using the data from 2015 to 2017 and evaluated its predictive ability using the data in 2018 based on four metrics, including mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination (R2). In addition, the SHapley Additive exPlanations (SHAP) framework was applied to provide explanation for the prediction of our stacking model. RESULTS Our proposed model outperformed all the base learners and long short-term memory (LSTM) on two datasets. Particularly, compared with the optimal results obtained by individual models, the MAE, RMSE, and MAPE of the stacking model decreased by 13.9%, 12.7%, and 5.8%, respectively, and the R2 improved by 6.8% on CD dataset. The model explanation demonstrated that environmental features played a role in further improving the model performance and identified that high temperature and high concentrations of gaseous air pollutants might strongly associate with an increased risk of CD. CONCLUSIONS Our stacking model considering environmental exposure is efficient in predicting daily HAs for CD and has practical value in early warning and healthcare resource allocation.
Collapse
Affiliation(s)
- Xiaoya Lu
- School of Computer Science and Engineering, University of Electronic Science and Technology of China, No.2006, Xiyuan Ave, West Hi-Tech Zone, 611731, Chengdu, Sichuan, People's Republic of China
| | - Hang Qiu
- School of Computer Science and Engineering, University of Electronic Science and Technology of China, No.2006, Xiyuan Ave, West Hi-Tech Zone, 611731, Chengdu, Sichuan, People's Republic of China.
- Big Data Research Center, University of Electronic Science and Technology of China, Chengdu, China.
| |
Collapse
|
3
|
Lin Y, Wan C, Li S, Xie S, Gan Y, Lu Y. Prediction of women and Children's hospital outpatient numbers based on the autoregressive integrated moving average model. Heliyon 2023; 9:e14845. [PMID: 37089366 PMCID: PMC10114184 DOI: 10.1016/j.heliyon.2023.e14845] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/04/2022] [Revised: 03/16/2023] [Accepted: 03/20/2023] [Indexed: 03/29/2023] Open
Abstract
Objective To evaluate the predictive value of the autoregressive integrated moving average (ARIMA) product seasonal model for the daily outpatient volume of paediatric internal medicine departments in hospitals. Methods The daily outpatient volume of paediatric internal medicine recorded by the hospital information system of the Chengdu Women's and Children's Central Hospital from 1 January 2011 to 31 December 2020 was collected. Using the data from 1 January 2011 to 31 December 2019, the seasonal summation ARIMA model of the time product was established by fitting the tseries program in the R-3.6.3 software. The monthly outpatient volume from January to December 2020 was predicted, and the prediction effect was evaluated according to the mean absolute percentage error (MAPE) between the predicted value and the actual value. Results The outpatient volume of paediatric internal medicine in the hospital from 2011 to 2019 showed an upward trend, with obvious seasonal fluctuations. The optimal model was the ARIMA model ([3,4], 1,2) × (1,1,0) 12, with an Akaike information criterion of 3116.656 and a Bayesian information criterion of 3217.412. The model's residual was a white noise sequence (x2 = 7.56, P = 0.819), and the MAPE between the predicted value and the actual value of the model was 9.56%. Within a 95% confidence interval of the predicted value, the prediction accuracy of the model was high. Conclusion The ARIMA multiplicative seasonal model established in this study is suitable for the short-term prediction of the outpatient volume.
Collapse
|
4
|
Lee W, Schwartz N, Bansal A, Khor S, Hammarlund N, Basu A, Devine B. A Scoping Review of the Use of Machine Learning in Health Economics and Outcomes Research: Part 2-Data From Nonwearables. VALUE IN HEALTH : THE JOURNAL OF THE INTERNATIONAL SOCIETY FOR PHARMACOECONOMICS AND OUTCOMES RESEARCH 2022; 25:2053-2061. [PMID: 35989154 DOI: 10.1016/j.jval.2022.07.011] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 02/07/2022] [Revised: 06/10/2022] [Accepted: 07/10/2022] [Indexed: 06/15/2023]
Abstract
OBJECTIVES Despite the increasing interest in applying machine learning (ML) methods in health economics and outcomes research (HEOR), stakeholders face uncertainties in when and how ML can be used. We reviewed the recent applications of ML in HEOR. METHODS We searched PubMed for studies published between January 2020 and March 2021 and randomly chose 20% of the identified studies for the sake of manageability. Studies that were in HEOR and applied an ML technique were included. Studies related to wearable devices were excluded. We abstracted information on the ML applications, data types, and ML methods and analyzed it using descriptive statistics. RESULTS We retrieved 805 articles, of which 161 (20%) were randomly chosen. Ninety-two of the random sample met the eligibility criteria. We found that ML was primarily used for predicting future events (86%) rather than current events (14%). The most common response variables were clinical events or disease incidence (42%) and treatment outcomes (22%). ML was less used to predict economic outcomes such as health resource utilization (16%) or costs (3%). Although electronic medical records (35%) were frequently used for model development, claims data were used less frequently (9%). Tree-based methods (eg, random forests and boosting) were the most commonly used ML methods (31%). CONCLUSIONS The use of ML techniques in HEOR is growing rapidly, but there remain opportunities to apply them to predict economic outcomes, especially using claims databases, which could inform the development of cost-effectiveness models.
Collapse
Affiliation(s)
- Woojung Lee
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA.
| | - Naomi Schwartz
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA
| | - Aasthaa Bansal
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA
| | - Sara Khor
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA
| | - Noah Hammarlund
- Department of Health Services Research, Management & Policy, University of Florida, Gainesville, FL, USA
| | - Anirban Basu
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA
| | - Beth Devine
- The Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, School of Pharmacy, University of Washington, Seattle, WA, USA
| |
Collapse
|
5
|
Zhao D, Zhang H, Cao Q, Wang Z, He S, Zhou M, Zhang R. The research of ARIMA, GM(1,1), and LSTM models for prediction of TB cases in China. PLoS One 2022; 17:e0262734. [PMID: 35196309 PMCID: PMC8865644 DOI: 10.1371/journal.pone.0262734] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/30/2021] [Accepted: 01/04/2022] [Indexed: 11/25/2022] Open
Abstract
Background and objective Tuberculosis (Tuberculosis, TB) is a public health problem in China, which not only endangers the population’s health but also affects economic and social development. It requires an accurate prediction analysis to help to make policymakers with early warning and provide effective precautionary measures. In this study, ARIMA, GM(1,1), and LSTM models were constructed and compared, respectively. The results showed that the LSTM was the optimal model, which can be achieved satisfactory performance for TB cases predictions in mainland China. Methods The data of tuberculosis cases in mainland China were extracted from the National Health Commission of the People’s Republic of China website. According to the TB data characteristics and the sample requirements, we created the ARIMA, GM(1,1), and LSTM models, which can make predictions for the prevalence trend of TB. The mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) were applied to evaluate the effects of model fitting predicting accuracy. Results There were 3,021,995 tuberculosis cases in mainland China from January 2018 to December 2020. And the overall TB cases in mainland China take on a downtrend trend. We established ARIMA, GM(1,1), and LSTM models, respectively. The optimal ARIMA model is the ARIMA (0,1,0) × (0,1,0)12. The equation for GM(1,1) model was X(k+1) = -10057053.55e(-0.01k) + 10153178.55 the Mean square deviation ratio C value was 0.49, and the Small probability of error P was 0.94. LSTM model consists of an input layer, a hidden layer and an output layer, the parameters of epochs, learning rating are 60, 0.01, respectively. The MAE, RMSE, and MAPE values of LSTM model were smaller than that of GM(1,1) and ARIMA models. Conclusions Our findings showed that the LSTM model was the optimal model, which has a higher accuracy performance than that of ARIMA and GM (1,1) models. Its prediction results can act as a predictive tool for TB prevention measures in mainland China.
Collapse
Affiliation(s)
- Daren Zhao
- Department of Medical Administration, Sichuan Provincial Orthopedics Hospital, Chengdu, Sichuan, P.R. China
| | - Huiwu Zhang
- Department of Medical Administration, Sichuan Provincial Orthopedics Hospital, Chengdu, Sichuan, P.R. China
| | - Qing Cao
- Department of Medical Administration, Sichuan Academy of Medical Sciences & Sichuan Provincial People’s Hospital, Chengdu, Sichuan, P.R. China
| | - Zhiyi Wang
- Department of Medical Administration, Sichuan Cancer Hospital & Institute, Chengdu, Sichuan, P.R. China
| | - Sizhang He
- Department of Information and Statistics, The Affiliated Hospital of Southwest Medical University, Luzhou, Sichuan, P.R. China
| | - Minghua Zhou
- Department of Medical Administration, Luzhou People’s Hospital, Luzhou, Sichuan, P.R. China
| | - Ruihua Zhang
- School of Management, Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan, P.R. China
- * E-mail:
| |
Collapse
|
6
|
Sai Prashanthi G, Molugu N, Kammari P, Vadapalli R, Das AV. Forecast of Outpatient Visits to a Tertiary Eyecare Network in India Using the EyeSmart Electronic Medical Record System. Healthcare (Basel) 2021; 9:healthcare9060749. [PMID: 34207070 PMCID: PMC8233739 DOI: 10.3390/healthcare9060749] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/28/2021] [Revised: 06/14/2021] [Accepted: 06/15/2021] [Indexed: 01/10/2023] Open
Abstract
India is home to 1.3 billion people. The geography and the magnitude of the population present unique challenges in the delivery of healthcare services. The implementation of electronic health records and tools for conducting predictive modeling enables opportunities to explore time series data like patient inflow to the hospital. This study aims to analyze expected outpatient visits to the tertiary eyecare network in India using datasets from a domestically developed electronic medical record system (eyeSmart™) implemented across a large multitier ophthalmology network in India. Demographic information of 3,384,157 patient visits was obtained from eyeSmart EMR from August 2010 to December 2017 across the L.V. Prasad Eye Institute network. Age, gender, date of visit and time status of the patients were selected for analysis. The datapoints for each parameter from the patient visits were modeled using the seasonal autoregressive integrated moving average (SARIMA) modeling. SARIMA (0,0,1)(0,1,7)7 provided the best fit for predicting total outpatient visits. This study describes the prediction method of forecasting outpatient visits to a large eyecare network in India. The results of our model hold the potential to be used to support the decisions of resource planning in the delivery of eyecare services to patients.
Collapse
Affiliation(s)
- Gumpili Sai Prashanthi
- Department of EyeSmart EMR & AEye, L. V. Prasad Eye Institute, Hyderabad 500034, India; (G.S.P.); (P.K.); (R.V.)
- Indian Health Outcomes, Public Health and Economics Research Center, L. V. Prasad Eye Institute, Hyderabad 500034, India
| | - Nareen Molugu
- LVPEI Center for Innovation, L.V. Prasad Eye Institute, Banjara Hills, Hyderabad 500034, India;
| | - Priyanka Kammari
- Department of EyeSmart EMR & AEye, L. V. Prasad Eye Institute, Hyderabad 500034, India; (G.S.P.); (P.K.); (R.V.)
| | - Ranganath Vadapalli
- Department of EyeSmart EMR & AEye, L. V. Prasad Eye Institute, Hyderabad 500034, India; (G.S.P.); (P.K.); (R.V.)
| | - Anthony Vipin Das
- Department of EyeSmart EMR & AEye, L. V. Prasad Eye Institute, Hyderabad 500034, India; (G.S.P.); (P.K.); (R.V.)
- Indian Health Outcomes, Public Health and Economics Research Center, L. V. Prasad Eye Institute, Hyderabad 500034, India
- Correspondence: ; Tel.: +040-68102367
| |
Collapse
|
7
|
Wen C, Huang X, Feng L, Chen L, Hu W, Lai Y, Hao Y. High-resolution age-specific mapping of the two-week illness prevalence rate based on the National Health Services Survey and geostatistical analysis: a case study in Guangdong province, China. Int J Health Geogr 2021; 20:20. [PMID: 33941201 PMCID: PMC8094611 DOI: 10.1186/s12942-021-00273-1] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/21/2020] [Accepted: 04/23/2021] [Indexed: 12/19/2022] Open
Abstract
BACKGROUND The two-week illness prevalence rate is an important and comparable indicator of health service needs. High-spatial-resolution, age-specific risk mapping of this indicator can provide valuable information for health resource allocation. The age-prevalence relationships may be different among areas of the study region, but previous geostatistical models usually ignored the spatial-age interaction. METHODS We took Guangdong province, the province with the largest population and economy in China, as a study case. We collected two-week illness data and other potential influencing predictors from the fifth National Health Services Survey in 2013 and other open-access databases. Bayesian geostatistical binary regression models were developed with spatial-age structured random effect, based on which, high-resolution, age-specific two-week illness prevalence rates, as well as number of people reporting two-week illness, were estimated. The equality of health resource distribution was further evaluated based on the two-week illness mapping results and the health supply data. RESULTS The map across all age groups revealed that the highest risk was concentrated in the central (i.e., Pearl River Delta) and northern regions of the province. These areas had a two-week illness prevalence > 25.0%, compared with 10.0-20.0% in other areas. Age-specific maps revealed significant differences in prevalence between age groups, and the age-prevalence relationships also differed across locations. In most areas, the prevalence rates decrease from age 0 to age 20, and then increase gradually. Overall, the estimated age- and population-adjusted prevalence was 16.5% [95% Bayesian credible interval (BCI): 14.5-18.6%], and the estimated total number of people reporting illness within the two-week period was 17.5 million (95% BCI: 15.5-19.8 million) in Guangdong Province. The Lorenz curve and the Gini coefficient (resulted in 0.3526) showed a moderate level of inequality in health resource distribution. CONCLUSIONS We developed a Bayesian geostatistical modeling framework with spatial-age structured effect to produce age-specific, high-resolution maps of the two-week illness prevalence rate and the numbers of people reporting two-week illness in Guangdong province. The methodology developed in this study can be generalized to other global regions with available relevant survey data. The mapping results will support plans for health resource allocation.
Collapse
Affiliation(s)
- Chuchun Wen
- School of Public Health, Sun Yat-sen University, Guangzhou, 510080, China
| | - Xiaoliang Huang
- Department of Statistics, Government Affairs Service Center of Health Commission of Guangdong Province, Guangzhou, 510060, China
| | - Lifen Feng
- Department of Statistics, Government Affairs Service Center of Health Commission of Guangdong Province, Guangzhou, 510060, China
| | - Long Chen
- Department of Statistics, Government Affairs Service Center of Health Commission of Guangdong Province, Guangzhou, 510060, China
| | - Wei Hu
- Department of Statistics, Government Affairs Service Center of Health Commission of Guangdong Province, Guangzhou, 510060, China
| | - Yingsi Lai
- School of Public Health, Sun Yat-sen University, Guangzhou, 510080, China.
| | - Yuantao Hao
- School of Public Health, Sun Yat-sen University, Guangzhou, 510080, China
| |
Collapse
|