Using machine learning tools to investigate factors associated with trends in 'no-shows' in outpatient appointments

Dynamic Price Application to Prevent Financial Losses to Hospitals Based on Machine Learning Algorithms

Hospitals that are considered non-profit take into consideration not to make any losses other than seeking profit. A model that ensures that hospital price policies are variable due to hospital revenues depending on patients with appointments is presented in this study. A dynamic pricing approach is presented to prevent patients who have an appointment but do not show up to the hospital from causing financial loss to the hospital. The research leverages three distinct machine learning (ML) algorithms, namely Random Forest (RF), Gradient Boosting (GB), and AdaBoost (AB), to analyze the appointment status of 1073 patients across nine different departments in a hospital. A mathematical formula has been developed to apply the penalty fee to evaluate the reappointment situations of the same patients in the first 100 days and the gaps in the appointment system, considering the estimated patient appointment statuses. Average penalty cost rates were calculated based on the ML algorithms used to determine the penalty costs patients will face if they do not show up, such as 22.87% for RF, 19.47% for GB, and 14.28% for AB. As a result, this study provides essential criteria that can help hospital management better understand the potential financial impact of patients missing appointments and can be considered when choosing between these algorithms.

Decision analysis framework for predicting no-shows to appointments using machine learning algorithms

BACKGROUND

No-show to medical appointments has significant adverse effects on healthcare systems and their clients. Using machine learning to predict no-shows allows managers to implement strategies such as overbooking and reminders targeting patients most likely to miss appointments, optimizing the use of resources.

METHODS

In this study, we proposed a detailed analytical framework for predicting no-shows while addressing imbalanced datasets. The framework includes a novel use of z-fold cross-validation performed twice during the modeling process to improve model robustness and generalization. We also introduce Symbolic Regression (SR) as a classification algorithm and Instance Hardness Threshold (IHT) as a resampling technique and compared their performance with that of other classification algorithms, such as K-Nearest Neighbors (KNN) and Support Vector Machine (SVM), and resampling techniques, such as Random under Sampling (RUS), Synthetic Minority Oversampling Technique (SMOTE) and NearMiss-1. We validated the framework using two attendance datasets from Brazilian hospitals with no-show rates of 6.65% and 19.03%.

RESULTS

From the academic perspective, our study is the first to propose using SR and IHT to predict the no-show of patients. Our findings indicate that SR and IHT presented superior performances compared to other techniques, particularly IHT, which excelled when combined with all classification algorithms and led to low variability in performance metrics results. Our results also outperformed sensitivity outcomes reported in the literature, with values above 0.94 for both datasets.

CONCLUSION

This is the first study to use SR and IHT methods to predict patient no-shows and the first to propose performing z-fold cross-validation twice. Our study highlights the importance of avoiding relying on few validation runs for imbalanced datasets as it may lead to biased results and inadequate analysis of the generalization and stability of the models obtained during the training stage.

Attendance at remote versus in-person outpatient appointments in an NHS Trust

INTRODUCTION

With the growing use of remote appointments within the National Health Service, there is a need to understand potential barriers of access to care for some patients. In this observational study, we examined missed appointments rates, comparing remote and in-person appointments among different patient groups.

METHODS

We analysed adult outpatient appointments at Imperial College Healthcare NHS Trust in Northwest London in 2021. Rates of missed appointments per patient were compared between remote versus in-person appointments using negative binomial regression models. Models were stratified by appointment type (first or a follow-up).

RESULTS

There were 874,659 outpatient appointments for 189,882 patients, 29.5% of whom missed at least one appointment. Missed rates were 12.5% for remote first appointments and 9.2% for in-person first appointments. Remote and in-person follow-up appointments were missed at similar rates (10.4% and 10.7%, respectively). For remote and in-person appointments, younger patients, residents of more deprived areas, and patients of Black, Mixed and 'other' ethnicities missed more appointments. Male patients missed more in-person appointments, particularly at younger ages, but gender differences were minimal for remote appointments. Patients with long-term conditions (LTCs) missed more first appointments, whether in-person or remote. In follow-up appointments, patients with LTCs missed more in-person appointments but fewer remote appointments.

DISCUSSION

Remote first appointments were missed more often than in-person first appointments, follow-up appointments had similar attendance rates for both modalities. Sociodemographic differences in outpatient appointment attendance were largely similar between in-person and remote appointments, indicating no widening of inequalities in attendance due to appointment modality.

Predicting missed health care visits during the COVID-19 pandemic using machine learning methods: evidence from 55,500 individuals from 28 European countries

BACKGROUND

Pandemics such as the COVID-19 pandemic and other severe health care disruptions endanger individuals to miss essential care. Machine learning models that predict which patients are at greatest risk of missing care visits can help health administrators prioritize retentions efforts towards patients with the most need. Such approaches may be especially useful for efficiently targeting interventions for health systems overburdened during states of emergency.

METHODS

We use data on missed health care visits from over 55,500 respondents of the Survey of Health, Ageing and Retirement in Europe (SHARE) COVID-19 surveys (June - August 2020 and June - August 2021) with longitudinal data from waves 1-8 (April 2004 - March 2020). We compare the performance of four machine learning algorithms (stepwise selection, lasso, random forest, and neural networks) to predict missed health care visits during the first COVID-19 survey based on common patient characteristics available to most health care providers. We test the prediction accuracy, sensitivity, and specificity of the selected models for the first COVID-19 survey by employing 5-fold cross-validation, and test the out-of-sample performance of the models by applying them to the data from the second COVID-19 survey.

RESULTS

Within our sample, 15.5% of the respondents reported any missed essential health care visit due to the COVID-19 pandemic. All four machine learning methods perform similarly in their predictive power. All models have an area under the curve (AUC) of around 0.61, outperforming random prediction. This performance is sustained for data from the second COVID-19 wave one year later, with an AUC of 0.59 for men and 0.61 for women. When classifying all men (women) with a predicted risk of 0.135 (0.170) or higher as being at risk of missing care, the neural network model correctly identifies 59% (58%) of the individuals with missed care visits, and 57% (58%) of the individuals without missed care visits. As the sensitivity and specificity of the models are strongly related to the risk threshold used to classify individuals, the models can be calibrated depending on users' resource constraints and targeting approach.

CONCLUSIONS

Pandemics such as COVID-19 require rapid and efficient responses to reduce disruptions in health care. Based on characteristics available to health administrators or insurance providers, simple machine learning algorithms can be used to efficiently target efforts to reduce missed essential care.

A spotlight on lymphoedema Did Not Attends: Demographics and workforce costs

This unique evaluation aimed to estimate, the financial impact of non-attendance on a nation-wide hospital lymphoedema service. Along with gaining some understanding of patient characteristics of those who Did Not Attend (DNA) and were subsequently discharged. The evaluation design interrogated existing performance data from 2012 to 2022. This information was used to estimate the costs incurred based on national published sources and pay scales. Staffing costs of over £1.1 m in one decade related to the financial impact of over 23 000 unattended lymphoedema appointments. The characteristics of 870 patients from 2019/2020 were also evaluated suggesting that those with a wound alongside complex lymphoedema were less likely to DNA appointments. Two-thirds of patients were managing two or more comorbidities-obesity, cardiac conditions and diabetes being the most common. It seems likely that some DNAs are avoidable by adapting appointment administrative processes and greater understanding of patients' perception of value. However, the reasons for DNA are likely to be varied and nuanced so potentially a small proportion are unavoidable. Modernising appointment processes and identifying patient value may help minimise DNA costs in the future.

No-Show in Medical Appointments with Machine Learning Techniques: A Systematic Literature Review

No-show appointments in healthcare is a problem faced by medical centers around the world, and understanding the factors associated with no-show behavior is essential. In recent decades, artificial intelligence has taken place in the medical field and machine learning algorithms can now work as an efficient tool to understand the patients’ behavior and to achieve better medical appointment allocation in scheduling systems. In this work, we provide a systematic literature review (SLR) of machine learning techniques applied to no-show appointments aiming at establishing the current state-of-the-art. Based on an SLR following the PRISMA procedure, 24 articles were found and analyzed, in which the characteristics of the database, algorithms and performance metrics of each study were synthesized. Results regarding which factors have a higher impact on missed appointment rates were analyzed too. The results indicate that the most appropriate algorithms for building the models are decision tree algorithms. Furthermore, the most significant determinants of no-show were related to the patient’s age, whether the patient missed a previous appointment, and the distance between the appointment and the patient’s scheduling.

Application of Machine Learning Techniques to Predict a Patient’s No-Show in the Healthcare Sector

The health sector faces a series of problems generated by patients who miss their scheduled appointments. The main challenge to this problem is to understand the patient’s profile and predict potential absences. The goal of this work is to explore the main causes that contribute to a patient’s no-show and develop a prediction model able to identify whether the patient will attend their scheduled appointment or not. The study was based on data from clinics that serve the Unified Health System (SUS) at the University of Vale do Itajaí in southern Brazil. The model obtained was tested on a real collected dataset with about 5000 samples. The best model result was performed by the Random Forest classifier. It had the best Recall Rate (0.91) and achieved an ROC curve rate of 0.969. This research was approved and authorized by the Ethics Committee of the University of Vale do Itajaí, under opinion 4270,234, contemplating the General Data Protection Law.