Statistical clues to postoperative blood loss: Moving averages applied to medical data

With the advent of computerized databases, medical data has become easy to accumulate; however, effective use of this data continues to pose significant problems. In other circumstances, smoothing algorithms have been used to uncover non-obvious correlations, trends and relationships in noisy data. We have applied four such algorithms to a large dataset of postoperative blood replacement in cardiopulmonary bypass patients. When applied to this dataset, one of the algorithms proved surprisingly effective. It confirmed several previously observed correlations, and also provided an additional series of counterintuitive and apparently unrelated associations. These associations have been explored in an accompanying paper.     

Simulation model of renal replacement therapy: predicting future demand in England.

BACKGROUND: The demand for renal replacement therapy (RRT) in England has risen steadily, although from a lower base than many other developed countries. Predicting the future demand for RRT and the impact of factors such as the acceptance rate, transplant supply and patient survival, is required in order to inform the planning of such services. METHODS: A discrete event simulation model estimates the future demand for RRT in England in 2010 for a range of scenarios. The model uses current prevalence and current and projected future acceptance rates, survival rates and the transitions between modalities to predict future patient numbers. National population and mortality data, published literature and data from the UK Renal Registry and UK Transplant, are used to estimate unmet need for RRT, the impact of changing demography and incidence of Type 2 diabetes, patient haemodialysis (HD) survival and transplant supply. RESULTS: By 2010 the predicted prevalence will have increased from about 30,000 in 2000 to between 42 and 51,000 (900-1000 p.m.p.), an average annual growth of 4.5-6%. Changing transplant supply has a small effect on overall numbers but changes the proportion of patients with functioning graft by up to 8%. Even with an optimistic increase in transplant supply (11% p.a. for 5 years), numbers on HD will continue to rise substantially, especially in the elderly. The factors most influencing future patient numbers are the acceptance rate and dialysis survival. CONCLUSION: This model predicts a substantial growth in the RRT population to 2010 to a rate approaching 1000 p.m.p., particularly in the elderly and those on HD, with a steady state not being reached for at least 25 years.     

Introducing expert systems to medical students using esta, Expert System Teaching Aid

Medical students are allocated little curriculum time for exposure to expert systems. ESTA, a computer model of an expert system, was developed to make best use of this time. The nature of the students' interaction with ESTA is described, and their reactions to the expert system concepts and the place of expert systems in medicine are presented. A discussion of these reactions draws some conclusions about teaching expert systems in particular, and computers in general, in the basic medical course.     

Prototype system for an MR simulation of intracavitary brachytherapy for cervical cancer: Preliminary results in six cases

We report a newly developed MR simulation system for intracavitary brachytherapy for cervical cancer and subsequent treatment results. MR simulation was performed on six patients. The spatial relationship of the tumor to the bladder, rectum, bowel, applicators, etc. was depicted well. Doses to the tumor and surrounding normal tissues were read from isodose curves superimposed on the T2-weighted sagittal image. This system promises to be useful in customizing the dose distribution.     

Development of computer-assisted simulation procedure to analyze receptor modulatory processes

Summary Experimentally induced changes in neurotransmitter receptors have been analyzed by means of a computer assisted simulation procedure over a wide range of free ligand concentrations. This approach allows to evaluate, for a given range of ligand concentrations, the relative influence of simultaneous variations in binding parameters (i.e. dissociation constant or K_d and in maximal number of binding sites or B_m) and to predict the net and final effect of the experimental condition on the receptor-mediated transmission line. The function representing the changes in bound values versus the respective free ligand concentrations, has been studied analytically on the basis of all the possible values that the percent changes in both K_d and B_m parameters induced by a given experimental condition can assume. A well characterized change in the pattern of bound radioligand could in this way be defined. This approach, developed to show in an immediate and clear way treatment-induce changes in receptor populations or to fit directly rough experimental data expressed as differences in bound values versus free ligand concentrations, seems to be an useful complement to the widely used saturation analysis of binding data.     

RF refocused echoes of J-coupled spin systems: effects on RARE-based spectroscopic imaging.

A numerical simulation tool was developed to calculate the echo amplitudes of J-coupled resonances within a series of radiofrequency (RF) refocused echoes. The signal modulation due to J-coupling in rapid acquisition with relaxation enhancement (RARE) is suppressed only when the inverse of the pulse interval (tau) is large compared to both the chemical shift (CS) difference (Deltadelta) of the coupled spins and the coupling constant. In contrast, the echo amplitudes in ultrafast low-flip-angle RARE (U-FLARE) oscillate around a quasi-steady-state value that is greater than zero (neglecting relaxation and diffusion) even when Deltadelta > 1/tau. The flip-angle distribution over the measured slice caused by the use of Gaussian-shape slice-selective refocusing pulses further reduces the echo oscillations. When the pulse interval falls short of the fast pulse rate regime, spectroscopic U-FLARE provides an improved spatial impulse response in the phase-encoding (PE) direction compared to spectroscopic RARE.     

AN-132 block of cardiac sodium channels: A gate-related receptor analysis

Summary Based on the gate-related receptor hypothesis, an analysis of kinetics of AN-132, a new antiarrhythmic agent, blockade of cardiac sodium channels and the gate-related receptor which is bound by the drug was performed by computer simulation. Model-predicted apparent rates of onset of AN-132 (30 μmol/L) blocking were 0.051, 0.038, and 0.034 AF⁻¹ at stimulation frequencies of 1.0, 2.0 and 3.0 Hz, respectively. The time constant of recovery from block by AN-132 at resting potential -90 mV was 39.5 s. These findings are in agreement with those experimental data documented. The analysis of gate-related receptor shows that AN-132 binds the inactivation gate-related receptor, and the binding and unbinding are modulated by the inactivation process.     

头低位模拟失重状态对前庭功能的影响

为研究头低位模拟失重对运动病症状、垂直视动眼震（ＶＯＫＮ）及体液重新分配的影响，在头低位－１０°的模拟失重状态下，采用大视野的垂直视动刺激，观察１８名正常人的运动病症状、ＶＯＫＮ、激素（ＡＶＰ、ＶＩＰ、ＣＯＲＴ、ＡＬＤＯ）的反应特点。结果表明，头低位－１０°状态下的大视野垂直视动刺激可以诱发出明显的运动病症状，头低位－１０°的垂直视动刺激比坐位更容易诱发运动病。坐位状态ＶＯＫＮ慢相速度有明显的方向性不对称，敏感组ＶＯＫＮ方向性不对称有显著差异（Ｐ＜０．０５）。头低位－１０°时ＶＯＫＮ的不对称现象不明显，向下方向运动的ＶＯＫＮ慢相速度显著增加。分析指出，头低位－１０°状态下垂直视动刺激比坐位和秋千刺激的贡献率大。尿中ＣＯＲＴ（皮质醇）在秋千和头低位的垂直视动刺激前后有显著性增加。提示：大视野的垂直视动刺激与头低位－１０°两种刺激的结合可能成为预测空间运动病的方法之一．     

Simulation for population analysis of Michaelis-Menten elimination kinetics

A simulation study was conducted to compare the cost and performance of various models for population analysis of the steady state pharmacokinetic data arising from a one-compartment model with Michaelis-Menten elimination. The usual Michaelis-Menten model (MM) and its variants provide no estimate of the volume of distribution, and generally give poor estimates of the maximal elimination rate and the Michaelis-Menten constant. The exact solution to the Michaelis-Menten differential equation (TRUE) requires a precise analysis method designed for estimation of population pharmacokinetic parameters (the first-order conditional estimation method) and also considerable computational time to estimate population mean parameters accurately. The one-compartment model with dose-dependent clearance (DDCL), in conjunction with the first-order conditional estimation or Laplacian method, ran approximately 20-fold faster than TRUE and gave accurate population mean parameters for a drug having a long biological half-life relative to the dosing interval. These findings suggest that the well-known MM and its variants should be used carefully for the analysis of blood concentrations of a drug with Michaelis-Menten elimination kinetics, and that TRUE, in conjunction with a precise analysis method, should be considered for estimating population pharmacokinetic parameters. In addition, DDCL is a promising alternative to TRUE with respect to computation time, when the dosing interval is short relative to the biological half-life of a drug. This work was supported in part by the Epilepsy Research Foundation, the Nakatomi Foundation, and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan.     

Numerical method for weights adjustment in minimax multi‐model LQ‐control

The minimax linear quadratic problem, where ‘max’ is taken over a finite set of indices (models) and ‘min’ is taken over the set of admissible controls, is considered. The solution is obtained by the robust optimal control application. The control turns out to be a linear combination of the controls optimal for each individual model. This paper develops a numerical method for the optimal weights adjustment. An example shows a quick convergence of the proposed procedure. Copyright © 2007 John Wiley & Sons, Ltd.