SEARCH · PubMed Health

Results for “Explainability”

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 73 records · Page 4Linked to original sources

PubMed1989

Ordered appearance of antigenic variants of African trypanosomes explained in a mathematical model based on a stochastic switch process and immune-selection against putative switch intermediates.

Antigenic variation of African trypanosomes results from the periodic activation of a single new variant cell surface glycoprotein (VSG) gene out of a repertoire of about a 1000 VSG genes. In spite of the apparently random genetic basis of the process of antigenic variation, the relapsing parasitemias are characterized by an as yet unexplained order of appearance of major VSG variants. Here we mathematically test hypotheses concerning the blood-based parasitemia. In our model the antigenic switches occur at random at the DNA level. A variable proportion of the switches has a short intermediate phase in which two different VSGs simultaneously occur on the cell surface. We show that, in a theoretical population of 230 single expressor variants in an immunocompetent or in an immunodeficient host, it is not possible to explain the ordered appearance of variants by affecting the growth coefficients of single expressors or double expressors or by affecting the antigen switch probabilities. Rather, a realistic parasitemia can be obtained if the majority of switches has a double expressor switch-intermediate phase and if the double expressors have a differential susceptibility to the immune control. This study is significant in providing a theoretical basis for the ordered appearance of variants and in explaining previously unresolved discrepancies between the rate of appearance of new variants in culture and in vivo. In addition, testable predictions as to the development of the infections, switch rate of variants, fraction of double expressors, and parasite mortality coefficients are generated.

Animals↗

PubMed1988

Explaining adolescent drug use: an elaboration strategy for structural equations modeling.

We report a series of analyses designed to estimate increasingly elaborated theoretical models that explain adolescent drug use. Each of the successive elaborations adds a theoretical construct to the explanatory model in order to increase our understanding of drug use by specifying in greater detail the nature of the structural relationships among the latent variables. The more detailed specification is accomplished by 1) specifying new direct effects that increase explained variance in drug use, 2) decomposing direct effects through the interpolation of hypothesized intervening variables, 3) specifying antecedents of variables that modify their direct effects, and 4) exposing suppressor effects. Where indicated, we evaluate alternative explanations of the observed relationships. We do this by controlling for common antecedent effects to reduce spuriousness or by examining different specifications of causal linkages among the explanatory constructs.

Adolescent↗

PubMed1995

Does digestibility of meat protein help explain large bowel cancer risk?

An association between meat eating and large bowel cancer risk has been shown in a variety of epidemiologic studies. One reason could be that meat is less well digested than other protein foods and leads to greater amounts of protein entering the large bowel and being metabolized by colonic bacteria to potential carcinogens. To test this hypothesis, five subjects with ileostomies were fed, for five days, a basal diet to which were added test meals of cheese, a small or a large fried beef steak, and a large steak with resistant starch (RS). Ileal true nitrogen digestibility was similar for all five diets: control, 86.3%; cheese, 89.4%; low beef, 88.6%; high beef, 89.6%; and high beef + RS, 88.7%. Beef, at both low and high intake levels, was as well digested as cheese, suggesting that poor digestibility of meat does not explain the association between meat intake and large bowel cancer risk. Ileal starch output on the high beef + RS diet was 27% greater than expected on the basis of the measurement of dietary RS in vitro (p = 0.005 for linear trend), and this was confirmed by a meta analysis with eight other published studies. The relation between meat and large bowel cancer may reflect higher protein intakes in meat eaters or may be explained by other mechanisms.

Animals↗

PubMed2025

SetBERT: the deep learning platform for contextualized embeddings and explainable predictions from high-throughput sequencing.

MOTIVATION: High-throughput sequencing (HTS) is a modern sequencing technology used to profile microbiomes by sequencing thousands of short genomic fragments from the microorganisms within a given sample. This technology presents a unique opportunity for artificial intelligence to comprehend the underlying functional relationships of microbial communities. However, due to the unstructured nature of HTS data, nearly all computational models are limited to processing DNA sequences individually. This limitation causes them to miss out on key interactions between microorganisms, significantly hindering our understanding of how these interactions influence the microbial communities as a whole. Furthermore, most computational methods rely on post-processing of samples which could inadvertently introduce unintentional protocol-specific bias. RESULTS: Addressing these concerns, we present SetBERT, a robust pre-training methodology for creating generalized deep learning models for processing HTS data to produce contextualized embeddings and be fine-tuned for downstream tasks with explainable predictions. By leveraging sequence interactions, we show that SetBERT significantly outperforms other models in taxonomic classification with genus-level classification accuracy of 95%. Furthermore, we demonstrate that SetBERT is able to accurately explain its predictions autonomously by confirming the biological-relevance of taxa identified by the model. AVAILABILITY AND IMPLEMENTATION: All source code is available at https://github.com/DLii-Research/setbert. SetBERT may be used through the q2-deepdna QIIME 2 plugin whose source code is available at https://github.com/DLii-Research/q2-deepdna.

Deep Learning↗

PubMed1991

Multilevel modelling of longitudinal cephalometric data explained for orthodontists.

Multilevel modelling of longitudinal data is an important new statistical technique. In this article some of the basic concepts and ideas of multilevel modelling are explained. The model is introduced by showing how individual and average growth can be modelled. The intercept, linear and quadratic coefficient, between and within variance, fixed and random part, and other concepts of multilevel modelling are explained. Attention is also given to the reading of statistical tables of the results of multilevel analysis. In the conclusion some of the advantages of multilevel modelling of cephalometric data are mentioned.

Aging↗

PubMed1990

Can changes in the unemployment rates explain the recent changes in suicide rates in developed countries?

Data were collected on unemployment and suicide rates in 16 developed countries for 1973 and 1983 (suicide rates were three-year averages). Unemployment rates rose appreciably in men and women in all countries. Among men suicide rates rose in 14 of the countries whereas among women they did so in only seven. A mathematical model was developed to investigate, for those countries with increased suicide rates, how much of the increase could be contributed by an increase in the numbers unemployed. It was found that the proportion of the increase that could be 'explained' varied considerably between countries. In general the amount of the increase explained was small, and often a considerable increase in the suicide rates among those employed would be required to account for the observed increase in the suicide in the whole population. It is concluded that unemployment shows an inconsistent relationship with suicide. Further, if a relationship does exist in some countries the effect may be as much a generalized one on the whole population as a specific effect on the unemployed. Finally the possible effects of unemployment on suicide differ appreciably between men and women.

Adolescent↗

PubMed1995

Ways of seeing: explaining variations in use of acute hospital services.

BACKGROUND: In the US Medicare programme, hospitals are paid directly by activity. To provide incentives for efficiency, the US Federal Government has sought objective measures of inpatients' need for resources. In the UK National Health Service, resources are allocated for acute hospital services as part of a global budget to purchasers, who then contract with hospitals. To provide equity in resource allocation, the Department of Health in England, has sought objective measures of populations' need for resources. METHODS: Examination of policy and technology that has used variations in utilization of resources to derive objective measures of efficiency and equity. RESULTS: The technology of developing empirical measures of resources needed by patients has lacked information on outcomes, which is vital for measures of efficiency. The technology of developing empirical measures of resources needed by populations has relied on aggregate data. Analyses of specific procedures and conditions consistently find that these variations are explained by differences in medical practice and not by need. CONCLUSIONS: There is scope for multidisciplinary research to explain small area variations for specific procedures and conditions in resources used by populations. It seems unlikely, however, that governments will be interested in findings from this research.

Health Care Rationing↗

PubMed1994

Are racial differences in squamous cell esophageal cancer explained by alcohol and tobacco use?

BACKGROUND: In the United States, incidence rates of squamous cell esophageal cancer are more than five times higher among black men than among white men. Reasons that might explain this large racial disparity are being sought. PURPOSE: We evaluated whether differential use of alcohol and tobacco can fully account for the excess of squamous cell esophageal cancer among U.S. blacks. METHODS: We conducted a population-based, case-control study with in-person interviews with 373 squamous cell esophageal cancer case patients (124 white males and 249 black males) and 1364 control subjects (750 white males and 614 black males) from three U.S. geographic areas. Histologically confirmed cases of squamous cell esophageal cancer newly diagnosed from August 1, 1986, through April 30, 1989, among white and black men aged 30-79 years were included. RESULTS: Alcohol use of more than one drink per day and/or current cigarette use of at least one pack per day accounted for 92.7% (95% confidence interval [CI] = 86.8%-98.5%) of the squamous cell esophageal cancers in blacks, versus 86.3% (95% CI = 75.5%-97.1%) in whites, and for 94% of the difference between the black and white annual incidence rates. The interaction between race and the continuous drinking/smoking variable in a logistic regression analysis was statistically significant (two-sided, P = .02). Exposure rates among controls at all levels of combined alcohol and tobacco use examined were slightly higher among blacks and accounted for a small portion of the racial differences in incidence rates. CONCLUSION: Although the vast majority of esophageal cancers in both blacks and whites in our data can be explained by use of alcohol and tobacco, it is not clear why heavy consumption of alcohol and/or tobacco is responsible for 14.9 per 100,000 per year more cases of squamous cell esophageal cancer among blacks than among whites. The differences in the odds ratios appear to account for more of the racial differences in incidence rates than do the prevalences of exposure to alcohol and tobacco alone. The reasons for this apparent racial difference in carcinogenic risk from the same level of alcohol and tobacco use are unknown, but they may include qualitative differences in alcohol consumption, differences in other environmental exposures that interact with alcohol and/or tobacco to modify risks, or differences in susceptibility to these factors.

Adult↗

PubMed1995

A challenge to the concept of tubal reflux to explain the rise and fall of CA125 in serum during the first trimester.

Although amniotic fluid concentrations of cancer antigen (CA) 125 rise during the first two trimesters of pregnancy, the serum concentrations of CA125 peak during the first trimester and drop to non-pregnant values in the second and third trimester. A previous hypothesis to explain this phenomenon was that in the early first trimester decidual CA125 gains access to the maternal compartment via 'tubal reflux' and subsequent absorption by peritoneal lymphatics. However, as pregnancy advances, the decidua capsularis fuses with the decidua parietalis, thus obliterating the endometrial cavity at 10-12 weeks; the Fallopian tubes thus become functionally obstructed. To test this hypothesis, we evaluated early first trimester CA125 concentrations in women conceiving by in-vitro fertilization (IVF) and embryo transfer with patent tubes (group 1) and in those conceiving by IVF and embryo transfer with bilateral tubal occlusion (group 2). We also compared those conceiving with human menopausal gonadotrophin therapy for ovulation induction without assisted reproduction (group 3) and those conceiving without fertility drugs in assisted reproduction (group 4). Mean CA125 concentrations were similar in groups 1-3; the mean CA125 concentration in group 4 was lower but this difference was not statistically significant, probably due to the small sample size. These data do not support the concept that tubal reflux explains the rise and fall of serum concentrations of CA125, since these were equal in IVF conceptions with or without tubal patency.

CA-125 Antigen↗

PubMed1985

Increased daily sperm production in the breeding season of stallions is explained by an elevated population of spermatogonia.

Seasonal variation in number of spermatogonia and germ cell degeneration was evaluated to determine which mechanism might explain seasonal differences in daily sperm production per testis (DSP/testis) or per g parenchyma (DSP/g) in stallions. Comparing 28 adult stallions (4 to 20 yr old) in each of the nonbreeding (December-January) and breeding (June-July) seasons, the population of type A spermatogonia was more than two times greater (P less than 0.01) in the breeding season. While the number of type B spermatogonia also was elevated (P less than 0.01) in the breeding season, the number of type B spermatogonia/type A spermatogonium was similar (P greater than 0.05) between seasons. Daily sperm production/testis based on each cell type from type B spermatogonia to spermatids with elongated nuclei was lower (P less than 0.01) in the nonbreeding season. Based on DSP/g, there was significant degeneration during the meiotic divisions in the nonbreeding season. However, this reduction in potential spermatozoan production was not significant (P greater than 0.05) when considering DSP/testis. Significant germ cell degeneration also occurred in the breeding season between type B spermatogonia and primary spermatocytes. However, the type A spermatogonial population was sufficiently elevated to override this degeneration and to explain elevated production of sperm in the breeding season of stallions.

Animals↗

PubMed1996

What does not explain the variation in the direct costs of graduate medical education.

BACKGROUND: There is considerable variation in the costs of training residents across hospitals. Previous studies have reported training costs that ranged for $7,500 to $200,000 per resident, with means in the $50,000 to $60,000 range. This paper examines the factors associated with the variation in the direct costs of residency education across hospitals. METHODS: Hospital costs, hospital payment rates, various cost-of-living indices, and hospital characteristics for all hospitals in the United States receiving Medicare funds for residency education in fiscal year 1991 were obtained from various public sources. Bivariate and multivariate analyses were performed to determine whether organizational structure of residency training, specialty mix of residents, quality of training, cost of living in the geographic area, patient mix of the hospital, or other factors could explain some or all of the cost variation. RESULTS: Only a small proportion of the variation in the costs of training residents or payments for residency education could be explained by the factors analyzed. CONCLUSION: Much of the current variation in residency costs and payments is attributable to accounting and not structural or locational factors. Institutions looking for ways to become more efficient could compare costs of training across institutions without sophisticated adjusters. Federal policymakers should consider a national payment rate per resident to replace the current system based on hospital-specific costs. This proposal would redistribute dollars across residency programs. The proposal would affect payments for direct medical education in the Medicare program and could form the basis for payment for residency education by all payers.

Accounting↗

PubMed1995

'Dynamics of neuronal interactions' cannot be explained by 'neuronal transients'.

In a recent paper, Vaadia et al. demonstrated that patterns of firing correlation between single neurons in the cortex of behaving monkeys can be modified within a fraction of a second. These changes occur in relation to sensory stimuli and behavioral events, and even without modulations of the neurons' firing rates. These findings call for a revision of prevailing models of neural coding that solely rely on single neuron firing rates. In a defense of these models, Friston put forward an alternative explanation, proposing that the observed correlation dynamics emerge solely from co-modulations of the firing rates of each of the neurons, while the strength of their interaction remains constant. To test this possibility we re-examined the data, adopting Friston's 'neuronal transients' model, and the associated equations and procedures. We found that, to explain the dynamic correlation between a pair of neurons, the alternative interpretation requires that each neuron's response to a single stimulus is composed of a relatively large number of independent components, which co-vary with their counterparts in the companion neuron. This large number of components and their shapes lead us to conclude that, although in principle possible, the neuronal transients model: (i) does not provide a simpler explanation of the experimental results; and (ii) cannot explain these results without itself deviating significantly from most rate code models.

Animals↗

PubMed2004

Genomic regionality in rates of evolution is not explained by clustering of genes of comparable expression profile.

In mammalian genomes, linked genes show similar rates of evolution, both at fourfold degenerate synonymous sites (K4) and at nonsynonymous sites (KA). Although it has been suggested that the local similarity in the synonymous substitution rate is an artifact caused by the inclusion of disparately evolving gene pairs, we demonstrate here that this is not the case: after removal of disparately evolving genes, both (1) linked genes and (2) introns from the same gene have more similar silent substitution rates than expected by chance. What causes the local similarity in both synonymous and nonsynonymous substitution rates? One class of hypotheses argues that both may be related to the observed clustering of genes of comparable expression profile. We investigate these hypotheses using substitution rates from both human-mouse and mouse-rat comparisons, and employing three different methods to assay expression parameters. Although we confirm a negative correlation of expression breadth with both K4 and KA, we find no evidence that clustering of similarly expressed genes explains the clustering of genes of comparable substitution rates. If gene expression is not responsible, what about other causes? At least in the human-mouse comparison, the local similarity in KA can be explained by the covariation of KA and K4. As regards K4, our results appear consistent with the notion that local similarity is due to processes associated with meiotic recombination.

Animals↗

PubMed1993

Degree of breast emptying explains changes in the fat content, but not fatty acid composition, of human milk.

We compared within and between breastfeed changes in milk fat to short-term rates of milk synthesis and degree of breast emptying (measured using the Computerized Breast Measurement system) over two 24 h periods for five lactating mothers. The fat content (f) of fore and hind milk samples increased more steeply as the breast was progressively emptied by the infant (degree of emptying, d, range 0-1; f = 21.59 + 9.38d + 70.99d2; P < 0.0001; r2 = 0.68; n = 154). For the nine individual breasts, between 41-95% of the variance of the fat content of milk was explained by degree of breast emptying. We argue that this relationship explains differences in the circadian rhythm of the fat content of milk and allows the accurate calculation of the average fat content of milk consumed by infants (37-66 g/l for the nine individual breasts). The fatty acid composition of the fore and hind milk samples was determined for four of the mothers. We observed within and between breastfeed variability in the relative proportions of the seven major fatty acids of milk fat and these changes are discussed with reference to the control of fat synthesis in the human mammary gland.

Breast↗

PubMed1991

Explaining outputs of primary health care: population and practice factors.

OBJECTIVE: To examine whether variations in the activities of general practice among family health service authorities can be explained by the populations characteristics and the organisation and resourcing of general practice. DESIGN: The family health services authorities were treated as discrete primary health care systems. Nineteen performance indicators reflecting the size, distribution, and characteristics of the population served; the organisation of general practice (inputs); and the activities generated by general practitioners and their staff (output) were analysed by stepwise regression. SETTING: 90 family health services authorities in England. MAIN OUTCOME MEASURES: Rates of cervical smear testing, immunisation, prescribing, and night visiting. RESULTS: 53% of the variation in uptake of cervical cytology was accounted for by Jarman score (t = -3.3), list inflation (-0.41), the proportion of practitioners over 65 (-0.64), the number of ancillary staff per practitioner (2.5), and 70% of the variation in immunisation rates by standardised mortality ratios (-6.6), the proportion of practitioners aged over 65 (-4.8), and the number of practice nurses per practitioner (3.5). Standardised mortality ratios (8.4), the number of practitioners (2.3), and the proportion over 65 (2.2), and the number of ancillary staff per practitioner (-3.1) accounted for 69% of variation in prescribing rates. 54% of the variation in night visiting was explained by standardised mortality ratios (7.1), the proportion of practitioners with lists sizes below 1000 (-2.2), the proportion aged over 65 (-0.4), and the number of practice nurses per practitioner (-2.5). CONCLUSIONS: Family health services authorities are appropriate systems for studying output of general practice. Their performance indicators need to be refined and to be linked to other relevant factors, notably the performance of hospital, community, and social services.

England↗

PubMed1993

Explaining socioeconomic differences in sickness absence: the Whitehall II Study.

OBJECTIVE: To describe and explain the socioeconomic gradient in sickness absence. DESIGN: Analysis of questionnaire and sickness absence data collected from the first phase of the Whitehall II study. Grade of employment was used as a measure of socioeconomic status. SETTING: 20 civil service departments in London. SUBJECTS: 6900 male and 3414 female civil servants aged 35-55 years. MAIN OUTCOME MEASURES: Rates of short spells (< or = 7 days) and long spells (> 7 days) of sickness absence. RESULTS: A strong inverse relation between grade of employment and sickness absence was evident. Men in the lowest grade had rates of short and long spells of absence 6.1 (95% confidence interval 5.3 to 6.9) and 6.1 (4.8 to 7.9) times higher than those in the highest grade. For women the corresponding rate ratios were 3.0 (2.3 to 3.9) and 4.2 (2.5 to 6.8) respectively. Several risk factors were identified, including health related behaviours (smoking and frequent alcohol consumption), work characteristics (low levels of control, variety and use of skills, work pace, and support at work), low levels of job satisfaction, and adverse social circumstances outside work (financial difficulties and negative support). These risk factors accounted for about one third of the grade differences in sickness absence. CONCLUSION: Large grade differences in sickness absence parallel socioeconomic differences in morbidity and mortality found in other studies. Identified risk factors accounted for a small proportion of the grade differences in sickness absence. More accurate measurement of the risk factors may explain some of the remaining differences in sickness absence but other factors, as yet unrecognised, are likely to be important.

Absenteeism↗

PubMed1994

Changes in risk factors explain changes in mortality from ischaemic heart disease in Finland.

OBJECTIVES: To estimate the extent to which changes in the main coronary risk factors (serum cholesterol concentration, blood pressure, and smoking) explain the decline in mortality from ischaemic heart disease and to evaluate the relative importance of change in each of these risk factors. DESIGN: Predicted changes in ischaemic heart disease mortality were calculated by a logistic regression model using the risk factor levels assessed by cross sectional population surveys, in 1972, 1977, 1982, 1987, and 1992. These predicted changes were compared with observed changes in mortality statistics. SETTING: North Karelia and Kuopio provinces, Finland. SUBJECTS: 14,257 men and 14,786 women aged 30-59 randomly selected from the national population register. MAIN OUTCOME MEASURES: Levels of the risk factors and predicted and observed changes in mortality from ischaemic heart disease. RESULTS: The observed changes in the risk factors in the population from 1972 to 1992 predicted a decline in mortality from ischaemic heart disease of 44% (95% confidence interval 37% to 50%) in men and 49% (37% to 59%) in women. The observed decline was 55% (51% to 58%) and 68% (61 to 74) respectively. CONCLUSION: An assessment of the data on the risk factors for ischaemic heart disease and mortality suggests that most of the decline in mortality from ischaemic heart disease can be explained by changes in the three main coronary risk factors.

Adult↗

PubMed1995

Do changes in cardiovascular risk factors explain changes in mortality from stroke in Finland?

OBJECTIVES: To estimate the extent to which the changes in the main cardiovascular risk factors (blood pressure, smoking, and serum cholesterol concentration) can explain the observed changes in mortality from stroke in Finland during the past 20 years. DESIGN: Predicted changes in mortality from cerebrovascular disease mortality were calculated by a proportional hazards model from data obtained in cross sectional population surveys in 1972, 1977, 1982, 1987, and 1992. Predicted changes were compared with the observed changes in mortality statistics. SETTING: North Karelia and Kuopio provinces, Finland. SUBJECTS: 16,741 men and 16,389 women aged 30-59 randomly selected from the national population register, of whom 14,054 men and 14,546 women participated. MAIN OUTCOME MEASURES: Levels of risk factors and predicted and observed changes in mortality from cerebrovascular disease. RESULTS: The observed changes in diastolic blood pressure, total serum cholesterol concentration, and smoking in the population from 1972 to 1992 predicted a 44% fall in mortality from stroke in men and changes in diastolic blood pressure and smoking predicted a 34% fall in women. The observed fall in mortality from stroke was 66% in men and 60% in women. CONCLUSIONS: Two thirds of the fall in mortality from stroke in men and half in women can be explained by changes in the three main cardiovascular risk factors.

Adult↗