Gestational diabetes is a type of diabetes that develops during pregnancy in women who didn't previously have it. It occurs when the body can't properly use insulin, leading to high blood sugar levels. Here is a breakdown of my analytical process:
Objectives:
To understand the impact of GD on maternal and fetal health.
Does race and ethnicity have an impact on GD?
Do women with high BP before or during pregnancy are at a higher risk of developing gestational diabetes?
Does Age and BMI play a role in developing GD?
Do women who are diagnosed with gestational diabetes later in their pregnancy are at a higher risk of complications?
Data Cleaning and modelling:
The original table has 259 columns and 600 rows. Used power query and power bi to perform cleaning and modelling.
Removed the duplicate columns in the data like Date of Data Collection.1, Miscarriage, V1 Creatinine.1, Date of Delivery, Screened 10 3, Miscarriage before 28/40
Removed columns with more than 60% null values, like ['Medications V1', 'V1 25OHD value (nmol/L)', 'Date 25OHD result received',
Irrelevant columns with respect to Weight, BMI, Height, HbA1c, Hb, Date, Lab values, and other columns were removed.
Removed the columns that we felt would not help with our objectives.
In the final stage of data cleaning, we focused on handling outliers, null values, and unique entries. Lab values and blood markers with null values were preserved for later calculations. Columns indicating medical diagnoses stored as "Yes/No" were replaced with "NR" for "Not Recorded" in the case of null values. Test or procedure columns with null values were set to "No" to indicate non-administration.
We standardized the Ethnicity column to match the provided codes and cleaned columns that duplicated data already found in dedicated separate columns.
After this process, there are 95 columns and 565 rows.
The different types of noise that we encountered in this dataset were:
Measurement errors – Some data was recorded incorrectly and we used Power Query Editor to replace those records.
Missing Data – The data collection was incomplete, so in order to fill in the missing information, certain columns were replaced with either a value of 0 or 'Not recorded', while other columns were imputed using the mean value of the available data, as necessary.
Bias – The data given comprises mostly white patients and very less records from other races.
Outliers – Some of the columns had outliers, but those columns were deleted as most of them had more than 60% null values.
Periodic Noise - Lots of repetition was encountered as columns representing the same values were available in both Yes/No and 1/0 format.
After cleaning the data did the modelling, which helps to organize and enrich data, optimize performance, and create insightful visualizations.
Analysis using Power BI:
1.Demographic Analysis:
In our Dataset we have 74 GDM Patients out of 565 total patients.
While comparing the significantly larger group without GDM to the smaller GDM group might yield misleading conclusions, analyzing a subset can offer valuable insights.
In our investigation, we discovered that patients diagnosed with GDM exhibited a higher incidence of pregnancy-related complications.
2. Physical Health of the patient as observed during visit 1:
Patients with high BP were almost twice as likely to be diagnosed with GDM (8% vs 4.8%)
Patients with a high HBA1C value at Visit 1 were 10x more likely to be diagnosed with GDM
Patients with a BMI of over 30 at visit 1 were 1.5x more likely to be diagnosed with GDM.
Hypercalcemia in pregnancy is an uncommon event that can cause major maternal morbidity and/or fetal or neonatal morbidity and mortality. The incidence of hypercalcemia in patients with GDM was 3x higher than those who did not present with GDM(19% vs 6%)
Patients with low levels of Vitamin D were 7% more likely to be diagnosed with GDM.
In the case of physical health, we cannot prove causation of GDM through these markers, as we do not have an understanding of individual risk factors. However, we found strong correlations between the physical markers listed above and a diagnosis of GDM later in pregnancy.
3. Health of the Liver and Kidneys:
Increased C-Reactive Protein levels are a sign of low-grade inflammation which may increase the risk of fetal growth restrictions and neonatal complications. Patients with GDM were 1.5 times more likely to have high levels of CRP (47% Vs 36.5%)
When liver cells are damaged, they release ALT into the bloodstream. High levels of ALT in the bloodstream may be a sign of a liver injury or disease. Patients with GDM were 3 times more likely to have signs of liver damage than those with no GDM (31% vs 12%)
High levels of Creatinine and Albumin can be an early sign of pregnancy-induced hypertension, which may increase the risk of pre-clampsia. Lab results show that patients with GDM were more than twice as likely to have high levels of both inflammatory markers (6.7% vs 3.1%)
4. Labor and Birth Related Complications:
47.3% of all pregnancies with GDM needed a C-Section Vs 31.03% in patients without GDM. Making C-Sections 1.5 times more likely if GDM was present.
Emergency procedures were twice as likely if a patient had GDM (17.6% Vs 8.8%)
A pregnancy was 1.5 times more likely to be classified as high risk if the patient was diagnosed with GDM (19% Vs 11%)
5.BMI Analysis:
BMI is really important because it helps us understand how our weight relates to our heart health. My analysis looked at people who had BMIs outside the recommended range.
I discovered that over 60% of individuals with BMIs outside the recommended ranges were admitted for emergencies.
This tells us that there's a strong link between BMI and the risk of needing urgent medical care.
I have also looked and found that more than 65% of them needed a cesarean section when they had a baby.
So, this analysis shows us just how important it is to keep an eye on our BMI and try to stay in the healthy range. It's not just about how we look; it's about our overall health and well-being."
Conclusion:
Effectively managing gestational diabetes (GDM) in women involves closely monitoring and controlling key factors such as blood pressure, HbA1c, vitamin D levels, BMI, and calcium. Keeping these factors in check not only helps prevent complications but also lowers the risk of serious health issues.
Early detection and treatment of GDM play a pivotal role in averting complications and promoting better outcomes for both the mother and the baby.
Comments