SEARCH · PubMed Health

Results for “Logistic Models”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 667 records · Page 37Linked to original sources

PubMed2005

Equally valid models gave divergent predictions for mortality in acute myocardial infarction patients in a comparison of logistic [corrected] regression models.

OBJECTIVE: Models that predict mortality after acute myocardial infarction (AMI) contain different predictors and are based on different populations. We studied the agreement and validity of predictions for individual patients. STUDY DESIGN AND SETTING: We compared predictions from five predictive logistic regression models for short-term mortality after AMI. Three models were developed previously, and two models were developed in the GUSTO-I data, where all five models were applied (n =40,830, 7.0% 30-day mortality). Agreement was studied with weighted kappa statistics of categorized predictions. Validity was assessed by comparing observed frequencies with predictions (indicating calibration) and by the area under the receiver operating characteristic curve (AUC), indicating discriminative ability. RESULTS: The predictions from the five models varied considerably for individual patients, with low agreement between most (kappa <0.6). Risk predictions from the three previously developed models were on average too high, which could be corrected by re-calibration of the model intercept. The AUC ranged from 0.76-0.78 and increased to 0.78-0.79 with re-estimated regression coefficients that were optimal for the GUSTO-I patients. The two more detailed GUSTO-I based models performed better (AUC approximately 0.82). CONCLUSION: Models with different predictors may have a similar validity while the agreement between predictions for individual patients is poor. The main concerns in the applicability of predictive models for AMI should relate to the selected predictors and average calibration.

Aged↗

PubMed1984

Random-effects models for serial observations with binary response.

This paper presents a general mixed model for the analysis of serial dichotomous responses provided by a panel of study participants. Each subject's serial responses are assumed to arise from a logistic model, but with regression coefficients that vary between subjects. The logistic regression parameters are assumed to be normally distributed in the population. Inference is based upon maximum likelihood estimation of fixed effects and variance components, and empirical Bayes estimation of random effects. Exact solutions are analytically and computationally infeasible, but an approximation based on the mode of the posterior distribution of the random parameters is proposed, and is implemented by means of the EM algorithm. This approximate method is compared with a simpler two-step method proposed by Korn and Whittemore (1979, Biometrics 35, 795-804), using data from a panel study of asthmatics originally described in that paper. One advantage of the estimation strategy described here is the ability to use all of the data, including that from subjects with insufficient data to permit fitting of a separate logistic regression model, as required by the Korn and Whittemore method. However, the new method is computationally intensive.

Air Pollution↗

PubMed2003

Taxonomic level as a determinant of the shape of the Phanerozoic marine biodiversity curve.

Key aims of recent paleobiological research have been the construction of Phanerozoic global biodiversity patterns and the formulation of models and mechanisms of diversification describing such patterns. Two conflicting theories of global diversification have been equilibrium versus expansionist growth of taxonomic diversity. These models, however, rely on accurate empirical data curves, and it is not clear to what extent the taxonomic level at which the data are analyzed controls the resulting pattern. Global Phanerozoic marine diversity curves are constructed at ordinal, familial, and generic levels using several fossil-range data sets. The fit of a single logistic model reduces from ordinal through familial to generic level, while conversely, that of an exponential growth model increases. Three sequential logistic equations, fitted to three time periods during which diversity appears to approach or reach an equilibrium state, provide the best description of the data at familial and generic levels. However, an exponential growth curve describes the diversification of marine life since the end-Permian extinction equally as well as a logistic. A species-level model of global Phanerozoic marine diversification, constructed by extrapolation of the trends from familial to generic level, suggests growth in numbers of marine species was broadly exponential. When smaller subsets of the data are analyzed, the effect of taxonomic level on the shape of the diversity curve becomes more pronounced. In the absence of species data, a consistent signal at more than one higher taxonomic level is required to predict a species-level pattern.

Ecosystem↗

PubMed2006

Promoting breastfeeding in Bolivia: do social networks add to the predictive value of traditional socioeconomic characteristics?

This study tested whether the prediction of health-related knowledge (correct breastfeeding practices in this case) could be improved by including information about the composition of an individual's personal network above and beyond that predicted by his/her socioeconomic or demographic characteristics. Few studies have tested the predictive value of social networks, especially for population-based studies, despite an increased use of social networks in the past few years in several fields of health research, especially in research relating to prevention of HIV/AIDS and design of HIV/AIDS programmes. Promotion of breastfeeding practices that enhance child survival is important in Bolivia because of high infant morbidity and mortality in the country. Data on a cross-sectional urban probability sample of 2,354 women and men aged 15-49 years were collected from seven urban areas in Bolivia. Model building and the log likelihood ratio criteria were used for assessing the significance of variables in a logistic model. Results showed that the network variables added significantly (p < 0.05 for knowledge of breastfeeding only with no other liquids and for knowledge of breastfeeding only with no solids p < 0.01) to the predictive power of the socioeconomic variables. These results may also hold for other health research areas, increasingly using social network analysis, such as that of HIV/AIDS.

Adolescent↗

PubMed2002

General health status and changes in chewing ability in older Canadians over seven years.

OBJECTIVES: The purpose of this study was to describe the onset of and recovery from chewing problems in an older adult population over a seven-year period and to describe factors associated with these changes. Of particular interest was the relationship between general health and changes in oral functioning. METHODS: The data came from a longitudinal study of community-dwelling individuals who were aged 50 years and older when first recruited. Data were collected at baseline (n = 907) and at three (n = 611) and seven-year (n = 425) follow-ups. Oral function was assessed by means of a six-item index of chewing ability. Data were weighted to account for loss to follow-up using weights derived from the seven-year response proportions for dentate and edentulous subjects. Logistic regression analysis using backward stepwise selection was used to identify predictors of onset and recovery. RESULTS: At baseline, 25 percent of subjects reported a problem chewing. This rose to 26 percent at three years and 34 percent at seven years. The seven-year incidence of chewing dysfunction was 19 percent. Of those with a chewing problem at baseline, 21 percent did not have a problem at seven years. A logistic regression model predicting the seven-year incidence of chewing problems indicated that subjects aged 65 years or older, the edentulous, those rating their oral health as poor, those without dental insurance and those without a regular source of dental care were more likely to be an incident case. In addition, a variable denoting the number of chronic medical conditions at baseline also entered the model. A logistic regression model predicting recovery indicated that older subjects, the edentulous, those from low-income households, and those with limitations in activities of daily living were less likely to recover over the observation period. CONCLUSION: The results of this study indicate a marked increase in the prevalence of chewing problems in this older adult population over the seven-year observation period. Poorer general health at baseline increased the probability of the onset of a chewing problem and decreased the probability of recovery.

Aged↗

PubMed1986

An insight on the use of multiple logistic regression analysis to estimate association between risk factor and disease occurrence.

Multiple logistic regression is an accepted statistical method for assessing association between an anticedant characteristic (risk factor) and a quantal outcome (probability of disease occurrence), statistically adjusting for potential confounding effects of other covariates. Yet the method has potential drawbacks which are not generally recognized. This article considers one important drawback of logistic regression. Specifically the so-called main effect logistic model assumes that the probability of developing disease is linearly and additively related to the risk factors on the logistic scale. This assumption stipulates that for each risk factor, the odds ratio is constant over all reference exposure levels, and that the odds ratio exposed to two or more factors is equal to the product of individual risk factor odds ratios. If the observed odds ratios in the data follow this pattern, the model-predicted odds ratios will be accurate, and the meaning of the odds ratio for each risk factor will be straightforward. But if the observed odds ratios deviate from the model assumption, the model will not fit the data accurately, and the model-predicted odds ratios will not reflect those in the data. Although satisfactory fit can always be achieved by adding to the model polynomial and product terms derived from the original risk factors, the odds ratios estimated by such an interaction logistic model are difficult to interpret, viz., the odds ratio for each risk factor depends not only on the reference exposure levels of that factor, but also on the exposure level in other factors.(ABSTRACT TRUNCATED AT 250 WORDS)

Environmental Exposure↗

PubMed1990

High-normal blood pressure and early diabetic nephropathy.

The relationship between high-normal blood pressure (BP) levels and early diabetic nephropathy is currently unknown. Blood pressure levels were checked longitudinally for a mean of 6.6 years in 230 subjects to determine their relationship to early diabetic nephropathy as monitored by microalbuminuria. High-normal BP level correlated with the presence of microalbuminuria. Microalbuminuria was 2.8 times as common in subjects with high-normal BP levels compared with those subjects with BP levels below the 90th percentile for their age. The elevated microalbumin excretion was primarily associated with high-normal diastolic BP levels. Our data suggest that either microalbuminuria or high-normal BP levels can precede the other. In a logistics model, diastolic BP and mean HbA1 (over 6.6 years) entered the model at similar levels, followed by duration of diabetes. When the influence of mean HbA1 was removed using logistic regression, the diastolic BP level remained a significant associated factor for the presence of microalbuminuria.

Adolescent↗

PubMed2004

Crack cocaine and sex.

The impact of crack cocaine use on number of sex partners was examined using bivariate analyses and a logistic model on a national treatment cohort of 4939 individuals. Number of sex partners over the last 12 months was dichotomized as none/one versus multiple partners for the logistic analyses. The model included 11 independent variables not including prostitution or use of crack cocaine. For both genders, the bivariate analyses showed significant positive associations between crack use and number of partners regardless of type of sexual activity; those who used crack had more partners for all sexual activities queried, compared to those who did not. Cocaine, whether in powder or crack form, was positively associated with prostitution for both genders. For men the odds ratio for crack cocaine use ranged from 1.6 (heterosexual anal) to 5.5 (homosexual anal) and for women from 2.9 (heterosexual oral) to 4.1 (homosexual oral). If prostitution is added to the model the odds ratios are reduced for homosexual activities for men and reduced dramatically for all types of sexual activity for women.

Administration, Inhalation↗

PubMed2006

Patterns of malaria-related hospital admissions and mortality among Malawian children: an example of spatial modelling of hospital register data.

BACKGROUND: Malaria is a leading cause of hospitalization and in-hospital mortality among children in Africa, yet, few studies have described the spatial distribution of the two outcomes. Here spatial regression models were applied, aimed at quantifying spatial variation and risk factors associated with malaria hospitalization and in-hospital mortality. METHODS: Paediatric ward register data from Zomba district, Malawi, between 2002 and 2003 were used, as a case study. Two spatial models were developed. The first was a Poisson model applied to analyse hospitalization and minimum mortality rates, with age and sex as covariates. The second was a logistic model applied to individual level data to analyse case-fatality rate, adjusting for individual covariates. RESULTS AND CONCLUSION: Rates of malaria hospitalization and in-hospital mortality decreased with age. Case fatality rate was associated with distance, age, wet season and increased if the patient was referred to the hospital. Furthermore, death rate was high on first day, followed by relatively low rate as length of hospital stay increased. Both outcomes showed substantial spatial heterogeneity, which may be attributed to the varying determinants of malaria risk, health services availability and accessibility, and health seeking behaviour. The increased risk of mortality of children referred from primary health facilities may imply inadequate care being available at the referring facility, or the referring facility are referring the more severe cases which are expected to have a higher case fatality rate. Improved prognosis as the length of hospital stay increased suggest that appropriate care when available can save lives. Reducing malaria burden may require integrated strategies encompassing availability of adequate care at primary facilities, introducing home or community case management as well as encouraging early referral, and reinforcing interventions to interrupt malaria transmission.

Adolescent↗

PubMed1993

World and regional populations.

Non-linear regression indicates that the logistic model provides excellent fits to world and continental human population data of the period 1950-1985. If the present logistic regime persists, the world population is likely to double in 47 years. Moreover, the eventual world population would be close to its final value of 23.8 billion by the year 2200, with about 11 billion in Asia and 10.5 billion in Africa. An approximately 95% confidence interval for the eventual world population is found to be between 11.9 and 35.7 billion. The logistic estimates of future populations are much higher than those considered most likely by the United Nations.

Humans↗

PubMed2004

[The five-word test in memory deficiency].

The 5-Word test is a bedside memory test with free and cued selective recollection. Here, we evaluated its reliability Belgian French speakers. Five groups were studied : normal subjects, depressive patients, patients with AD, patients with vascular dementia and 47 patients for a validation of a logistic model. The results of the 5-Word test with the MMSE, both values were included as variables in a logistic regression model. The probability of diagnosis of AD and dementia can be calculated and used prospectively for assessment of new cases in a validation study. The importance of the study of free and cued response during immediate and delayed recall was assessed by variance analysis in 4 groups. A third order interaction (p < 0.001) was observed between the normal and AD groups, the normal and vascular dementia (VD) groups and between AD and VD groups. With this initial model, 98% of the 62 normal subjects and the 57 patients with AD were correctly classified (p < 0.001). For AD diagnosis at the onset, the sensitivity of the MMSE is increased when it is associated with the 5-Word. This low-tech bedside method associating the 5-Word test and MMSE has a good specificity for the diagnosis of AD. It is very useful for screening a population of patients at the onset of cognitive and memory deficiency and for selecting the patients requiring an examination at a memory clinic.

Aged↗

PubMed2000

[Quantitative characteristics of rare and endangered Castanopsis kawakamii population].

In this paper, a new self-adaptive model for population growth, i.e. S = exp(alpha ln2(1 + ce-rt) + beta ln(1 + ce-rt) + gamma) was established, which combined Logistic, Smith, Gompertz, Gui-Lawson, Zhang-Logistic and Liu-Logistic models. Estimations with genetic algorithm showed that the new model was more suitable than other models to simulate the actual growth trend of Castanopsis kawakamii population.

China↗

PubMed1997

Childhood adversity and adult psychiatric disorder in the US National Comorbidity Survey.

BACKGROUND: Survey data are presented on the associations between retrospectively reported childhood adversities and subsequent onset and persistence of DSM-III-R disorders. METHODS: Data come from the US National Comorbidity Survey, a large survey of the US household population. RESULTS: Twenty-six adversities were considered, including loss events (e.g. parental divorce), parental psychopathologies (e.g. maternal depression), interpersonal traumas (e.g. rape) and other adversities (e.g. natural disaster). These adversities were consistently associated with onset, but not persistence, of DSM-III-R mood disorders, anxiety disorders, addictive disorders and acting out disorders. Most bivariate associations with onset attenuated in models that controlled for clustering of adversities and for lifetime co-morbidities among psychiatric disorders. Multivariate effects of adversities in logistic models were additive, which means that they have multiplicative effects on probability of disorder onset. Adversities showed little specificity. An analysis of time decay showed that the effects of childhood adversities on disorder onset persist beyond childhood. CONCLUSIONS: The existence of strong clustering among childhood adversities and lifetime co-morbidity among adult disorders means that caution is needed in interpreting the results of previous single-adversity single-disorder studies as documenting unique effects of specific childhood adversities on specific adult disorders. Future studies need to assess a broader range of adversities and disorders and to explore the existence and effects of commonly occurring adversity clusters. Replication is needed to verify that the effects of childhood adversities are mostly on first onset rather than on the creation of vulnerabilities that lead to increased risk of persistence.

Adolescent↗

PubMed1991

Bacteremia in febrile patients. A clinical model for diagnosis.

Among 244 patients aged 18 to 98 years who were consecutively hospitalized in a department of internal medicine because of a febrile disease, 52 (21%) were bacteremic. On a logistic regression analysis, five variables known within 24 hours of admission were found to be associated both significantly and independently with bacteremia: low serum albumin level, low premorbid performance status, chills, renal failure, and an assumptive diagnosis of urinary tract infection on admission. The logistic model was used to divide patients into three groups. In group 1, the percentage of bacteremic patients was 5%, in group 2, 40%, and in group 3, 83%. The percentage of deaths in the three groups was 0%, 23%, and 50%, respectively. The model was validated in a second group of 257 patients. The percentage of bacteremia was 1% in group 1, 23% in group 2, and 65% in group 3. The death rate in three groups was 3%, 4%, and 35%, respectively. The accuracy of the attending physician in diagnosing bacteremia within 24 hours of hospitalization was compared with that of the model. Use of the model could have improved the diagnostic accuracy in 5% of the patients in group 1 and in 18% of patients in group 3.

Aged↗

PubMedDate not supplied

Impact of the Salmonella status of market-age pigs and the pre-slaughter process on Salmonella caecal contamination at slaughter.

The aim of this study was to evaluate the impact of the pre-slaughter process on Salmonella caecal contamination of pigs at slaughter. An observational study was carried out in 2001 on 101 conventional farrow-to-finish pig farms. On each farm, one batch of contemporary pigs was followed from the end of the fattening period until slaughter. The Salmonella bacteriological status of the batches was assessed by environmental samples of faecal material. The serological Salmonella status was obtained on 30 individually identified market-age pigs using an indirect ELISA test. At the slaughterhouse, 25 g of caecal contents were taken from 10 of the identified pigs. Faecal and caecal material were analysed according to a classical bacteriological method. A questionnaire was designed to obtain information about the type of feeding during the fattening period (dry versus wet), the duration of fasting on the farm before leaving for the slaughterhouse, the duration of transport between the farm and the slaughterhouse, the holding time in lairage at the slaughterhouse and loading and unloading conditions on the farm and at the slaughterhouse. To assess the relationships between these factors and the Salmonella caecal status of the pigs and the batches, two logistic models were fitted at the individual and at the batch level, respectively. The first analysis was performed using a random effects logistic regression model. The second analysis was based on a cumulative logit model with a positive caecal rate classified into three classes as the outcome variable. The results showed that the Salmonella status of market-age pigs assessed on the farm either by serological or bacteriological examinations and the time spent in lairage before slaughtering played a crucial role on caecal contamination. In the light of these results, actions should be considered both on the farm and at the slaughterhouse to decrease the risk of Salmonella contamination of the caecal contents.

Abattoirs↗

PubMed2005

The effect of time to peak ankle torque on balance stability boundary: experimental validation of a biomechanical model.

Pai and Patton (1997), using a biomechanical model, determined a set of feasible center of mass (CM) velocity-position combinations (balance stability boundary) that guarantee upright stability. In their study, the magnitude of the restoring ankle torque was used to study the subject's ability to recover balance. Recent studies have suggested that the ability to maintain a stable posture depends not only on the magnitude of the restoring torque but also on the time to generate this torque. The objectives of the present study were: (1) to build a biomechanical model that predicts the balance stability boundary which includes time to peak ankle torque, (2) to determine the capability of the model to predict successful and failed experimental balance recovery trials, and (3) to compare the predictive capability of the biomechanical model with that of a statistical model (logistic regression). A single-link-plus-foot biomechanical model was used to determine a set of balance stability boundaries, computed from the combination of maximum CM velocity and related CM position, for various times to peak ankle torque. An experiment was conducted to validate the biomechanical model. The participants self-initiated a forward destabilization and were asked to regain balance using an ankle feet-in-place strategy. Also, a forward stepwise logistic regression (predictors: CM position and velocity and time to peak ankle torque) was used to discriminate between successful and failed experimental trials. (1) The outcomes of the biomechanical model confirmed that the time to peak ankle torque drastically constrained the stability boundaries. (2) The biomechanical model predicted 79.9% of the failed experimental trials and 74.5% of the successful experimental trials. (3) The stepwise logistic regression included all independent variables and predicted 57.2% of the failed and 93.7% of the successful experimental trials. Hence, the biomechanical model showed better predictive capability than the statistical model for identifying unsuccessful balance recovery. It is noteworthy that the balance stability boundaries constrained by the speed of ankle torque development predicted the outcome of the experimental trial earlier in the time series than balance stability boundary constrained by constant ankle torque. Overall, the present biomechanical model may serve as an assessment tool to develop specific interventions towards improving a patient's speed of ankle torque development and to possibly reduce falling frequency.

Accidental Falls↗

PubMed1998

The case-time-control design: further assumptions and conditions.

The case-time-control design is a strategy that was developed to tackle the problem of confounding by indication in the nonexperimental assessment of intended or known effects of drugs. By using subjects as their own controls, the case-time-control design, under an explicitly defined model, eliminates the biasing effect of unmeasured confounding factors in the situation where exposure varies over time. The correct application of this design is based on a specific model that contains inherent assumptions and imposes certain conditions for the approach to be valid. In a recent article, Greenland questioned the validity of the case-time-control design by presenting several "counterexamples." In this paper, we review the assumptions inherent to the validity of the case-time-control model. We show that the presumed counterexamples are not what they are claimed to be, simply because they do not conform to the logistic model explicitly underlying the case-time-control approach. These examples are shown to arise from an alternative model that includes a confounder by period interaction, a term expressly avoided in the case-time-control model. When the data from these examples are modified to satisfy the correct model, the resulting case-time-control estimates of the treatment odds ratio are exactly 2, the true treatment effect. We clarify the necessity of this assumption in the context of matching in epidemiology. We also discuss briefly the assumptions of conditional independence and carryover effects.

Bias↗

PubMed2003

Gender and outcomes after coronary artery bypass grafting: a propensity-matched comparison.

OBJECTIVE: Our objective is to determine whether gender is a marker or a causal influence for poor outcomes after coronary revascularization. METHODS: Propensity-modeling techniques were used to investigate whether gender adversely impacts outcomes after coronary revascularization. A parsimonious explanatory model was developed by bootstrap bagging with variable selection from 64 baseline and 37 operative variables. Propensity scores were calculated from a logistic model that included the parsimonious model and additional baseline variables. Greedy matching techniques were applied to match female and male patients to the nearest propensity scores. Comparisons were made among the propensity-matched women and men. RESULTS: Of the 15,597 patients undergoing isolated coronary artery bypass graft surgery, only 26% of the 3596 women were matched on propensity scores with men. Distribution of covariates among the matched pairs was, on average, equal. Postoperative mortality (P =.76), neurologic morbidity (global deficit P =.07, focal deficit P =.51), infection (sepsis P =.88), mediastinitis (P =.18), renal failure (P =.84), intra-aortic balloon pump usage (P =.61), and reoperation for bleeding (P =.10) were similar among women and men. Occurrence of Q-wave myocardial infarction (P = <.01), postoperative inotropic usage (P = <.01), and prolonged ventilatory support (P =.02) were more common in women compared with propensity-matched men. CONCLUSIONS: The preoperative profiles of women and men are markedly different. Propensity matching women and men was difficult, because only 26% of women were able to be matched with men. However, in well-matched patients, female gender was not associated with increased mortality and had minimal impact on morbidity after coronary artery bypass grafting.

Aged↗