Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Predictive”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,351 records · Page 75Linked to original sources
PubMed1999

Repeatability of ultrasound-predicted percentage of intramuscular fat in feedlot cattle.

We used data from 144 bulls, heifers, and steers to determine the repeatability of ultrasound-predicted percentage of intramuscular fat and to study the effect of repeated measurements on the standard error of prediction. Animals were scanned at an average age of 433 d by a certified technician. Individual bulls, heifers, and steers were scanned five to six times each with two Aloka 500-V machines, and the percentage of intramuscular fat was predicted from two regions of interest within an image. Variance components and repeatability values were computed for the overall data and by machine, region of interest, and sex. Animals were broadly divided into two groups based on mean ultrasound-predicted percentage of intramuscular fat. Variance components and repeatability values were then estimated within each group. The overall repeatability of ultrasound-predicted percentage of intramuscular fat was .63 +/- .03. Differences in the repeatability values between machines and between regions of interest were not different from zero (P > .05). Bulls showed a lower within-animal SD of .82% as compared to .97 and 1.02% for steers and heifers, respectively. However, steer ultrasound-predicted percentage of intramuscular fat measures were more repeatable (P < .05) than those of bulls and heifers. The difference in repeatability between bull and heifer measures was not important (P > .05). Animals with mean ultrasound-predicted percentage of intramuscular fat less than 4.79% showed less repeatable measures (P < .05) than those with means above 4.79%. The image variance contributed to nearly 70% of the total variance of observations within an animal. Standard error of animal mean measures showed a 50% reduction when the number of images per animal increased to four. Therefore, we concluded that increasing the number of images per animal plays a more significant role in reducing the standard error of prediction than taking multiple measurements within a single image.

Adipose Tissue↗
PubMed2001

Predicting rates of inbreeding for livestock improvement schemes.

This article presents a deterministic method to predict rates of inbreeding (deltaF) for typical livestock improvement schemes. The method is based on a recently developed general theory to predict rates of inbreeding, which uses the concept of long-term genetic contributions. A typical livestock breeding population was modeled, with overlapping generations, BLUP selection, and progeny testing of male selection candidates. Two types of selection were practiced: animals were either selected by truncation on estimated breeding values (EBV) across age classes, or the number of parents selected from each age class was set to a fixed value and truncation selection was practiced within age classes. Bulmer's equilibrium genetic parameters were obtained by iterating on a pseudo-BLUP selection index and deltaF was predicted for the equilibrium situation. Predictions were substantially more accurate than predictions from other available methods, which ignore the effect of selection on deltaF. Predictions were accurate for schemes with up to 20 sires. Predicted deltaF was somewhat too low for schemes with more than 20 sires, which was due to the use of simple linear models to predict genetic contributions. The present method provides a computationally feasible (i.e., deterministic) tool to consider both the rate of inbreeding and the rate of genetic gain when optimizing livestock improvement schemes.

Animals↗
PubMed2006

Evaluation of average daily gain prediction by level one of the 1996 National Research Council beef model and development of net energy adjusters.

Two data sets were developed to evaluate and refine feed energy predictions with the beef National Research Council (NRC, 1996) model level 1. The first data set included pen means of group-fed cattle from 31 growing trials (201 observations) and 17 finishing trials (154 observations) representing over 7,700 animals fed outside in dirt lots. The second data set consisted of 15 studies with individually fed cattle (916 observations) fed in a barn. In each data set, actual ADG was compared with ADG predicted with the NRC model level 1, assuming thermoneutral environmental conditions. Next, the observed ADG (kg), TDN intake (kg/d), and TDN concentration (kg/kg of DM) were used to develop equations to adjust the level 1 predicted diet NEm and NEg (diet NE adjusters) to be applied to more accurately predict ADG. In both data sets, the NRC (1996) model level 1 inaccurately predicted ADG (P < 0.001 for slope = 1; intercept = 0 when observed ADG was regressed on predicted ADG). The following nonlinear relationships to adjust NE based on observed ADG, TDN intake, and TDN concentration were all significant (P < 0.001): NE adjuster = 0.7011 x 10(-0.8562 x ADG) + 0.8042, R2 = 0.325, s(y.x) = 0.136 kg; NE adjuster = 4.795 10(-0.3689 x TDN intake) + 0.8233, R2 x = 0.714, s(y.x) = 0.157 kg; and NE adjuster = 357 x 10(-5.449 x TDN concentration) + 0.8138, R2 = 0.754, s(y.x) = 0.127 kg. An NE adjuster < 1 indicates overprediction of ADG. The average NE adjustment required for the pen-fed finishing trials was 0.820, whereas the (P < 0.001) adjustment of 0.906 for individually fed cattle indicates that the pen-fed environment increased NE requirements. The use of these equations should improve ADG prediction by the NRC (1996) model level 1, although the equations reflect limitations of the data from which they were developed and are appropriate only over the range of the developmental data set. There is a need for independent evaluation of the ability of the equations to improve ADG prediction by the NRC (1996) model level 1.

Animal Nutritional Physiological Phenomena↗
PubMed2001

Prediction of energy balance in high yielding dairy cows with test-day information.

This study used a previously developed model to predict herd mean energy balance of the first 12 wk of lactation from test-day information. The predictions were compared with calculated energy balance based on feed analysis and to changes in body weight. Seven independent feeding trials including 43 diets (519 lactations, 254 cows; 1987 to 1996) were used. Conventional diets were discriminated from nonconventional diets by significant differences between mean calculated energy balance of subtrial diets versus control diets. The total difference between group means of predicted minus calculated energy balance was positive throughout the observed lactation period. It was lowest (5 to 9 MJ of net energy for lactation) during negative energy balance of the conventional diets in wk 2 to 7 when 18 to 50% of the total difference was due to random variation. Because of this difference, both predicted and calculated energy balances were compared to body weight change as a reference for true energy balance. Body weight change was adjusted for rumen fill. While calculated energy balance tended to be negative at times when cows gained weight, predicted energy balance was positive. Cows fed nonconventional diets gained weight, while calculated energy balance was extremely negative, whereas predicted energy balance based on test-day information was positive. We concluded that the prediction difference was relatively small when standard rations were used, and that nonconventional rations biased predicted energy balance to a lesser extent than calculated energy balance. Estimating energy balance based on test-day information appears feasible.

Animal Feed↗
PubMed2004

Prediction of body lipid change in pregnancy and lactation.

A simple method to predict the genetically driven pattern of body lipid change through pregnancy and lactation in dairy cattle is proposed. The rationale and evidence for genetically driven body lipid change have their basis in evolutionary considerations and in the homeorhetic changes in lipid metabolism through the reproductive cycle. The inputs required to predict body lipid change are body lipid mass at calving (kg) and the date of conception (days in milk). Body lipid mass can be derived from body condition score and live weight. A key assumption is that there is a linear rate of change of the rate of body lipid change (dL/dt) between calving and a genetically determined time in lactation (T') at which a particular level of body lipid (L') is sought. A second assumption is that there is a linear rate of change of the rate of body lipid change (dL/dt) between T' and the next calving. The resulting model was evaluated using 2 sets of data. The first was from Holstein cows with 3 different levels of body fatness at calving. The second was from Jersey cows in first, second, and third parity. The model was found to reproduce the observed patterns of change in body lipid reserves through lactation in both data sets. The average error of prediction was low, less than the variation normally associated with the recording of condition score, and was similar for the 2 data sets. When the model was applied using the initially suggested parameter values derived from the literature the average error of prediction was 0.185 units of condition score (+/- 0.086 SD). After minor adjustments to the parameter values, the average error of prediction was 0.118 units of condition score (+/- 0.070 SD). The assumptions on which the model is based were sufficient to predict the changes in body lipid of both Holstein and Jersey cows under different nutritional conditions and parities. Thus, the model presented here shows that it is possible to predict genetically driven curves of body lipid change through lactation in a simple way that requires few parameters and inputs that can be derived in practice. It is expected that prediction of the cow's energy requirements can be substantially improved, particularly in early lactation, by incorporating a genetically driven body energy mobilization.

Animals↗
PubMed2004

Prediction of daily and lactation yields of milk, fat, and protein using an autoregressive repeatability test day model.

We evaluated the accuracy of an autoregressive multiple-lactation test day (ATD) model to predict missing test day yields of milk, fat, and protein to obtain cumulative 305-d records for cows with incomplete or in-progress lactations. The data consisted of more than one million observations of daily yields on test days in the first 3 lactations of over 75,000 Portuguese Holstein cows. Differences between actual (estimates from complete lactations using the test interval method) and ATD-predicted 305-d yields were negligible and smaller than those predicted by the test interval method. The ATD procedure tended to slightly underestimate cumulative lactation yields, whereas the test interval method substantially overestimated them. Smaller differences obtained by the ATD procedure resulted in less biased estimates of lactation yield, which also implies greater accuracy. As expected, the correlations between actual and predicted lactation yields increased with the number of test days from 0.831 to 0.997. Average correlations (by parity) between actual and ATD-predicted yields ranged from 0.977 to 0.984. Correlations between actual test day yields and corresponding predicted yields exceeded 0.5 for up to 7 time-intervals from the last test day yield used to predict cumulative yield of projected lactations. These correlations indicate the good predictive ability of the ATD method. From a producer's viewpoint, these advantages underwrite management because most on-farm selection decisions are based on the producing abilities of cows. Implementation of ATD methodology does not require special computing capability and is easily transferable to the farm level.

Animals↗
PubMed2005

Accounting for pregnancy diagnosis in predicting days open.

The system for estimating days open for cows with no subsequent lactation was examined to determine if estimates should vary depending on pregnancy diagnosis. Pregnancy diagnosis information was unavailable when the original prediction system was developed, but collection was begun in 2002. New prediction equations were estimated from nearly 1.1 million Holstein lactations for 20-d intervals from 110 to 250 days in milk (DIM). Use of pregnancy diagnosis improved accuracy compared with the original system. The improvement was particularly evident for lactations of cows confirmed to be open in the 130-to-149 DIM interval, where predicted days open increased by > 96 d. For lactations of cows with a confirmed pregnancy, predicted days open decreased by 18 d for the same interval. Prediction errors decreased with increasing DIM. Jersey lactations averaged fewer days open, but in most cases Holstein solutions provided adequate predictions. Specific adjustments were generated for Jersey lactations with no breedings reported. Those adjustments reduced the predicted days open averaged across parity by an amount that increased from 9 to 27 d with DIM interval. The new prediction equations were implemented for November 2004 evaluations for daughter pregnancy rate.

Animals↗
PubMed2006

Development and evaluation of models to predict the feed intake of dairy cows in early lactation.

Inaccurate prediction of dry matter intake (DMI) limits the ability of current models to anticipate the technical and economic consequences of adopting different strategies for production management on individual dairy farms. The objective of the present study was to develop an accurate, robust, and broadly applicable prediction model and to compare it with the current NRC model for dairy cows in early lactation. Among various functions, an exponential model was selected for its best fit to DMI data of dairy cows in early lactation. Daily DMI data (n = 8,547) for 3 groups of Holstein cows (at Illinois, New Hampshire, and Pennsylvania) were used in this study. Cows at Illinois and New Hampshire were fed totally mixed diets for the first 70 d of lactation. At Pennsylvania, data were for the first 63 d postpartum. Data from Illinois cows were used as the developmental dataset, and the other 2 datasets were used for model evaluation and validation. Data for BW, milk yield, and milk composition were only available for Illinois and New Hampshire cows; therefore, only these 2 datasets were used for model comparisons. The exponential model, fitted to the individual cow daily DMI data, explained an average of 74% of the total variation in daily DMI for Illinois data, 49% of the variation for New Hampshire data, 67% of the variation for Pennsylvania data, and 64% of the variation overall. Based on all model selection criteria used in this study, the exponential model for prediction of weekly DMI of individual cows was superior to the current NRC equation. The exponential model explained 85% of the variation in weekly mean DMI compared with 42% for the NRC equation. Compared with the relative prediction error of 6% for the exponential model, that associated with prediction using the NRC equation was 14%. The overall mean square prediction error value for individual cows was 5-fold higher for the NRC equation than for the exponential model (10.4 vs. 2.0 kg2/d2). The consistently accurate and robust prediction of DMI by the exponential model for all data-sets suggested that it could safely be used for predicting DMI in many circumstances.

Animal Nutritional Physiological Phenomena↗
PubMed2006

Predictions of type 2 diabetes and complications in Greenland in 2014.

OBJECTIVES: The objective of this study was to predict the prevalence of type 2 diabetes and the associated burden to the health care system in Greenland posed by diabetic complications by 2014. The predictions were based on changes in demographic variables and obesity. STUDY DESIGN: Projection model based on two cross-sectional population surveys from 1993 and 1999. METHODS: The development in BMI was described and projected to 2014 under two assumptions: 1) distribution of BMI is constant from 1999, and 2) the trend in BMI found in the surveys will continue until 2014. The prevalence of type 2 diabetes was predicted under these assumptions and based on the observed association between BMI and type 2 diabetes. The prevalence of complications was estimated using the 2nd assumption, as was the prevalence of hypertension, dyslipidemia, Ischemic Heart Disease (IHD) and stroke in the non-diabetic population in 2014. RESULTS: The prevalence of type 2 diabetes was not predicted to increase by 2014 under the 1st assumption. It was predicted to increase from 11% to 23% for women, but not for men under the 2nd assumption. Approximately half of the cases of cardiovascular disease and cardiovascular risk factors predicted by 2014 were attributable to diabetes. CONCLUSIONS: The prevalence of type 2 diabetes was predicted to increase in Greenland, and the number of complications was predicted to double from 1999 to 2014. Both prophylactic and treatment initiatives are needed to deal with the extra burden posed by type 2 diabetes to the Greenlandic health care system in 2014.

Adult↗
PubMed2005

Prediction of HLA-A2-restricted CTL epitope specific to HCC by SYFPEITHI combined with polynomial method.

AIM: To predict the HLA-A2-restricted CTL epitopes of tumor antigens associated with hepatocellular carcinoma (HCC). METHODS: MAGE-1, MAGE-3, MAGE-8, P53 and AFP were selected as objective antigens in this study for the close association with HCC. The HLA-A*0201 restricted CTL epitopes of objective tumor antigens were predicted by SYFPEITHI prediction method combined with the polynomial quantitative motifs method. The threshold of polynomial scores was set to -24. RESULTS: The SYFPEITHI prediction values of all possible nonamers of a given protein sequence were added together and the ten high-scoring peptides of each protein were chosen for further analysis in primary prediction. Thirty-five candidates of CTL epitopes (nonamers) derived from the primary prediction results were selected by analyzing with the polynomial method and compared with reported CTL epitopes. CONCLUSION: The combination of SYFPEITHI prediction method and polynomial method can improve the prediction efficiency and accuracy. These nonamers may be useful in the design of therapeutic peptide vaccine for HCC and as immunotherapeutic strategies against HCC after identified by immunology experiment.

Amino Acid Sequence↗
PubMed2002

When wrong predictions provide more support than right ones.

Correct predictions of rare events are normatively more supportive of a theory or hypothesis than correct predictions of common ones. In other words, correct bold predictions provide more support than do correct timid predictions. Are lay hypothesis testers sensitive to the boldness of predictions? Results reported here show that participants were very sensitive to boldness, often finding incorrect bold predictions more supportive than correct timid ones. Participants were willing to tolerate inaccurate predictions only when predictions were bold. This finding was demonstrated in the context of competing forecasters and in the context of competing scientific theories. The results support recent views of human inference that postulate that lay hypothesis testers are sensitive to the rarity of data. Furthermore, a normative (Bayesian) account can explain the present results and provides an alternative interpretation of similar results that have been explained using a purely descriptive model.

Adult↗
PubMedDate not supplied

Topographic predicted corneal acuity with intrastromal corneal ring segments.

PURPOSE: To evaluate predicted optical quality of the central anterior corneal surface before and after the intrastromal corneal ring segment (ICRS) refractive procedure using a clinical videokeratoscope and software index developed for that purpose. METHODS: Predicted corneal acuity, a topographically derived index provided with the EyeSys System 2000 videokeratscope, representing potential optical quality of the cornea, was assessed preoperatively and at postoperative month 3 in 94 eyes that received an ICRS to treat -1.00 to -6.00 D of myopia. Predicted corneal acuity was calculated by determining the difference between a measured cornea and its best-fit ellipses for reflected ring circumferences within the central 3 mm diameter zone. RESULTS: Preoperative predicted corneal acuity was 20/10 in 92 of 94 eyes (98%). At month 3 after the ICRS procedure, 48 (51%) of moderately myopic eyes were corrected to 20/20 or better, 96% (90 eyes) were corrected to 20/40 or better, and 98% of eyes (92 eyes) had a predicted corneal acuity of 20/10. For the eyes with a predicted corneal acuity of 20/10, spectacle-corrected visual acuity was normally distributed between 20/10 and 20/25. CONCLUSION: Predicted corneal acuity did not change significantly from baseline in eyes with an ICRS. This suggests that topographic irregularities in the central 3 mm of the cornea detectable by predicted corneal acuity software were not induced in the central cornea with the ICRS.

Corneal Stroma↗
PubMed2000

Validation and analysis of modeled predictions of growth of Bacillus cereus spores in boiled rice.

The growth of psychrotrophic Bacillus cereus 404 from spores in boiled rice was examined experimentally at 15, 20, and 30 degrees C. Using the Gompertz function, observed growth was modeled, and these kinetic values were compared with kinetic values for the growth of mesophilic vegetative cells as predicted by the U.S. Department of Agriculture's Pathogen Modeling Program, version 5.1. An analysis of variance indicated no statistically significant difference between observed and predicted values. A graphical comparison of kinetic values demonstrated that modeled predictions were "fail safe" for generation time and exponential growth rate at all temperatures. The model also was fail safe for lag-phase duration at 20 and 30 degrees C but not at 15 degrees C. Bias factors of 0.55, 0.82, and 1.82 for generation time, lag-phase duration, and exponential growth rate, respectively, indicated that the model generally was fail safe and hence provided a margin of safety in its growth predictions. Accuracy factors of 1.82, 1.60, and 1.82 for generation time, lag-phase duration, and exponential growth rate, respectively, quantitatively demonstrated the degree of difference between predicted and observed values. Although the Pathogen Modeling Program produced reasonably accurate predictions of the growth of psychrotrophic B. cereus from spores in boiled rice, the margin of safety provided by the model may be more conservative than desired for some applications. It is recommended that if microbial growth modeling is to be applied to any food safety or processing situation, it is best to validate the model before use. Once experimental data are gathered, graphical and quantitative methods of analysis can be useful tools for evaluating specific trends in model prediction and identifying important deviations between predicted and observed data.

Bacillus cereus↗
PubMed2006

Predicting the duration of sickness absence for patients with common mental disorders in occupational health care.

OBJECTIVES: This study attempted to determine the factors that best predict the duration of absence from work among employees with common mental disorders. METHODS: A cohort of 188 employees, of whom 102 were teachers, on sick leave with common mental disorders was followed for 1 year. Only information potentially available to the occupational physician during a first consultation was included in the predictive model. The predictive power of the variables was tested using Cox's regression analysis with a stepwise backward selection procedure. The hazard ratios (HR) from the final model were used to deduce a simple prediction rule. The resulting prognostic scores were then used to predict the probability of not returning to work after 3, 6, and 12 months. Calculating the area under the curve from the ROC (receiver operating characteristic) curve tested the discriminative ability of the prediction rule. RESULTS: The final Cox's regression model produced the following four predictors of a longer time until return to work: age older than 50 years [HR 0.5, 95% confidence interval (95% CI) 0.3-0.8], expectation of duration absence longer than 3 months (HR 0.5, 95% CI 0.3-0.8), higher educational level (HR 0.5, 95% CI 0.3-0.8), and diagnosis depression or anxiety disorder (HR 0.7, 95% CI 0.4-0.9). The resulting prognostic score yielded areas under the curves ranging from 0.68 to 0.73, which represent acceptable discrimination of the rule. CONCLUSIONS: A prediction rule based on four simple variables can be used by occupational physicians to identify unfavorable cases and to predict the duration of sickness absence.

Adjustment Disorders↗
PubMed1994

A clinical rule to predict preserved left ventricular ejection fraction in patients after myocardial infarction.

OBJECTIVE: To derive and validate a clinical prediction rule that identifies patients after myocardial infarction who have preserved left ventricular systolic function. DESIGN: Retrospective analysis of a prospective cohort study, with a derivation set to generate a clinical prediction rule and a validation set to test the prediction rule. SETTING: Urban tertiary care hospital. PATIENTS: 314 consecutive patients admitted with myocardial infarction who had one or more of the following tests to determine left ventricular ejection fraction: transthoracic echocardiography, contrast left ventriculography, or radionuclide ventriculography. MEASUREMENTS: Left ventricular ejection fractions were determined by transthoracic echocardiography, contrast left ventriculography, and gated blood pool scan. RESULTS: Multivariate analysis of patients in the derivation set yielded the following rule: The left ventricular ejection fraction is predicted to be 40% or more in patients who have 1) an interpretable electrocardiogram, 2) no previous Q-wave myocardial infarction, 3) no history of congestive heart failure, and 4) an index myocardial infarction that is not a Q-wave anterior infarction. In the derivation and the validation sets, the positive predictive value of the prediction rule was more than 0.98. CONCLUSIONS: A simple clinical prediction rule using easily obtained historical and electrocardiographic data reliably identifies a substantial percentage of patients after myocardial infarction (40% in our hospital) who are likely to have preserved left ventricular systolic function. If validated in other patient populations, application of this prediction rule in clinical practice could result in a substantial decrease in the cost of treating uncomplicated myocardial infarction.

Aged↗
PubMed1999

Length of pregnancy in African Americans: validation of a new predictive rule.

This study evaluated whether a new predictive rule is more accurate for estimating the length of pregnancy in African Americans than Nägele's rule, the accepted standard. After identifying women in early pregnancy, telephone interviews were conducted to obtain information about 16 previously established determinants of gestational length. Based on these data, a linear multivariate regression model was used to predict an estimated delivery date (EDD) for each mother. In addition, the EDD was determined using Nägele's rule. Later, the actual delivery date was compared with the EDD predicted by the new rule and with the EDD predicted by Nägele's rule. Each pregnancy was assigned to its better prediction group, either the new rule's group or the Nägele's rule group. Fifty-seven pregnancies were identified prospectively and monitored. The new rule predicted the actual delivery date more accurately in 66% (37/56) of pregnancies, Nägele's rule was a better predictor in 34% (19/56) of pregnancies, and both rules were equally accurate in predicting the delivery date for one pregnancy. The new rule was more precise than Nägele's rule (P = .022) when the binomial distribution was used. When using the linear regression model rule, a more accurate EDD can be determined for African-American women. Moreover, it is possible to predict the risk of preterm delivery (those occurring > 3 weeks earlier than the EDD).

Adult↗
PubMed2000

Conscious and unconscious processing of nonverbal predictability in Wernicke's area.

The association of nonverbal predictability and brain activation was examined using functional magnetic resonance imaging in humans. Participants regarded four squares displayed horizontally across a screen and counted the incidence of a particular color. A repeating spatial sequence with varying levels of predictability was embedded within a random color presentation. Both Wernicke's area and its right homolog displayed a negative correlation with temporal predictability, and this effect was independent of individuals' conscious awareness of the sequence. When individuals were made aware of the underlying sequential predictability, a widespread network of cortical regions displayed activity that correlated with the predictability. Conscious processing of predictability resulted in a positive correlation to activity in right prefrontal cortex but a negative correlation in posterior parietal cortex. These results suggest that conscious processing of predictability invokes a large-scale cortical network, but independently of awareness, Wernicke's area processes predictive events in time and may not be exclusively associated with language.

Adult↗
PubMedDate not supplied

Prediction of plasma levels of aminoglycoside antibiotic in patients with severe illness by means of an artificial neural network simulator.

PURPOSE: The purpose of this work was to predict plasma peak and trough levels of an aminoglycoside antibiotic in patients with severe illness in an intensive care unit by a novel approach. Plasma levels were predicted based on the values of 15 physiological measurements using an artificial neural network (ANN) simulator. METHOD: A data set of 15 physiological measurements for 30 patients was used to develop the model. The ANN structure consisted of three layers: an input layer comprised of 15 processing elements, a hidden layer comprised of 10 processing elements with a sigmoid function as an activation function, and an output layer of two processing elements (peak and trough levels). The weight between neurons was trained according to the delta rule back-propagation of errors algorithm. Predicted values were obtained by "leave-one-out" experiments by both ANN and multiple linear regression analysis (MLRA). RESULTS: The correlation coefficients between observed and predicted values obtained by ANN prediction using standardized data sets were r=0.825 and r=0.854 for peak and trough levels, respectively. The correlation coefficients obtained by MLRA were r=0. 037 and r=0.276 for peak and trough levels, respectively. These results indicate that ANN shows better performance in prediction of aminoglycoside plasma levels from patients' physiological measurements than MLRA. CONCLUSIONS: Prediction of plasma levels of antibiotic in patients with severe illness by ANN was superior to the standard statistical method. Standardization of input data was found to be important for better prediction. ANN has some advantages over standard statistical methods, as it can recognize complex relationships in the data.

Aminoglycosides↗