Statistical notes. Regression towards the mean.
Explore the source record for details and available documents.
Adult↗
SEARCH · Search PubMed
Results for “Statistical Regression”
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.
[Media optimization for the novel antimicrobial peptide by Bacillus sp. fmbJ224].
The novel antimicrobial peptide in submerged fermentation by Bacillus sp. fmbJ224 is strongly influenced by many internal and external factors, namely medium constituents and fermentation conditions. In this study, Plackett-Burman design was undertaken to evaluate the effects of the seventeen factors. By the statistical regression analysis, the significant factors affecting the novel antimicrobial peptide in submerged fermentation by Bacillus sp. fmbJ224 were determined as follows: glucose, NH4NO3, glutamic acid, CaCl2, MnSO4. In the second phase of the optimization process, a response surface methodology (RSM) was used to optimize the above critical internal factors, and to find out the optimization concentraction levels and the relationships between these factors. By solving the quadratic regression model equation using appropriate statistic methods, the optimal concentration of the variables were determined as: 8.13 g/L glucose, 6.14 g/L NH4NO3, 4.2 g/L glutamic acid, 3.98 mg/L CaCl2, 4.87 mg/L MnSO4. The content of the novel antimicrobial peptide was increased from 1304.21 microg/mL to 1487.58 microg/mL. The experimental data under various conditions have validated the theoretical values.
Anti-Infective Agents↗
Learning anchor verbs for biological interaction patterns from published text articles.
Much of knowledge modeling in the molecular biology domain involves interactions between proteins, genes, various forms of RNA, small molecules, etc. Interactions between these substances are typically extracted and codified manually, increasing the cost and time for modeling and substantially limiting the coverage of the resulting knowledge base. In this paper, we describe an automatic system that learns from text interaction verbs; these verbs can then form the core of automatically retrieved patterns which model classes of biological interactions. We investigate text features relating verbs with genes and proteins, and apply statistical tests and a logistic regression statistical model to determine whether a given verb belongs to the class of interaction verbs. Our system, AVAD, achieves over 87% precision and 82% recall when tested on an 11 million word corpus of journal articles. In addition, we compare the automatically obtained results with a manually constructed database of interaction verbs and show that the automatic approach can significantly enrich the manual list by detecting rarer interaction verbs that were omitted from the database.
Artificial Intelligence↗
Validity of regression-based norms: an empirical test of the comprehensive norms with older adults.
This study empirically tested Fastenau and Adams' (1996) concerns about using regression-based norms (RBN) to statistically correct for demographic influences. A widely used RBN system (Heaton, Grant, & Matthews, 1991) was applied to scores from 63 healthy older adults for the Trail Making Test, Boston Naming Test, and Wisconsin Card Sorting Test (WCST). Age influences on all tests were virtually eliminated by the RBNs, whereas education influences were created by the RBN (better educated older adults penalized by the norms). Using RBNs from the Revised WCST Manual also created distortions, far overcorrecting for age. These findings cast considerable suspicion on norms that are generated by statistical regression when there are not adequate numbers of people supporting each demographic cell. Clinically, these norms can lead to higher rates of false negatives (or missed diagnoses) in older adults, especially those with less education and especially women in their 60s.
Adult↗
Sociologic factors are major determinants of prolonged hospital stay after abdominal aneurysm repair.
BACKGROUND: With increasing pressure to optimize the utilization of hospital resources, it is important to identify patients who may have prolonged hospital length of stay (LOS). The purpose of this report was to identify the preoperative variables that are predictive of prolonged postoperative hospital LOS for patients undergoing elective infrarenal abdominal aneurysm repair and to discuss strategies that might assist in minimizing LOS for these patients. METHODS: Three hundred sixty-five consecutive patients underwent elective infrarenal abdominal aneurysm repair between 1989 and 1994. The relationship between 13 preoperative variables and LOS was analyzed by using both univariate (Kaplan-Meier) and multivariate (Cox regression) statistical techniques. RESULTS: By using Cox regression a model was developed to estimate LOS (p < 0.001 for model). The independent predictors for prolonged LOS were (1) age older than 70 years and (2) absence of a spouse. CONCLUSIONS: Knowledge of the predictive factors that are associated with prolonged LOS should identify those patients who may require prompt and efficient discharge planning, early consultation with a home care nurse, or transfer to a convalescent facility. This approach may significantly improve the utilization of hospital resources.
Age Factors↗
Cytomegalovirus infection in infants: an example of a chronopharmacological approach.
A case of a male infant at the age of 3 months with cytomegalovirus (CMV) infection, diagnosed 40 days after the first symptoms, was discussed. While the right dosage and schedule during the initial treatment with ganciclovir (Cymevene) were agreed on, the right time of further application of anti-CMV IgG (Cytotest) and ganciclovir was unclear. Daily mean temperature was taken as an overall measure based on 7 single readings (6:00, 9:00, 12:00, 15:00, 18:00, 21:00 and 24:00 h +/- 10 min). Dynamics of daily mean temperature was studied by descriptive statistics, regression analysis, autocorrelation and periodogram regression analysis (software package "6-D Statistics" PC ver.4.5-98 by B. P. Komitov) with the purpose of identifying the time periods of viral DNA replications and CMV population growth inclinations. A five-step procedure was applied: (1) description of a tendency in the daily mean temperature; (2) studying variations in time series of the daily mean temperature; (3) decomposition of cyclic variations; (4) reconstruction of time series; (5) best model determination and forecasting. Three main cycles in variations of daily mean temperature were revealed (period T approximately 2.25-3.25, 5.5 and 12.25 days, p < 0.05) until the 37th day of therapy, when a strongly decreasing trend of daily mean temperature emerged. The cycle of 2.25-3.25 days disappeared and the daily mean temperature continued to decrease significantly on the background of improvement of the clinical status. It was concluded that (1) cyclic patterns in daily mean temperature (periods T = 2.25-3.25 and 12.25 days) during the etiologic medication stages could be due to the life-cycle of 2-4 days of the cytomegalovirus and a cycle in the viral population growth, respectively. The above findings confirmed previous results on the period of viral DNA replication from in vitro studies. It was possible to forecast the right moment of replication and viral load to adjust the treatment schemes and improve the outcome; (2) the complex time-series approach has shown to be very useful and effective in analyzing and forecasting the temporal dynamics of daily mean temperature in order to optimize the clinical management in this particular case of an infant with CMV infection.
Antiviral Agents↗
Heart rate variability and particulate exposure in vehicle maintenance workers: a pilot study.
The association between occupational exposure to PM(2.5) and heart rate variability was investigated in a repeated measures, longitudinal study of vehicle maintenance workers occupationally exposed to automobile emissions. Five subjects were monitored for occupational exposure to fine particulate matter (PM(2.5)) on 6 workdays using an aerosol photometer, validated with side-by-side sampling with a gravimetric method. End-of-day heart rate variability statistics were derived using short-term electrocardiogram recordings for each participant. Workplace carbon monoxide and outdoor, ambient fine particulate matter were also monitored. Regression statistics were used to investigate associations between same-day PM(2.5) levels and heart rate variability statistics using mixed-effects multiple regression of pooled data. No statistically significant associations were observed between occupational PM(2.5) and measures of heart rate variability. A statistically significant increase in total spectral power was associated with ambient PM(2.5) (p < 0.05). The data suggest a threshold below which no degradation in cardiac autonomic control of healthy workers occurs when challenged by occupational PM(2.5) exposure. This study was limited in population, exposure level, and type of particulate exposures. Additional studies are recommended on broader occupational populations.
Adult↗
Elements of a predictive model for determining beach closures on a real time basis: the case of 63rd Street Beach Chicago.
Data on hydrometeorological conditions and E. coli concentration were simultaneously collected on 57 occasions during the summer of 2000 at 63rd Street Beach, Chicago, Illinois. The data were used to identify and calibrate a statistical regression model aimed at predicting when the bacterial concentration of the beach water was above or below the level considered safe for full body contact. A wide range of hydrological, meteorological, and water quality variables were evaluated as possible predictive variables. These included wind speed and direction, incoming solar radiation (insolation), various time frames of rainfall, air temperature, lake stage and wave height, and water temperature, specific conductance, dissolved oxygen, pH, and turbidity. The best-fit model combined real-time measurements of wind direction and speed (onshore component of resultant wind vector), rainfall, insolation, lake stage, water temperature and turbidity to predict the geometric mean E. coli concentration in the swimming zone of the beach. The model, which contained both additive and multiplicative (interaction) terms, accounted for 71% of the observed variability in the log E. coli concentrations. A comparison between model predictions of when the beach should be closed and when the actual bacterial concentrations were above or below the 235 cfu 100 ml(-1) threshold value, indicated that the model accurately predicted openings versus closures 88% of the time.
Bathing Beaches↗
Comparison of a commercial method for total protein with a candidate reference method.
The biuret method for total protein has been compared on the Boehringer Mannheim Diagnostics (BMD) Hitachi 717 Analyzer with a candidate reference method in efforts to standardize the BMD method. The methods were compared using 115 paired serum specimens collected during morning rounds at the Roger Williams Hospital. Results from the test method were compared to the reference method using a statistical procedure for quantifying bias between analytical methods. Linear regression statistics were also calculated. The Hitachi 717 biuret method shows no bias when compared to the reference method (z = -0.90), and there is acceptable correlation between the two methods (y = 0.86, m = 0.86, r = 0.896). The Hitachi method is linear to 15.0 g per dL and demonstrates excellent precision (CV less than or equal to 2.14 percent, N = 140). The Hitachi 717 biuret method has been found to be excellent in all respects and its use is recommended as a convenient and accurate means of measuring total protein.
Blood Chemical Analysis↗
Centring in regression analyses: a strategy to prevent errors in statistical inference.
Regression analyses are perhaps the most widely used statistical tools in medical research. Centring in regression analyses seldom appears to be covered in training and is not commonly reported in research papers. Centring is the process of selecting a reference value for each predictor and coding the data based on that reference value so that each regression coefficient that is estimated and tested is relevant to the research question. Using non-centred data in regression analysis, which refers to the common practice of entering predictors in their original score format, often leads to inconsistent and misleading results. There is very little cost to unnecessary centring, but the costs of not centring when it is necessary can be major. Thus, it would be better always to centre in regression analyses. We propose a simple default centring strategy: (1) code all binary independent variables +1/2; (2) code all ordinal independent variables as deviations from their median; (3) code all 'dummy variables' for categorical independent variables having m possible responses as 1 - 1/m and -1/m instead of 1 and 0; (4) compute interaction terms from centred predictors. Using this default strategy when there is no compelling evidence to centre protects against most errors in statistical inference and its routine use sensitizes users to centring issues.
Bias↗
Recombinant human erythropoietin and overall survival in cancer patients: results of a comprehensive meta-analysis.
BACKGROUND: Anemia associated with cancer and cancer therapy is an important clinical and economic factor in the treatment of malignant diseases. METHODS: We conducted a systematic literature review to assess the efficacy of erythropoietin to prevent or treat anemia in cancer patients with regard to red blood cell transfusions, hematologic response, adverse events, and overall survival. We searched the Cochrane Library, Medline, EMBASE, and other databases for relevant articles published from January 1985 to December 2001. We included all randomized controlled trials that compared the use of recombinant human erythropoietin (plus transfusion, if needed) with no erythropoietin treatment (plus transfusion, if needed). Relative risks (RRs) and 95% confidence intervals (CIs) were calculated under a fixed-effects model. Clinical and statistical heterogeneity were examined with sensitivity analyses and meta-regression. Statistical tests for effect estimates were two-sided. RESULTS: We identified 27 trials involving 3287 adult patients. Patients treated with erythropoietin had a lower relative risk of having a blood transfusion than untreated patients (RR = 0.67, 95% CI = 0.62 to 0.73). Erythropoietin-treated patients with baseline hemoglobin levels lower than 10 g/dL were more likely to have a hematologic response than untreated patients (RR = 3.60, 95% CI = 3.07 to 4.23). The relative risk for thromboembolic complications after erythropoietin treatment was not statistically significantly increased (RR = 1.58, 95% CI = 0.94 to 2.66) compared with that of untreated patients. There is suggestive but inconclusive evidence that erythropoietin may improve overall survival (adjusted data: hazard ratio [HR] = 0.81, 95% CI = 0.67 to 0.99; unadjusted data: HR = 0.84, 95% CI = 0.69 to 1.02). CONCLUSIONS: Erythropoietin treatment may reduce the risk for blood transfusions and improve hematologic response in cancer patients. However, our favorable survival outcome is in contrast to two large (N = 351 and 939) recently published randomized controlled trials in which erythropoietin-treated patients had statistically significantly worse survival than untreated patients. Possible reasons for the disparity with our results include differences in study population and design, higher target hemoglobin levels and higher risk of thromboembolic complications, and concerns that erythropoietin may stimulate tumor growth.
Anemia↗
Evolution and generalization of a single neurone. III. Primitive, regularized, standard, robust and minimax regressions.
We show that during training the single layer perceptron, one can obtain six conventional statistical regressions: a primitive, regularized, standard, the standard with the pseudo-inversion of the covariance matrix, robust, and minimax (support vector). The complexity of the regression equation increases with an increase in the number of iterations. The generalization accuracy depends on the type of the regression obtained during the training, on the data, learning-set size, and, in certain cases, on the distribution of components of the weight vector. For small intrinsic dimensionality of the data and certain distributions of components of the weight vector the single layer perceptron can be trained even with very short learning sequences. The type of the regression obtained in SLP training should be controlled by the sort of cost function as well as by training parameters (the number of iterations, learning step, etc.). Whitening data transformation prior to training the perceptron is a tool to incorporate a prior information into the prediction rule design, and helps both to diminish the generalization error and the training time.
Biological Evolution↗
Regression models of emergency medical service demand for different types of emergencies.
Second-order statistical regression models are developed for the rates of demand for different types of emergency medical services (EMS) as they relate to various socioeconomic, demographic, and other characteristics of a service area. The model parameters are estimated by the recent technique of ridge regression, which is shown to provide superior estimates to those of the ordinary least squares regression method, because of the presence of substantial multicollinearity (nonorthogonality) in the exogenous data set. The ridge results are also compared with those derived from a related Bayesian approach. The resulting models provide substantial fits to the empirical data for the EMS system of the city of Atlanta, GA.
Catchment Area, Health↗
A comparison of myocardial perfusion gated SPECT studies obtained at rest and after stress.
OBJECTIVE: To perform a comparison between post-stress and rest gated single-photon emission computer tomography (SPECT) myocardial perfusion imaging (MPI) studies to assess post-stress stunning and the variables underlying this phenomenon. SUBJECTS AND METHODS: This was a prospective study of 318 consecutive adult patients undergoing stress and rest gated SPECT using a 2-day (99m)Tc-tetrofosmin protocol. Bruce protocol treadmill stress (n = 93) or i.v. dipyridamole pharmacologic stress (n = 225) were used as stressors. Ejection fractions (EF) and left ventricular (LV) end-diastolic (EDV) and end-systolic volumes (ESV) were calculated using the Cedars Sinai Quantitative Gated SPECT software. Perfusion defects were visually scored using a 20-segment model to obtain summed stress scores. Statistical analysis was performed by applying paired t test and multi-regression. Statistically significant (p < 0.05) paired differences between post-stress and rest EF and ESV and type of stressing were noted. RESULTS: Analysis of patient groups based on the type of stress showed significantly low EF on post-stress studies for both treadmill and dipyridamole stressing and also significantly high ESV on post-stress studies for dipyridamole. Multi-regression analysis using differences in post-stress and rest EF, EDV and ESV as dependent and summed stress scores, type of stress, delay time, previous myocardial infarction and size of LV as independent variables showed statistically significant associations between high summed stress scores (>13) and size of post-stress LV for EDV and ESV differences. High volumes were noted on post-stress studies. The magnitudes of the observed differences were well within the reproducibility of LV volume calculations. CONCLUSIONS: Patients showing stunning had significant ischemia or a dilated LV. The stunning manifested as high EDV and ESV differences. The magnitude of the differences observed in EF and LV volumes were not clinically significant.
Adult↗
Is perioperative blood transfusion a risk factor for mortality or infection after hip fracture?
OBJECTIVES: To assess whether allogeneic blood transfusion in the perioperative period is associated with changes in mortality or complication rates in patients undergoing surgical treatment for hip fracture (proximal femoral fracture). DESIGN: Retrospective case-control series, all patients followed up for 1 year or until death. SETTING: District General Hospital in Peterborough, UK. PATIENTS PARTICIPANTS: Three thousand six hundred twenty-five consecutive patients admitted and operated for hip fracture (proximal femoral fracture) during July 1989 to January 2002 (151 months); 1068 (29.9%) received a perioperative allogeneic blood transfusion. MAIN OUTCOME MEASURES: Thirty- 120-, and 365-day mortality, deep and superficial wound infection rates. RESULTS: Overall mortality for all patients at 1 year post fracture was 28.2% (1007 patients). Transfusion was associated with a statistically significant increase in mortality from 120 days onward after hip fracture. However, when this was adjusted with a statistical regression model for baseline characteristics and confounding variables, this difference became statistically insignificant (P = 0.17). Infection rates in the transfusion group were 2.0% for superficial infection and 0.9% for deep infection compared with 1.9% and 0.6%, respectively, in the nontransfusion group. These figures were not statistically significantly different. Other complications of deep venous thrombosis, chest infection, and congestive cardiac failure showed no statistically significant increase in those patients who received transfusion. CONCLUSIONS: Our data suggest that transfusion is not associated with a change in mortality or infection rates in the hip-fracture patient.
Aged↗
Risk factors for cerebral palsy: a case-control study in Greece.
The purpose of this study was to investigate the relation between a series of maternal, antenatal, perinatal, socioeconomic and environmental variables and the occurrence of cerebral palsy (CP) in a setting different from those in which previous analytic epidemiologic studies had been undertaken. The study was of case-control design and included 103 children with cerebral palsy born between 1984 and 1988 and residents of the Greater Athens area at any time during 1991 and 1992. Controls were chosen among the neighbors of the index case or were healthy siblings of children with neurological diseases other than CP seen by the same neurologists as the children with CP; a total of 254 control children were eventually included. Statistical analysis was done by modeling the data through unconditional logistic regression. Statistically significant (p < 0.05) risk factors of potential causal importance were: twin membership (OR = 10.2), gestational age (OR = 0.5 per 4 weeks), birth weight conditional on gestational age (OR = 0.9 per 100 g), congenital malformations (OR = 7.5), unhealthy placenta (OR = 6.6), placenta previa (6 cases, no controls), abnormal amniotic fluid (OR = 3.6), head circumference more than 36 cm (OR = 9.0), general anesthesia during labor (OR = 4.3), forceps delivery (OR = 6.8), and birth trauma (OR = 11.5). Among children with no identifiable prenatal risk factors there was no excess prevalence of one or more perinatal risk factors in CP cases compared to controls, which implies that the latter factors impart their effect through interactions with co-existing prenatal or other risk factors.
Case-Control Studies↗
Graph valence shells as molecular descriptors.
We have introduced a new simple structural descriptor for molecules that is based on the count of the valence shells for vertices in molecular graphs. The construction of the new descriptor is illustrated on 2,3-dimethylhexane and is reported for the 18 octane isomers. The relationship of the new descriptor to the path numbers of a graph is discussed. It can be seen that the path counts and the count of valence of neighbor shells are related for paths of length two (and shells of range two). There is no appreciable correlation between the count of the longer paths and the count of the corresponding neighbor valence shells at larger separations. Use of the neighbor valence shells as molecular descriptors is illustrated on the boiling point, the entropy, and the density of octanes. An intriguing situation is observed for regressions involving considered properties of n-octane isomers C8H18 in that the paths of length two, three, and four and the shells of the range two, three, and four give identical multivariate regression statistics. An explanation for this somewhat unusual aspect of MRA (multiple regression analysis) is offered.
Journal Article↗
The impact of minor perioperative anesthesia-related incidents, events, and complications on postanesthesia care unit utilization.
UNLABELLED: The German Society of Anesthesiology and Intensive Care Medicine evaluates the standardized and routine reporting of perioperative anesthesia-related incidents, events, and complications (IEC). As part of the long-term project's definitions, IECs are graded according to severity and to their clinical consequence on further postanesthesia monitoring and treatment demands. The adult study population of our department comprised 37,079 patients recovering from anesthesia in a tertiary university hospital from July 1992 through June 1997. Cardiac, obstetric, craniotomy, thoracotomy, laparotomy, and emergency operations were excluded. Multivariate regression statistics were used to calibrate the impact of minor graded IECs on necessary postanesthesia care unit (PACU) utilization. Minor and severe IECs appeared in 22.1% and 0.2% of the patients. A minor IEC occurrence was a statistically significant (P < 0.001) predictor of PACU utilization in a multivariate regression model. The mean difference of PACU length of stay for patients with minor IECs was prolonged by a range of 6%-26% when adjusted for coexisting severity features such as age, gender, ASA physical status, and type and duration of anesthesia and surgery. We conclude that the IEC methodology integrates epidemiologic information about perioperative anesthesia outcome. Minor but frequently occurring IECs have an impact on PACU utilization and are thus important to measure and follow. IMPLICATIONS: It is desirable to know how anesthesia-related incidents, events, and complications influence postanesthesia care. Analyses of standardized and routine perioperative outcome data, as proposed by the German anesthesia quality project, can show that even minor events consume relevant resources and are thus important to measure and follow.
Adolescent↗