scholarly journals EXTERNAL VALIDATION OF THE IMPROVING PARTIAL RISK ADJUSTMENT IN SURGERY (PRAIS2) MODEL FOR 30-DAY MORTALITY AFTER PEDIATRIC CARDIAC SURGERY

2020 ◽  
Author(s):  
Lucia Cocomello ◽  
Massimo Caputo ◽  
Rosie Cornish ◽  
Deborah A. Lawlor
Keyword(s):  
Cardiac Surgery ◽  
Clinical Practice ◽  
Risk Adjustment ◽  
Heart Surgery ◽  
Risk Model ◽  
External Validation ◽  
Risk Function ◽  
P Value ◽  
List Type ◽  
Single Centre

ABSTRACTObjectiveRisk stratification in paediatric patients undergoing heart surgery remains a challenge. The improving partial risk adjustment in surgery (PRAIS2) is a risk model predicting 30-day mortality which has been recently developed and validated using a UK-based cohort from April 2009-March 2015. We aimed to perform an independent temporal external validation to explore its generalisability and clinical utility.MethodsPRAIS2 validation was carried out using a single centre (Bristol, UK) cohort from April 2004 to March 2009 and April 2015 to July 2019. For each subject PRAIS2 score was calculated according to the original formula. PRAIS2 performance was assessed in terms of discrimination by means of ROC curve analysis and calibration by using the calibration belt method.ResultsA total of 1330 (2004-2009) and 1187 (2015-2019) paediatric cardiac surgical procedures were included in the first and second independent validation, respectively (median age at the procedure 6.0 and 6.9 months). PRAIS2 score showed excellent discrimination for both independent validations (AUC 0.72 (95%CI: 0.65 to 0.80) and 0.87 (95%CI: 0.82 to 0.93), respectively). While PRAIS2 was only marginally calibrated in the first validation, with a tendency to underestimate risk P-value = 0.051), the second validation showed good calibration with 95% confidence belt containing the bisector for predicted mortality (P-value = 0.15); We also observed good performance in the subgroup of patients undergoing non-elective procedures (N = 482; AUC 0.78 (95%CI 0.68 to 0.87); Calibration belt containing the bisector (P-value=0.61).ConclusionsIn a single centre UK-based cohort, PRAIS2 showed excellent discrimination and calibration in predicting 30-day mortality in paediatric cardiac surgery including in those undergoing non-elective procedures. Our results support a wider adoption of PRAIS2 score in the clinical practice.Strengths and limitations of this studyA strength of the present study is that data were prospectively collected as part of the UK National Congenital Heart Disease Audit and as such they undergo continuous and inclusive systematic validation that includes the review of a sample of case notes by external auditors to ensure coding accuracy.We used a recently proposed method (calibration belt) which does not require patients to be categorised according to risk percentile but rather provides a risk function across all risk value with relative uncertainty measure (95% CI)A key limitation of this study is that the sample size is relatively small and considerably smaller than the cohort used to develop PRAIS2Key questionsWhat is already known about this subject? The improving partial risk adjustment in surgery (PRAIS2) is a risk model predicting 30-day mortality which has been recently developed and validated using a UK-wide cohort.What does this study add? The present study reported the first independent external validation of the PRAIS2 using a single centre cohort which confirmed excellent performance of the model and for the first time showed that it also accurately predicts mortality in patients undergoing non-elective proceduresHow might this impact on clinical practice? Our results support a wider adoption of the PRAIS2 in the clinical practice.

scholarly journals External validation of the improving partial risk adjustment in surgery (PRAIS-2) model for 30-day mortality after paediatric cardiac surgery

BMJ Open ◽  
2020 ◽  
Vol 10 (11) ◽  
pp. e039236
Author(s):  
Lucia Cocomello ◽  
Massimo Caputo ◽  
Rosie Cornish ◽  
Deborah Lawlor
Keyword(s):  
Cardiac Surgery ◽  
Risk Adjustment ◽  
Characteristic Curve ◽  
External Validation ◽  
Secondary Outcome ◽  
P Value ◽  
Good Prediction ◽  
Single Centre ◽  
Paediatric Cardiac Surgery ◽  
Good Prediction Accuracy

ObjectiveIndependent temporal external validation of the improving partial risk adjustment in surgery model (PRAIS-2) to predict 30-day mortality in patients undergoing paediatric cardiac surgery.DesignRetrospective analysis of prospectively collected data.SettingPaediatric cardiac surgery.InterventionPRAIS-2 validation was carried out using a two temporally different single centre (Bristol, UK) cohorts: Cohort 1 surgery undertaken from April 2004 to March 2009 and Cohort 2 from April 2015 to July 2019. For each subject PRAIS-2 score was calculated according to the original formula.ParticipantsA total of 1352 (2004-2009) and 1197 (2015-2019) paediatric cardiac surgical procedures were included in the Cohort 1 and Cohort 2, respectively (median age at the procedure 6.3 and 7.1 months).Primary and secondary outcome measuresPRAIS-2 performance was assessed in terms of discrimination by means of ROC (receiver operating characteristic) curve analysis and calibration by using the calibration belt method.ResultsPRAIS-2 score showed excellent discrimination for both cohorts (AUC 0.72 (95%CI: 0.65 to 0.80) and 0.88 (95%CI: 0.82 to 0.93), respectively). While PRAIS-2 was only marginally calibrated in Cohort 1, with a tendency to underestimate risk in lowrisk and overestimate risk in high risk procedures (P-value = 0.033), validation in Cohort 2 showed good calibration with the 95% confidence belt containing the bisector for predicted mortality (P-value = 0.143). We also observed good prediction accuracy in the non-elective procedures (N = 483;AUC 0.78 (95%CI 0.68 to 0.87); Calibration belt containing the bisector (P-value=0.589).ConclusionsIn a single centre UK-based cohort, PRAIS-2 showed excellent discrimination and calibration in predicting 30-day mortality in paediatric cardiac surgery including in those undergoing non-elective procedures. Our results support a wider adoption of PRAIS-2 score in the clinical practice.

scholarly journals Incidence of Permanent Pacemaker Implantation after Cardiac Surgery: A Single Centre Experience

2019 ◽  
Vol 73 (4) ◽  
pp. 364-367
Author(s):  
Mārtiņš Kalējs ◽  
Edgars Prozorovskis ◽  
Kaspars Kupics ◽  
Ivars Brečs ◽  
Uldis Strazdiņš ◽  
...  
Keyword(s):  
Cardiac Surgery ◽  
Aortic Valve ◽  
Heart Surgery ◽  
Open Heart Surgery ◽  
Pacemaker Implantation ◽  
Intervention Group ◽  
Permanent Pacemaker ◽  
Open Heart ◽  
Single Centre ◽  
Permanent Pacemaker Implantation

Abstract Permanent pacemaker implantation (PPI) after open heart surgery is required in 0.4–8.5% of patients. The aim of our study was to determine the incidence of PPI after cardiac surgery at Pauls Stradiņš Clinical University Hospital and to assess its influence on intrahospital outcomes. This was a single-centre retrospective study. We reviewed all patients who underwent either open heart surgery or transcatheter aortic valve implantation (TAVI) between the years 2015 and 2017. Included were all patients with PPI postoperatively before discharge. We compared the patient demographics, and perioperative state, incidence of PPI and intrahospital stay among groups. After cardiac surgery a total of 135 (4.2%) patients received a PPI. The PPI incidence was highest in the tricuspid valve intervention group — 8.8% followed by aortic valve replacement (AVR) patients with 3.3%. After TAVI incidence of PPI was 4.0% after Sapien valve and 8% after CoreValve implantations, respectively. Incidence of PPI after TAVI with the Sapien valve was not significantly higher when compared to conventional AVR, but it was significantly higher after TAVI with CoreValve. Regardless of the initial procedure a need for PPI significantly increased the total length of hospital stay.

A retrospective study of perioperative clinical seizures and epilepsy in children after operation for CHD

2021 ◽  
pp. 1-7
Author(s):  
Takeshi Ikegawa ◽  
Shin Ono ◽  
Kouji Yamamoto ◽  
Mikihiro Shimizu ◽  
Sadamitsu Yanagi ◽  
...  
Keyword(s):  
Risk Factors ◽  
Cardiac Surgery ◽  
White Matter ◽  
Risk Adjustment ◽  
Congenital Heart ◽  
Heart Surgery ◽  
Congenital Heart Surgery ◽  
White Matter Integrity ◽  
Partial Regression Coefficient ◽  
Clinical Seizure

Abstract This study investigated the incidence and risk factors of perioperative clinical seizure and epilepsy in children after operation for CHD. We included 777 consecutive children who underwent operation from January 2013 to December 2016 at Kanagawa Children’s Medical Center, Kanagawa, Japan. Perinatal, perioperative, and follow-up medical data were collected. Elastic net regression and mediation analysis were performed to investigate risk factors of perioperative clinical seizure and epilepsy. Anatomic CHD classification was performed based on the preoperative echocardiograms; cardiac surgery was evaluated using Risk Adjustment in Congenital Heart Surgery 1. Twenty-three (3.0%) and 15 (1.9%) patients experienced perioperative clinical seizure and epilepsy, respectively. Partial regression coefficient with epilepsy as the objective variable for anatomical CHD classification, Risk Adjustment in Congenital Heart Surgery 1, and the number of surgeries was 0.367, 0.014, and 0.142, respectively. The proportion of indirect effects on epilepsy via perioperative clinical seizure was 22.0, 21.0, and 33.0%, respectively. The 15 patients with epilepsy included eight cases with cerebral infarction, two cases with cerebral haemorrhage, and three cases with hypoxic-ischaemic encephalopathy; white matter integrity was not found. Anatomical complexity of CHD, high-risk cardiac surgery, and multiple cardiac surgeries were identified as potential risk factors for developing epilepsy, with a low rate of indirect involvement via perioperative clinical seizure and a high rate of direct involvement independently of perioperative clinical seizure. Unlike white matter integrity, stroke and hypoxic-ischaemic encephalopathy were identified as potential factors for developing epilepsy.

Abstract 3044: The Preoperative Utilization of Levosimendan (Preoperative Optimization) Reduced Mortality and Low Output Syndrome After Cardiac Surgery in Patients With Low Ejection Fraction (EF<25%).

Circulation ◽  
2007 ◽  
Vol 116 (suppl_16) ◽  
Author(s):  
Ricardo L Levin ◽  
Marcela A Degrange ◽  
Rafael Porcile ◽  
Flavio Salvagio ◽  
Norberto Blanco ◽  
...  
Keyword(s):  
Cardiac Surgery ◽  
Ejection Fraction ◽  
Heart Surgery ◽  
Open Heart Surgery ◽  
Control Group ◽  
Placebo Control ◽  
P Value ◽  
Open Heart ◽  
Preoperative Optimization ◽  
Low Ejection Fraction

Background: Patients with low ejection fraction (EF<25%) present high-risk of mortality and development of low output state (LOS) after cardiac surgery. The objective of this research was to evaluate the preoperative use of the calcium sensitizer Levosimendan (Levo) in patients with EF<25%, underwent open heart surgery. Thirty-day mortality and development of postoperative LOS were the primary end-points of the study. Methods: Patients with EF<25% and hemodynamic parameters of LOS (cardiac index<2.2 L/min/m2 and pulmonary artery occlusion pressure>15 mm Hg), underwent coronary bypass surgery between 12/01/2002 and 02/01/2007 were randomized to: preoperative infusion of Levo 0.1 mcg/Kg/min, 24 hours before surgery (Levo group-preoperative optimization), or placebo (Control group). LOS postoperative was defined for the same hemodynamic variables. A P value < 0.05 was considered significant Results: Two-hundred and twenty one patients fulfilled the inclusion criteria, being randomized 111 of them to Levo, and 110 patients to placebo. Both groups were comparable in their general and surgical characteristics. No withdrawal of Levo was required during the preoperative administration, with 8 patients showing hypotension episodes which was resolved with fluid infusions. There were not ventricular arrhythmias, supraventricular arrhythmias (with heart rate over 125) or preoperative ischemic events. The 30-day mortality was 3 patients in the Levo group (2.7%) versus 12 patients in the Control group (10.9%, P value 0.001, OR 0.23, IC95 0.05– 0.90). Seven patients in the Levo group developed postoperative LOS (6.3%) against 20 patients in the Control group (18.2%, P value <0.001, OR 0.30, IC95 0.11– 0.80) Conclusion: The preoperative optimization with Levosimendan reduced the operative mortality and the development of postoperative LOS in patients with EF<25% underwent open heart surgery. The infusion was safety no needing to withdraw it in any case. These findings could represent a new strategy to reduce the operative risk in this group of patients.:

Clinical response to arginine vasopressin therapy after paediatric cardiac surgery

2012 ◽  
Vol 23 (3) ◽  
pp. 387-393 ◽  
Author(s):  
Christopher W. Mastropietro ◽  
Maria C. Davalos ◽  
Shivaprakash Seshadri ◽  
Henry L. Walters ◽  
Ralph E. Delius
Keyword(s):  
Blood Pressure ◽  
Cardiac Surgery ◽  
Arginine Vasopressin ◽  
Heart Surgery ◽  
Open Heart Surgery ◽  
P Value ◽  
Haemodynamic Instability ◽  
Post Operation ◽  
The Mean ◽  
Haemodynamic Improvement

AbstractObjectiveTo describe the haemodynamic response of children who receive arginine vasopressin for haemodynamic instability after cardiac surgery and to identify clinical variables associated with a favourable response.Materials and MethodsWe reviewed patients less than or equal to 6 years undergoing open heart surgery in our institution between January, 2009 and July, 2010 who received arginine vasopressin during the first 7 days post operation. Favourable responders were defined as those in whom blood pressure was increased or maintained and catecholamine score was decreased, or blood pressure was increased by greater than or equal to 10% of baseline and catecholamine score was unchanged at 6 hours following arginine vasopressin initiation.ResultsOf the 34 patients identified, 17 (50%) patients responded favourably to arginine vasopressin. At 6 hours, the mean blood pressure was increased by 32.2% in responders as compared with 4.6% in non-responders, with a p-value less than 0.001. The mean catecholamine score decreased by 30.1% in responders and increased by 7.6% in non-responders, with a p-value less than 0.001. Anthropometric, demographic, and intra-operative variables were similar in both groups, as was maximum dose of arginine vasopressin. The median time after arrival to the intensive care unit at which arginine vasopressin was initiated, however, was later in those who responded, 20 hours as compared with those who did not, 6 hours, with a p-value equal to 0.032.ConclusionsArginine vasopressin therapy led to haemodynamic improvement in only half of the children in this study, and improvement was more likely to occur if arginine vasopressin was initiated after the post-operative night.

Stratification of complexity: The Risk Adjustment for Congenital Heart Surgery-1 Method and The Aristotle Complexity Score – past, present, and future

2008 ◽  
Vol 18 (S2) ◽  
pp. 163-168 ◽  
Author(s):  
Marshall Lewis Jacobs ◽  
Jeffrey Phillip Jacobs ◽  
Kathy J. Jenkins ◽  
Kimberlee Gauvreau ◽  
David R. Clarke ◽  
...  
Keyword(s):  
Cardiac Surgery ◽  
Quality Of Care ◽  
Risk Adjustment ◽  
Congenital Heart ◽  
Heart Surgery ◽  
Congenital Heart Surgery ◽  
Case Mix ◽  
Objective Data ◽  
Complexity Score

AbstractMeaningful evaluation of quality of care must account for variations in the population of patients receiving treatment, or “case-mix”. In adult cardiac surgery, empirical clinical data, initially from tens of thousands, and more recently hundreds of thousands of operations, have been used to develop risk-models, to increase the accuracy with which the outcome of a given procedure on a given patient can be predicted, and to compare outcomes on non-identical patient groups between centres, surgeons and eras.In the adult cardiac database of The Society of Thoracic Surgeons, algorithms for risk-adjustment are based on over 1.5 million patients undergoing isolated coronary artery bypass grafting and over 100,000 patients undergoing isolated replacement of the aortic valve or mitral valve. In the pediatric and congenital cardiac database of The Society of Thoracic Surgeons, 61,014 operations are spread out over greater than 100 types of primary procedures. The problem of evaluating quality of care in the management of pediatric patients with cardiac diseases is very different, and in some ways a great deal more challenging, because of the smaller number of patients and the higher number of types of operations.In the field of pediatric cardiac surgery, the importance of the quantitation of the complexity of operations centers on the fact that outcomes analysis using raw measurements of mortality, without adjustment for complexity, is inadequate. Case-mix can vary greatly from program to program. Without stratification of complexity, the analysis of outcomes for congenital cardiac surgery will be flawed. Two major multi-institutional efforts have attempted to measure the complexity of pediatric cardiac operations: the Risk Adjustment in Congenital Heart Surgery-1 method and the Aristotle Complexity Score. Both systems were derived in large part from subjective probability, or expert opinion. Both systems are currently in wide use throughout the world and have been shown to correlate reasonably well with outcome.Efforts are underway to develop the next generation of these systems. The next generation will be based more on objective data, but will continue to utilize subjective probability where objective data is lacking. A goal, going forward, is to re-evaluate and further refine these tools so that, they can be, to a greater extent, derived from empirical data. During this process, ideally, the mortality elements of both the Aristotle Complexity Score and the Risk Adjustment in Congenital Heart Surgery-1 methodology will eventually unify and become one and the same. This review article examines these two systems of stratification of complexity and reviews the rationale for the development of each system, the current use of each system, the plans for future enhancement of each system, and the potential for unification of these two tools.

P5709External validation of the ACEF II operative risk model in a cardiac surgery population: an interim evaluation

2019 ◽  
Vol 40 (Supplement_1) ◽  
Author(s):  
M Georgievska ◽  
R Saiti ◽  
D Popevski ◽  
T Gramosli ◽  
E Stoicovski ◽  
...  
Keyword(s):  
Mechanical Ventilation ◽  
Cardiac Surgery ◽  
High Risk ◽  
Mortality Risk ◽  
Risk Model ◽  
Prolonged Mechanical Ventilation ◽  
P Value ◽  
Medium Risk ◽  
Risk Patients ◽  
Mechanical Ventilation Time

Abstract Background The ACEF II score has been proposed as a parsimonious, alternative, operative mortality risk prediction model for cardiac surgery. External validation is warranted to establish its use. Aim The primary goal was to evaluate the ACEF II model performance for cardiac surgery mortality risk stratification. We also tested the discriminatory power to classify patients in need of prolonged postoperative respiratory support and hospitalisation. Methods We evaluated 743 Cardiac Surgery patients – median age 65 (range 20–80 years), 27.4% females - operated between November 2017 and October 2018. Receiver Operating Curves (ROC) were generated based on a dichotomous outcome, “yes/no”, for intrahospital mortality, prolonged mechanical ventilation time (>24 hours), ICU length-of-stay (>48 hours) and postoperative hospitalisation (>7 days). The ACEF II was compared to the ACEF I and the EuroSCORE II (ESII). The DeLong method was used to test the statistical significance of the difference between the areas under different dependent ROC curves. Results The median ACEF II scores for low risk (= ESII <2%), medium risk (= ESII ≥2–≤5%) and high-risk patients (= ESII >5%) were 1.24 (IQR 1.05–1.505), 1.48 (IQR 1.28–1.928) and 2.240 (IQR 1.560–2.933), respectively. The observed mortality for low risk, medium risk and high-risk patients were 1.48% (5/337), 3.26% (9/275) and 19.23% (25/130), respectively. ACEF II outperformed the ACEF I but was similar to the EuroSCORE II in discriminating intrahospital mortality cases and patients in need of prolonged mechanical ventilation (Table 1). All risk models lacked sufficient power to classify patients requiring prolonged ICU-LOS and postoperative hospitalisation time (AUC <0.7). Table 1. Pairwise comparison of ROC Risk Score Model AUC + CI95% – Intrahospital Mortality Area difference when compared to ACEF II AUC + CI95% p-value ACEF II 0.766 [0.733 to 0.796] ACEF I 0.645 [0.609 to 0.679] 0.121 [0.0288 to 0.212] 0.0100 EuroSCORE II 0.809 [0.778 to 0.836] 0.0429 [−0.0431 to 0.129] 0.3284 AUC + CI95% – Prolonged MVT Area difference when compared to ACEF II AUC + CI95% p-value ACEF II 0.721 [0.687 to 0.753] ACEF I 0.632 [0.596 to 0.667] 0.0891 [0.0224 to 0.156] 0.0088 EuroSCORE II 0.721 [0.687 to 0.753] 0.000128 [−0.0732 to 0.0735] 0.9973 AUC = Area Under the Curve, DeLong et al., 1988 – Binomial exact CI95% for the AUC, MVT = Mechanical Ventilation time. Conclusion The ACEF II risk model has a fair discriminative capacity to classify intrahospital mortality cases and patients who will require prolonged mechanical respiratory support following cardiac surgery. Acknowledgement/Funding None

scholarly journals Improving risk adjustment in the PRAiS (Partial Risk Adjustment in Surgery) model for mortality after paediatric cardiac surgery and improving public understanding of its use in monitoring outcomes

2017 ◽  
Vol 5 (23) ◽  
pp. 1-164 ◽  
Author(s):  
Christina Pagel ◽  
Libby Rogers ◽  
Katherine Brown ◽  
Gareth Ambler ◽  
David Anderson ◽  
...  
Keyword(s):  
Congenital Heart Disease ◽  
Heart Disease ◽  
Risk Adjustment ◽  
Congenital Heart ◽  
Heart Surgery ◽  
Cross Validation ◽  
Risk Model ◽  
Final Model ◽  
The Public ◽  
Validation Set

BackgroundIn 2011, we developed a risk model for 30-day mortality after children’s heart surgery. The PRAiS (Partial Risk Adjustment in Surgery) model uses data on the procedure performed, diagnosis, age, weight and comorbidity. Our treatment of comorbidity was simplistic because of data quality. Software that implements PRAiS is used by the National Congenital Heart Disease Audit (NCHDA) in its audit work. The use of PRAiS triggered the temporary suspension of surgery at one unit in 2013. The public anger that surrounded this illustrated the need for public resources around outcomes monitoring.Objectives(1) To improve the PRAiS risk model by incorporating more information about comorbidities. (2) To develop online resources for the public to help them to understand published mortality data.DesignObjective 1 The outcome measure was death within 30 days of the start of each surgical episode of care. The analysts worked with an expert panel of clinical and data management representatives. Model development followed an iterative process of clinical discussion of risk factors, development of regression models and assessment of model performance under cross-validation. Performance was measured using the area under the receiving operator characteristic (AUROC) curve and calibration in the cross-validation test sets. The final model was further assessed in a 2014–15 validation data set.Objective 2 We developed draft website material that we iteratively tested through four sets of two workshops (one workshop for parents of children who had undergone heart surgery and one workshop for other interested users). Each workshop recruited new participants. The academic psychologists ran two sets of three experiments to explore further understanding of the web content.DataWe used pseudonymised NCHDA data from April 2009 to April 2014. We later unexpectedly received a further year of data (2014–15), which became a prospective validation set.ResultsObjective 1The cleaned 2009–14 data comprised 21,838 30-day surgical episodes, with 539 deaths. The 2014–15 data contained 4207 episodes, with 97 deaths. The final regression model included four new comorbidity groupings. Under cross-validation, the model had a median AUROC curve of 0.83 (total range 0.82 to 0.83), a median calibration slope of 0.92 (total range 0.64 to 1.25) and a median intercept of –0.23 (range –1.08 to 0.85). In the validation set, the AUROC curve was 0.86 [95% confidence interval (CI) 0.83 to 0.89], and its calibration slope and intercept were 1.01 (95% CI 0.83 to 1.18) and 0.11 (95% CI –0.45 to 0.67), respectively. We recalibrated the final model on 2009–15 data and updated the PRAiS software.Objective 2We coproduced a website (http://childrensheartsurgery.info/) that provides interactive exploration of the data, two animations and background information. It was launched in June 2016 and was very well received.LimitationsWe needed to use discharge status as a proxy for 30-day life status for the 14% of overseas patients without a NHS number. We did not have sufficient time or resources to extensively test the usability and take-up of the website following its launch.ConclusionsThe project successfully achieved its stated aims. A key theme throughout has been the importance of collaboration and coproduction. In particular for aim 2, we generated a great deal of generalisable learning about how to communicate complex clinical and mathematical information.Further workExtending our codevelopment approach to cover many other aspects of quality measurement across congenital heart disease and other specialised NHS services.FundingThe National Institute for Health Research Health Services and Delivery Research programme.

scholarly journals Development of an accessible 10-year Digital CArdioVAscular (DiCAVA) risk assessment: a UK Biobank study

2021 ◽  
Author(s):  
Nikola Dolezalova ◽  
Angus B Reed ◽  
Aleksa Despotovic ◽  
Bernard Dillon Obika ◽  
Davide Morelli ◽  
...  
Keyword(s):  
Clinical Practice ◽  
Proportional Hazards ◽  
Risk Model ◽  
External Validation ◽  
Cox Proportional Hazards ◽  
Machine Learning Techniques ◽  
Uk Biobank ◽  
Framingham Score ◽  
Cvd Risk ◽  
Remote Setting

Abstract Background Cardiovascular diseases (CVDs) are among the leading causes of death worldwide. Predictive scores providing personalised risk of developing CVD are increasingly used in clinical practice. Most scores, however, utilise a homogenous set of features and require the presence of a physician. Objective The aim was to develop a new risk model (DiCAVA) using statistical and machine learning techniques that could be applied in a remote setting. A secondary goal was to identify new patient-centric variables that could be incorporated into CVD risk assessments. Methods Across 466,052 participants, Cox proportional hazards (CPH) and DeepSurv models were trained using 608 variables derived from the UK Biobank to investigate the 10-year risk of developing a CVD. Data-driven feature selection reduced the number of features to 47, after which reduced models were trained. Both models were compared to the Framingham score. Results The reduced CPH model achieved a c-index of 0.7443, whereas DeepSurv achieved a c-index of 0.7446. Both CPH and DeepSurv were superior in determining the CVD risk compared to Framingham score. Minimal difference was observed when cholesterol and blood pressure were excluded from the models (CPH: 0.741, DeepSurv: 0.739). The models show very good calibration and discrimination on the test data. Conclusion We developed a cardiovascular risk model that has very good predictive capacity and encompasses new variables. The score could be incorporated into clinical practice and utilised in a remote setting, without the need of including cholesterol. Future studies will focus on external validation across heterogeneous samples.

Sign in / Sign up

Export Citation Format

Share Document