Artificial Intelligence-Based Framework for Analyzing Health Care Staff Security Practice: Mapping Review and Simulation Study.

BACKGROUND: Blocklisting malicious activities in health care is challenging in relation to access control in health care security practices due to the fear of preventing legitimate access for therapeutic reasons. Inadvertent prevention of legitimate access can contravene the availability trait of the confidentiality, integrity, and availability triad, and may result in worsening health conditions, leading to serious consequences, including deaths. Therefore, health care staff are often provided with a wide range of access such as a "breaking-the-glass" or "self-authorization" mechanism for emergency access. However, this broad access can undermine the confidentiality and integrity of sensitive health care data because breaking-the-glass can lead to vast unauthorized access, which could be problematic when determining illegitimate access in security practices. OBJECTIVE: A review was performed to pinpoint appropriate artificial intelligence (AI) methods and data sources that can be used for effective modeling and analysis of health care staff security practices. Based on knowledge obtained from the review, a framework was developed and implemented with simulated data to provide a comprehensive approach toward effective modeling and analyzing security practices of health care staff in real access logs. METHODS: The flow of our approach was a mapping review to provide AI methods, data sources and their attributes, along with other categories as input for framework development. To assess implementation of the framework, electronic health record (EHR) log data were simulated and analyzed, and the performance of various approaches in the framework was compared. RESULTS: Among the total 130 articles initially identified, 18 met the inclusion and exclusion criteria. A thorough assessment and analysis of the included articles revealed that K-nearest neighbor, Bayesian network, and decision tree (C4.5) algorithms were predominantly applied to EHR and network logs with varying input features of health care staff security practices. Based on the review results, a framework was developed and implemented with simulated logs. The decision tree obtained the best precision of 0.655, whereas the best recall was achieved by the support vector machine (SVM) algorithm at 0.977. However, the best F1-score was obtained by random forest at 0.775. In brief, three classifiers (random forest, decision tree, and SVM) in the two-class approach achieved the best precision of 0.998. CONCLUSIONS: The security practices of health care staff can be effectively analyzed using a two-class approach to detect malicious and nonmalicious security practices. Based on our comparative study, the algorithms that can effectively be used in related studies include random forest, decision tree, and SVM. Deviations of security practices from required health care staff's security behavior in the big data context can be analyzed with real access logs to define appropriate incentives for improving conscious care security practice.

Can HbA1c detect undiagnosed diabetes in acute medical hospital admissions?

OBJECTIVE: To study hyperglycaemia in acute medical admissions to Irish regional hospital. RESEARCH DESIGN AND METHODS: From 2005 to 2007, 2061 white Caucasians, aged >18 years, were admitted by 1/7 physicians. Those with diabetes symptoms/complications but no previous record of hyperglycaemia (n=390), underwent OGTT with concurrent HbA1c in representative subgroup (n=148). Comparable data were obtained for 108 primary care patients at risk of diabetes. RESULTS: Diabetes was diagnosed immediately by routine practice in 1% (22/2061) [aged 36 (26-61) years (median IQ range)/55% (12/22) male] with pre-existing diabetes/dysglycaemia present in 19% (390/2061) [69 (58-80) years/60% (235/390) male]. Possible diabetes symptoms/complications were identified in 19% [70 (59-79) years/57% (223/390) male] with their HbA1c similar to primary care patients [54 (46-61) years], 5.7 (5.3-6.0)%/39 (34-42)mmol/mol (n=148) vs 5.7 (5.4-6.1)%/39 (36-43)mmol/mol, p=0.35, but lower than those diagnosed on admission, 10.2 (7.4-13.3)%/88 (57-122)mmol/mol, p<0.001. Their fasting plasma glucose (FPG) was similar to primary care patients, 5.2 (4.8-5.7) vs 5.2 (4.8-5.9) mmol/L, p=0.65, but 2hPG higher, 9.0 (7.3-11.4) vs 5.5 (4.4-7.5), p<0.001. HbA1c identified diabetes in 10% (15/148) with 14 confirmed on OGTT but overall 32% (48/148) were in diabetic range on OGTT. The specificity of HbA1c in 2061 admissions was similar to primary care, 99% vs 96%, p=0.20, but sensitivity lower, 38% vs 93%, p<0.001 (63% on FPG/23% on 2hPG, p=0.037, in those with possible symptoms/complications). CONCLUSION: HbA1c can play a diagnostic role in acute medicine as it diagnosed another 2% of admissions with diabetes but the discrepancy in sensitivity shows that it does not reflect transient/acute hyperglycaemia resulting from the acute medical event.

Timing of access to secondary healthcare services and lower extremity amputations in patients with diabetes: a case-control study.

OBJECTIVE: To investigate the association between timing of patient access to secondary healthcare services for diabetes management and lower extremity amputation (LEA) among patients with diabetes. RESEARCH DESIGN AND METHODS: A case-control study was conducted in the secondary healthcare system in Ireland. Cases were 116 patients with diabetes who underwent a first major non-traumatic LEA between 2006 and 2012. Controls were 348 patients with diabetes, over 45 years, admitted to the same hospital as an emergency or electively, frequency-matched for gender, type of diabetes, and year. Data were collected for 7 years prior to the event year. ORs for LEA in patients with diabetes comparing early versus late referral from primary to secondary healthcare were calculated. RESULTS: Statistically significant risk factors associated with LEA in patients with diabetes included being single, chronic kidney disease, hypertension, and hyperglycemia. Documented retinopathy was a significant protective factor. In unconditional logistic regression analysis adjusted for potential confounders, there was no evidence of a reduced risk of LEA among patients referred earlier to secondary healthcare for diabetes management. CONCLUSIONS: Specialist referral may need to occur earlier than the 7-year cut-off used to demonstrate an effect on reducing LEA risk. Documented retinopathy was associated with reduced risk of LEA, most likely as a proxy for better self-care. Variation in the management of diabetes in primary care may also be impacting on outcomes. Efforts to develop more integrated care between primary and secondary services may be beneficial, rather than focusing on timing of referral to secondary healthcare.

Timing of access to secondary healthcare services for diabetes management and lower extremity amputation in people with diabetes: a protocol of a case-control study.

BACKGROUND: Lower extremity amputation (LEA) is a complication of diabetes and a marker of the quality of diabetes care. Clinical and sociodemographic determinants of LEA in people with diabetes are well known. However, the role of service-related factors has been less well explored. Early referral to secondary healthcare is assumed to prevent the occurrence of LEA. The objective of this study is to investigate a possible association between the timing of patient access to secondary healthcare services for diabetes management, as a key marker of service-related factors, and LEA in patients with diabetes. METHODS/DESIGN: This is a case-control study. The source population is people with diabetes. Cases will be people with diabetes who have undergone a first major LEA, identified from the hospital discharge data at each of three regional centres for diabetes care. Controls will be patients with diabetes without LEA admitted to the same centre either electively or as an emergency. Frequency-matching will be applied for gender, type of diabetes, year and centre of LEA. Three controls per case will be selected from the same population as the cases. With a power of 90% to detect OR of 0.4 for an association between 'good quality care' and LEA in people with diabetes, 107 cases and 321 controls are required. Services involved in diabetes management are endocrinology, ophthalmology, renal, cardiology, vascular surgery and podiatry; timing of first contact with any of these services is the main exploratory variable. Using unconditional logistic regression, an association between this exposure and the outcome of major LEA in people with diabetes will be explored, while adjusting for confounders. ETHICS AND DISSEMINATION: Ethical approval was granted by the Clinical Research Ethics Committee of the Cork Teaching Hospitals, Ireland. Results will be presented at conferences and published in peer-reviewed journals.