Heterogeneity of reasons for attendance in frequent attenders of emergency departments and its relationship to future attendance

Using machine learning to identify frequent attendance at accident and emergency services in Lanarkshire

Background

Frequent attenders to accident and emergency (A&E) services pose complex challenges for healthcare providers, often driven by critical clinical needs. Machine learning (ML) offers potential for predictive approaches to managing frequent attendance, yet its application in this area is limited. Existing studies often focus on specific populations or models, raising concerns about generalisability. Identifying risk factors for frequent attendance and high resource use is crucial for effective prevention strategies.

Objectives

This research aims to evaluate the strengths and weaknesses of ML approaches in predicting frequent A&E attendance in NHS Lanarkshire, Scotland, identify associated risk factors and compare findings with existing research to uncover commonalities and differences.

Method

Health and social care data were collected from 17,437 A&E patients in NHS Lanarkshire (2021-2022), including clinical, social and demographic information. Five classification models were tested: multinomial logistic regression (LR), random forests (RF), support vector machine (SVM) classifier, k-nearest neighbours (k-NN) and multi-layer perceptron (MLP) classifier. Models were evaluated using a confusion matrix and metrics such as precision, recall, F1 and area under the curve. Shapley values were used to identify risk factors.

Results

MLP achieved the highest F1 score (0.75), followed by k-NN, RF and SVM (0.72 each), and LR (0.70). Key health conditions and risk factors consistently predicted frequent attendance across models, with some variation highlighting dataset-specific characteristics.

Conclusions

This study underscores the utility of combining ML models to enhance prediction accuracy and identify risk factors. Findings align with existing research but reveal unique insights specific to the dataset and methodology.

Association between social support and ambulance use among older people in Japan: an empirical cross-sectional study

BACKGROUND

Ambulance service demand and utilization are increasing worldwide. To address this issue, the factors that affect ambulance use must be identified. Few studies have examined factors that can intervene and thus reduce the frequency of ambulance use. This study aimed to examine the association between social support and ambulance use among older adults in Japan. We hypothesize that social support is associated with reduced ambulance use.

METHODS

This cross-sectional study was conducted as part of the Japan Gerontological Evaluation Study. In December 2019 and January 2020, we collaborated with individuals aged 65 years or above with no long-term care needs. A total of 24,581 participants were included in the analysis. The objective and explanatory variables were ambulance use and social support, respectively. Binomial regression analysis was used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs).

RESULTS

Social support was associated with ambulance use. People who had no one to listen to their complaints or worries were significantly more likely to use ambulance services than those who did (OR [95% CI] = 1.26 [1.03-1.53]). People with no one to take care of them when they were ill were also significantly more likely to use ambulance services than those who had someone to provide care (1.15 [1.01-1.31]). Moreover, the results of binomial logistic regression analysis indicated that individuals who called an ambulance but were not hospitalized had significantly lower social support compared to those who did not call an ambulance.

CONCLUSIONS

The results suggest that the presence and quality of social support play a significant role in ambulance use among older adults in Japan. Our findings can help policymakers to plan and implement strategies for reducing the burden on emergency medical care.

Supporting High-impAct useRs in Emergency Departments (SHarED) quality improvement: a mixed-method evaluation

BACKGROUND

The need to better manage frequent attenders or high-impact users (HIUs) in hospital emergency departments (EDs) is widely recognised. These patients often have complex medical needs and are also frequent users of other health and care services. The West of England Academic Health Science Network launched its Supporting High impAct useRs in Emergency Departments (SHarED) quality improvement programme to spread a local HIU intervention across six other EDs in five Trusts.

AIM

SHarED aimed to reduce ED attendance and hospital admissions by 20% for enrolled HIUs. To evaluate the implementation of SHarED, we sought to learn about the experience of staff with HIU roles and their ED colleagues and assess the impact on HIU attendance and admissions.

METHODS

We analysed a range of data including semistructured interviews with 10 HIU staff; the number of ED staff trained in HIU management; an ED staff experience survey; and ED attendances and hospital admissions for 148 HIUs enrolled in SHarED.

RESULTS

Staff with HIU roles were unanimously positive about the benefits of SHarED for both staff and patients. SHarED contributed to supporting ED staff with patient-centred recommendations and provided the basis for more integrated case management across the health and care system. 55% of ED staff received training. There were improvements in staff views relating to confidence, support, training and HIUs receiving more appropriate care. The mean monthly ED attendance per HIU reduced over time. Follow-up data for 86% (127/148) of cases showed a mean monthly ED attendances per HIU reduced by 33%, from 2.1 to 1.4, between the 6 months pre-enrolment and post-enrolment (p<0.001).

CONCLUSION

SHarED illustrates the considerable potential for a quality improvement programme to promote more integrated case management by specialist teams across the health and care system for particularly vulnerable individuals and improve working arrangements for hard-pressed staff.