Secondary Objectives 1. To evaluate the prevalence of dyslipidemia among T2DM patients categorized by BMI class (normal, overweight, obesel, obese II). 2. To evaluate the prevalence of dyslipidemia among T2DM patients stratified by WHR (normal vs. high WHR). 3. To compare the strength of correlation of BMI and WHR with lipid profile parameters. These are my secondary objectives for irb How can I make tables for the same in my dissertation and are they easily feasible?
Searching the Web
dyslipidemia prevalence T2DM BMI WHR lipid profile dissertation table design
Searching PubMed
dyslipidemia T2DM BMI WHR lipid profile
Here is a complete, practical guide for all three secondary objectives.
Are These Objectives Feasible?
Yes, all three are highly feasible - this is a standard cross-sectional study design. You need:
- T2DM patient records with fasting lipid profile (TC, LDL, HDL, TG)
- Anthropometric measurements (height, weight for BMI; waist and hip circumference for WHR)
- A working definition of dyslipidemia (e.g., NCEP-ATP III or IDF criteria)
The data collection is straightforward outpatient/OPD-based. Prevalence studies of this type are commonly published with sample sizes of 100-300 patients.
Table Designs for Each Objective
Objective 1 - Dyslipidemia Prevalence by BMI Class
Table 1: Prevalence of Dyslipidemia Among T2DM Patients by BMI Category
| Dyslipidemia Component | Normal BMI (<18.5–24.9) n=__ | Overweight (25–29.9) n=__ | Obese I (30–34.9) n=__ | Obese II (≥35) n=__ | Total n=__ | p-value |
|---|---|---|---|---|---|---|
| High Total Cholesterol (>200 mg/dL) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| High LDL-C (>130 mg/dL) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| Low HDL-C (<40 M / <50 F mg/dL) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| High Triglycerides (>150 mg/dL) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| Any Dyslipidemia (overall) | n (%) | n (%) | n (%) | n (%) | n (%) |
Statistical test: Chi-square test (or Fisher's exact if cell counts <5) for comparing proportions across BMI groups.
Supplementary table option: Add mean lipid values (±SD) per BMI group and test with one-way ANOVA.
Objective 2 - Dyslipidemia Prevalence by WHR
Table 2: Prevalence of Dyslipidemia Among T2DM Patients Stratified by Waist-Hip Ratio
| Dyslipidemia Component | Normal WHR* n=__ | High WHR* n=__ | p-value |
|---|---|---|---|
| High Total Cholesterol | n (%) | n (%) | |
| High LDL-C | n (%) | n (%) | |
| Low HDL-C | n (%) | n (%) | |
| High Triglycerides | n (%) | n (%) | |
| Any Dyslipidemia (overall) | n (%) | n (%) |
*WHR cut-offs: Normal = <0.90 (male), <0.85 (female) per WHO; High = ≥0.90 (male), ≥0.85 (female)
Statistical test: Chi-square test / independent t-test for proportions; you can also calculate Odds Ratio with 95% CI for the "any dyslipidemia" row - this strengthens the table considerably for IRB and publication.
Objective 3 - Correlation of BMI and WHR with Lipid Parameters
This is the strongest table analytically. Use Pearson correlation (if lipid values are normally distributed) or Spearman's rank correlation (if skewed - common with TG).
Table 3: Correlation of BMI and WHR with Lipid Profile Parameters in T2DM Patients
| Lipid Parameter | BMI | WHR | ||
|---|---|---|---|---|
| r | p-value | r | p-value | |
| Total Cholesterol (mg/dL) | ||||
| LDL Cholesterol (mg/dL) | ||||
| HDL Cholesterol (mg/dL) | ||||
| Triglycerides (mg/dL) | ||||
| TC/HDL ratio |
r = Pearson/Spearman correlation coefficient
You can then add a brief interpretive note: "A positive r indicates direct correlation; negative r indicates inverse correlation. |r| >0.3 = moderate, >0.5 = strong."
Optional extension: Run multiple linear regression for each lipid parameter with BMI and WHR as predictors simultaneously - this controls for confounders (age, sex, duration of DM, medications) and dramatically elevates the analytical quality of Objective 3.
Practical Tips for Your IRB/Dissertation
-
Define dyslipidemia upfront in your methodology - state which guideline cut-offs you're using (NCEP-ATP III is most commonly cited; IDF is acceptable too). Be consistent across all three tables.
-
BMI classification: Use WHO standard or Indian-specific cut-offs (overweight ≥23, obese ≥25 kg/m²) - since Asian populations have metabolic risk at lower BMIs, many Indian dissertations use the Indian ICMR/WHO Asia-Pacific criteria. State which you use.
-
Sample size: For correlation analysis (Objective 3), a minimum of ~84 patients gives 80% power to detect r ≥ 0.3 at α=0.05. For prevalence comparisons (Objectives 1 and 2), aim for ≥100-150 to have meaningful numbers in each BMI subgroup.
-
Software: SPSS or MedCalc handles all these analyses. For the correlation table, a scatter plot per lipid parameter (BMI vs. LDL, WHR vs. TG, etc.) makes a strong visual complement.
-
IRB feasibility argument: These are all observational measurements (no intervention, no extra blood draw beyond routine care), so ethical risk is minimal and approval is typically straightforward.
Literature supports high prevalence of dyslipidemia in T2DM (ranging 62-91% across published studies), so you will almost certainly find significant results - making this a publishable study.
Isn't it much cumbersome?
Fair concern. Let me clarify what's actually simple vs. what looks complicated on paper.
What's Actually Simple Here
Data collection is the same for all 3 objectives - you measure:
- Height, weight (→ BMI)
- Waist, hip circumference (→ WHR)
- Fasting lipid profile (TC, LDL, HDL, TG)
You collect this once per patient. All three tables come from the same dataset. There's no extra data collection burden for having 3 objectives vs. 1.
What Feels Cumbersome vs. What It Actually Is
| What looks complex | Reality |
|---|---|
| 3 separate tables | Same data, just sliced 3 ways in SPSS |
| BMI categories (4 groups) | One variable, pre-coded once |
| Correlation table | 5 minutes in SPSS - Analyze → Correlate → Bivariate |
| Chi-square tests | Auto-generated when you run the crosstabs |
Once your data is entered in SPSS, all 3 tables take under 30 minutes to generate.
The Real Work is Data Collection
The only genuinely effortful part is:
- Recruiting enough patients (aim for 100-150)
- Ensuring waist/hip measurements are done correctly and consistently
- Making sure fasting lipid values are available
If you're working in an OPD or ward where lipid profiles are routinely done, even step 3 is just record extraction.
Bottom line: The tables look elaborate in design, but the actual analysis is almost entirely automated by SPSS once the data is entered. The objectives are proportionate to a standard MD/MS dissertation - not excessive.
Is the concern more about the data collection side or the analysis/writing side?
Analysis side
Good - then let me break it down to exactly what you click in SPSS for each objective.
Objective 1 - Prevalence by BMI Class
What you do in SPSS:
- Create a variable
BMI_category(1=Normal, 2=Overweight, 3=Obese I, 4=Obese II) - Create a variable
dyslipidemia(0=No, 1=Yes) for each component - Analyze → Descriptive Statistics → Crosstabs
- Row: dyslipidemia type
- Column: BMI_category
- Tick: Chi-square + Row/Column percentages
- SPSS gives you the table and p-value automatically
That's it. 5 clicks.
Objective 2 - Prevalence by WHR
Identical process to Objective 1, just replace
BMI_category with WHR_category (0=Normal, 1=High).Since it's only 2 groups instead of 4, it's actually simpler than Objective 1.
Same 5 clicks.
Objective 3 - Correlation of BMI and WHR with Lipid Parameters
- Analyze → Correlate → Bivariate
- Move BMI, WHR, TC, LDL, HDL, TG into the box
- Select Pearson (or Spearman if TG is skewed)
- Click OK
SPSS produces a full correlation matrix with r values and p-values for every combination in one single run.
You just copy the relevant cells (BMI vs each lipid, WHR vs each lipid) into your Table 3.
Under 2 minutes.
Total Analysis Time (Realistic Estimate)
| Task | Time |
|---|---|
| Coding variables in SPSS | 1-2 hours (done once) |
| Objective 1 analysis | 10 min |
| Objective 2 analysis | 5 min |
| Objective 3 analysis | 5 min |
| Formatting tables for dissertation | 30-45 min |
| Total | ~2-3 hours |
The analysis itself is genuinely not the hard part. If you know someone who has used SPSS once, they can walk you through all three in a single sitting. The objectives are well-matched to the level of analysis expected in an MD dissertation - not too simple to be rejected, not so complex that it becomes a burden.
Should I simplify to just 1-2 objectives instead of 3?
Honestly, no - keep all 3. Here's why:
The Case for Keeping All 3
Objective 3 is actually your strongest objective. Comparing the strength of correlation of BMI vs. WHR with lipid parameters is what makes your study original and answerable. It gives you a clear conclusion - "WHR correlates better with dyslipidemia than BMI in T2DM patients" (or vice versa) - which is a publishable finding.
Objectives 1 and 2 are essentially prerequisite steps to Objective 3. You're already calculating prevalence by BMI group (Obj 1) and WHR group (Obj 2) as part of describing your data. Dropping them doesn't save you work - the data is already there.
What Dropping an Objective Would Cost You
| If you drop | What you lose |
|---|---|
| Objective 1 | Can't show how dyslipidemia burden changes across BMI grades - weakens your narrative |
| Objective 2 | No WHR-stratified prevalence - makes Objective 3 feel unsupported |
| Objective 3 | Lose the most analytically interesting finding - the comparison |
The Real Question to Ask
Rather than reducing objectives, ask: does your guide/IRB committee expect secondary objectives to be this specific?
Some institutions are happy with broader phrasing like:
"To study the association of anthropometric indices (BMI and WHR) with dyslipidemia in T2DM patients"
That single line covers all 3 of your current objectives without listing them separately - and reduces perceived complexity at the IRB stage while keeping your full analysis plan intact.
Recommendation: Keep all 3 objectives as written. They're well-constructed, mutually supportive, and the analysis is genuinely quick. Reducing to 1-2 would thin out your dissertation without saving meaningful effort.
Create a simplified objective statement combining all 3 into one broader aim
Here are a few versions depending on your preferred tone/style:
Option A - Clean and concise (recommended):
"To assess the prevalence of dyslipidemia and its association with anthropometric indices (BMI and waist-hip ratio) in patients with type 2 diabetes mellitus."
Option B - Slightly more specific, still tight:
"To evaluate the prevalence of dyslipidemia across BMI and waist-hip ratio categories and to compare the correlation of these anthropometric indices with lipid profile parameters in type 2 diabetes mellitus patients."
Option C - IRB-formal language:
"To study the relationship between anthropometric measures of obesity (body mass index and waist-hip ratio) and dyslipidemia in type 2 diabetes mellitus patients, including prevalence estimation across adiposity categories and correlation analysis with individual lipid parameters."
Which to Use
- Option A - best for IRB title page / objectives section where brevity is expected
- Option B - best if your guide wants the analytical intent (correlation comparison) visible
- Option C - best for the formal IRB protocol document where precise language matters
All three cover your original 3 objectives without listing them separately, and none will raise questions from an IRB reviewer about analytical complexity.
So should I keep 3 or this according to feasibility
Keep the 3 separate objectives - here's the direct reasoning:
Feasibility Verdict
The combined single statement is useful only for presentation (title page, synopsis cover sheet). For the actual IRB protocol and dissertation objectives section, 3 separate objectives is better because:
- IRB committees want to see each objective is measurable and testable - 3 clear objectives does that better than one broad statement
- Your guide can tick off each objective against your results chapter - easier to evaluate
- Word count-wise, 3 short bullet objectives takes barely more space than one long sentence
Best of Both Worlds - What Most Dissertations Do
Use the combined statement as your AIM, then list the 3 as secondary objectives under it:
Secondary Aim: To assess the prevalence of dyslipidemia and its association with anthropometric indices in T2DM patients.Secondary Objectives:
- To evaluate the prevalence of dyslipidemia by BMI category (normal, overweight, obese I, obese II)
- To evaluate the prevalence of dyslipidemia by WHR (normal vs. high)
- To compare the correlation of BMI and WHR with lipid profile parameters
This structure looks organized, is IRB-friendly, and keeps the analytical clarity of 3 objectives - without making it feel like an overwhelming list. Most MD dissertation guides prefer exactly this format.
Should I include vldl or not?
Short answer: No, leave VLDL out.
Why Not
- VLDL is not directly measured in a standard lipid panel - it is calculated from TG using the Friedewald estimate (VLDL = TG ÷ 5), so it carries no independent information beyond what TG already tells you
- Including it means you have two variables saying the same thing - reviewers and guides will question the redundancy
- VLDL is not part of dyslipidemia diagnostic criteria in NCEP-ATP III or IDF guidelines - so it doesn't strengthen your prevalence analysis
- Most published T2DM dyslipidemia studies report TC, LDL, HDL, TG - and stop there
When VLDL Would Be Worth Including
- If your lab directly measures VLDL by ultracentrifugation (almost never done in routine OPD settings)
- If your primary objective was specifically about atherogenic dyslipidemia or metabolic syndrome components
- If your guide specifically asks for it
What to Report Instead
If you want a fifth lipid variable that adds real value, consider:
Non-HDL Cholesterol (= TC - HDL)
It's calculated from existing values, is recommended by ACC/AHA as a better atherogenic marker than LDL alone, and is increasingly reported in T2DM lipid studies. That would genuinely strengthen your correlation table.
Stick with TC, LDL, HDL, TG - and optionally add Non-HDL. Skip VLDL.
This is a shared conversation. Sign in to Orris to start your own chat.