Analysis of the imputed female urinary incontinence data for the evaluation of expert system parameters.

We evaluated parameters for an expert system which will be designed to aid the differential diagnosis of female urinary incontinence by using knowledge discovered from data. To allow the statistical analysis, we applied means, regression and Expectation-Maximization (EM) imputation methods to fill in missing values. In addition, complete-case analysis was performed. Logistic regression results from the imputed data were reasonable. The significant parameters were mostly those that are important in the diagnostic work-up. Moreover, directions of relations between the parameters and the stress, mixed and sensory urge diagnoses were as expected. Analysis with the complete reduced data set gave clearly insufficient results. Imputed values had a moderate agreement, but odds ratios and classification accuracies of logistic regression equations were similar. Results suggest that with these data, simpler methods may be used to allow multivariate analysis and knowledge discovery, when better methods, such as EM imputation, are unavailable. Cluster analysis detected clusters corresponding to the small normal class, but was unable to clearly separate the larger incontinence classes.

Usefulness of imputation for the analysis of incomplete otoneurologic data.

The usefulness of imputation in the treatment of missing values of an otoneurologic database for the discriminant analysis was evaluated on the basis of the agreement of imputed values and the analysis results. The data consisted of six patient groups with vertigo (N=564). There were 38 variables and 11% of the data was missing. Missing values were filled in with the means, regression and Expectation-Maximisation (EM) imputation methods and a random imputation method provided the baseline results. Means, regression and EM methods agreed on 41-42% of the imputed missing values. The level of agreement between these and the random method was 20-22%. Despite the moderate agreement between the means, regression and EM methods, the discriminant functions were similar and accurate (prediction accuracy 83-99%). The discriminant functions obtained from the randomly imputed data were also accurate having prediction accuracy 88-97%. Imputation seems to be a useful method for treating the missing data in this database. However, a lot of data was missing in otoneurologic tests, which are likely to be of less importance in the diagnosis of vertiginous patients. Consequently, the disagreement of the methods did not affect clearly the discriminant analysis, and, therefore, future research requires more complete data and advanced imputation methods.

Comparison of genetic algorithms and other classification methods in the diagnosis of female urinary incontinence.

Galactica, a newly developed machine-learning system that utilizes a genetic algorithm for learning, was compared with discriminant analysis, logistic regression, k-means cluster analysis, a C4.5 decision-tree generator and a random bit climber hill-climbing algorithm. The methods were evaluated in the diagnosis of female urinary incontinence in terms of prediction accuracy of classifiers, on the basis of patient data. The best methods were discriminant analysis, logistic regression, C4.5 and Galactica. Practically no statistically significant differences existed between the prediction accuracy of these classification methods. We consider that machine-learning systems C4.5 and Galactica are preferable for automatic construction of medical decision aids, because they can cope with missing data values directly and can present a classifier in a comprehensible form. Galactica performed nearly as well as C4.5. The results are in agreement with the results of earlier research, indicating that genetic algorithms are a competitive method for constructing classifiers from medical data.

Palatal morphology and type of clefting.

In this study, the variation in palatal dimensions, particularly for height, width, and depth, were evaluated in subjects with different types of cleft palate using the moiré technique. The sample consisted of 95 subjects with cleft palate and 68 noncleft individuals. Effects of cleft type, gender, developmental stage of the dentition, missing teeth, and 11 skeletal and pharyngeal cephalometric variables on palatal dimensions were assessed using multiple-regression analyses. The typing of clefts had an effect on the all palatal dimensions. The palate was shallower in subjects with clefts involving the secondary palate than in other types of clefts or in noncleft individuals. The palatal width was decreased in subjects with cleft in the primary palate, and the palatal depth was affected by all cleft types except submucous cleft. Gender, developmental stage of the dentition, missing teeth, and some of the cephalometric variables also had a certain effect on the palatal dimensions.

A study in 47,XYY men of the expression of sex-chromosome anomalies in dental occlusion.

Occlusal traits were determined for 47,XYY men and compared with previous determinations of occlusal morphology in other sex-chromosome anomalies and in normal women and men. The 47,XYY men, like 47,XXY men, tended to have a mesial molar occlusion and a mandibular overjet more often than did other groups, while 45,X women (Turner patients) clearly had the highest frequency of distal occlusion and large overjet. The 47,XXY men had the highest frequency of most occlusal anomalies. As a whole, these and earlier findings suggest that the number of X or Y chromosomes is associated negatively with distal occlusion and lateral crossbite, whereas a positive association was found with mesial molar occlusion and scissors bite.

Developmental stage of the dentition and speech sound production in a series of first-grade schoolchildren.

This study was undertaken to investigate the effect of developmental stage of the dentition on speech sound production in children with and without articulatory speech disorders. In the whole sample, there were 281 subjects, mean age 7.5 years (SD 0.4). The stage of the eruption of the permanent incisors was calculated as the percentage height of each tooth of the mean crown height of the corresponding tooth in a group of 252 young adults. Eruptional stage of the first permanent molars was classified into three categories. Articulatory disorders were diagnosed by one experienced speech therapist and phoniatrician using the Remes Articulation Test [Remes, 1975] for the Finnish language. The results showed a wide interindividual variation in developmental stage of the dentition in a group of 7-year-old children with a tendency of the first permanent molars but not the incisors to erupt earlier in girls than in boys. Boys had a higher frequency of misarticulations estimated to need treatment than did girls. The present results indicate that during eruption of permanent teeth it is more important to draw attention to symmetrical eruption and to individual timing pattern rather than to the mean values of tooth eruption in subjects of the same age. The findings also suggest a decreasing risk of faulty speech sound production with advancing eruption of some permanent teeth still present in 7-year-old children.

Incidence of clinically confirmed farmer's lung in Finland and its relation to meteorological factors.

In Finland, data about the incidence of clinically confirmed (according to predefined criteria) farmer's lung have been collected in successive years since 1980. All special hospitals for pulmonary diseases and departments of pulmonary diseases in other hospitals (23 clinical units throughout Finland) took part in the study. This report deals with the results collected from 1980 to 1982 (inclusive), during which 512 new cases of farmer's lung were recorded, 186 in 1980, 153 in 1981, and 173 in 1982. The mean age of the patients was 47 years (range 14-75), 45 years for men and 48 years for women. About two-thirds of the patients were women. The mean annual incidence of farmer's lung among the entire farming population (standardized for age and sex to the total population in Finland in 1975) was 44 per 100,000 persons in farming. Both the standardized mean annual incidence (range 8-60 per 100,000 persons in farming) and the female/male ratio of the standardized incidence showed distinct regional differences. Most cases of farmer's lung occurred during the indoor feeding season for cattle. The highest frequency of new cases was observed in April and the lowest in October. The incidence of farmer's lung was positively correlated with measures of daily rainfall and negatively correlated with days without rainfall and with sunshine during the haymaking period preceding the diagnosis. A rainy haymaking period calls for artificial drying of hay in order to reduce the incidence of farmer's lung.

Occlusal morphology in 45,X females.

Seventy-two 45,X females (Turner syndrome) were examined for occlusal anomalies. Prevalence comparisons were made with first-degree female relatives and female population controls. The method of Björk et al [1964] was used for determination of malocclusion. The results clearly suggest an increased prevalence of occlusal anomalies in 45,X females. The most common malocclusions are cross bite, large maxillary overjet, distal molar occlusion, and a tendency to open bite. These anomalies reflect an imbalanced growth of the craniofacial skeleton in three dimensions. These findings suggest that genes on the human X chromosome are of importance in the harmonious growth and development of the craniofacial skeleton and ultimately normal occlusal morphology and relations.

Statistical evidence for the relation between citrate and carbonate in human cortical bone.

Bone specimens from the anterior iliac crest were collected post-mortem from 128 subjects whose death had been primarily caused by acute coronary disease or accidents. The cortical bone was separated and used for the following determinations: citrate by an enzymatic method, carbon dioxide by microdiffusion, chloride and phosphorus by spectrophotometry, fluoride by the ion-selective electrode technique, and Ca, Mg, Zn by atomic absorption spectrophotometry. The citrate concentration (+/- SD) was 10.9 +/- 3.1 mg/g (dry weight) and was not age-dependent. A statistically significant negative correlation was obtained between citrate and carbon dioxide, however, which was confirmed by regression analysis. The only statistically significant differences in carbon dioxide and fluoride concentrations between three citrate groups were obtained between the lowest group (3.2-8.7 mg/g) and the highest (13.1-16.8 mg/g). In conclusion an inverse relationship may suggest either an effect of citrate on the binding of carbonate to apatite structures, or that the carbonate concentration is dependent to a slight degree on citrate metabolism.

Palatal dimensions in 45,X-females.

Seventy-two females with 45,X-chromosome complement were examined for palatal dimensions, and the results of the measurements were compared to those of first-degree normal female relatives of the study subjects and population control females. Hard stone casts were prepared for measuring widths and lengths of the maxillary alveolar arch and palatal height between or at the level of the permanent canines, first and second premolars, and first molars with a palatometer and sharp-pointed vernier calipers (0.5 mm). According to the analyses of covariance (history of orthodontic treatment and loss of permanent teeth as cofactors and age as covariate), the differences in palatal dimensions between the groups were statistically significant or highly significant for all dimensions except for palatal height in the posterior segments. The group of the 45,X-females had the highest mean value for palatal height at the level of the canines and systematically the lowest mean values for palatal width in all segments, while no clear difference could be found in the length of the alveolar arch between the 45,X-females and the relatives, the population control group showing the lowest value in the posterior segments. The findings of this study indicate that the narrowed palate rather than the high palate is a frequent but not definite feature in 45,X-females. This may reflect the effect of sex-chromosomes on width but not other dimensions of the palate, resulting in deficiency of transversal growth of the palate possibly through decreased growth of the palatal shelves or through disturbances in the growth of the nasal septum, sutural growth, or--to which the exostosis on the palatal alveolar plates would refer--disturbances in apposition-resorption growth changes of the maxilla.

Factors predicting caries risk in children.

Selection of caries-risk groups is difficult and unreliable. In this study we analyzed the known correlates of caries and their interrelations. The subjects (n = 153) were seventh graders (13 yr old) from a school in Kuopio. Data were collected by mail survey (n = 150), from the clinical registers of the local health center (n = 135), and from a saliva test. In the first stage we computed the bivariate correlates between the D-index and its traditionally defined determinants. In the second stage the predictors of caries risk were ranked using discriminant and multiple regression analyses. The bivariate correlates of caries were: DMF-index of the previous year (r = 0.69), D-index of the previous year (r = 0.59), lactobacillus count (r = 0.26), buffer capacity (r = 0.20), and toothbrushing frequency (r = 0.07). In discriminant analyses the DMF-index among girls and the frequency of consumption of sweets among boys were the only significant discriminators. In the multiple regression analyses, however, the final model (P less than 0.001) for girls included: DMF-index of the previous year (beta = 0.64) and lactobacillus count (beta = 0.31). The corresponding model for boys (P less than 0.001) included: DMF-index of the previous year (beta = 0.55) and frequency of consumption of sweets (beta = 0.24). Using these variables, it is difficult to develop a reliable method for determining the caries susceptible groups, but of them, the DMF-index seems the most reliable for predicting caries risk among children.

Incidence of farmer's lung leading to hospitalization and its relation to meteorological observations in Finland.

The standardized annual incidence of farmer's lung (FL) leading to hospitalization during 1980 in Finland was 50 per 100 000 persons employed in farming. Using preset uniform criteria, data were collected from all pulmonary disease units in the country. There were 186 new cases of FL in 1980. The mean age was 46 years, and 64% were women. Using the national population in 1975 as standard, the rates were standardized for age and sex. There were differences both in the incidence and in the sex ratio of FL among the farming population of the six districts of Finland. These differences are closely related to prevalent proportions of cattle raising among farmers and local cultural traditions in the various districts. Seasonal variation in the incidence of FL was statistically significant (Roger's test, p less than 0.001), the peak occurring in April. This confirms an earlier observation. The incidence rates were statistically significantly correlated with the sum and the mean of daily rainfall during the haymaking period preceding the diagnosis of FL.

Fluoride content related to the elemental composition, mineral density and strength of bone in healthy and chronically diseased persons.

The fluoride (F) content of bone was determined and compared to the elemental composition, mineral density and compressive strength in bone specimens from the crista iliac of 88 subjects who had died suddenly, and from 50 subjects who had died as the result of chronic immobilizing diseases. The elemental composition of bone was determined using an atomic absorption spectrophotometer, mineral density using gamma-ray attenuation, compressive strength using a strain transducer, and F content using an ion-selective electrode. The increase of F content with age was not statistically significantly different between the groups. The Ca content of bone decreased more with age in post-menopausal women who had suffered from chronic diseases than in women who had died suddenly. The same was true for the negative correlation coefficient between Ca and F. A similar trend was found for mineral density. This does not support the idea that the F concentration obtained produces any effect that would clearly prevent bone loss in post-menopausal women when compared to those who had died suddenly. In both groups F content was correlated with Zn and Mn; in the sudden-death group Zn concentration explained 12.9% of the F concentration. The relation between Zn and F may be important in the formation and structure of biological apatite.

Cadmium content of human cancellous bone.

The cadmium content of human cancellous bone was related to age, sex, bone loss, physical properties, and elemental composition. Bone specimens from the anterior iliac crest were collected from 889 cadavers with a normal mineral status, and from 50 cadavers which had bone loss from chronic diseases and immobilization. The element concentrations were analyzed using atomic absorption spectrophotometry. Bone fluoride levels ere determined with the ion specific electrode, the mineral density with the gamma ray attenuation method, and the compressive strength with a strain transducer. The data were analyzed using multiple linear regression analysis. The mean cadmium content of 0.22 +/- 9.16 micrograms/g dry weight (+/- SD) in the samples did not change with age and its content was slightly greater in males than in females. Furthermore, no statistically significant relation was found in cadmium content to bone loss changes or to the calcium content of bone. The cadmium content had a high statistically significant positive correlation with the strontium and nickel content.

Interaction between Li, Ni and Sr content in human cancellous bone.

Li, Sr and Ni content of human cancellous bone and their relation to elemental composition was studied in specimens obtained from 88 subjects with a normal mineral status and from 50 subjects having osteoporotic changes due to chronic immobilizing diseases. The concentrations of Ca, Mg, Cd, Co, Cu, Mn, Ni, Li, Pb, Sr and Zn were analyzed with an atomic absorption spectrophotometer and fluoride with an ion selective electrode. The data was analyzed with the linear multiple regression analysis. Ni had a statistically significant inverse relation to Li. There was a significant positive correlation between Sr and Ni contents. Chronic immobilizing diseases did not affect significantly Li, Sr and Ni contents. Also no age dependent changes of these elements were found. The mean concentration S.D. of Li was 0.23 +/- 0.14 microgram/g, of Sr 65.5 +/- 40.7 micrograms/g and Ni 1.29 +/- 0.83 micrograms/g, respectively. The strong interrelation between Li, Sr and Ni suggest that these elements should be analyzed together in studying the role of these elements in the mineralization.

Zn and Cu content in human cancellous bone.

Duplicate cancellous bone samples from the iliac crest were obtained from autopsies of 88 persons with a normal mineral status and 50 having osteoporotic change due to chronic immobilizing diseases. The following parameters were determined: the minimal density with gamma-ray attenuation, the compressive strength with a strain transducer, the concentrations of Ca, Mg, Cd, Co, Cu, Mn, Ni, Li, Pb, Sr and Zn with an atomic absorption spectrophotometer and the concentration of F with an ion-selective electrode. The data were analyzed with linear multiple regression analysis. Altogether 34.8 per cent of the Zn concentration was accounted for in all the variables. The selected variables, after the nonsignificant and those with very low explanatory power were removed, explained 38.3 per cent. The two most important variables were the F concentration (16.1 per cent) and Cu concentration (13.7 per cent). The third variable was Mn concentration (8.5 per cent). All variables explained 41.7 per cent of the Cu concentration and five selected variables 40.7 per cent, respectively. The concentration of Zn explained itself as much (20.1 per cent) as four other variables (Ca 7.7 pe cent, Mn 5.9 per cent, Sr 3.9 per cent and Co 3.2 per cent) together. Age and sex did not significantly affect the concentration of Zn and Cu. No remarkable difference was found between the bone samples representing a normal mineral status and osteoporosis.

Zn concentration of human subgingival calculus related to F, Mg and Cu contents.

F and Zn content of human subgingival calculus was studied with respect to its Ca, Mg, Cu, Fe and Ph contents. The material consisted of 60 subgingival calculus samples from 60 persons. The element concentrations were determined with an atomic absorption spectrophotometer except for the F content, which was measured with an ion-selective electrode. The data were analyzed using the multiple linear regression analysis. 24% of F content was explained by Zn and none of the other elements were significantly related to F content. Altogether, 55.4% of Zn content was accounted for in all variables. The selected variables after the removal of nonsignificant ones explained 54.8%. The most important variable was F (24.0%) and the explanation power of Cu was 18.3% and of Mg 12.5%. The regression coefficients of these variables were positive. The mean +/- s.d. concentration of F in the samples was 221.2 +/- 174.1 micrograms/g and of Zn 498.6 +/- 499.1 micrograms/g, respectively.

Alveolar bone loss in relation to periodontal treatment need, socioeconomic status and dental health.

The distance from cemento-enamel junction to alveolar crest was studied in relation to periodontal treatment need, socioeconomic factors and dental health. The subjects (315) of the study were employees of a paper mill in Finland. Alveolar crest height was measured from orthopantomograms, periodontal treatment need was assessed by the Periodontal Treatment Need System (PTNS). Number of teeth, interproximal fillings and caries were determined clinically, and background information was obtained by questionnaires. On an average, the rate of alveolar bone loss was 0.6 mm in 10 years in the study population and 1.2 mm in subjects aged 30 to 45 years. The bone loss increased with decreasing number of teeth. Subjects with plaque retention had somewhat more severe bone loss than those with no inflammation or gingival inflammation only. Periodontal treatment need, age and number of remaining teeth altogether explained 42.8% of the bone loss when tested by multiple linear regression analysis.