A review of machine learning in obesity

Rich sources of obesity-related data arising from sensors, smartphone apps, electronic medical health records and insurance data can bring new insights for understanding, preventing and treating obesity. For such large datasets, machine learning provides sophisticated and elegant tools to describe, classify and predict obesity-related risks and outcomes. Here, we review machine learning methods that predict and/or classify such as linear and logistic regression, artificial neural networks, deep learning and decision tree analysis. We also review methods that describe and characterize data such as cluster analysis, principal component analysis, network science and topological data analysis. We introduce each method with a high-level overview followed by examples of successful applications. The algorithms were then applied to National Health and Nutrition Examination Survey to demonstrate methodology, utility and outcomes. The strengths and limitations of each method were also evaluated. This summary of machine learning algorithms provides a unique overview of the state of data analysis applied specifically to obesity.

Interfacial tensiometry and dilational surface visco-elasticity of biological liquids in medicine

Dynamic surface tensions and dilational visco-elasticity are easy accessible parameters of liquids. For human body liquids, such as urine, blood serum, amniotic fluid, gastric juice, saliva and others, these parameters are very characteristic for the health status of people. In case of a disease the composition of certain liquids specifically changes and the measured characteristics of dynamic surface tension of the dilational surface elasticity and viscosity reflect these changes in a clear way. Thus, this kind of physico-chemical measurements represent sensitive tools for evaluating the severity of a disease and can serve as control tool for the efficiency of applied therapies. The overview summarises the results of a successful work over about 25years on this subject and gives specific insight into a number of diseases for which the diagnostics as well as the therapy control have been significantly improved by the application of physico-chemical experimental techniques.

Cardiac Pathology in Patients with End-Stage Renal Disease Maintained on Hemodialysis

We report the spectrum of cardiovascular pathology found on autopsy examination in 106 consecutive patients with end-stage renal disease. Cardiovascular pathology was present in nearly all patients, and was the most frequent cause of death (36%), with acute myocardial infarction accounting for 15%. Particularly frequent pathologic findings were left ventricular hypertrophy, coronary and aortic atherosclerosis, pericarditis with effusion, myocardial fibrosis, and valvular dilatation. Cardiovascular death rate was higher during the first year than after the fifth year of dialysis. Nearly all patients had a history of hypertension. The nature of the underlying cause of renal failure and pre-existing cardiovascular disease, specifically diabetes mellitus and hypertension, were the principal predictors of cardiovascular mortality rather than maintenance hemodialysis therapy per se.

Creation of Realistic Appearing Simulated Patient Cases Using the INTERNIST-1/QMR Knowledge Base and Interrelationship Properties of Manifestations

The Internist-1/0uick Medical Reference (OMR) knowledge base (KB) describes the clinical manifestations of some 600 diseases in the domain of internal medicine. This KB, while not representing deep causal modelling of disease processes, is nonetheless effective iri providing medical diagnostic assistance through the OMR medical decision support system. One potential application ofthis extensive KB is the generation of simulated patient cases for use in educating health professionals. However, the “flat” KB is not adequate for this because the clinical manifestations used in the disease descriptions are not mutually independent. While it is theoretically possible to construct disease descriptions which embody pathophysiologic mechanisms of disease causality, it is not practical from the standpoint of resource utilization. Short of constructing a causal knowledge base, the authors herein describe the generation of realistic appearing simulated patient case data using existing information in the knowledge base. This existing information in the KB is in the form of properties which represent a shallow form of interrelationships of the manifestations. The authors conclude that this ability to generate simulated cases represents another view in which to look at an extensive knowledge base, as well as having application to constructing intelligent tutoring systems for health professionals in training.

Use of the Quick Medical Reference (QMR) Program as a Tool for Medical Education

The original goal of the INTERNIST-1 project, as formulated in the early 1970s, was to develop an expert consultant program for diagnosis in general internal medicine. By the early 1980s, it was recognized that the most valuable product of the project was its medical knowledge base (KB). The INTERNIST-1/QMR KB comprehensively summarizes information contained in the medical literature regarding diagnosis of disorders seen in internal medicine. The QMR program was developed to enable its users to exploit the contents ofthe INTERNIST-1/ QMR KB in educationally, clinically, and computationally useful ways. Utilizing commonly available microcomputers, the program operates at three levels - as an electronic textbook, as an intermediate level spreadsheet for the combination and exploration of simple diagnostic concepts, and as an expert consultant program. The electronic textbook contains an average of 85 findings and 8 associated disorders relevant to the diagnosis of approximately 600 disorders in internal medicine. Inverting the disease profiles creates extensive differential diagnosis lists for the over 4250 patient findings known to the system. Unlike a standard printed medical textbook, the QMR knowledge base can be manipulated “on the fly” to format displays that match the information needs of users. Preliminary use of the program for education of medical students and medical house officers at several sites has met with an enthusiastic response.

Local recurrence after 'standard' abdominoperineal resection: do we really need ELAPE?

Introduction Low rectal cancers requiring abdominoperineal resection tend to have a worse prognosis than higher tumours, which may be treated by anterior resection. One of the reasons for this may be inadequate local surgery, in particular the narrow waist of the resection specimen of a standard abdominoperineal resection may be associated with a high positive circumferential resection margin. The extralevator abdominoperineal excision (ELAPE) aims to improve the R0 resection rate but carries significant morbidity. We examined our own results of standard abdominoperineal resection to assess the need for a change of policy. Methods We operformed a retrospective analysis of consecutive standard abdominoperineal resections for rectal cancer in a single centre from June 2002 to December 2011. Results A total of 102 patients underwent standard abdominoperineal resection with curative intent; 19 had no preoperative treatment, 42 had short course radiotherapy, 9 had long course radiotherapy and 32 had neoadjuvant chemotherapy followed by long course chemoradiotherapy. In 17/102(16.6%), there was a positive circumferential resection margin. Over a median follow up of 32 months, 20 patients developed recurrence of any type. Local recurrence occurred in five patients (two of which also had distant recurrence), of whom two had a positive circumferential resection margin (P = 0.10). Actuarial two-year local only recurrence was 3.4% and any recurrence was 17.7%. Overall five-year cancer specific survival was 77%. Conclusions In this series we found low rates of local recurrence after standard abdominoperineal resection even with a circumferential margin rate positivity of 16.6%.Performing an ELAPE in selected cases may improve these results further but is not necessarily required for all patients.

Adsorption of C14EO8 at the interface between its aqueous solution drop and air saturated by different alkanes vapor

The dynamic and equilibrium surface tension for drops of aqueous C₁₄EO₈ solutions at the interface to pure air or pentane, hexane, heptane and toluene saturated air, and the dynamic surface tension of pure water at these interfaces are presented. Two theoretical models were employed: both assuming a diffusion controlled adsorption of the surfactant, and either a diffusion or kinetic barrier governed adsorption of the alkanes. The experimental results are best described by the model which implies a diffusion control for the C₁₄EO₈ molecules and the existence of a kinetic barrier for the alkane molecules. The desorption of alkanes from the surface layer after equilibration and their subsequent removal from the measuring cell was studied as well. The desorption process was shown to be slow for heptane and hexane. However, for the pentane vapor the desorption is quite rapid, and after the desorption commences the surface tension becomes equal to that at the interface with pure air.

Prevalence and disease associations in feline thrombocytopenia: a retrospective study of 194 cases

OBJECTIVES

To assess the prevalence of thrombocytopenia in a referral population of cats in the UK, to identify disease processes associated with thrombocytopenia and to assess the proportion of thrombocytopenic cats that tested positive for feline leukaemia virus or feline immunodeficiency virus.

MATERIALS AND METHODS

Retrospective analysis of medical records at a UK referral hospital. Cats were grouped by mechanism of thrombocytopenia and disease process (where known).

RESULTS

Prevalence of thrombocytopenia was 5·9%. The most common disease processes associated with thrombocytopenia were haematological or infectious disease and neoplasia; 11% of thrombocytopenic cats tested were positive for feline leukaemia virus, which is lower than reported previously. Cats presenting with unexplained haemorrhage had significantly lower platelet counts than other thrombocytopenic cats. Primary immune-mediated thrombocytopenia was less commonly diagnosed than in dogs and associated with the most severe platelet depletion in this study.

CLINICAL SIGNIFICANCE

Thrombocytopenia in cats may be more prevalent than previously reported and severe thrombocytopenia may be associated with spontaneous haemorrhage. Severe thrombocytopenia in cats appears less commonly immune-mediated than in dogs. Thrombocytopenia did not appear to be associated with retroviral infections.

A study of factors influencing surgical cesarean delivery times in an academic tertiary center

BACKGROUND

Knowledge of hospital-specific average cesarean delivery operative times, and factors influencing length of surgery, can serve as a guide for anesthesiologists when choosing the optimal anesthetic technique. The aim of this study was to determine operative times and the factors influencing those times for cesarean delivery.

METHODS

We conducted a retrospective review of all 1348 cesarean deliveries performed at an academic hospital in 2011. The primary outcome was mean operative time for first, second, third and fourth or more cesarean deliveries. The secondary goal was to identify factors influencing operative time. Variables included age, body mass index, previous surgery, gestational age, urgency of cesarean delivery, anesthesia type, surgeon's seniority, layers closed, and performance of tubal ligation.

RESULTS

Mean (standard deviation) operative times for first (n=857), second (n=353), third (n=108) and fourth or more (n=30) cesarean deliveries were 56 (19), 60 (19), 69 (28) and 82 (31) minutes, respectively (P <0.0001, all groups different). Emergency status of the case and later gestational age were associated with shorter operative times. Higher body mass index, a less senior surgeon, the number of layers closed, and tubal ligation, increased operative times. These factors accounted for 18% of the variability.

CONCLUSIONS

Third and fourth cesarean delivery or the presence of other factors that could increase operative time may warrant catheter-based anesthetic techniques or the addition of adjunctive medications to prolong spinal anesthetic block. Institutional and individual surgeon factors may play an even more important role in determining surgical time.