Analyzing a new diet intervention using Bland Altman analysis

Bland Altman analysis can be used to evaluate the effectiveness of a new diet intervention in patients with a family history of heart disease. A physician conducted a study involving 16 patients who were placed on a diet for six months. The patients' weights and triglyceride (TG) levels were measured before and after the diet. The physician aims to determine whether the diet led to significant changes in either the patients' weight or their triglyceride levels.

This analysis utilizes the dietstudy.sav dataset, which contains data on patient demographics, weight, and triglyceride levels measured at multiple time points before and after the diet intervention. By using Bland Altman analysis, the physician can assess the agreement between the pre and postdiet measurements and determine if any systematic bias exists between the two sets of measurements.

Consider that the physician needs to evaluate the effectiveness of a diet intervention by comparing the pre and postdiet measurements of weight and triglyceride (TG) levels in patients with a family history of heart disease. Specifically, they aim the following objectives:
  1. Assess agreement: Use Bland Altman analysis to assess the agreement between prediet and postdiet measurements for both weight and triglyceride levels.
  2. Identify bias: Identify any systematic bias between the pre and postdiet measurements by calculating the mean differences and visualizing the limits of agreement.
  3. Provide insight: Determine if the diet intervention led to statistically significant changes in weight and triglyceride levels by evaluating the distribution of differences and agreement limits.

This analysis helps the physician determine the effectiveness of the diet intervention and assess whether any meaningful changes occurred in the patients' health metrics over the study period.

Dataset used for analysis

The dataset used in this case study is dietstudy.sav, which contains data from a study involving 16 patients with a family history of heart disease. The dataset includes demographic information as well as measurements of weight and triglyceride levels taken before and after a 6-month diet intervention. The columns in the dataset are as follows:
patid
Unique patient identifier.
age
Age of the patient (in years).
gender
Gender of the patient (0 = Male, 1 = Female).
tg0 - tg4
Triglyceride (TG) levels at different time points:
  • tg0: Triglyceride level before the diet.
  • tg1 - tg4: Triglyceride levels at various intervals after starting the diet.
wgt0 - wgt4
Weight measurements at different time points:
  • wgt0: Weight before the diet
  • wgt1 - wgt4: Weight measurements at various intervals after starting the diet.
Figure 1. Variables used in Bland Altman Analysis

Each row represents a patient. The measurements track changes in weight and triglyceride levels over time during the diet intervention. For the Bland Altman analysis, compare the values for pre-diet (tg0, wgt0) and post-diet (tg4, wgt4) measurements to evaluate any differences that may indicate the diet's impact.