Predicting Hospital Readmission in Heart Failure Patients in Iran: A Comparison of Various Machine Learning Methods / 대한의료정보학회지

Objectives@#Heart failure (HF) is a common disease with a high hospital readmission rate. This study considered class imbalance and missing data, which are two common issues in medical data. The current study’s main goal was to compare the performance of six machine learning (ML) methods for predicting hospital readmission in HF patients. @*Methods@#In this retrospective cohort study, information of 1,856 HF patients was analyzed. These patients were hospitalized in Farshchian Heart Center in Hamadan Province in Western Iran, from October 2015 to July 2019. The support vector machine (SVM), least-square SVM (LS-SVM), bagging, random forest (RF), AdaBoost, and naïve Bayes (NB) methods were used to predict hospital readmission. These methods’ performance was evaluated using sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. Two imputation methods were also used to deal with missing data. @*Results@#Of the 1,856 HF patients, 29.9% had at least one hospital readmission. Among the ML methods, LS-SVM performed the worst, with accuracy in the range of 0.57–0.60, while RF performed the best, with the highest accuracy (range, 0.90–0.91). Other ML methods showed relatively good performance, with accuracy exceeding 0.84 in the test datasets. Furthermore, the performance of the SVM and LS-SVM methods in terms of accuracy was higher with the multiple imputation method than with the median imputation method. @*Conclusions@#This study showed that RF performed better, in terms of accuracy, than other methods for predicting hospital readmission in HF patients.

Mean platelet volume as a predictor of one-year major adverse cardiac events following elective percutaneous coronary interventions

Mean platelet volume [MPV] correlates with platelet activity. The relation between MPV and long-term outcome in patients undergoing percutaneous coronary intervention [PCI] has been investigated in several studies. The aim of the present study was to investigate the utility of MPV in prognosticating the long-term outcome after elective PCI. The study cohort included 2627 patients undergoing elective PCI between September 2008 and June 2010, whose baseline MPV measurements before PCI were available. The patients were divided into three groups of MPV < 9.1 fL, MPV = 9.1 to 10 fL, and MPV > 10 fL, and they were assessed for developing major adverse cardiac events [MACE], comprising death, myocardial infarction [MI], target vessel revascularization [TVR], and target lesion revascularization [TLR] over a one-year follow-up. Of 2539 patients, major adverse cardiac events [MACE] at one year occurred in 77 [3.0%] patients, including mortality in 26 [1.0%]. The patients in the highest tertile [MPV > 10 fL] had no increased frequency of MACE compared to those in the mid [9.1 to 10 fL] and lowest [< 9.1 fL] tertiles [3.3%, 2.2%, and 3.8%, respectively; p value = 0.14]. No significant differences were found for each of the primary endpoints among the MPV tertiles. In multivariate logistic regression, we investigated the association between high MPV and total MACE [OR = 1.10, 95%CI: 0.69-1.77; p value = 0.68], death [OR = 1.14, 95%CI: 0.51-2.54; p value = 0.74], and non-fatal MI [OR = 1.85, 95%CI: 0.73-4.67; p value = 0.19] at one year's follow-up but MPV did not remain in the model in any of the cases. In the diabetic patients, the one-way analysis of variance demonstrated that mortality was 1.6% [4 patients] in the highest tertile, 0.8% [2 patients] in the mid tertile, and 0.5% [one patient] in the lowest tertile. There was no direct correlation between pre-procedural MPV and MACE in elective PCI. MPV can only be considered as an appropriate factor for predicting mortality in diabetic patients undergoing elective PCI

Advanced cardiac life support training by problem-based method: effect on the trainee's skills, knowledge and evaluation of trainers

Cardiopulmonary-cerebral resuscitation [CPCR] training is essential for all hospital workers, especially junior residents who might become the manager of the resuscitation team. In our center, the traditional CPCR knowledge training curriculum for junior residents up to 5 years ago was lecture-based and had some faults. This study aimed to evaluate the effect of a problem-based method on residents' CPCR knowledge and skills as well as their evaluation of their CPCR trainers. This study, conducted at Tehran University of Medical Sciences, included 290 first-year residents in 2009-2010 - who were trained via a problem-based method [the problem-based group] - and 160 first-year residents in 2003-2004 - who were trained via a lecture-based method [the lecture-based group]. Other educational techniques and facilities were similar. The participants self-evaluated their own CPCR knowledge and skills pre and post workshop and also assessed their trainers' efficacy post workshop by completing special questionnaires. The problem-based group, trained via the problem-based method, had higher self-assessment scores of CPCR knowledge and skills post workshop: the difference as regards the mean scores between the problem-based and lecture-based groups was 32.36 +/- 19.23 vs. 22.33 +/- 20.35 for knowledge [p value = 0.003] and 10.13 +/- 7.17 vs. 8.19 +/- 8.45 for skills [p value = 0.043]. The residents' evaluation of their trainers was similar between the two study groups [p value = 0.193], with the mean scores being 15.90 +/- 2.59 and 15.46 +/- 2.90 in the problem-based and lecture-based groups, respectively. The problem-based method increased our residents' self-evaluation score of their own CPCR knowledge and skills

Impact of side-branch flow in coronary bifurcation intervention

The optimal strategy in percutaneous coronary intervention [PCI] for coronary artery bifurcation lesions has yet to be agreed upon. We compared a strategy for stenting the main vessel to provide a complete perfusion flow in the side branch, namely thrombolysis in myocardial infarction [TIMI] - III, with a strategy for intervention in both the main vessel and the side branch [MV + SB]. This retrospective study utilized data on 258 consecutive patients with bifurcation lesions scheduled for PCI at Tehran Heart Center between March 2003 and March 2008. The patients were followed up for 12 months, and the primary end point was a major adverse cardiac event [MACE], i.e. cardiac death, myocardial infarction, target-vessel revascularization, and target-lesion revascularization during the 12-month follow-up period. A total of 52.7% of the patients underwent PCI on the main vessel of the bifurcation lesions [MV group] and 47.3% with a similar lesion type received a percutaneous intervention on both the main vessel and the side branch [MV + SB group]. The total rate of MACE during the follow-up was 4.3% [11 patients]; the rate was not significantly different between the MV and MV + SB groups [3.7% vs. 4.9%, respectively; p value = 0.622]. There was no association between MACE in performing a simple or complex interventional strategy to treat coronary bifurcation lesions when drawing the TIMI- III flow as a goal in a simple technique.

Effectiveness of two-year versus one-year use of dual antiplatelet therapy in reducing the risk of very late stent thrombosis after drug-eluting stent implantation

Impact of 12 months' versus 24 months' use of dual antiplatelet therapy on the prevalence of stent thrombosis in patients undergoing percutaneous coronary intervention [PCI] with the drug-eluting stent [DES] is not clear. As a result, duration of dual antiplatelet therapy is still under debate among interventionists. From March 2007 until August 2008, all consecutive patients with successful PCI who received at least one DES and were treated with dual antiplatelet therapy [Clopidogrel + Aspirin] were included. All the patients were followed up for more than 24 months [mean = 35.27 +/- 6.91 months] and surveyed for very late stent thrombosis and major cardiovascular events. From 961 patients eligible for the study, 399 [42%] discontinued Clopidogrel after 12 months and 562 [58%] continued Clopidogrel for 24 months. The clinical and procedural variables were compared between the two groups. In the 12 months' use group, two cases of definite thrombosis occurred at 18 and 13 months post PCI. In the 24 months' use group, 2 cases of definite thrombosis occurred at 14 and 28 months post PCI, one of them with stenting in a bifurcation lesion. Five cases of probable stent thrombosis were detected at 21, 28, 32, 33, and 34 months after the procedure. It is of note that amongst the 10 cases of stent thrombosis, only 1[10%] thrombosis occurred when the patient was on Clopidogrel and Aspirin and all the other 9 [90%] cases of thrombosis appeared after the discontinuation of the dual antiplatelet therapy. Extended use of dual antiplatelet therapy [for more than 12 months] was not significantly more effective than Aspirin monotherapy in reducing the risk of myocardial infarction or stent thrombosis, death from cardiac cause, and stroke

Effects of cardiac rehabilitation program on right ventricular function after coronary artery bypass graft surgery

Cardiac rehabilitation has been recognized as one of the most effective strategies for managing cardiovascular indices as well as controlling the cardiovascular risk profile, in particular after coronary artery bypass graft surgery [CABG]. However, the effect of this program on right ventricular function following CABG is unclear.The aim of this study was to evaluate the impact of cardiac rehabilitation on the right ventricular [RV] function in a cohort of patients who underwent CABG. A total of 28 patients who underwent CABG and participated consecutively in an 8-week cardiac rehabilitation program at Tehran Heart Center were studied. The control group consisted of 39 patients who refused to attend cardiac rehabilitation and only received postoperative medical treatment after registration in the Cardiac Rehabilitation Clinic. Two-dimensional and Doppler echocardiography was performed to assess the RV function in both groups at the three time points of before surgery, at the end of surgery, and at the end of the rehabilitation program. Significant increase of RV function parameters were observed in both rehabilitation group [RG] and control group [CG] at the end of the rehabilitation program compared with post-CABG evaluation in terms of tricuspid annular plane systolic execution [RG: 12.50 mm to 14.18 mm; CG: 13.41 mm to 14.56 mm], tricuspid annular peak systolic velocity [RG: 8.55 cm/s to 9.14 cm/s; CG: 9.03 cm/s to 9.26 cm/s], and tricuspid annular late diastolic velocity [RG: 8.93 cm/s to 9.39 cm/s; CG: 9.26 cm/s to 9.60 cm/s].The parameters of the RV function did improve in both groups, but this improvement was not associated with participation in the complete cardiac rehabilitation program. The R Vfunction parameters gradually improved after CABG; this progress, however, was independent of the exercise-based cardiac rehabilitation program

Demographics and angiographic findings in patients under 35 years of age with acute ST elevation myocardial infarction

ST-elevation myocardial infarction [STEMI] is a major cause of cardiovascular mortality worldwide. There are differences between very young patients with STEMI and their older counterparts. This study investigates the demographics and clinical findings in very young patients with STEMI. Through a review of the angiography registry, 108 patients aged 35 years [Group II] who underwent coronary angiography after STEMI. Group I patients were more likely to be male [92.6%], smokers, and have a family history of cardiovascular diseases [34.6%]. The prevalence of diabetes, dyslipidemia, and hypertension was higher in the old patients. Triglyceride and hemoglobin were significantly higher in Group I. Normal coronary angiogram was reported in 18.5% of the young patients, and in 2.1% of the older patients. The prevalence of single-vessel and multi-vessel coronary artery disease was similar in the two groups [34.3% vs. 35.2%]. The younger subjects were more commonly candidates for medical treatment and percutaneous coronary intervention [PCI] [84.2%], while coronary artery bypass grafting [CABG] was considered for the 39.5% of their older counterparts. In the young adults with STEMI, male gender, smoking, family history, and high triglyceride level were more often observed. A considerable proportion of the young patients presented with multi-vessel coronary disease. PCI or medical treatment was the preferred treatment in the younger patients; in contrast to their older counterparts, in whom CABG was more commonly chosen for revascularization

Success rate, procedural complications and clinical outcomes of coronary interventions in octogenarians: a case-control study

Clinical trials of revascularization have routinely under-enrolled elderly subjects. Thus, symptom relief and improved survival might not apply to elderly patients, in whom the risk of mortality and disability from revascularization procedures seems to be high and co-morbidity is more prevalent. The present case control study was performed to draw a comparison in terms of the procedural success, procedural and in-hospital complications, and major adverse cardiac events [MACE] in a one-year follow-up of octogenarians [age >/= 80 years] with a selected matched younger control group in the Tehran Heart Center Angioplasty Registry. According to the Tehran Heart Center Interventional Registry of 9, 250 patients with a minimum follow-up period of one year between April 1993 and February 2010, 157 percutaneous coronary intervention [PCI] procedures were performed in 112 octogenarians. Additionally, 336 younger patients [459 PCI procedures] were selected from the database as the propensity-score matched controls. There were 147 [93.6%] and 441 [96.1%] successful PCI procedures in the elderly group and control group, respectively [p value = 0.204]. Procedural complications were seen in 5 [3.2%] of the elderly group and 16 [3.5%] of the control group [p value = 0.858]. Totally, 7 [6.3%] in-hospital complications occurred in the elderly group and 22 [6.8%] in the control group [p value = 0.866]. One-year MACE was seen in 9 [9.1%] of the elderly and 18 [5.8%] of the control group [p value = 0.26]. Procedural success and complications, in-hospital complications, and one-year MACE were not significantly different between our two study groups. Therefore, age alone should not be used as the sole criterion when considering revascularization procedures. Furthermore, PCI should not be refused in octogenarians if indicated

Outcomes of percutaneous coronary intervention on saphenous vein graft and native coronary vessels

The optimal target for revascularization in patients with history of coronary artery bypass graft surgery [CABG] is unclear. This study was designed to compare the outcome of percutaneous coronary intervention [PCI] on saphenous vein grafts [SVG] and that on native vessels in patients with previous CABG in terms of major adverse cardiac events [MACE]. The study drew upon data on consecutive patients hospitalized for PCI and MACE rate during a nine-month follow- up period. The patients were divided according to the target vessel for PCI into two groups: SVG and native vessel. Between 2003 and 2007, 226 patients underwent PCI 6.57 +/- 4.55 years after CABG. Their mean age was 59.52 +/- 9.38 years, and 176 [77.9%] were male. PCI was performed on the SVG in 63 [27.9%] patients and on the native coronary artery in the rest. During a nine-month follow-up period, 9 [4%] patients suffered MACE; the prevalence of MACE was not significantly different between the SVG group [4.8%] and the native vessel group [4.9%], [p value = 0.999]. PCI on grafted and native vessels did not affect MACE in patients undergoing PCI after CABG

Efficacy of two streptokinase formulations in acute myocardial infarction: a double-blind randomized clinical trial

We sought to evaluate the efficacy and safety of the different trade forms of streptokinase available in our country, namely Heberkinasa [Heberbiotec, Havana, Cuba] and Streptase [Aventis Behring GmbH, Marburg, Germany]. We conducted a double-blind randomized clinical trial to compare the two streptokinase formulations, i.e. Heberkinasa [HBK] or Streptase [STP], in patients with acute myocardial infarction who needed thrombolysis. Thrombolysis success was evaluated angiographically and/or clinically. Clinical follow-up was done 30 days after thrombolysis. We randomly allocated 221 patients with a mean age of 56.9 +/- 10.8 years [males: 88.2%] to HBK [n = 119] and STP [n=102] groups. Baseline clinical and demographic characteristics were similar between the two groups, and the two groups were not significantly different in terms of door-to-needle and pain-to-needle intervals. The rate of complications was not significantly different between the groups [44.1% [HBK] vs. 42% [STP]]. Angiography was done for 158 [71.5%] patients in the first 24 hours [9%] and in the first 72 hours [38.8%] after thrombolysis. Lesion morphology and lesion/patient ratio were not significantly different between the two groups [1.87[HBK] vs. 1.67[STP]]. The two groups were similar with respect to angiographic patency rate [67.5% [HBK] vs. 67.6% [STP]]. The study groups were also similar as regards clinical outcome and complications of both streptokinase formulations. The present study demonstrated that Heberkinasa is as effective and as safe as a standard streptokinase, namely Streptase, in a clinical setting