Previous Emergency Admissions Research Articles

Artificial intelligence informing clinical decision making on the risk of hospital readmissions in multi-morbid patients: a systematic review

Abstract Introduction Multi-morbid patients have complex care needs and are more likely to utilise healthcare services, which represents a financial burden on healthcare organisations. 30-day readmission of patients is a method of measuring the quality of the services provided within healthcare.[1] Artificial intelligence (AI) models can flag multi-morbidity patients at increased risk of readmission.[2] However, it is unclear which common predictors have been used to develop these models. Aim To systematically review the medical literature to identify predictors that have been used to develop AI models that predicted unplanned 30-day hospital readmissions of multi-morbid patients, and whether these predictors were modifiable or non-modifiable. Methods Four large databases: Medline, Embase, Web of Science and Cumulative Index to Nursing and Allied Health Literature (CINAHL) were searched on 15th November 2022. Only publications that developed a machine learning algorithm to predict unplanned 30-day hospital readmission for patients with multi-morbidity were included. Articles published in English language were eligible for inclusion. A narrative synthesis of all eligible studies was undertaken. Key findings were identified through inductive strategy, which includes simplifying raw data in the included articles into a comprehensive format. This systematic review was registered with the PROSPRO database (CRD42022373937) and followed PRISMA guidelines. Results There were 1,906 articles extracted from all databases, 18 of which met our inclusion criteria. A total of 669 predictors of hospital readmission were found, and were divided into 103 modifiable (i.e., action could be taken to change the outcome) and 566 non-modifiable (i.e. age, gender, number of previous emergency admissions. The average number of modifiable predictors per prediction model was six, with the most common being length of hospital stay, multi-morbidity, obesity, hypertension, diabetes, depression, and anaemia. The most commonly used AI algorithm was the gradient boosting algorithm. This algorithm generally showed a better performance than other models reported. The performance of all models in the included studies was compared; 13 studies showed an average sensitivity of 72%. The average specificity was calculated at 73% from 11 studies. The highest AUC reported was above 0.9, which demonstrates an excellent performing model, while five models did not report a value for the AUC. Conclusion We identified modifiable and non-modifiable predictors used in previously developed AI algorithms that predict 30-day hospital readmissions of multi-morbid patients. Modifiable predictors in particular can help guide clinical decision-making by allowing early actions to be taken in primary care to potentially reduce the risk of readmission. However, one limitation of this review is that some included studies used balanced (i.e. synthetic) data which could have introduced a level of bias.

Risk of winter hospitalisation and death from acute respiratory infections in Scotland: national retrospective cohort study.

We undertook a national analysis to characterise and identify risk factors for acute respiratory infections (ARIs) resulting in hospitalisation during the winter period in Scotland. A population-based retrospective cohort analysis. Scotland. The study involved 5.4 million residents in Scotland. Cox proportional hazard models were used to estimate adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for the association between risk factors and ARI hospitalisation. Between 1 September 2022 and 31 January 2023, there were 22,284 (10.9% of 203,549 with any emergency hospitalisation) ARI hospitalisations (1759 in children and 20,525 in adults) in Scotland. Compared with the reference group of children aged 6-17 years, the risk of ARI hospitalisation was higher in children aged 3-5 years (aHR = 4.55; 95% CI: 4.11-5.04). Compared with those aged 25-29 years, the risk of ARI hospitalisation was highest among the oldest adults aged ≥80 years (aHR = 7.86; 95% CI: 7.06-8.76). Adults from more deprived areas (most deprived vs. least deprived, aHR = 1.64; 95% CI: 1.57-1.72), with existing health conditions (≥5 vs. 0 health conditions, aHR = 4.84; 95% CI: 4.53-5.18) or with history of all-cause emergency admissions (≥6 vs. 0 previous emergency admissions, aHR = 7.53; 95% CI: 5.48-10.35) were at a higher risk of ARI hospitalisations. The risk increased by the number of existing health conditions and previous emergency admission. Similar associations were seen in children. Younger children, older adults, those from more deprived backgrounds and individuals with greater numbers of pre-existing conditions and previous emergency admission were at increased risk for winter hospitalisations for ARI.

Risk factors for early revision after total hip and knee arthroplasty: National observational study from a surgeon and population perspective.

AimsTo identify predictors of early revision (within 3 years of the index operation) for hip and knee replacement (HR, KR) from both surgeon and population perspectives.Patients and methodsHierarchical logistic regression on national administrative data for England for index procedures between April 2009 and March 2014.ResultsThere were 315,273 index HR procedures and 374,530 index KR procedures for analysis. Three-year revision rates were 2.1% for HR and 2.2% for KR. The highest odds ratios for HR were for 3+ previous emergency admissions, drug abuse, Parkinson’s disease, resurfacing and ages under 60; for KR these were patellofemoral or partial joint replacement, 3+ previous emergency admissions, paralysis and ages under 60. Smaller effects were found for other comorbidities such as obesity (HR) and diabetes (KR). From a population perspective, the only population attributable fractions over 5% were for male gender, uncemented total hip replacements and partial knee or patellofemoral replacements.ConclusionsMeeting the rising demand for revision surgery is a challenge for healthcare leaders and policymakers. Our findings suggest optimising patients pre-operatively and improving patient selection for primary arthroplasty may reduce the burden of early revision of arthroplasty. Our study gives useful information on the additional risks of various comorbidities and procedures, which enables a more informed consent process.Clinical relevanceSurgeons should make patients with certain risk factors such as age and procedure type aware of their higher revision risk as part of shared decision-making.

Death within 1 year among emergency medical admissions to Scottish hospitals: incident cohort study

BackgroundIt is increasingly recognised that large numbers of hospital inpatients have entered the last year of their lives.AimTo establish the likelihood of death within 12 months of admission to hospital;...

Evidence for Supported Self-Care at Scale

Context: Liverpool Clinical Commissioning Group (LCCG) is committed to provide health and social care for Liverpool that is person-centred, supports people to stay well and provides the very best care. The work described here is part of a UK government sponsored programme in Liverpool to deliver assisted living technologies at scale. It recognises that the only way in which increasing demand can be managed with decreasing resources is by solutions that focus face-to-face care where and when it is needed, and support people to manage their long term conditions (LTCs) through technology.Methods: Over 3 years 2234 patients with COPD, HF or diabetes were recruited through case finding from GP practice lists. They received a programme of tele-monitoring through TV or tablet with the support of a clinical hub and a structured programme of case management, monitoring (vital signs and questionnaires), education (hints, tips and videos) and telephone coaching.Using a dataset of more than 480,000 people, a study cohort of 1808 patients is matched to a pseudonymised control cohort. Matching is done in terms of disease, age, previous emergency admissions, future emergency admissions risk, deprivation and polypharmacy. There was no change in the usual care for the individuals in the control group. Analysis is performed to control for the duration of the intervention, season and risk of one or more emergency admissions in the following 12 months.Results: The intervention cohort showed a broad distribution in emergency admission risk. The average risk is 25% and the 10-20% was the single largest risk-band. On average people remained on service 4.5 months. People with high risk were 4 times more likely to be on service than people with low risk and they remained on service longer. Patient satisfaction was high with 90% of patients reporting that they felt more in control, had gained confidence and/or felt better able to cope with their condition. The high risk group showed reductions in emergency admissions and secondary care costs in comparison with the control group ranging from 22% to 32% (p<0.05)

Population-based cohort study of outcomes following cholecystectomy for benign gallbladder diseases.

The aim was to describe the management of benign gallbladder disease and identify characteristics associated with all-cause 30-day readmissions and complications in a prospective population-based cohort. Data were collected on consecutive patients undergoing cholecystectomy in acute UK and Irish hospitals between 1 March and 1 May 2014. Potential explanatory variables influencing all-cause 30-day readmissions and complications were analysed by means of multilevel, multivariable logistic regression modelling using a two-level hierarchical structure with patients (level 1) nested within hospitals (level 2). Data were collected on 8909 patients undergoing cholecystectomy from 167 hospitals. Some 1451 cholecystectomies (16·3 per cent) were performed as an emergency, 4165 (46·8 per cent) as elective operations, and 3293 patients (37·0 per cent) had had at least one previous emergency admission, but had surgery on a delayed basis. The readmission and complication rates at 30 days were 7·1 per cent (633 of 8909) and 10·8 per cent (962 of 8909) respectively. Both readmissions and complications were independently associated with increasing ASA fitness grade, duration of surgery, and increasing numbers of emergency admissions with gallbladder disease before cholecystectomy. No identifiable hospital characteristics were linked to readmissions and complications. Readmissions and complications following cholecystectomy are common and associated with patient and disease characteristics.

Appropriateness of diagnostic effort in hospital emergency room attention for episodes of COPD exacerbation.

To assess the adequacy of diagnostic effort in the emergency departments of Spanish hospitals with respect to episodes of exacerbation of chronic obstructive pulmonary disease (COPD). A descriptive cross-sectional study, conducted between 2007 and 2010 in 15 hospitals in Andalusia, Catalonia, Madrid and the Basque Country. The study population included cases of COPD exacerbation attended at the emergency departments of the participating hospitals. Diagnostic efforts were considered sufficient and appropriate when the emergency room conducted a clinical evaluation including electrocardiogram, chest X-ray, arterial blood gas analysis and spirometry. 2852 episodes of COPD exacerbation attended in hospital emergency departments were assessed. 91.4% of the patients were male, with a mean age of 72.8 (SD 9.5) years, and 45.6% had had a previous emergency admission. The diagnostic effort was considered adequate in 60.1% of the episodes (95% CI: 58.3-61.9). The inter-hospital range of variation(25-75) was 1.67 and the coefficient of variation was 28.3%. In multivariate analysis, adjusting for hospital, date of admission and previous hospitalization, among the male patients, the OR for adequate diagnostic effort was 1.38 (95% CI: 1.04-1.84) CONCLUSION: With respect to diagnostic effort, inequities were observed in our assessment of episodes of COPD exacerbation attended in the emergency departments of Spanish public hospitals. In a high percentage of cases (40%), proper assessment was not conducted. Moreover, inter-individual and inter-hospital differences were observed.

Enhanced clinical pharmacy service targeting tools: risk-predictive algorithms.

This study aimed to determine the value of using a mix of clinical pharmacy data and routine hospital admission spell data in the development of predictive algorithms. Exploration of risk factors in hospitalized patients, together with the targeting strategies devised, will enable the prioritization of clinical pharmacy services to optimize patient outcomes. Predictive algorithms were developed using a number of detailed steps using a 75% sample of integrated medicines management (IMM) patients, and validated using the remaining 25%. IMM patients receive targeted clinical pharmacy input throughout their hospital stay. The algorithms were applied to the validation sample, and predicted risk probability was generated for each patient from the coefficients. Risk threshold for the algorithms were determined by identifying the cut-off points of risk scores at which the algorithm would have the highest discriminative performance. Clinical pharmacy staffing levels were obtained from the pharmacy department staffing database. Numbers of previous emergency admissions and admission medicines together with age-adjusted co-morbidity and diuretic receipt formed a 12-month post-discharge and/or readmission risk algorithm. Age-adjusted co-morbidity proved to be the best index to predict mortality. Increased numbers of clinical pharmacy staff at ward level was correlated with a reduction in risk-adjusted mortality index (RAMI). Algorithms created were valid in predicting risk of in-hospital and post-discharge mortality and risk of hospital readmission 3, 6 and 12 months post-discharge. The provision of ward-based clinical pharmacy services is a key component to reducing RAMI and enabling the full benefits of pharmacy input to patient care to be realized.

Prediction of 30-day cardiac-related-emergency-readmissions using simple administrative hospital data

BackgroundControl and reduction of cardiovascular-disease-related readmissions is clinically, logistically and politically challenging. Recent strategies focus on 30-day readmissions. A screening tool for the detection of potential cases is necessary to make further case management more efficient. MethodsCohort study. Hospital administrative data were analyzed in order to obtain information about cardiac-related hospitalizations from 2003 to 2009 at a Spanish academic tertiary care center. Predictor-variables of admissions that presented or did not present 30-day cardiac-related readmission were compared. A prediction model was constructed and tested on a validation sample. Model performance was assessed for all cardiac diseases and for 24 main-cardiac-disease-sets. ResultsThe study sample was 35531 hospital-admissions. The model included 11 predictors: number of previous emergency admission in 180days, residence out of area, no procedure applied during hospitalization, major or minor therapeutic procedure applied during hospitalization, anemia, hypertensive disease, acute coronary syndrome, congestive heart failure, diabetes and renal disease.The performance indicators applied on all cardiac diseases were: C-statistic=0.75, Sensitivity=0.66, Specificity=0.70, Positive predictive value=0.10, Negative predictive value=0.98, Positive likelihood ratio=2.21 and Negative likelihood ratio=0.48. Diseases for discriminative prediction are: stenting, circulatory disorders, acute myocardial infarction and defibrillator and pacemaker implantation. ConclusionsThis study provides a prediction model for 30-day cardiac-related diseases based on available administrative data ready to be integrated as a screening tool. It has reasonable validity and can be used to increase the efficiency of case management.