Gestão do conhecimento usando data mining: estudo de caso na Universidade Federal de Lavras

A gestão do conhecimento abrange toda a forma de gerar, armazenar, distribuir e utilizar o conhecimento, tornando necessária a utilização de tecnologias de informação para facilitar esse processo, devido ao grande aumento no volume de dados. A descoberta de conhecimento em banco de dados é uma metodologia que tenta solucionar esse problema e o data mining é uma técnica que faz parte dessa metodologia. Este artigo desenvolve, aplica e analisa uma ferramenta de data mining, para extrair conhecimento referente à produção científica das pessoas envolvidas com a pesquisa na Universidade Federal de Lavras. A metodologia utilizada envolveu a pesquisa bibliográfica, a pesquisa documental e o método do estudo de caso. As limitações encontradas na análise dos resultados indicam que ainda é preciso padronizar o modo do preenchimento dos currículos Lattes para refinar as análises e, com isso, estabelecer indicadores. A contribuição foi gerar um banco de dados estruturado, que faz parte de um processo maior de desenvolvimento de indicadores de ciência e tecnologia, para auxiliar na elaboração de novas políticas de gestão científica e tecnológica e aperfeiçoamento do sistema de ensino superior brasileiro.

The relationships between selected anthropometric and socio-economic data in schoolchildren from different social strata in Rio de Janeiro, Brazil

The nutritional status according to anthropometric data was assessed in 756 schoolchildren from 5 low-income state schools and in one private school in the same part of Rio de Janeiro, Brazil. The prevalence of stunting and wasting (cut-off point: <90% ht/age and <80% wt/ht) ranged in the public schools from 6.2 to 15.2% and 3.3 to 24.0%, respectively, whereas the figures for the private school were 2.3 and 3.5%, respectively. Much more obesity was found in the private school (18.0%) than in the state schools (0.8 - 6.2%). Nutritional problems seem to develop more severely in accordance with the increasing age of the children. Therefore it appears advisable to assess schoolchildren within the context of a nutritional surveillance system.

Trends in the modes of delivery and their impact on perinatal mortality rates

OBJECTIVE: To determine changes in the incidence of vaginal deliveries, cesarean sections, and forceps deliveries and their potential association with fetal, early neonatal, and perinatal mortality rates over time. METHODS: A retrospective study was carried out and the occurrence of deliveries supervised by university services between January 1991 and December 2000 was determined. Data regarding fetal, early neonatal, and perinatal deaths were assessed using obstetric and pediatric records and autopsy reports. RESULTS: Of a total of 33,360 deliveries, the incidence of vaginal deliveries, cesarean sections, and forceps deliveries was relatively steady (around 60, 30, and 10%, respectively) while, at the same time, there was a marked reduction in fetal mortality (from 33.3 to 13.0‰), early neonatal mortality (from 30.6 to 9.0‰), and perinatal mortality (from 56.4 to 19.3‰). CONCLUSIONS: The marked reduction in perinatal mortality rates seen during the study period without an increase in cesarean sections indicates that the decrease in perinatal mortality was not impacted by cesarean section rates. The plausible hypothesis seems to be that the reduction in perinatal mortality of deliveries performed under the supervision of university services was more likely to be associated with better neonatal care rather than the mode of delivery.

Suspected acute toxoplasmosis in pregnant women

OBJECTIVE: To determine the prevalence of reagent serology for suspected acute toxoplasmosis in pregnant women and to describe clinical, laboratory and therapeutic profiles of mothers and their children. METHODS: A retrospective study was conducted with IgM-anti-Toxoplasma gondii-reagent pregnant women and their children who attended the public health system in the state of Paraná, Southern Brazil, from January 2001 to December 2003. Information were obtained from clinical, laboratory (ELISA IgM/IgG) and ultrasonographic data and from interviews with the mothers. To test the homogeneity of the IgM indices in relation to the treatment used, the Pearson's Chi-square test was applied. Comparisons were considered significant at a 5% level. RESULTS: Two hundred and ninety (1.0%) cases of suspected IgM-reagent infection were documented, with a prevalence of 10.7 IgM-reagent women per 1,000 births. Prenatal care started within the first 12 weeks for 214/290; 146/204 were asymptomatic. Frequent complaints included headaches, visual disturbance and myalgia. Ultrasonography revealed abnormalities in 13 of 204 pregnancies. Chemoprophylaxis was administered to 112/227; a single ELISA test supported most decisions to begin treatment. Pregnant women with IgM indices =2.000 tended to be treated more often. Among exposed children, 44/208 were serologically followed up and all were IgG-reagent, and three IgM-reagent cases showed clinical symptoms. CONCLUSIONS: The existence of pregnant women with laboratorially suspected acute toxoplasmosis who were not properly followed up, and of fetuses that were not adequately monitored, shows that basic aspects of the prenatal care are not being systematically observed. There is need of implementing a surveillance system of pregnant women and their children exposed to T. gondii.

Agreement between data obtained from repeated interviews with a six-years interval

The objective of the study was to compare information collected through face-to-face interviews at first time and six years later in a city of Southeastern Brazil. In 1998, 32 mothers (N=32) of children aged 20 to 30 months answered a face-to-face interview with structured questions regarding their children's brushing habits. Six years later this same interview was repeated with the same mothers. Both interviews were compared for overall agreement, kappa and weighted kappa. Overall agreement between both interviews varied from 41 to 96%. Kappa values ranged from 0.00 to 0.65 (very bad to good) without any significant differences. The results showed lack of agreement when the same interview is conducted six years later, showing that the recall bias can be a methodological problem of interviews.

Risk factors for sudden infant death syndrome in a developing country

OBJECTIVE: To analyze whether previously identified risk factors for sudden death syndrome have a significant impact in a developing country. METHODS: Retrospective longitudinal case-control study carried out in Porto Alegre, Southern Brazil. Cases (N=39) were infants born between 1996 and 2000 who died suddenly and unexpectedly at home during sleep and were diagnosed with sudden death syndrome. Controls (N=117) were infants matched by age and sex who died in hospitals due to other conditions. Data were collected from postmortem examination records and questionnaires answers. A conditional logistic model was used to identify factors associated with the outcome. RESULTS: Mean age at death of cases was 3.2 months. The frequencies of infants regarding gestational age, breastfeeding and regular medical visits were similar in both groups. Sleeping position for most cases and controls was the lateral one. Supine sleeping position was found for few infants in both groups. Maternal variables, age below 20 years (OR=2, 95% CI: 1.1; 5.1) and smoking of more than 10 cigarettes per day during pregnancy (OR=3, 95% CI: 1.3; 6.4), significantly increased the risk for the syndrome. Socioeconomic characteristics were similar in both groups and did not affect risk. CONCLUSIONS: Infant-maternal and socioeconomic profiles of cases in a developing country closely resembled the profile described in the literature, and risk factors were similar as well. However, individual characteristics were identified as risks in the population studied, such as smoking during pregnancy and maternal age below 20 years.

Spatial analysis of urban violence based on emergency room data

OBJECTIVE: To estimate the spatial intensity of urban violence events using wavelet-based methods and emergency room data. METHODS: Information on victims attended at the emergency room of a public hospital in the city of São Paulo, Southeastern Brazil, from January 1, 2002 to January 11, 2003 were obtained from hospital records. The spatial distribution of 3,540 events was recorded and a uniform random procedure was used to allocate records with incomplete addresses. Point processes and wavelet analysis technique were used to estimate the spatial intensity, defined as the expected number of events by unit area. RESULTS: Of all georeferenced points, 59% were accidents and 40% were assaults. There is a non-homogeneous spatial distribution of the events with high concentration in two districts and three large avenues in the southern area of the city of São Paulo. CONCLUSIONS: Hospital records combined with methodological tools to estimate intensity of events are useful to study urban violence. The wavelet analysis is useful in the computation of the expected number of events and their respective confidence bands for any sub-region and, consequently, in the specification of risk estimates that could be used in decision-making processes for public policies.

Metabolic changes associated with antiretroviral therapy in HIV-positive patients

OBJECTIVE: To evaluate metabolic changes associated with highly active antiretroviral therapy (HAART) in HIV-positive patients, and to identify risk factors associated. METHODS: Retrospective study that included 110 HIV-positive patients who where on HAART in the city of Porto Alegre (Southern Brazil) between January 2003 and March 2004. Data on demographic variables, cigarette smoking, diabetes mellitus, cholesterol and triglyceride levels, stage of HIV infection, antiretroviral therapy and HCV coinfection were collected. General linear models procedure for repeated measures was used to test the interaction between HAART and HCV coinfection or protease inhibitor treatment. RESULTS: Total cholesterol, triglycerides, and glucose levels significantly increased after receiving HAART (p<0.001 for all variables), but no interaction with protease inhibitors was seen for total cholesterol, glucose and triglyceride levels (interaction treatment*protease inhibitors p=0.741, p=0.784, and p=0.081, respectively). An association between total cholesterol levels and HCV coinfection was found both at baseline and follow-up (effect of HCV coinfection, p=0.011). Glucose levels were increased by HAART (treatment effect, p=0.036), but the effect was associated to HCV coinfection (treatment*HCV effect, p=0.018). Gender, smoking habit, intravenous drug use and age were not significantly associated with cholesterol, triglyceride and glucose changes. CONCLUSIONS: HCV-infected patients at baseline were significantly less likely to develop hypercholesterolemia. The results provide further evidence of the role of HAART for the development of metabolic disturbances.

Risk factors of HIV-related oral lesions in adults

OBJECTIVE: To assess the risk factors in the occurrence of oral lesions in HIV-positive adults. METHODS: A retrospective analytical-descriptive survey was conducted using the medical/dental records of 534 patients with oral lesions associated with HIV. The data were collected from five referral centers for managing HIV and associated comorbidities in the city of Porto Alegre, Southern Brazil, between 1996 and 2011. Using a standardized form, socio-demographic and clinical data were recorded. Exclusively and definitively diagnosed oral pathologies were included and classified according to ECC criteria on Oral Problems Related to HIV Infection. For data analysis cross-tabulations, Chi-squared tests and logistic regression models were used where appropriate. RESULTS: CD4+ counts lower than 350 cells/mm³ (p < 0.001), alcohol consumption (p = 0.011) and female gender (p = 0.031) were predisposing factors for oral candidiasis. The occurrence of hairy leukoplakia was independently associated with CD4+ counts below 500 cells/mm³, (p = 0.029) a viral load above 5,000 copies/mm³ (p = 0.003) and smoking (p = 0.005). CONCLUSIONS: Moderate and severe degrees of immunodeficiency and detectable viral loads were risk factors for the onset of oral lesions. Smoking and alcohol consumption also increased susceptibility to the development of opportunistic infections in HIV-positive adults from Porto Alegre, irrespective of the use of antiretroviral therapy.

Direct treatment costs of HIV/AIDS in Portugal

OBJECTIVE To analyze the direct medical costs of HIV/AIDS in Portugal from the perspective of the National Health Service. METHODS A retrospective analysis of medical records was conducted for 150 patients from five specialized centers in Portugal in 2008. Data on utilization of medical resources during 12 months and patients’ characteristics were collected. A unit cost was applied to each care component using official sources and accounting data from National Health Service hospitals. RESULTS The average cost of treatment was 14,277 €/patient/year. The main cost-driver was antiretroviral treatment (€ 9,598), followed by hospitalization costs (€ 1,323). Treatment costs increased with the severity of disease from € 11,901 (> 500 CD4 cells/µl) to € 23,351 (CD4 count ≤ 50 cells/ µl). Cost progression was mainly due to the increase in hospitalization costs, while antiretroviral treatment costs remained stable over disease stages. CONCLUSIONS The high burden related to antiretroviral treatment is counterbalanced by relatively low hospitalization costs, which, however, increase with severity of disease. The relatively modest progression of total costs highlights that alternative public health strategies that do not affect transmission of disease may only have a limited impact on expenditure, since treatment costs are largely dominated by constant antiretroviral treatment costs.

Intra-hospital organ and tissue donation coordination project: cost-effectiveness and social benefits

OBJECTIVE To evaluate the viability of a professional specialist in intra-hospital committees of organ and tissue donation for transplantation. METHODS Epidemiological, retrospective and cross-sectional study (2003-2011 and 2008-2012), which was performed using organ donation for transplants data in the state of Sao Paulo, Southeastern Brazil. Nine hospitals were evaluated (hospitals 1 to 9). Logistic regression was used to evaluate the differences in the number of brain death referrals and actual donors (dependent variables) after the professional specialist started work (independent variable) at the intra-hospital committee of organ and tissue donation for transplantation. To evaluate the hospital invoicing, the hourly wage of the doctor and registered nurse, according to the legislation of the Consolidation of Labor Laws, were calculated, as were the investment return and the time elapsed to do so. RESULTS Following the nursing specialist commencement on the committee, brain death referrals and the number of actual donors increased at hospital 2 (4.17 and 1.52, respectively). At hospital 7, the number of actual donors also increased from 0.005 to 1.54. In addition, after the nurse started working, hospital revenues increased by 190.0% (ranging 40.0% to 1.955%). The monthly cost for the nurse working 20 hours was US$397.97 while the doctor would cost US$3,526.67. The return on investment was 275% over the short term (0.36 years). CONCLUSIONS This paper showed that including a professional specialist in intra-hospital committees for organ and tissue donation for transplantation proved to be cost-effective. Further economic research in the area could contribute to the efficient public policy implementation of this organ and tissue harvesting model.

Dynamic network data envelopment analysis for university hospitals evaluation

ABSTRACT OBJECTIVE To develop an assessment tool to evaluate the efficiency of federal university general hospitals. METHODS Data envelopment analysis, a linear programming technique, creates a best practice frontier by comparing observed production given the amount of resources used. The model is output-oriented and considers variable returns to scale. Network data envelopment analysis considers link variables belonging to more than one dimension (in the model, medical residents, adjusted admissions, and research projects). Dynamic network data envelopment analysis uses carry-over variables (in the model, financing budget) to analyze frontier shift in subsequent years. Data were gathered from the information system of the Brazilian Ministry of Education (MEC), 2010-2013. RESULTS The mean scores for health care, teaching and research over the period were 58.0%, 86.0%, and 61.0%, respectively. In 2012, the best performance year, for all units to reach the frontier it would be necessary to have a mean increase of 65.0% in outpatient visits; 34.0% in admissions; 12.0% in undergraduate students; 13.0% in multi-professional residents; 48.0% in graduate students; 7.0% in research projects; besides a decrease of 9.0% in medical residents. In the same year, an increase of 0.9% in financing budget would be necessary to improve the care output frontier. In the dynamic evaluation, there was progress in teaching efficiency, oscillation in medical care and no variation in research. CONCLUSIONS The proposed model generates public health planning and programming parameters by estimating efficiency scores and making projections to reach the best practice frontier.

Trends of T. cruzi infection based on data from blood bank screening

Between October 1988 and April 1989 a cross-sectional survey was carried out in six out of eight blood banks of Goiânia, Central Brazil. Subjects attending for first-time blood donation in the mornings of the study period (n = 1358) were interviewed and screened for T. cruzi infection as a part of a major study among blood donors. Tests to anti-T. cruzi antibodies were performed, simultaneously, by indirect hem agglutination test (IHA) and complement fixation test (CFT). A subject was considered seropositive when any one of the two tests showed a positive result. Information on age, sex, place of birth, migration and socio-economic level was recorded. Results from this survey were compared with seroprevalence rates obtained in previous studies in an attempt to analyse trend of T. cruzi infection in an endemic urban area. The overall seroprevalence of T. cruzi infection among first-time donors was found to be 3.5% (95% confidence interval 2.5%-4.5% ). The seroprevalence rate increased with age up to 45 years and then decreased. Migrants from rural areas had higher seroprevalence rates than subjects from urban counties (1.8%-16.2% vs. 0%-3.6%). A four fold decrease in prevalence rates was observed when these rates were compared with those of fifteen years ago. Two possible hypotheses to explain this difference were suggested: 1. a cohort effect related with the decrease of transmission in rural areas and/or 2. a differential proportion of people of rural origin among blood donors between the two periods. The potential usefulness of blood banks as a source of epidemiological information to monitor trends of T. cruzi infection in an urban adult population was stressed.

Reduced schedule of human anti-rabies immunization with Fuenzalida & Palacios vaccine: additional data

It was reevaluated a reduced schedule for anti-rabies post-exposure immunization with newborn mice nervous tissue vaccine (Fuenzalida 8c Palacios) in a group of 30 non exposed volunteers. The vaccine was administered by intramuscular injections on days zero, 2, 4, 16 and 27, in the deltoid area. Antibody levels were determinated by a simplified serum neutralization microtest on days zero, 16 and 37. On days 16 and 37 the antibody levels of the whole group was >0.5 IU/ml and >1.0 IU/ml, respectively. The cell mediated immunity was precociously detected (on day 4) by the delayed type hipersensitivity skin test. Our results show that this reduced schedule elicited an early and effective humoral and cellular immune response. However it is necessary other studies with larger groups of vaccinees in order to obtain definitive conclusion.