Population Health Management to identify and characterise ongoing health need for high-risk individuals shielded from COVID-19: a cross-sectional cohort study

Use of linked patient data to assess the effect of Long-COVID on system-wide healthcare utilisation

Background: Within the relatively early stages of the COVID-19 pandemic, there had been an awareness of the potential longer-term effects of infection (so called Long-COVID) but little was known of the ongoing demands such patients may place on healthcare services. Objective: To investigate whether COVID-19 illness is associated with increased post-acute healthcare utilisation. Method: Using linked data from primary care, secondary care, mental health and community services, activity volumes were compared across the 3 months preceding and proceeding COVID-19 diagnoses for 7,791 individuals, with a distinction made between whether or not patients were hospitalised for treatment. Differences were assessed against those of a control group containing individuals who had not received a COVID-19 diagnosis. All data were sourced from the authors' healthcare system in South West England. Results: For hospitalised COVID-19 cases, a statistically significant increase in non-elective admissions was identified for males and females <65 years. For non-hospitalised cases, statistically significant increases were identified in GP Doctor and Nurse attendances and GP prescriptions (males and females, all ages); Emergency Department attendances (females <65 years); Mental Health contacts (males and females ≥65 years); and Outpatient consultations (males ≥65 years). Conclusion: There is evidence of an association between positive COVID-19 diagnosis and increased post-acute activity within particular healthcare settings. Linked patient-level data provides information that can be useful to understand ongoing healthcare needs resulting from Long-COVID, and support the configuration of Long-COVID pathways of care.

A scoping review of the clinical application of machine learning in data-driven population segmentation analysis

OBJECTIVE

Data-driven population segmentation is commonly used in clinical settings to separate the heterogeneous population into multiple relatively homogenous groups with similar healthcare features. In recent years, machine learning (ML) based segmentation algorithms have garnered interest for their potential to speed up and improve algorithm development across many phenotypes and healthcare situations. This study evaluates ML-based segmentation with respect to (1) the populations applied, (2) the segmentation details, and (3) the outcome evaluations.

MATERIALS AND METHODS

MEDLINE, Embase, Web of Science, and Scopus were used following the PRISMA-ScR criteria. Peer-reviewed studies in the English language that used data-driven population segmentation analysis on structured data from January 2000 to October 2022 were included.

RESULTS

We identified 6077 articles and included 79 for the final analysis. Data-driven population segmentation analysis was employed in various clinical settings. K-means clustering is the most prevalent unsupervised ML paradigm. The most common settings were healthcare institutions. The most common targeted population was the general population.

DISCUSSION

Although all the studies did internal validation, only 11 papers (13.9%) did external validation, and 23 papers (29.1%) conducted methods comparison. The existing papers discussed little validating the robustness of ML modeling.

CONCLUSION

Existing ML applications on population segmentation need more evaluations regarding giving tailored, efficient integrated healthcare solutions compared to traditional segmentation analysis. Future ML applications in the field should emphasize methods' comparisons and external validation and investigate approaches to evaluate individual consistency using different methods.

Machine learning reveals sex differences in clinical features of acute exacerbation of chronic obstructive pulmonary disease: A multicenter cross-sectional study

Introduction

Intrinsically, chronic obstructive pulmonary disease (COPD) is a highly heterogonous disease. Several sex differences in COPD, such as risk factors and prevalence, were identified. However, sex differences in clinical features of acute exacerbation chronic obstructive pulmonary disease (AECOPD) were not well explored. Machine learning showed a promising role in medical practice, including diagnosis prediction and classification. Then, sex differences in clinical manifestations of AECOPD were explored by machine learning approaches in this study.

Methods

In this cross-sectional study, 278 male patients and 81 female patients hospitalized with AECOPD were included. Baseline characteristics, clinical symptoms, and laboratory parameters were analyzed. The K-prototype algorithm was used to explore the degree of sex differences. Binary logistic regression, random forest, and XGBoost models were performed to identify sex-associated clinical manifestations in AECOPD. Nomogram and its associated curves were established to visualize and validate binary logistic regression.

Results

The predictive accuracy of sex was 83.930% using the k-prototype algorithm. Binary logistic regression revealed that eight variables were independently associated with sex in AECOPD, which was visualized by using a nomogram. The AUC of the ROC curve was 0.945. The DCA curve showed that the nomogram had more clinical benefits, with thresholds from 0.02 to 0.99. The top 15 sex-associated important variables were identified by random forest and XGBoost, respectively. Subsequently, seven clinical features, including smoking, biomass fuel exposure, GOLD stages, PaO2, serum potassium, serum calcium, and blood urea nitrogen (BUN), were concurrently identified by three models. However, CAD was not identified by machine learning models.

Conclusions

Overall, our results support that the clinical features differ markedly by sex in AECOPD. Male patients presented worse lung function and oxygenation, less biomass fuel exposure, more smoking, renal dysfunction, and hyperkalemia than female patients with AECOPD. Furthermore, our results also suggest that machine learning is a promising and powerful tool in clinical decision-making.

Emotional and Mental Wellbeing Following COVID-19 Public Health Measures on People Living With Dementia and Carers

BACKGROUND

To date, there appears to be no evidence on the longer-term impacts caused by COVID-19 and its related public health restrictions on some of the most vulnerable in our societies. The aim of this research was to explore the change in impact of COVID-19 public health measures on the mental wellbeing of people living with dementia (PLWD) and unpaid carers.

METHOD

Semi-structured, follow-up telephone interviews were conducted with PLWD and unpaid carers between June and July 2020. Participants were asked about their experiences of accessing social support services during the pandemic, and the impact of restrictions on their daily lives.

RESULTS

20 interviews were conducted and thematically analyzed, which produced 3 primary themes concerning emotional responses and impact to mental health and wellbeing during the course of the pandemic: 1) Impact on mental health during lockdown, 2) Changes to mental health following easing of public health, and 3) The long-term effect of public health measures.

CONCLUSIONS

The findings from this research shed light on the longer-term psychological impacts of the UK Government's public health measures on PLWD and their carers. The loss of social support services was key in impacting this cohort mentally and emotionally, displaying a need for better psychological support, for both carers and PLWD.