Novel statistical classification model of type 2 diabetes mellitus patients for tailor-made prevention using data mining algorithm

Improvement of grey wolf optimizer with adaptive middle filter to adjust support vector machine parameters to predict diabetes complications

In medical science, collecting and classifying data from various diseases is a vital task. The confused and large amounts of data are problems that prevent us from achieving acceptable results. One of the major problems for diabetic patients is a failure to properly diagnose the disease. As a result of this mistake in diagnosis or failure in early diagnosis, the patient may suffer from complications such as blindness, kidney failure, and cutting off the toes. Nowadays, doctors diagnose the disease by relying on their experience and knowledge and performing complex and time-consuming tests. One of the problems with current diabetic, diagnostic methods is the lack of appropriate features to diagnose the disease and consequently the weakness in its diagnosis, especially in its early stages. Since diabetes diagnosis relies on large amounts of data with many parameters, it is necessary to use machine learning methods such as support vector machine (SVM) to predict the complications of diabetes. One of the disadvantages of SVM is its parameter adjustment, which can be accomplished using metaheuristic algorithms such as particle swarm optimization algorithm (PSO), genetic algorithm, or grey wolf optimizer (GWO). In this paper, after preprocessing and preparing the dataset for data mining, we use SVM to predict complications of diabetes based on selected parameters of a patient acquired by laboratory test using improved GWO. We improve the selection process of GWO by employing dynamic adaptive middle filter, a nonlinear filter that assigns appropriate weight to each value based on the data value. Comparison of the final results of the proposed algorithm with classification methods such as a multilayer perceptron neural network, decision tree, simple Bayes, and temporal fuzzy min–max neural network (TFMM-PSO) shows the superiority of the proposed method over the comparable ones.

Mining data when technology is applied to support patients and professional on the control of chronic diseases: the experience of the METABO platform for diabetes management

This chapter provides an overview of how healthcare institution could benefit from the usage of technologies and personal health systems. Clinical, Usage and Technical data are mined in different ways and with different methods to support users (patients, health professionals and informal caregivers) in taking decisions. As a case study, the solutions and the techniques adopted in a research project focused on the delivery of technologies to improve diabetes management are described.

Application of data mining techniques to explore predictors of HCC in Egyptian patients with HCV-related chronic liver disease

BACKGROUND

Hepatocellular carcinoma (HCC) is the second most common malignancy in Egypt. Data mining is a method of predictive analysis which can explore tremendous volumes of information to discover hidden patterns and relationships. Our aim here was to develop a non-invasive algorithm for prediction of HCC. Such an algorithm should be economical, reliable, easy to apply and acceptable by domain experts.

METHODS

This cross-sectional study enrolled 315 patients with hepatitis C virus (HCV) related chronic liver disease (CLD); 135 HCC, 116 cirrhotic patients without HCC and 64 patients with chronic hepatitis C. Using data mining analysis, we constructed a decision tree learning algorithm to predict HCC.

RESULTS

The decision tree algorithm was able to predict HCC with recall (sensitivity) of 83.5% and precession (specificity) of 83.3% using only routine data. The correctly classified instances were 259 (82.2%), and the incorrectly classified instances were 56 (17.8%). Out of 29 attributes, serum alpha fetoprotein (AFP), with an optimal cutoff value of ≥50.3 ng/ml was selected as the best predictor of HCC. To a lesser extent, male sex, presence of cirrhosis, AST>64U/L, and ascites were variables associated with HCC.

CONCLUSION

Data mining analysis allows discovery of hidden patterns and enables the development of models to predict HCC, utilizing routine data as an alternative to CT and liver biopsy. This study has highlighted a new cutoff for AFP (≥50.3 ng/ml). Presence of a score of >2 risk variables (out of 5) can successfully predict HCC with a sensitivity of 96% and specificity of 82%.

Developing screening services for colorectal cancer on Android smartphones

INTRODUCTION

Colorectal cancer (CRC) is an important health problem in Western countries and also in Asia. It is the third leading cause of cancer deaths in both men and women in Taiwan. According to the well-known adenoma-to-carcinoma sequence, the majority of CRC develops from colorectal adenomatous polyps. This concept provides the rationale for screening and prevention of CRC. Removal of colorectal adenoma could reduce the mortality and incidence of CRC. Mobile phones are now playing an ever more crucial role in people's daily lives. The latest generation of smartphones is increasingly viewed as hand-held computers rather than as phones, because of their powerful on-board computing capability, capacious memories, large screens, and open operating systems that encourage development of applications (apps).

SUBJECTS AND METHODS

If we can detect the potential CRC patients early and offer them appropriate treatments and services, this would not only promote the quality of life, but also reduce the possible serious complications and medical costs. In this study, an intelligent CRC screening app on Android™ (Google™, Mountain View, CA) smartphones has been developed based on a data mining approach using decision tree algorithms. For comparison, the stepwise backward multivariate logistic regression model and the fecal occult blood test were also used.

RESULTS

Compared with the stepwise backward multivariate logistic regression model and the fecal occult blood test, the proposed app system not only provides an easy and efficient way to quickly detect high-risk groups of potential CRC patients, but also brings more information about CRC to customer-oriented services.

CONCLUSIONS

We developed and implemented an app system on Android platforms for ubiquitous healthcare services for CRC screening. It can assist people in achieving early screening, diagnosis, and treatment purposes, prevent the occurrence of complications, and thus reach the goal of preventive medicine.

Factors affecting length of stay in the pediatric emergency department

BACKGROUND

A large volume of visits can cause an emergency department (ED) to become overcrowded, resulting in a longer length of stay (LOS). The objective of this study was to analyze factors affecting the LOS in the pediatric ED.

METHODS

Records of all visits to the pediatric ED of the study hospital, from July 1, 2006 to June 31, 2007, were retrospectively retrieved. Data were collected from the hospital's computerized records system. Eta-squared correlation ratio and Cramer's V test evaluated the associations between variables. Two-thirds of the database was randomized for the classification and regression tree (CART) model-building dataset, and one-third was used for the validation dataset.

RESULTS

A total of 29,035 patients visited the pediatric ED during the evaluation period. Of the total visits, 61.1% were due to complaints of fever. The mean LOS was 2.6 ± 4.67 hours, and 74.3% of visits had an LOS of shorter than 2 hours. The CART analysis selected five factors (waiting time for hospitalization, laboratory tests, door-to-physician time, gastrointestinal symptoms, and patient outcome) to produce a total of nine subgroups of patients. The mean LOS of the model-building dataset closely correlated with that of the validation dataset (r(2) = 0.999).

CONCLUSION

Patients who were waiting for hospitalization for less than 8 hours or were not admitted, those without any laboratory tests, those having door-to-physician time less than 60 minutes, and those without any gastrointestinal symptoms had the shortest LOS. Patients who waited for hospitalization for more than 16 hours had the longest LOS.

The assessment of data mining for the prediction of therapeutic outcome in 3719 Egyptian patients with chronic hepatitis C

INTRODUCTION

Decision-tree analysis; a core component of data mining analysis can build predictive models for the therapeutic outcome to antiviral therapy in chronic hepatitis C virus (HCV) patients.

AIM

To develop a prediction model for the end virological response (ETR) to pegylated interferon PEG-IFN plus ribavirin (RBV) therapy in chronic HCV patients using routine clinical, laboratory, and histopathological data.

PATIENTS AND METHODS

Retrospective initial data (19 attributes) from 3719 Egyptian patients with chronic HCV presumably genotype-4 was assigned to model building using the J48 decision tree-inducing algorithm (Weka implementation of C4.5). All patients received PEG-IFN plus RBV at Cairo-Fatemia Hospital, Cairo, Egypt in the context of the national treatment program. Factors predictive of ETR were explored and patients were classified into seven subgroups according to the different rates of ETR. The universality of the decision-tree model was subjected to a 10-fold cross-internal validation in addition to external validation using an independent dataset collected of 200 chronic HCV patients.

RESULTS

At week 48, overall ETR was 54% according to intention to treat protocol. The decision-tree model included AFP level (<8.08 ng/ml) which was associated with high probability of ETR (73%) followed by stages of fibrosis and Hb levels according to the patients' gender followed by the age of patients.

CONCLUSION

In a decision-tree model for the prediction for antiviral therapy in chronic HCV patients, AFP level was the initial split variable at a cutoff of 8.08 ng/ml. This model could represent a potential tool to identify patients' likelihood of response among difficult-to-treat presumably genotype-4 chronic HCV patients and could support clinical decisions regarding the proper selection of patients for therapy without imposing any additional costs.

Model incorporating the ITPA genotype identifies patients at high risk of anemia and treatment failure with pegylated-interferon plus ribavirin therapy for chronic hepatitis C

This study aimed to develop a model for predicting anemia using the inosine triphosphatase (ITPA) genotype and to evaluate its relationship with treatment outcome. Patients with genotype 1b chronic hepatitis C (n = 446) treated with peg-interferon alpha and ribavirin (RBV) for 48 weeks were genotyped for the ITPA (rs1127354) and IL28B (rs8099917) genes. Data mining analysis generated a predictive model for anemia (hemoglobin (Hb) concentration <10 g/dl); the CC genotype of ITPA, baseline Hb <14.0 g/dl, and low creatinine clearance (CLcr) were predictors of anemia. The incidence of anemia was highest in patients with Hb <14.0 g/dl and CLcr <90 ml/min (76%), followed by Hb <14.0 g/dl and ITPA CC (57%). Patients with Hb ≥ 14.0 g/dl and ITPA AA/CA had the lowest incidence of anemia (17%). Patients with two predictors (high-risk) had a higher incidence of anemia than the others (64% vs. 28%, P < 0.0001). At baseline, the IL28B genotype was a predictor of a sustained virological response [adjusted odds ratio 9.88 (95% confidence interval 5.01-19.48), P < 0.0001]. In patients who achieved an early virological response, the IL28B genotype was not associated with a sustained virological response, while a high risk of anemia was a significant negative predictor of a sustained virological response [0.47 (0.24-0.91), P = 0.026]. For high-risk patients with an early virological response, giving >80% of the planned RBV dose increased sustained virological responses by 24%. In conclusion, a predictive model incorporating the ITPA genotype could identify patients with a high risk of anemia and reduced probability of sustained virological response.

Dietary patterns analysis using data mining method. An application to data from the CYKIDS study

Data mining is a computational method that permits the extraction of patterns from large databases. We applied the data mining approach in data from 1140 children (9-13 years), in order to derive dietary habits related to children's obesity status. Rules emerged via data mining approach revealed the detrimental influence of the increased consumption of soft dinks, delicatessen meat, sweets, fried and junk food. For example, frequent (3-5 times/week) consumption of all these foods increases the risk for being obese by 75%, whereas in children who have a similar dietary pattern, but eat >2 times/week fish and seafood the risk for obesity is reduced by 33%. In conclusion patterns revealed from data mining technique refer to specific groups of children and demonstrate the effect on the risk associated with obesity status when a single dietary habit might be modified. Thus, a more individualized approach when translating public health messages could be achieved.

Data mining model using simple and readily available factors could identify patients at high risk for hepatocellular carcinoma in chronic hepatitis C

BACKGROUND & AIMS

Assessment of the risk of hepatocellular carcinoma (HCC) development is essential for formulating personalized surveillance or antiviral treatment plan for chronic hepatitis C. We aimed to build a simple model for the identification of patients at high risk of developing HCC.

METHODS

Chronic hepatitis C patients followed for at least 5 years (n=1003) were analyzed by data mining to build a predictive model for HCC development. The model was externally validated using a cohort of 1072 patients (472 with sustained virological response (SVR) and 600 with nonSVR to PEG-interferon plus ribavirin therapy).

RESULTS

On the basis of factors such as age, platelet, albumin, and aspartate aminotransferase, the HCC risk prediction model identified subgroups with high-, intermediate-, and low-risk of HCC with a 5-year HCC development rate of 20.9%, 6.3-7.3%, and 0-1.5%, respectively. The reproducibility of the model was confirmed through external validation (r(2)=0.981). The 10-year HCC development rate was also significantly higher in the high-and intermediate-risk group than in the low-risk group (24.5% vs. 4.8%; p<0.0001). In the high-and intermediate-risk group, the incidence of HCC development was significantly reduced in patients with SVR compared to those with nonSVR (5-year rate, 9.5% vs. 4.5%; p=0.040).

CONCLUSIONS

The HCC risk prediction model uses simple and readily available factors and identifies patients at a high risk of HCC development. The model allows physicians to identify patients requiring HCC surveillance and those who benefit from IFN therapy to prevent HCC.

Sequences in the interferon sensitivity-determining region and core region of hepatitis C virus impact pretreatment prediction of response to PEG-interferon plus ribavirin: data mining analysis

The aim of the present study was to clarify the significance of viral factors for pretreatment prediction of sustained virological response to pegylated-interferon (PEG-IFN) plus ribavirin (RBV) therapy for chronic hepatitis C using data mining analysis. Substitutions in the IFN sensitivity-determining region (ISDR) and at position 70 of the HCV core region (Core70) were determined in 505 patients with genotype 1b chronic hepatitis C treated with PEG-IFN plus RBV. Data mining analysis was used to build a predictive model of sustained virological response in patients selected randomly (n = 304). The reproducibility of the model was validated in the remaining 201 patients. Substitutions in ISDR (odds ratio = 9.92, P < 0.0001) and Core70 (odds ratio = 1.92, P = 0.01) predicted sustained virological response independent of other covariates. The decision-tree model revealed that the rate of sustained virological response was highest (83%) in patients with two or more substitutions in ISDR. The overall rate of sustained virological response was 44% in patients with a low number of substitutions in ISDR (0-1) but was 83% in selected subgroups of younger patients (<60 years), wild-type sequence at Core70, and higher level of low-density lipoprotein cholesterol (LDL-C) (≥ 120 mg/dl). Reproducibility of the model was validated (r(2) = 0.94, P < 0.001). In conclusion, substitutions in ISDR and Core70 of HCV are significant predictors of response to PEG-IFN plus RBV therapy. A decision-tree model that includes these viral factors as predictors could identify patients with a high probability of sustained virological response.

Data-mining technologies for diabetes: a systematic review

BACKGROUND

The objective of this study is to conduct a systematic review of applications of data-mining techniques in the field of diabetes research.

METHOD

We searched the MEDLINE database through PubMed. We initially identified 31 articles by the search, and selected 17 articles representing various data-mining methods used for diabetes research. Our main interest was to identify research goals, diabetes types, data sets, data-mining methods, data-mining software and technologies, and outcomes.

RESULTS

The applications of data-mining techniques in the selected articles were useful for extracting valuable knowledge and generating new hypothesis for further scientific research/experimentation and improving health care for diabetes patients. The results could be used for both scientific research and real-life practice to improve the quality of health care diabetes patients.

CONCLUSIONS

Data mining has played an important role in diabetes research. Data mining would be a valuable asset for diabetes researchers because it can unearth hidden knowledge from a huge amount of diabetes-related data. We believe that data mining can significantly help diabetes research and ultimately improve the quality of health care for diabetes patients.

Pretreatment prediction of anemia progression by pegylated interferon alpha-2b plus ribavirin combination therapy in chronic hepatitis C infection: decision-tree analysis

BACKGROUND

This study aimed to develop a model to predict the development of severe anemia during pegylated interferon alpha-2b plus ribavirin combination therapy.

METHODS

Data were collected from 1081 genotype 1b chronic hepatitis C patients who were treated at 6 hospitals in Japan. These patients were randomly assigned to a model-building group (n = 691) or an internal validation group (n = 390). Factors predictive of severe anemia (hemoglobin, Hb < 8.5 g/dl) were explored using data-mining analysis.

RESULTS

Hb values at baseline, creatinine clearance (Ccr), and an Hb concentration decline by 2 g/dl at week 2 were used to build a decision-tree model, in which the patients were divided into 5 subgroups based on variable rates of severe anemia ranging from 0.4 to 11.8%. The reproducibility of the model was confirmed by the internal validation group (r² = 0.96). The probability of severe anemia was high in patients whose Hb value was <14 g/dl before treatment (6.5%), especially (a) in those whose Ccr was <80 ml/min (11.8%) and (b) those whose Ccr was ≥ 80 ml/min but whose Hb concentration decline at week 2 was ≥ 2 g/dl (11.5%). The probability of severe anemia was low in the other patients (0.4-2.5%).

CONCLUSIONS

The decision-tree model that included Hb values at baseline, Ccr, and an Hb concentration decline by 2 g/dl at week 2 was useful for predicting the probability of severe anemia, and has the potential to support clinical decisions regarding early dose reduction of ribavirin.

Pretreatment prediction of response to peginterferon plus ribavirin therapy in genotype 1 chronic hepatitis C using data mining analysis

BACKGROUND

This study aimed to develop a model for the pre-treatment prediction of sustained virological response (SVR) to peg-interferon plus ribavirin therapy in chronic hepatitis C.

METHODS

Data from 800 genotype 1b chronic hepatitis C patients with high viral load (>100,000 IU/ml) treated by peg-interferon plus ribavirin at 6 hospitals in Japan were randomly assigned to a model building (n = 506) or an internal validation (n = 294). Data from 524 patients treated at 29 hospitals in Japan were used for an external validation. Factors predictive of SVR were explored using data mining analysis.

RESULTS

Age (<50 years), alpha-fetoprotein (AFP) (<8 ng/mL), platelet count (≥ 120 × 10(9)/l), gamma-glutamyltransferase (GGT) (<40 IU/l), and male gender were used to build the decision tree model, which divided patients into 7 subgroups with variable rates of SVR ranging from 22 to 77%. The reproducibility of the model was confirmed by the internal and external validation (r (2) = 0.92 and 0.93, respectively). When reconstructed into 3 groups, the rate of SVR was 75% for the high probability group, 44% for the intermediate probability group and 23% for the low probability group. Poor adherence to drugs lowered the rate of SVR in the low probability group, but not in the high probability group.

CONCLUSIONS

A decision tree model that includes age, gender, AFP, platelet counts, and GGT is useful for predicting the probability of response to therapy with peg-interferon plus ribavirin and has the potential to support clinical decisions regarding the selection of patients for therapy.

A predictive model of response to peginterferon ribavirin in chronic hepatitis C using classification and regression tree analysis

AIM

Early disappearance of serum hepatitis C virus (HCV) RNA is the prerequisite for achieving sustained virological response (SVR) in peg-interferon (PEG-IFN) plus ribavirin (RBV) therapy for chronic hepatitis C. This study aimed to develop a decision tree model for the pre-treatment prediction of response.

METHODS

Genotype 1b chronic hepatitis C treated with PEG-IFN alpha-2b and RBV were studied. Predictive factors of rapid or complete early virological response (RVR/cEVR) were explored in 400 consecutive patients using a recursive partitioning analysis, referred to as classification and regression tree (CART) and validated.

RESULTS

CART analysis identified hepatic steatosis (<30%) as the first predictor of response followed by low-density-lipoprotein cholesterol (LDL-C) (>/=100 mg/dL), age (<50 and <60 years), blood sugar (<120 mg/dL), and gamma-glutamyltransferase (GGT) (<40 IU/L) and built decision tree model. The model consisted of seven groups with variable response rates from low (15%) to high (77%). The reproducibility of the model was confirmed by the independent validation group (r(2) = 0.987). When reconstructed into three groups, the rate of RVR/cEVR was 16% for low probability group, 46% for intermediate probability group and 75% for high probability group.

CONCLUSIONS

A decision tree model that includes hepatic steatosis, LDL-C, age, blood sugar, and GGT may be useful for the prediction of response before PEG-IFN plus RBV therapy, and has the potential to support clinical decisions in selecting patients for therapy and may provide a rationale for treating metabolic factors to improve the efficacy of antiviral therapy.

Factors alleviating metabolic syndrome via diet-induced weight loss with or without exercise in overweight Japanese women

OBJECTIVE

Although a 5%-10% loss in the baseline weight has been associated with improvement in obesity-related disorders, only a few studies have explored the factors to alleviate metabolic syndrome (MS). This study aimed to determine the factors that alter MS components in overweight Japanese women.

METHODS

Between 1999 and 2006, 323 Japanese women aged 24-67 with body mass indices of 25-40 kg/m(2) and the presence of at least 1 component of MS were recruited from Ibaraki and Chiba. The participants were enrolled in a 3-month weight-loss program with a low-calorie diet with or without exercise. The factors to alleviate MS components were explored using classification and regression tree (CART) analyses.

RESULTS

Of the 323 participants, 309 completed the weight-loss program and were included in the analyses. The CART analyses revealed that a weight reduction of 8.1% in baseline body weight was sufficient to improve at least 1 component of MS. Similarly, classification trees were generated for improvement in abdominal obesity (essential factor: > or =13.0% weight loss), hypertension (essential factor: baseline age, < or =41.5 years), and hyperglycemia (essential factor: > or =13.2% weight loss).

CONCLUSION

These results suggest that moderate weight loss of 8%-13% contributes to improving the MS components in overweight Japanese women.

Applications of artificial intelligence systems in the analysis of epidemiological data

A brief review of the germane literature suggests that the use of artificial intelligence (AI) statistical algorithms in epidemiology has been limited. We discuss the advantages and disadvantages of using AI systems in large-scale sets of epidemiological data to extract inherent, formerly unidentified, and potentially valuable patterns that human-driven deductive models may miss.

Interactions between psychosocial problems and management of asthma: who is at risk of dying?

Adjustment for psychosocial and family problems is common in epidemiological research. Recursive partitioning algorithms, such as CHi Square Automatic Interaction Detection (CHAID), can be used to explore complex interactions between these factors and predictor and outcome variables. We investigated the nature of interactions between asthma management variables and psychosocial problems and how these interactions changed the risk of asthma mortality; 50 cases of asthma death and 201 emergency department controls were recruited. A validated questionnaire was used to collect data. An extended version of CHAID was used to identify statistically significant (p < or = 0.05) interactions controlling for asthma severity. Family problems were associated with increased risk of mortality for patients aged > 31 years (OR = 6.5; 95% CI 2.6-16.1) but not for younger patients. Males were at increased risk overall, but females with family problems (OR = 4.3; 95% CI 1.7-10.7) were at greater risk then males (OR = 3.1; 95% CI 1.2-7.9) with family problems. Alcohol use increased risk of mortality for individuals with verbal instructions (OR = 5.4; 95% CI 1.5-19.5) or without a written action plan (OR = 4.4; 95% CI 1.0-19.4). Individuals with severe asthma and who reported having lung function tests were at increased risk for mortality if family (OR = 8.2; 95% CI 1.6-41.6) or financial problems (OR = 11.5; 95% CI 2.0-65.9) were present. This analysis highlights some important interactions and the magnitude of additional risk for mortality associated with psychosocial or family problems. Psychosocial problems need to be identified and addressed as part of asthma management, because even with best practice, these problems place patients at an increased risk of dying.

High throughput multiple combination extraction from large scale polymorphism data by exact tree method

Single nucleotide polymorphisms (SNPs) are increasingly becoming important in clinical settings as useful genetic markers. For the evaluation of genetic risk factors of multifactorial diseases, it is not sufficient to focus on individual SNPs. It is preferable to evaluate combinations of multiple markers, because it allows us to examine the interactions between multiple factors. If all the combinations possible were evaluated round-robin, the number of calculations would rapidly explode as the number of markers analyzed increased. To overcome this limitation, we devised the exact tree method based on decision tree analysis and applied it to 14 SNP data from 68 Japanese stroke patients and 189 healthy controls. From the obtained tree models, we succeeded in extracting multiple statistically significant combinations that elevate the risk of stroke. From this result, we inferred that this method would work more efficiently in the whole genome study, which handles thousands of genetic markers. This exploratory data mining method will facilitate the extraction of combinations from large-scale genetic data and provide a good foothold for further verificatory research.