Electronic Healthcare Databases Research Articles

A Cross-Sectional Study of Unstable Housing and Housing-Related Symptom Content in People With Psychosis Admitted for Inpatient Treatment: A Clinical Record Interactive Search Study.

Rates of psychosis in the homeless population are markedly higher compared to the general population. Understanding potential psychological mechanisms underpinning links between psychosis and homelessness is important for the development of effective care pathways for this highly marginalised group. This study aimed to examine the housing status of a sample of people with psychosis admitted to psychiatric inpatient hospital in one UK mental health trust. We further aimed to examine the presence and relevance of symptoms which were directly related to housing (e.g., persecutory beliefs about neighbours). A cross-sectional study was conducted using an electronic healthcare database, containing anonymised patient records (Clinical Record Interactive Search). Information on housing status, symptoms, and content of symptoms relating to housing were extracted and independently double-coded from clinical notes at the person's admission. 351 service users were in the sample, which covered discharges over a 12-month period (1st April 2021 to 31st March 2022). There was a higher proportion of individuals living without a fixed address in the sample (10%) compared to population estimates (1%-2%). Housing-related thematic content of symptoms was evident in 34% of the sample (e.g., attributing voices to neighbours, feeling under surveillance at home). The only variable significantly associated with housing status was gender, with men at higher odds of unstable housing compared to women. Individuals with concerns directly related to their housing or housing instability were no more likely to have a longer length of admission than those without housing-related concerns. A significant proportion of individuals with psychosis admitted for psychiatric inpatient treatment, experienced housing instability. For some of the sample, symptom content directly related to housing. This raises important considerations for how social needs are assessed on admission to psychiatric hospital. Further research in this area is encouraged.

Submacular Hemorrhage Rates Following Anti-Vascular Endothelial Growth Factor Injections for Exudative Age-Related Macular Degeneration

Objective: To examine rates of submacular hemorrhage in patients undergoing anti-VEGF injections, comparing rates between specific anti-VEGF agents. Design: Retrospective clinical cohort study. Participants: All patients in the database from January 2015-November 2023 with a diagnosis of neovascular age-related macular degeneration (nAMD) and accompanying submacular hemorrhage (SMH). Methods: SMH prevalence and associated anti-VEGF injection type were analyzed in 140,915 eyes (of which 9,107 had SMH) in a nationwide aggregated electronic health care database using chi-square test of proportion. Visual acuity (VA) data was assessed using 2-sample independent t-tests. The primary outcome was rate of SMH per injection type. Secondary datapoints examined were time between SMH diagnosis and last anti-VEGF injection, number of injections before SMH, treatment interval at time of SMH, VA before and at 12 months after SMH, eyes undergoing pars plana vitrectomy (PPV) within 30 days of SMH, and VA before PPV and at 12 months after PPV. Results: The last injection type in eyes with SMH was bevacizumab in 3,439 (37.8%) eyes, brolucizumab-dbII in 46 (0.51%) eyes, aflibercept in 3,221 (35.4%) eyes, ranibizumab in 2,246 (24.7%) eyes, and faricimab-svoa in 155 (1.7%) eyes. Rates of SMH were significantly higher (p≤0.001) for last injection with bevacizumab compared to every other injection type. Rates of SMH were significantly lower (p=0.0004) for last injection with faricimab-svoa compared to every other injection type. Patients receiving either faricimab-svoa or ranibizumab injections each had significantly shorter (mean and standard deviation (SD) 48.9 (27.9), p<0.02; mean and standard deviation (SD) 59.6 (38.2), p=0.003, respectively) mean time between SMH diagnosis and last injection than did patients undergoing any other injection. Mean VA before SMH and at 12 months after SMH did not significantly differ by injection type among all patients. The number of patients who underwent PPV were 52 (1.51%) for bevacizumab, 4 (8.7%) for brolucizumab-dbII, 58 (1.8%) for aflibercept, 41 (1.8%) for ranibizumab, and 3 (1.9%) for faricimab-svoa. Mean VA before SMH and at 12 months after SMH did not significantly differ by injection type in patients undergoing PPV. Conclusions: Faricimab may be more protective than other anti-VEGF injections against SMH in patients with nAMD.

Medication adherence to inhalation therapy and the risk of COPD exacerbations: a systematic review with meta-analysis

BackgroundAssessing medication adherence is crucial in chronic obstructive pulmonary disease (COPD) management to prevent exacerbations. However, it is unclear whether this association between adherence and exacerbations is influenced by the...

Incidence and prevalence of asthma, COPD, and ILD across England, Wales, and South-East Scotland – a harmonised approach to data analysis

ObjectiveThis study aimed to create harmonised datasets for asthma, chronic obstructive pulmonary disease (COPD), and Interstitial Lung Disease (ILD), and subsequently use them to calculate the incidence and prevalence of these diseases across three UK nations (England, Wales, and South-East Scotland). ApproachHarmonised research-ready datasets were created using three electronic healthcare record databases. Records were selected between 01/01/2004 and 31/12/2019, with yearly age and sex standardised incidence and prevalence rates calculated. ResultsIn 2019, the incidence rates for asthma were: England, 0.89 (95%CI: 0.88–0.90); Wales, 0.66 (0.65–0.68); and South-East Scotland, 0.67 (0.64–0.71), which were all lower than in 2004. Incidence rates of COPD in England were 0.83 (0.82–0.85) and 0.67 (0.65–0.69) in Wales, which were both lower than in 2004, but Scotland remained similar at 1.06 (0.99–1.13). Incidence rates of ILD were: England, 3.27 (3.05–3.50); Wales, 1.39 (1.27–1.53); and Scotland, 1.63 (1.36–1.95), which were all higher than in 2004. Between 2004 and 2019, the prevalence of all three diseases increased across all three nations. ConclusionsSolving differences in coding between CPRD, SAIL Databank, and DataLoch databases enabled comparable cohorts to be created and the first harmonised UK estimates of incidence and prevalence for asthma, ILD, and COPD. ImplicationsHarmonised studies of asthma, COPD, and ILD across the UK can now be performed. The code produced is publicly available to facilitate future research, which can be adapted for other conditions and databases.

Effectiveness and safety of empagliflozin: final results from the EMPRISE study.

Limited evidence exists on the comparative safety and effectiveness of empagliflozin against alternative glucose-lowering medications in individuals with type 2 diabetes with the broad spectrum of cardiovascular risk. The EMPagliflozin compaRative effectIveness and SafEty (EMPRISE) cohort study was designed to monitor the safety and effectiveness of empagliflozin periodically for a period of 5 years with data collection from electronic healthcare databases. We identified individuals ≥18 years old with type 2 diabetes who initiated empagliflozin or dipeptidyl peptidase-4 inhibitors (DPP-4i) from 2014 to 2019 using US Medicare and commercial claims databases. After 1:1 propensity score matching using 143 baseline characteristics, we identified four a priori-defined effectiveness outcomes: (1) myocardial infarction (MI) or stroke; (2) hospitalisation for heart failure (HHF); (3) major adverse cardiovascular events (MACE); and (4) cardiovascular mortality or HHF. Safety outcomes included lower-limb amputations, non-vertebral fractures, diabetic ketoacidosis (DKA), acute kidney injury (AKI), severe hypoglycaemia, retinopathy progression, and short-term kidney and bladder cancers. We estimated HRs and rate differences (RDs) per 1000 person-years, overall and stratified by age, sex, baseline atherosclerotic cardiovascular disease (ASCVD) and heart failure. We identified 115,116 matched pairs. Compared with DPP-4i, empagliflozin was associated with lower risks of MI/stroke (HR 0.88 [95% CI 0.81, 0.96]; RD -2.08 [95% CI (-3.26, -0.90]), HHF (HR 0.50 [0.44, 0.56]; RD -5.35 [-6.22, -4.49]), MACE (HR 0.73 [0.62, 0.86]; RD -6.37 [-8.98, -3.77]) and cardiovascular mortality/HHF (HR 0.57 [0.47, 0.69]; RD -10.36 [-12.63, -8.12]). Absolute benefits were larger in older individuals and in those with ASCVD/heart failure. Empagliflozin was associated with an increased risk of DKA (HR 1.78 [1.44, 2.19]; RD 1.59 [1.08, 2.09]); decreased risks of AKI (HR 0.62 [0.54, 0.72]; RD -2.39 [-3.08, -1.71]), hypoglycaemia (HR 0.75 [0.67, 0.84]; RD -2.46 [-3.32, -1.60]) and retinopathy progression (HR 0.78 [0.63, 0.96)]; RD -9.49 [-16.97, -2.10]); and similar risks of other safety events. Empagliflozin relative to DPP-4i was associated with risk reductions of MI or stroke, HHF, MACE and the composite of cardiovascular mortality or HHF. Absolute risk reductions were larger in older individuals and in those who had history of ASCVD or heart failure. Regarding the safety outcomes, empagliflozin was associated with an increased risk of DKA and lower risks of AKI, hypoglycaemia and progression to proliferative retinopathy, with no difference in the short-term risks of lower-extremity amputation, non-vertebral fractures, kidney and renal pelvis cancer, and bladder cancer.

Safety profile of rivaroxaban in first-time users treated for venous thromboembolism in four European countries.

The European rivaroxaban post-authorization safety study evaluated bleeding risk among patients initiated on rivaroxaban or vitamin K antagonists for the treatment and secondary prevention of venous thromboembolism in routine clinical practice. Cohorts were created using electronic healthcare databases from the UK, the Netherlands, Germany and Sweden. Patients with a first prescription of rivaroxaban or vitamin K antagonist during the period from December 2011 (in the UK, January 2012) to December 2017 (in Germany, December 2016) for venous thromboembolism indication, with no record of atrial fibrillation or recent cancer history, were observed until the occurrence of each safety outcome (hospitalization for intracranial, gastrointestinal, urogenital or other bleeding), death or study end (December 2018; in Germany, December 2017). Crude incidence rates of each outcome per 100 person-years were computed. Overall, 44 737 rivaroxaban and 45 842 vitamin K antagonist patients were enrolled, mean age, 59.9-63.8 years. Incidence rates were similar between rivaroxaban and vitamin K antagonist users with some exceptions, including higher incidence rates for gastrointestinal bleeding in rivaroxaban users than in vitamin K antagonist users. Among rivaroxaban users, mortality and bleeding risk generally increased with age, renal impairment and diabetes. This study provides further data from routine clinical practice that broadly support safety profile of rivaroxaban for VTE indication and complement findings from previous randomized clinical trials.

Long-Term Safety of Roflumilast in Patients with Chronic Obstructive Pulmonary Disease, a Multinational Observational Database Cohort Study.

This study evaluated the long-term safety of roflumilast in patients with chronic obstructive pulmonary disease or chronic bronchitis using electronic healthcare databases from Germany, Norway, Sweden, and the United States (US). The study population consisted of patients aged ≥40 years who had been exposed to roflumilast and a matched cohort unexposed to roflumilast. The matching was based on sex, age, calendar year of cohort entry date (2010-2011, 2012, or 2013), and a propensity score that included variables such as demographics, markers of chronic obstructive pulmonary disease (COPD) severity and morbidity, and comorbidities. In comparison to the unexposed matched cohort (never use), three exposure definitions were used for the exposed matched cohort: ever use, use status (current, recent, past use), and cumulative duration of use. The main outcome was 5-year all-cause mortality. Cox regression models were used to estimate crude and adjusted hazard ratios (HRs) and 95% confidence intervals (CI). 112,541 unexposed and 23,239 exposed patients across countries were included. Some variables remained unbalanced after matching, indicating higher COPD disease severity among the exposed patients. Adjusted HRs of 5-year all-cause mortality for "ever use" of roflumilast, compared to "never use", were 1.12 (95% CI, 1.08-1.17) in Germany, 1.00 (95% CI, 0.92-1.08) in Norway, 0.98 (95% CI, 0.92-1.04) in Sweden, and 1.16 (95% CI, 1.12-1.20) in the US. Compared to never users, there was a decrease in 5-year mortality risk observed among "current users" in Germany (HR: 0.93, 95% CI: 0.88-0.98), Norway (HR: 0.77, 95% CI: 0.67-0.87), and Sweden (HR: 0.80, 95% CI: 0.73-0.88). There was no observed increase in 5-year mortality risk with the use of roflumilast in Sweden or Norway. A small increase in 5-year mortality risk was observed in Germany and the US in the ever versus never comparison, likely due to residual confounding by indication.

The impact of urgency of umbilical hernia repair on adverse outcomes in patients with cirrhosis: a population-based cohort study from England

IntroductionUmbilical hernia is common in patients with cirrhosis; however, there is a paucity of dedicated studies on postoperative outcomes in this group of patients. This population-based cohort study aimed to determine the outcomes after emergency and elective umbilical hernia repair in patients with cirrhosis.MethodsTwo linked electronic healthcare databases from England were used to identify all patients undergoing umbilical hernia repair between January 2000 and December 2017. Patients were grouped into those with and without cirrhosis and stratified by severity into compensated and decompensated cirrhosis. Length of stay, readmission, 90-day case fatality rate and the odds ratio of 90-day postoperative mortality were defined using logistic regression.ResultsIn total, 22,163 patients who underwent an umbilical hernia repair were included and 297 (1.34%) had cirrhosis. More patients without cirrhosis had an elective procedure, 86% compared with 51% of those with cirrhosis (P < 0.001). In both the elective and emergency settings, patients with cirrhosis had longer hospital length of stay (elective: 0 vs 1 day, emergency: 2 vs 4 days, P < 0.0001) and higher readmission rates (elective: 4.87% vs 11.33%, emergency:11.39% vs 29.25%, P < 0.0001) than those without cirrhosis. The 90-day case fatality rates were 2% and 0.16% in the elective setting, and 19% and 2.96% in the emergency setting in patients with and without cirrhosis respectively.ConclusionEmergency umbilical hernia repair in patients with cirrhosis is associated with poorer outcomes in terms of length of stay, readmissions and mortality at 90 days.

Increased risk of decompensated heart failure, acute coronary syndrome, arrythmias and ischemic stroke following exacerbation of COPD: results from a multi-database cohort study

Abstract Introduction The risk of acute coronary syndrome (ACS), stroke or death is increased following an exacerbation of chronic obstructive pulmonary disease (COPD). However, little is known about the risk of decompensated heart failure (HF) or acute arrhythmias. Purpose To quantify the risk of decompensated HF, ACS, acute arrhythmias and ischemic stroke following an exacerbation of COPD. Methods Individuals living with COPD were identified in electronic healthcare and claims databases between 2014 and 2018 in Canada (n=142,787), Germany (n=126,795), the Netherlands (NL, n=8,020) and Spain (n=24,393) and followed until a first non-fatal severe CV event of interest (defined as a hospitalisation for ACS, decompensated HF, arrythmias and ischemic stroke), death or censoring. For each non-fatal severe CV event, a multivariable time-dependent Cox model compared the risk of outcome in the 12 months following the onset of a COPD exacerbation (exposed period, split into granular sub-periods of time) with the risk of outcome during periods outside an exacerbation (unexposed period), providing adjusted hazard ratios (HR) and 95% confidence intervals. Results Mean age ranged between 65-68 years old and more than 50% of patients were male, in all countries (Table 1). Cardiovascular risk factors and diseases were common, with 47% to 91% of patients prescribed cardiovascular medications. During follow-up a non-fatal severe CV event occurred in 11% of patients in Canada, 23% in Germany, 8% in NL and 24% in Spain. Decompensated HF was the most frequent event of interest and accounted for 35% to 47% of non-fatal CV events, except in Canada where ACS accounted for 37% of events. In the first 7 days following an exacerbation, the HR ranged between 2.59 (95% CI, 2.32 - 2.89) and 72.34 (64.43 - 81.22) for decompensated HF; 5.59 (5.09 - 6.15) and 24.57 (21.70 - 27.83) for ACS; 9.16 (8.36 - 10.03) and 31.18 (26.84 - 36.21) for arrythmia (mainly driven by atrial fibrillation); and between 2.22 (1.88 - 2.63) and 16.12 (13.39 - 19.41) for ischemic stroke (Table 2). The risks remained elevated for 6 months for decompensated HF and 30 days for ACS in all countries except NL. Conclusion The risk of all severe CV events, including decompensated HF and arrhythmias, was increased following an exacerbation of COPD. Despite heterogeneity of findings across countries, our results indicate that an exacerbation of COPD should be considered as an important risk factor for cardiopulmonary events, emphasizing the need for optimized COPD management across specialties.

Incidence and risk prediction of post-Covid-19 cardiovascular disease in the general population: a multivariable prediction model development and validation study

Abstract Background Previous studies showed that relative risks of various cardiovascular events are increased after Covid-19, even in non-hospitalized individuals. The global burden of post-Covid-19 cardiovascular disease is large and not (yet) hospitalized Covid-19 infected adults may represent a potential target for prevention strategies to mitigate this burden. However, good estimates of individual absolute risks in the general population and risk stratification models are lacking, yet pivotal to assess its impact and direct health care policy. Purpose This study aimed to assess the absolute risk of post-Covid-19 cardiovascular events in the general population, and subsequently explore individual risk prediction in this domain. Methods Covid-19 diagnosed adults were identified from the United Kingdom’s Clinical Practice Research Datalink Aurum, a primary care electronic healthcare database. We assessed 180-day incidence rates of atherosclerotic events (acute coronary syndrome, stroke and transient ischaemic accident), venous thrombo-embolic events (pulmonary embolism and deep venous thrombosis), and newly diagnosed atrial fibrillation and heart failure. Multivariable logistic regression models using elastic net penalization were developed on a cohort of 220,751 adults with Covid-19 before 1 December 2020 (before the start of national vaccination campaigns). Candidate predictors included age, sex, body mass index, smoking, medical history and medication. Three models were fit for each endpoint, with increasing numbers of predictors. All models were validated in (1) 138,034 vaccinated and (2) 503,404 non-vaccinated adults with a first Covid-19 diagnosis after 1 December 2020. We evaluated discriminative and calibration performance upon internal cross-validation and in both validation cohorts. Results Post-Covid-19 cardiovascular event incidences ranged from 13 (venous thromboembolic events) down to 4 (heart failure de novo) per 1000 person-years. Incidences were relatively increased up to 50 days after Covid-19 infection. However, on an absolute scale, individual risks were very low: only less than one in twenty adults with Covid-19 had a predicted 180-day risk larger than 1% and these mainly included individuals aged 60 years or older. Even though the best models had high discriminative performance (AUC &amp;gt; 0.90), rare outcomes complicated predictive performance and calibration. Conclusions Risks of atherosclerotic events, venous thromboembolic events, and newly diagnosed atrial fibrillation and heart failure are increased after Covid-19 infection in the general population, notably within the first 50 days after infection and in individuals aged 60 years or older. However, individual risks are too low to be predictive or helpful to guide individual risk management. Risk mitigation strategies should therefore be aimed at population-based prevention (such as increasing vaccination uptake) and at individual management of Covid-19 hospitalized patients.Graphical abstractPredicted 180-day risk after Covid-19

Characteristics of patients initiating PCSK9i mAb following myocardial infarction and comparability of treatment groups in the Netherlands

Abstract Background The clinical benefits of lipid lowering treatment (LLT) for reducing risk of atherosclerotic cardiovascular disease (ASCVD) are well established. Recent findings support early initiation of proprotein convertase subtilisin/kexin type 9 inhibitors monoclonal antibodies (PCSK9i mAb) to maximize clinical benefit. However, when comparing treatment effects in a real-world setting, it is important to establish comparability of treatment groups and assess and adjust for confounding. Purpose To compare clinical characteristics and LLT use in patients initiating and not initiating PCSK9i mAb within 365 days following a myocardial infarction (MI). Methods MI events occurring in patients ≥18 years between 1 April 2016 and 31 December 2020 were identified in PHARMO Database Network, a population-based network of electronic healthcare databases linking data from general practitioners, out-patient and in-patient pharmacies and hospitals in the Netherlands. The database has about 3.3 million patients, representative of the Dutch population with respect to age and sex albeit with slight over-representation of those &amp;gt;80 years old. Characteristics of patients were presented by treatment groups: initiating PCSK9i mAb and not initiating PCSK9i mAb within 365 days post-MI. Characteristics and LLT use in the two treatment groups was evaluated. Results Of the 28,975 MI events, 157 (0.5%) initiated a PCSK9i mAb within 365 days. Patients who initiated a PCSK9i mAb within 365 days post-MI were younger than those who did not (median age of 63 versus 69 years), more likely to have a history of ASCVD prior to the MI event (56% versus 39%) and dyslipidaemia (55% versus. 28%), while non-initiators were more likely to have diabetes mellitus, chronic kidney disease and atrial fibrillation (Table). Patients who initiated PCSK9i mAb within 365 days post-MI had higher low-density lipoprotein cholesterol levels (3.7 versus 2.9 mmol/L), higher use of ezetimibe (25% versus 4%) and high-intensity statins (15% vs 11%, Figure). Conclusions In the Netherlands, despite being at very high risk of recurrent ASCVD, based on this dataset excluding Patient Support Programs, fewer than 1% of patients initiated a PCSK9i mAb within 365 days post-MI. Further analyses are needed to reveal if these treatment groups are comparable to assess differences and uncontrolled confounders in cardiovascular outcomes in the real world.

Prediction models for in-hospital deaths of patients with COVID-19 using electronic healthcare data

Objective Many models for predicting various disease prognoses have achieved high performance without laboratory test results. However, whether laboratory test results can improve performance remains unclear. This study aimed to investigate whether laboratory test results improve the model performance for coronavirus disease 2019 (COVID-19). Methods Prediction models were developed using data from the electronic healthcare record database in Japan. Patients aged ≥18 years hospitalized for COVID-19 after February 11, 2020, were included. Their age, sex, comorbidities, laboratory test results, and number of days from February 11, 2020, were collected. We developed a logistic regression, XGBOOST, random forest, and neural network analysis and compared the performance with and without laboratory test results. The performance of predicting in-hospital death was evaluated using the area under the curve (AUC). Results Data from 8,288 hospitalized patients (females, 46.5%) were analyzed. The median patient age was 71 years. A total of 6,630 patients were included in the training dataset, and 312 (4.7%) died. In the logistic regression model, the area under the curve was 0.88 (95% confidence interval [CI] = 0.83–0.93) and 0.75 (95% CI = 0.68–0.81) with and without laboratory test results, respectively. The performance was not fundamentally different between the model types, and the laboratory test results improved the performance in all cases. The variables useful for prediction were blood urea nitrogen, albumin, and lactate dehydrogenase. Conclusions Laboratory test results, such as blood urea nitrogen, albumin, and lactate dehydrogenase levels, along with background information, helped estimate the prognosis of patients hospitalized for COVID-19.

The role of oestrogen therapy in reducing risk of Alzheimer's disease: systematic review.

Studies have shown a relationship between oestrogen and Alzheimer's disease. However, there is neither clear nor strong evidence on the use of oestrogen-only therapy in reducing the risk of Alzheimer's disease. To assess the effects of oestrogen-only therapy on reducing the risk of Alzheimer's disease. Inclusion criteria was determined with the PICO framework. Outcome was cognitive function measured by neuropsychological tests and strict protocols. Exclusion criteria included non-Alzheimer's dementia, progesterone-only therapy and pre-menopausal women. Searches were conducted in nine electronic healthcare databases, last searched in July 2022. Quality assessments conducted on randomised controlled trials (RCTs) were performed with the GRADE assessment, and cohort studies and case-control studies were assessed with the Newcastle-Ottawa Scale. Extracted data were used to analyse participants, interventions and outcomes. Twenty-four studies satisfied the search criteria (four RCTs, nine cohort studies, 11 case-control studies). Fifteen studies showed positive associations for oestrogen-only therapy reducing the risk of Alzheimer's disease, and the remaining nine found no evidence of association. Fifteen studies showed that oestrogen-only therapy effectively reduced the risk of Alzheimer's disease, whereas nine showed no correlation. Studies also investigated oestrogen-related variables such as length of oestrogen exposure, being an apolipoprotein E ε4 carrier and concomitant use of non-steroidal anti-inflammatory drugs, and their role in neuroprotection. This review was limited by the limited ranges of duration of oestrogen treatment and type of oestrogen-only therapy used. In conclusion, oestrogen-only therapy has potential for use in preventing Alzheimer's disease, although current evidence is inconclusive and requires further study.

Dynamic relationship among immediate release fentanyl use and cancer incidence: A multivariate time-series analysis using vector autoregressive models

Background A substantial increase in the incidence of immediate release fentanyl (IRF) use was reported in Spain from 2012 to 2017. Purpose This study aimed to investigate the relationship dynamically with cancer incidence in order to provide empirical evidence of inappropriate use of IRF with respect to the pathology. Research design A vector autoregresive (VAR) model was constructed using data from a nationwide electronic healthcare record database in primary care in Spain (BIFAP) according to the following step procedure: (1) split data into training data for modelling and test for validation (2) assessing for time series stationarity; (3) selecting lag-length; (4) building the VAR model; (5) assessing residual autocorrelation; (6) checking stability of the VAR system; (7) evaluating Granger causality; (8) impulse response analysis and forecast error variance decomposition (9) prediction performance with validation data. Results The analysis showed a strong and linear correlation between IRF and cancer (Pearson correlation coefficient: 0.594 (95% CI: 0.420–0.726). Two VAR models, VAR (2) and VAR (11) were selected and compared. All tests performed for both models satisfied assumptions for stability, predictability and accuracy. Granger causality revealed cancer incidence is a good predictor for IRF use. VAR (2) seemed to be slightly more accurate, according to the RMSE of the test data. Conclusions This study demonstrates that using a robust and structured VAR modelling approach, is able to estimate dynamics associations, involving IRF use and cancer incidence.

Outcomes after emergency appendicectomy in patients with liver cirrhosis: a population-based cohort study from England

IntroductionThe mortality risk after appendicectomy in patients with liver cirrhosis is predicted to be higher than in the general population given the associated risk of perioperative bleeding, infections and liver decompensation. This population-based cohort study aimed to determine the 90-day mortality risk following emergency appendicectomy in patients with cirrhosis.MethodsAdult patients undergoing emergency appendicectomy in England between January 2001 and December 2018 were identified from two linked primary and secondary electronic healthcare databases, the clinical practice research datalink and hospital episode statistics data. Length of stay, re-admission, case fatality and the odds ratio of 90-day mortality were calculated for patients with and without cirrhosis, adjusting for age, sex and co-morbidity using logistic regression.ResultsA total of 40,353 patients underwent appendicectomy and of these 75 (0.19%) had cirrhosis. Patients with cirrhosis were more likely to be older (p < 0.0001) and have comorbidities (p < 0.0001). Proportionally, more patients with cirrhosis underwent an open appendicectomy (76%) compared with 64% of those without cirrhosis (p = 0.03). The 90-day case fatality rate was 6.67% in patients with cirrhosis compared with 0.56% in patients without cirrhosis. Patients with cirrhosis had longer hospital length of stay (4 (IQR 3–9) days versus 3 (IQR 2–4) days and higher readmission rates at 90 days (20% vs 11%, p = 0.019). Most importantly, their odds of death at 90 days were 3 times higher than patients without cirrhosis, adjusted odds ratio 3.75 (95% CI 1.35–10.49).ConclusionPatients with cirrhosis have a threefold increased odds of 90-day mortality after emergency appendicectomy compared to those without cirrhosis.

Application of machine learning approaches in predicting clinical outcomes in older adults – a systematic review and meta-analysis

BackgroundMachine learning-based prediction models have the potential to have a considerable positive impact on geriatric care.DesignSystematic review and meta-analyses.ParticipantsOlder adults (≥ 65 years) in any setting.InterventionMachine learning models for predicting clinical outcomes in older adults were evaluated. A random-effects meta-analysis was conducted in two grouped cohorts, where the predictive models were compared based on their performance in predicting mortality i) under and including 6 months ii) over 6 months.Outcome measuresStudies were grouped into two groups by the clinical outcome, and the models were compared based on the area under the receiver operating characteristic curve metric.ResultsThirty-seven studies that satisfied the systematic review criteria were appraised, and eight studies predicting a mortality outcome were included in the meta-analyses. We could only pool studies by mortality as there were inconsistent definitions and sparse data to pool studies for other clinical outcomes. The area under the receiver operating characteristic curve from the meta-analysis yielded a summary estimate of 0.80 (95% CI: 0.76 – 0.84) for mortality within 6 months and 0.81 (95% CI: 0.76 – 0.86) for mortality over 6 months, signifying good discriminatory power.ConclusionThe meta-analysis indicates that machine learning models display good discriminatory power in predicting mortality. However, more large-scale validation studies are necessary. As electronic healthcare databases grow larger and more comprehensive, the available computational power increases and machine learning models become more sophisticated; there should be an effort to integrate these models into a larger research setting to predict various clinical outcomes.

Diagnostic and prognostic performance of the SAFE score in non-alcoholic fatty liver disease.

The steatosis-associated fibrosis estimator (SAFE) score was developed to detect clinically significant liver fibrosis in patients with NAFLD in the United States. We compare the performance of the SAFE score and other non-invasive tests to diagnose liver fibrosis and to correlate the scores with liver-related outcomes in patients with NAFLD in Hong Kong. This was a retrospective cohort study involving two data sets. The first cohort was a biopsy cohort of NAFLD patients (n = 279), and the second was a territory-wide cohort of NAFLD patients (n = 4603) retrieved from a territory-wide electronic healthcare database in Hong Kong. In detecting significant fibrosis, liver stiffness measured by transient elastography had the highest area under the receiver operating characteristic curve (AUROC) (.844), followed by SAFE score (.773). SAFE score had the highest AUROC among blood-based algorithms (.773 vs. .746 for FIB-4, .697 for APRI). Based on cut-off values of SAFE score (0 and 100 points), 854 (18.6%), 1596 (34.6%) and 2153 (46.8%) were in the low-, intermediate- and high-risk groups, respectively, in the territory-wide cohort. Six (.7%), 15 (.9%) and 59 (2.7%) developed liver-related events in those three groups respectively. Among patients who had liver-related events at 5 years, using the high cut-off, SAFE score could predict 84.9% of patients accurately, compared to 40.9% for FIB-4 and 27.2% for APRI. The SAFE score performed well and better than other blood-based markers in diagnosing significant fibrosis and predicting liver-related events in Asian patients with NAFLD.

Artificial intelligence in foot and ankle surgery: current concepts.

The twenty-first century has proven that data are the new gold. Artificial intelligence (AI) driven technologies might potentially change the clinical practice in all medical specialities, including orthopedic surgery. AI has a broad spectrum of subcomponents, including machine learning, which consists of a subdivision called deep learning. AI has the potential to increase healthcare delivery, improve indications and interventions, and minimize errors. In orthopedic surgery. AI supports the surgeon in the evaluation of radiological images, training of surgical residents, and excellent performance of machine-assisted surgery. The AI algorithms improve the administrative and management processes of hospitals and clinics, electronic healthcare databases, monitoring the outcomes, and safety controls. AI models are being developed in nearly all orthopedic subspecialties, including arthroscopy, arthroplasty, tumor, spinal and pediatric surgery. The present study discusses current applications, limitations, and future prospective of AI in foot and ankle surgery.

Trends in Secondary Intraocular Lens Surgery among Vitreoretinal Surgeons

To identify changes in secondary lens techniques over time and to determine common complications of each technique. Retrospective cohort study. All patients in the database from January 2015-December 2021 who underwent secondary lens placement (anterior chamber IOL, scleral fixated IOL, or scleral sutured IOL). Rates of secondary IOL surgery technique were analyzed in 3597 participants in a nationwide aggregated electronic health care database using 2 sample independent T-tests. Rates of post-operative rhegmatogenous retinal detachment (RRD) following secondary IOL surgery were assessed using chi square test of proportion. Post-operative visual acuity was assessed using 2 sample independent T-tests. The primary outcome was change in IOL technique over time. Secondary datapoints examined were the development of post-operative rhegmatogenous retinal detachment (RRD) following secondary IOL surgery, visual acuity changes, the development of endophthalmitis, suture erosion, haptic erosion, or corneal edema following IOL surgery. ACIOL use decreased over the 7-year period from 93% of cases to 36% of cases (p<0.0001), while SFIOL use increased from 3% to 34% (p<0.0001), and SSIOL use increased from 4% to 30% (p<0.0001). Visual acuity increased for each surgical technique (ACIOL: 44.1 vs. 49.2 Early Treatment Diabetic Retinopathy Study (ETDRS) letters, p<0.001; SFIOL: 48.7 vs. 57.6 letters, p<0.001; SSIOL: 51.5 vs. 61.2 letters, p<0.001), with larger visual acuity gains seen in SFIOL and SSIOL use (ACIOL vs. SFIOL, p=0.004; ACIOL vs. SSIOL, p=0.002; SFIOL vs. SSIOL, p=0.64). Average RRD rates did not significantly differ between techniques. Rates of endophthalmitis, haptic erosion, and suture erosion were low and did not significantly differ between techniques. Rates of corneal edema were significantly higher in ACIOL cases (vs. SFIOL, p<0.0001; vs. SSIOL, p<0.0001). Rates of ACIOL implantation performed by vitreoretinal surgeons have decreased over time with more vitreoretinal surgeons electing to place either a SFIOL or SSIOL towards the end of the study period; complication profiles among the three techniques may be similar.

Drug safety signal detection in a regional healthcare database using the tree-based scan statistic and comparison to 3 other mining methods.

To evaluate and compare the relative performance of the tree-based scan statistic (TreeScan) with the crude cohort study, Bayesian confidence propagation neural network (BCPNN) and Gamma Poisson Shrinker (GPS) in detecting statin-related adverse events (AEs) in an electronic healthcare database. Data from a Chinese healthcare database from 2010 to 2016 were evaluated. We identified statin users based on prescription information in their out-/in-patient records, and AEs were defined according to the ICD-10 codes in patients' diagnosis records. TreeScan was applied to detect AE signals related to statin use and was compared with 3 other methods based on sensitivity, specificity, positive predictive value, negative predictive value, accuracy, the Youden index, area under the precision-recall curve and the area under the receiver operating characteristic curve. A total of 224 187 patients were enrolled and divided into 85 758 statin users and 138 429 nonusers. TreeScan generated 29 positive signals, of which 9 were known AEs. The sensitivities of TreeScan, BCPNN and GPS were all 69.2%, which was higher than that of the crude cohort study (46%). The specificity (82.3%), positive predictive value (31.0%), negative predictive value (95.9%), accuracy (81.0%), Youden index (51.5%) and area under the receiver operating characteristic curve (75.8%) of TreeScan were the highest among the 4 methods. TreeScan outperformed the crude cohort, BCPNN and GPS in detecting statin-related AEs in an electronic healthcare database. Therefore, it can be used as a complementary tool for other signal detection methods in drug safety surveillance.