Using machine learning models to predict falls in hospitalised adults.

BACKGROUND: Identifying patients at high risk of falling is crucial in implementing effective fall prevention programs. While the integration of information systems is becoming more widespread in the healthcare industry, it poses a significant challenge in analysing vast amounts of data to identify factors that could enhance patient safety. OBJECTIVE: To determine fall-associated factors and develop high-performance prediction tools for at-risk patients in acute and sub-acute care services in Australia. METHODS: A retrospective study of 672,400 patients admitted to acute and sub-acute care services within a large metropolitan tertiary health service in Victoria, Australia, between January 1, 2019, and December 31, 2021. Data were obtained from four sources: the Department of Health Victorian Admitted Episodes Dataset, RiskManTM, electronic health records, and the health workforce dataset. Machine learning techniques, including Random Forest and Deep Neural Network models, were used to analyse the data, predict patient falls, and identify the most important risk factors for falls in this population. Model performance was evaluated using accuracy, F1-score, precision, recall, specificity, Matthew's correlation coefficient, and the area under the receiver operating characteristic curve (AUC). RESULTS: The deep neural network and random forest models were highly accurate in predicting hospital patient falls. The deep neural network model achieved an accuracy of 0.988 and a specificity of 0.999, while the RF achieved an accuracy of 0.989 and a specificity of 1.000. The top 20 variables impacting falls were compared across both models, and 12 common factors were identified. These factors can be broadly classified into three categories: patient-related factors, staffing-related factors, and admission-related factors. Although not all factors are modifiable, they must be considered when planning fall prevention interventions. CONCLUSION: The study demonstrated machine learning's potential to predict falls and identify key risk factors. Further validation across diverse populations and settings is essential for broader applicability.

Reporting health research translation and impact in the curriculum vitae: a survey.

BACKGROUND: Increasingly, health researchers must demonstrate the impact and real-life applications of their research. We investigated how health researchers with expertise in knowledge translation report research translation activities and impact on their curriculum vitae (CV). METHODS: We conducted a cross-sectional survey of health researchers with expertise in knowledge translation as we anticipated best practices in CV reporting from this specialized group. Our survey asked participants about their reporting of research translation and impact activities on their CVs, intention to report, and barriers and facilitators to reporting such activities on their CVs. We calculated univariate descriptive statistics for all quantitative data. Linear regression models determined predictors of researchers' intention to report research translation and impact activities on their CVs. We analyzed open-ended qualitative responses using content analysis. RESULTS: One hundred and fifty-three health researchers responded to the survey (response rate = 29%). Most respondents were Canadian, were female, and had a doctoral degree. Eighty-two percent indicated they reported at least one research translation and/or impact indicator on their CVs. Of those, health researchers commonly reported the following: advisory/regulatory committee membership related to research program (83%), research translation award(s) (61%), and academic performance assessments (59%). Researchers least commonly indicated the following: citation metric scores (31%), summaries of impact (21%), and requests to use research materials and/or products (19%). Fewer than half of the health researchers intended to report knowledge translation (43%) and impact (33%) on their CVs. Strong beliefs about capabilities and consequences of reporting research translation and/or impact were significant predictors of intention. Main barriers were as follows: CV templates do not include research translation and impact activities, participants perceived employers do not value research translation and impact activities, and lack of metrics to evaluate research translation and impact. Ninety-six percent were unaware of a CV template formatted to include research translation and/or impact reporting. CONCLUSIONS: Knowledge translation and impact indicators on the CV are inconsistently reported by our sample of health researchers. Modifiable barriers should be addressed to support more consistent reporting of such activities, including providing a CV template that includes research translation and impact as well as clear metrics to quantify them.

Comparison of not for resuscitation (NFR) forms across five Victorian health services.

BACKGROUND: Within Australian hospitals, cardiac and respiratory arrests result in a resuscitation attempt unless the patient is documented as not for resuscitation. AIM: To examine the consistency of policies and documentation for withholding in-hospital resuscitation across health services. METHOD: An observational, qualitative review of hospital policy and documentation was conducted in June 2013 in three public and two private sector hospitals in metropolitan Melbourne. Not for resuscitation (NFR) forms were evaluated for physical characteristics, content, authorisation and decision-making. Hospital policies were coded for alerts, definition of futility and burden of treatment and management of discussions and dissent. RESULTS: There was a lack of standardisation, with each site using its own unique NFR form and accompanying site-specific policies. Differences were found in who could authorise the decision, what was included on the form, the role of patients and families, and how discussions were managed and dissent resolved. Futility and burden of treatment were not defined independently. These inconsistencies across sites contribute to a lack of clarity regarding the decision to withhold resuscitation, and have implications for staff employed across multiple hospitals. CONCLUSIONS: NFR forms should be reviewed and standardised so as to be clear, uniform and consistent with the legislative framework. We propose a two-stage process of documentation. Stage 1 facilitates discussion of patient-specific goals of care and consideration of limitations of treatment. Stage 2 serves to communicate a NFR order. Decisions to withhold resuscitation are inherently complex but could be aided by separating the decision-making process from the communication of the decision, resulting in improved end-of-life care.

Coping with Stressors in Late Adolescence Young Adulthood : a Salutogenic Perspective

Literature indicates that globally young people are increasingly experiencing life as hopeless and meaningless. It is thus beneficial to investigate factors facilitating well-being in young people; as they are required to cope with multiple stressors whilst simultaneously negotiating the transition from childhood to adulthood. Health practitioners need to establish ways to enhance adequate coping in young people in order to minimise stress and ward off negative consequences such as addiction; depression and other pathologies that compromise health. This study aimed to ascertain whether discovering meaning in life and developing a strong sense of coherence may predict coping with stressors in a group of male and female South Africans in late adolescence/young adulthood (N=258). Three self-report measures were used: the Purpose in Life Questionnaire to measure the extent to which one has found meaning; the Sense of Coherence Scale to assess the extent to which individuals view life as comprehensible; manageable and meaningful; and the Adolescent Coping Scale to determine ability to cope. Logistic regression analysis results indicate that the extent to which one has discovered meaning in life and developed a sense of coherence predict ability to cope in young people. This suggests that intervention strategies that encourage individuals to search for meaning and strengthen sense of coherence may be effective in facilitating coping and may contribute positively to the overall health of young people

Evanescent wave biosensors. Real-time analysis of biomolecular interactions.

Optical biosensors, based on evanescent wave technology, are analytical devices that measure the interactions between biomolecules in real time, without the need for any labels. Specific ligands are immobilized to a sensor surface, and a solution of receptor or antibody is injected over the top. Binding is measured by recording changes in the refractive index, caused by the molecules interacting near the sensor surface within the evanescent field. Evanescent wave-based biosensors are being used to study an increasing number of applications in the life sciences, including the binding and dissociation kinetics of antibodies and receptor-ligand pairs, protein-DNA and DNA-DNA interactions, epitope mapping, phage display libraries, and whole cell- and virus-protein interactions. There are currently four commercially available evanescent wave biosensors on the market. This article describes the technology behind their sensing techniques, as well as the range of applications in which they are employed.

Surface plasmon resonance studies of the interaction between factor VII and tissue factor. Demonstration of defective tissue factor binding in a variant FVII molecule (FVII-R79Q).

The blood coagulation cascade is initiated when vessel injury allows factor VII (FVII) to form a complex with tissue factor (TF). Complete deficiency of FVII causes a lethal bleeding diathesis, but individuals with moderately reduced FVII levels are often asymptomatic. Some of these individuals have circulating partially functional FVII, as a result of point missense mutations in critical parts of the molecule. One such mutation has been reported at position 79 in the first epidermal growth factor-like (EGF) domain of FVII, where an arginine residue has been replaced by glutamine. There is controversy as to whether or not this mutation reduces the affinity of the FVII/TF interaction compared to wild-type FVII. To address this problem, we have expressed recombinant FVII-R79Q and subjected it to detailed biochemical analysis. One-stage FVII:C assays show the variant FVII to have reduced activity with respect to the wild type. Rates of autoactivation and activation by FXa to the two-chain molecule were identical for wild-type and variant FVII. The Vmax for FX activation was lower for the mutant as measured using an amidolytic assay for FX activity. In contrast, the Km for FX was lower for the variant than the wild-type molecule. Peptidyl substrate hydrolysis was virtually identical for both variant and normal FVIIa in the presence and absence of TF. The variant has reduced affinity for TF as measured by surface plasmon resonance. FVII-R79Q has an association rate constant (kassoc) one-fifth of that of normal FVII, but a similar kdiss, resulting in a decrease in the affinity of the enzyme for its cofactor.(ABSTRACT TRUNCATED AT 250 WORDS)

Characterization of glycoprotein oligosaccharides using surface plasmon resonance.

Surface plasmon resonance was used to characterize the carbohydrate moieties of bovine fetuin. This technique, requiring no sample derivatization or labeling, identified the presence and composition of both N- and O-linked oligosaccharides, using a combination of lectin probes and in situ glycosidase digestion. A complete analysis was achieved using 1.4 micrograms of pure bovine fetuin, and was fully automated using Pharmacia's BIA-core. The presence or absence of specific oligosaccharide structures was determined by the binding of a panel of unlabeled lectins. Monosaccharide order and linkages were determined by sequential digestion using a range of specific exoglycosidases. This novel method implementing in situ digestion was achieved using a modification to the BIAcore software, allowing the flow of reagent over the sensor chip to be stopped for variable lengths of time, thereby permitting enzymatic digestion to occur. This technique can be applied to other commercial SPR biosensors currently available. Glycoanalysis by SPR uses approximately 100- to 1000-fold less protein than comparable analyses using alternative techniques such as gel permeation chromatography of released oligosaccharides, labeled lectin binding, or mass spectrometry.