logo

Assignment : Women and Diabetes

   

Added on  2020-03-16

7 Pages3690 Words56 Views
Women and DiabetesWOMEN AND DIABETESAuthor: ________________

Women and DiabetesAbstractThis paper focus to test hypotheses revolving around the development of diabetes for women based on several diagnostic among other factors such as the number of pregnancies and pedigree function. Also, a logistic regression model is developed to predict the probability of developing diabetes based on several factors included in the model. Keyword: BMI – Body Mass Index, GLM –Generalized Linear Models, CatBMI – Categorical variable for BMI <30 and ≤301.INTRODUCTIONDiabetes is a condition caused by increased levels of blood glucose and can be prevented andmanaged effectively when diagnosed [1] [2]. The elevated blood glucose levels in the bodyincrease the risk of having several health conditions such as; nerve damage, kidney problems, heartdiseases and eye problems among others. According to health surveillance statistics in 2014,around 9.3% of U.S citizens who are aged above 20 years have diabetes [3]. Despite the highprevalence of the condition, close to 25% of the disease does not know about their status. Thehealth department has raised concerns for the need of diagnostic tests for the population under riskof being diabetic to enable them to engagein preventive and management measures. Controlleddiabetes reduces the risk of developing the health conditions mentioned above. Therefore, it is veryimportant to sensitize people to undertake the required diagnostic measures to monitor their sugarlevels over time. There are two types of diabetes which include; type 1 and type 2. Type I diabetes ischaracterized by the body’s immune system destroying the cells responsible for insulin production.In this case, insufficient production of insulin leads to reduced sugar absorption in the body. Thepatients end up feeling weak because glucose is required for energy production. On the other hand,type 2 diabetes patients’ body is not able to utilize insulin in the right manner, hence resulting inthe same issues as type 1. Therefore, individuals suffering from type 1 diabetes will showindications of low blood sugars while those with type 2 diabetes will show high levels. Also,pregnant women are at risk of developing gestational diabetes which can only be tested 24 and 28thweeks of pregnancy. There is a chance that the patient will develop type 2 diabetes after givingbirth. This paper will focus on prediction on whether a woman above 21 years will developdiabetes based on some diagnostic measures. The variables to be considered include the number oftimes an individual has been pregnant, glucose levels, diastolic blood pressure, insulin, BMI, skinthickness, age and diabetes pedigree function. 1.1Research Questionsi.What is nature statistical relationships between the outcome variable [being diabetic] and the independent variables?ii.What is the set of predictor variables, which provide maximum evidence in connection to the outcome? 1.2Hypothesisi.Diabetic women have higher glucose levels than the non-diabeticii.Diabetic women have higher diabetes pedigree function levels than the non-diabeticiii.Blood pressures are higher for the diabetic than the non-diabetic. iv.Women with BMI greater than 30 have higher risks of developing diabetes2.DESCRIPTIVE ANALYSIS1.1Descriptive Statistics1.1.1PREDICTOR VARIABLESTable 1: Descriptive statistics for continuous variablesVariableMeanMinimum25th percentileMedian75thpercentileMaximumGlucose120.90.099117140.2199Blood pressure69.110.0627280122Skin Thickness20.540.00.0233299Insulin79.80.00.030.5127.2846BMI31.990.027.33236.667.1Diabetes Pedigree Function 0.47190.0780.24370.37250.62622.42Age33.242124294181The variables in the descriptive table 1 above are approximately normal because there isminimal deviation of the median from the mean. Figure 1: Boxplot for Glucose levels, Blood pressure, Skin thickness and Insulin variablesFigure 1 shows that glucose and blood pressure variables had some missing values that havebeen recorded as zeros. Blood pressure variable has outliers on the both side with values above120mmHg and below 40mmHg. Skin thickness values also had outliers and the data is slightlyskewed to the right. The two-hour serum insulin measure was highly skewed to the right withvalues above 300 mu U/ml. Figure 2: Boxplots for BMI, Diabetes pedigree function and Age variablesAccording to figure 2, BMI is approximately normal with some cases of outliers on the upper end and some missing values. Diabetes pedigree function and age variables have distributions that are skewed to the right because of the outliers. Table 2: Pregnancies distributionPregnancies012345678910111213141517Frequency111135103756857504538282411910211The variable of a number of pregnancies is a discrete variable ranging from 0 to 17. Table 2 shows that that the frequency of women reduces the number of pregnancies increase.

Women and DiabetesFigure 3: Distribution of number of pregnanciesAccording to figure 3 above, the distribution of pregnancies is skewed to the right by the indication of decreasing count as the number increases [2].1.1.2OUTCOME VARIABLETable 3: Table for outcome variableOutcomeDiabetesNo diabetesTotalCount268500768Percent34.9%65.1%Figure 4: Outcome pie chart3.LOGISTIC REGRESSION 3.1CorrelationsFigure 5: Correlations of predictors [No of pregnancies, glucose, Blood pressure, skin thickness, insulin and BMI] by OutcomeThere is a distinct difference in the number of pregnancies between women who emerged to bediabetic compared to the others. On average, the glucose levels for the diabetic women werehigher, which indicates that the body sugars were not effectively used energy production. Also, theblood pressure levels for the diabetic group were slightly higher compared to the healthy group.The distribution of skin thickness seems to be constant from 0 to around 30mm; however, thediabetic groups have slightly thicker skin on average. The non-diabetic group of women havehigher levels of insulin on average compared to the diabetics. People who were found to havehigher BMI levels were at higher risk of developing diabetes than their counterparts [4]. Figure 6: Boxplot for Diabetes pedigree function and Age by Outcome

End of preview

Want to access all the pages? Upload your documents or become a member.

Related Documents
Gestational diabetes Name of the university Author note
|2
|353
|376

Gestational Diabetes: Assignment
|12
|3545
|475

Graduate Certificate in Diabetic Management Plan | Report
|14
|2848
|20

Gestational Diabetes Management: Risk Factors, Interventions, and Follow-up
|9
|2271
|449

Risk of Type 2 Diabetes in Asian Women PDF
|6
|1354
|480

Risk Factors and Aetiology of Type 2 Diabetes
|4
|1666
|460