Machine Learning for Classification of DOS Attack in Smart Healthcare Networks

Healthcare industry has taken 360-degree change when it comes to managing, analyzing and leveraging healthcare data. With 5G technology increased data rates, more reliability and greater capacity, the healthcare system can provide remote services for patients. For remote monitoring of a patient’s health, real time data delivery is a must. The crucial requirement of the current time in the healthcare industry is the security of the patient’s sensitive and critical data against potential threats. Therefore, it is important that we have security mechanisms ensuring not only authorized parties have access to a patient’s sensitive data and medical information but also preserve its privacy and security. By 2022, cybercrime is predicted to cost \$6 trillion each year. Healthcare industry is continually changing and adopting new aspects in technological transformations. In recent years there has been a broad adoption of Machine learning approaches because of their high level performance in healthcare services starting from the prediction of heart arrest, to medical imaging for detection of tumors and even infections like COVID. AI could help the Healthcare Industry protect their patients’ data as well as secure their 5G network of computers across their organization. We have discussed Machine Learning techniques for Data security and privacy and a case study for detecting the intensity of DDOS attack using Decision Tree algorithm on Healthcare application network data. We have also proposed a model which includes 4 interconnecting lifecycle stages- collecting the data, storing the data, processing the data along with the analysis stage and creating knowledge from those data. © 2022, Springer Nature Switzerland AG.

A VIRTUAL MOCK INTERVIEW PROGRAM IMPROVES THE CONFIDENCE AND PREPAREDNESS FOR APPLICANTS TO FELLOWSHIP IN CARDIOVASCULAR DISEASE

Background: In 2020, the American College of Cardiology Fellows-in-Training (FIT) Section Leadership Council piloted a virtual mock interview (MI) initiative in response to the transition to cardiovascular disease (CVD) fellowship virtual recruitment during the Coronavirus Disease 2019 pandemic. The impact of the expanded MI initiative in 2021 was evaluated. Methods: Through ACC outreach, applicants voluntarily enrolled to participate in virtual 30-minute MI followed by a feedback session conducted by volunteer FIT. Pre- and post-MI surveys utilizing Likert scales were analyzed with paired Wilcoxon rank sum tests. Results: A total of 100 FIT interviewed 159 applicants (34% female, mean age 30 years). Applicants were of diverse racial and ethnic backgrounds (45% Asian, 28% White, 7% Black, 4% Hispanic, 7% multi-ethnic). 26% cited no cardiology-specific mentorship from their institution or outside institutions, and 65% had no prior experience with a virtual interview format. 129 applicants completed both preand post-MI surveys. Compared to pre-MI, applicants’ confidence, preparation, and comfort with a virtual platform improved significantly (p<0.001 for all, Figure). More than 85% of applicants agreed that FIT feedback during the MI helped identify strengths and weaknesses and enhanced their interview skills. Conclusion: The ACC FIT MI initiative improved confidence and subjective virtual interview skills in a diverse cohort of applicants to CVD fellowship. [Formula presented]

Evaluating resident involvement and the 'July effect' in parotidectomy.

OBJECTIVE: This study aimed to evaluate the effect of resident involvement and the 'July effect' on peri-operative complications after parotidectomy. METHOD: The American College of Surgeons National Surgical Quality Improvement Program database was queried for parotidectomy procedures with resident involvement between 2005 and 2014. RESULTS: There were 11 733 cases were identified, of which 932 involved resident participation (7.9 per cent). Resident involvement resulted in a significantly lower reoperation rate (adjusted odds ratio, 0.18; 95 per cent confidence interval, 0.05-0.73; p = 0.02) and readmission rate (adjusted odds ratios 0.30; 95 per cent confidence interval, 0.11-0.80; p = 0.02). However, resident involvement was associated with a mean 24 minutes longer adjusted operative time and 23.5 per cent longer adjusted total hospital length of stay (respective p < 0.01). No significant difference in surgical or medical complication rates or mortality was found when comparing cases among academic quarters. CONCLUSION: Resident participation is associated with significantly decreased reoperation and readmission rates as well as longer mean operative times and total length of stay. Resident transitions during July are not associated with increased risk of adverse peri-operative outcomes after parotidectomy.