Analysing and quantifying the effect of predictors of stroke direct costs in South Africa using quantile regression

Predicting stroke events with a proactive fusion system: a comprehensive study on imbalance class handling in computational biomechanics

Stroke, as a critical global health concern and the second leading cause of death, occurs when blood flow to the brain is interrupted. Although machine learning has advanced in medical safety, there is limited research on stroke prediction using information fusion systems. This study presents a fusion framework that combines multiple base classifiers and a Meta classifier to improve stroke prediction performance. The research utilizes Grid Search optimized models, such as Random Forest, Support Vector Machine, K Nearest Neighbors, AdaBoost, Gradient Boosting, Light Gradient Boosting, Categorical Boosting, and eXtreme Gradient Boosting for in-depth stroke analysis. Since stroke events are rare and unpredictable, classification outcomes can be deceptive due to imbalanced data. This article examines stroke probability by comparing three data balancing methods: over-sampling, under-sampling, and tomek-link synthetic minority over-sampling (SMOTE-TL) to enhance prediction accuracy. The findings reveal that AdaBoost as a meta-classifier attains the highest performance in the fusion framework, with a peak of 88.09% Recall and 83.66% F1 score. This innovative approach provides crucial insights into stroke prediction and can be a valuable resource for strengthening intervention efforts in advanced healthcare systems.

The healthcare system costs of hip fracture care in South Africa

Despite rapidly ageing populations, data on healthcare costs associated with hip fracture in Sub-Saharan Africa are limited. We estimated high direct medical costs for managing hip fracture within the public healthcare system in SA. These findings should support policy decisions on budgeting and planning of hip fracture services.

PURPOSE

We estimated direct healthcare costs of hip fracture (HF) management in the South African (SA) public healthcare system.

METHODS

We conducted a micro-costing study to estimate costs per patient treated for HF in five regional public sector hospitals in KwaZulu-Natal (KZN), SA. Two hundred consecutive, consenting patients presenting with a fragility HF were prospectively enrolled. Resources used including staff time, consumables, laboratory investigations, radiographs, operating theatre time, surgical implants, medicines, and inpatient days were collected from presentation to discharge. Counts of resources used were multiplied by unit costs, estimated from the KZN Department of Health hospital fees manual 2019/2020, in local currency (South African Rand, ZAR), and converted to 2020 US$ prices. Generalized linear models estimated total covariate-adjusted costs and cost predictors.

RESULTS

The mean unadjusted cost for HF management was US$6935 (95% CI; US$6401-7620) [ZAR114,179 (95% CI; ZAR105,468-125,335)]. The major cost driver was orthopaedics/surgical ward costs US$5904 (95% CI; 5408-6535), contributing to 85% of total cost. The covariate-adjusted cost for HF management was US$6922 (95% CI; US$6743-7118) [ZAR113,976 (95% CI; ZAR111,031-117,197)]. After covariate adjustment, total costs were higher in patients operated under general anaesthesia [US$7251 (95% CI; US$6506-7901)] compared to surgery under spinal anaesthesia US$6880 (95% CI; US$6685-7092) and no surgery US$7032 (95% CI; US$6454-7651).

CONCLUSION

Healthcare costs following a HF are high relative to the gross domestic product per capita and per capita spending on health in SA. As the population ages, this significant economic burden to the health system will increase.

Estimating the healthcare cost of overweight and obesity in South Africa

Background

Overweight and obesity are major risk factors for noncommunicable diseases. This presents a major burden to health systems and to society in South Africa. Collectively, these conditions are overwhelming public healthcare. This is happening when the country has embarked on a journey to universal health coverage, hence the need to estimate the cost of overweight and obesity.

Objective

Our objective was to estimate the healthcare cost associated with treatment of weight-related conditions from the perspective of the South African public sector payer.

Methods

Using a bottom-up gross costing approach, this study draws data from multiple sources to estimate the direct healthcare cost of overweight and obesity in South Africa. Population Attributable Fractions (PAF) were calculated and multiplied by each disease’s total treatment cost to apportion costs to overweight and obesity. Annual costs were estimated for 2020.

Results

The total cost of overweight and obesity is estimated to be ZAR33,194 million in 2020. This represents 15.38% of government health expenditure and is equivalent to 0.67% of GDP. Annual per person cost of overweight and obesity is ZAR2,769. The overweight and obesity cost is disaggregated as follows: cancers (ZAR352 million), cardiovascular diseases (ZAR8,874 million), diabetes (ZAR19,861 million), musculoskeletal disorders (ZAR3,353 million), respiratory diseases (ZAR360 million) and digestive diseases (ZAR395 million). Sensitivity analyses show that the total overweight and obesity cost is between ZAR30,369 million and ZAR36,207 million.

Conclusion

This analysis has demonstrated that overweight and obesity impose a huge financial burden on the public health care system in South Africa. It suggests an urgent need for preventive, population-level interventions to reduce overweight and obesity rates. The reduction will lower the incidence, prevalence, and healthcare spending on noncommunicable diseases.

Estimation and predictors of direct hospitalisation expenses and in-hospital mortality for patients who had a stroke in a low-middle income country: evidence from a nationwide cross-sectional study in Iranian hospitals

OBJECTIVE

Stroke is the second most prevalent cardiovascular disease in Iran. This study investigates the estimation and predictors of hospitalisation expenses and in-hospital mortality for patients who had a stroke in Iranian hospitals.

SETTING

Patients who had a stroke in Iran between 2019 and 2020 were identified through the data collected from the Iran Health Insurance Organization and the Ministry of Health and Medical Education. This study is the first to conduct a pervasive, nationwide investigation.

DESIGN

This is a cross-sectional, prevalence-based study. Generalised linear models and a multiple logistic regression model were used to determine the predictors of hospitalisation expenses and in-hospital mortality for patients who had a stroke.

PARTICIPANTS

A total of 19 150 patients suffering from stroke were studied.

RESULTS

Mean hospitalisation expenses per patient who had a stroke in Iran amounted to US$590.91±974.44 (mean±SD). Mean daily hospitalisation expenses per patient who had a stroke were US$55.18±37.89. The in-hospital mortality for patients who had a stroke was 18.80%. Younger people (aged ≤49 years) had significantly higher expenses than older patients. The OR of in-hospital mortality in haemorrhagic stroke was significantly higher by 1.539 times (95% CI, 1.401 to 1.691) compared with ischaemic and unspecified strokes. Compared with patients covered by the rural fund, patients covered by Iranian health insurance had significantly higher costs by 1.14 times (95% CI, 1.186 to 1.097) and 1.319 times (95% CI, 1.099 to 1.582) higher mortality. There were also significant geographical variations in patients who had a stroke's expenses and mortality rates.

CONCLUSION

Applying cost-effective stroke prevention strategies among the younger population (≤49 years old) is strongly recommended. Migration to universal health insurance can effectively reduce the inequality gap among all insured patients.

The direct and indirect effects of length of hospital stay on the costs of inpatients with stroke in Ningxia, China, between 2015 and 2020: A retrospective study using quantile regression and structural equation models

Importance

Length of hospital stay (LOHS) is the main cost-determining factor of hospitalization for stroke patients. However, previous analyses involving LOHS did not consider confounding or indirect factors, or the effects of other factors on LOHS and inpatient costs.

Objective

To investigate the direct and indirect effects of LOHS on the hospitalization costs of inpatients with ischemic and hemorrhagic stroke.

Design setting and participants

This was a population-based, retrospective, and observational study that analyzed data acquired from the Nationwide Inpatient Sample between 2015 and 2020 relating to ischemic and hemorrhagic stroke in Ningxia, China.

Main outcomes and measures

Hospitalizations were identified by the International Classification of Diseases 10th Revision (ICD-10). Inpatient costs were described by the median M (P25, P75). We used a quantile regression model to estimate the linear relationships between a group of independent variables X and the quantile of the explained variable hospitalization cost (Y). A structural equation model (SEM) was then used to investigate the direct and indirect effects of LOHS on inpatient costs.

Results

The study included 129,444 patients with ischemic stroke and 15,525 patients with hemorrhagic stroke. The median LOHS was 10 (8-13) days for ischemic stroke and 15 (10-22) days for hemorrhagic stroke. The median M (P₂₅, P₇₅) of inpatient costs was $1020 (742-1545) for ischemic stroke and 2813 (1576-6191) for hemorrhagic stroke. The total effect of LOHS on inpatient costs was 0.795 in patients with ischemic stroke. The effect of yearof discharge (X4) and CCI (X8) on inpatient costs was dominated by an indirect effect through the LOHS. The indirect effect was -0.071 (84.52% of the total effect value) and 0.034 (69.39% of the total effect value), respectively. The total effect of LOHS on inpatient costs in patients with hemorrhagic stroke was 0.754. The influence of CCI on inpatient costs was dominated by an indirect effect through LOHS; the indirect effect value was -0.028 (77.78% of the total effect value). The payment type, surgery, method of discharge, and hospital level also exerted an impact on inpatient costs by direct and indirect effects through the LOHS.

Conclusions and relevance

Length of hospital stay (LOHS) was identified as the main factor influencing hospitalization costs. However, other social factors were shown to indirectly influence hospitalization costs through the LOHS. Taking effective measures to further reduce hospitalization costs remains an effective way to control hospitalization costs for stroke patients.

Analysis of the Cost and Case-mix of Post-acute Stroke Patients in China Using Quantile Regression and the Decision-tree Models

Purpose

Post-acute care is fast developing in China, yet a payment system for post-acute care has not been established. As stroke is the leading cause of mortality and disability in China, patients constitute a large share of post-acute-care patients among all hospitalized patients. This study was to identify the cost determinants and establish a case-mix classification of the post-acute care system for stroke patients in China.

Patients and Methods

A total of 5401 post-acute stroke patients in seven hospitals of Jinhua City from January 2018 to December 2020 were selected. Demographic characteristics, medical status, functional measures (eg, the Barthel Index, Mini-Mental State Examination, Gugging Swallowing Screen, Hamilton Depression Scale), and cost data were extracted. Generalized linear model (GLM) and quantile regression (QR) were conducted to determine the predictors of cost, and a case-mix classification model was established using the decision-tree analysis.

Results

The GLM regression revealed that gender, tracheostomy, complication or comorbidity (CC), activities of daily living (ADL), and cognitive impairment were the main variables significantly affecting the hospitalization expenses of post-acute stroke patients. The QR model showed that the gender, tracheostomy and CC factors had a more significant impact on per diem costs on the upper quantiles. In contrast, cognitive impairment had a more substantial effect on the lower quantiles, and ADL significantly impacted the central quantile. Using tracheostomy, CC, and ADL as node variables of the regression tree, 12 classes were generated. The case-mix classification performed reliably and robustly, as measured by the reduction in the variation statistic (RIV=0.46) and class-specific coefficients of variation (CV less than 1.0; range: 0.18–0.81).

Conclusion

QR has strengths in comprehensively identifying cost predictors across cost groups. Tracheostomy, CC, and ADL significantly can predict the expenses of post-acute care for stroke patients. The established case-mix classification system can inform the future payment policy of post-acute care in China.