Search

Aim: Hip fractures are highly frequent and severe problem, increasing incidence and severity with aging and high disability and mortality rates. Hence, the ability to predict this condition can be of high value for preventive healthcare. The present work explores clinical biomarkers' application for hip fracture risk evaluation and prediction, using an information-theoretical methodology. Material and Method: A dataset on geriatric patients, with and without hip fracture, was analyzed, including blood biomarkers routinely available to physicians: albumin, urea, hemoglobin, calcium, and white blood cells. This research comprised geriatric patients hospitalized at the Shmuel Harofe Geriatric Medical Center, Israel. The patients' data, collected during 2012-2017, were accessed retrospectively. Normalized mutual information was utilized to establish correlations between the parameters, and the nearest neighbor rule with the weighted Hamming distance was used for the construction of a diagnostic decision rule for hip fracture risk evaluation. Results: We developed an algorithm (decision tree) for hip fracture risk group attribution for subjects under 80 years old. The algorithm provided the sensitivity of 0.581 with the 95% confidence interval (0.505, 0.653), and specificity of 0.540 with the 95% confidence interval (0.479, 0.604). This performance was comparable with the results of other common methods for hip fracture risk evaluation, yet the present method may be preferable in terms of data accessibility and the ability to determine the possible time of the fracture. Conclusion: The use of this method has been piloted in the clinic, and with further development and application, can help evaluate the risks of hip fracture in older subjects (aged 60 years or over) to optimize preventive interventions. [ABSTRACT FROM AUTHOR]

With the spread of coronavirus disease, reports have indicated that young patients are usually asymptomatic with a short convalescence period. The current study compared the time to resolution of infection in symptomatic and asymptomatic patients. Seventy-six patients aged 44.4 ± 23.3 years were admitted to the coronavirus disease 2019 (COVID-19) unit during the study period. Data were collected from patient records. Throat and nasal swabs for COVID-19 were collected for reverse transcriptase-polymerase chain reaction (RT-PCR). Time to resolution of infection was defined as the number of days from the date of the first COVID-19 positive outcome to the second consecutive negative PCR results. Most patients showed COVID-19 signs and symptoms (71.1%) between 1 and 6 days, and the rest were asymptomatic. No association was found between the time to resolution of infection and the presence of COVID-19 signs and symptoms (symptomatic: median [Md] 10.0, 95% confidence interval [CI] 8.4-11.6; asymptomatic: Md 15.0, 95% CI 10.5-15.5; P = 0.54). Age was not correlated with the number of COVID-19 signs and symptoms (r = 0.13, P = 0.37) or with the time to resolution of infection (r = 0.06, P = 0.58). In patients with mild to moderate symptoms, the time to resolution of infection from COVID-19 is not different from that in asymptomatic patients.