The dependent variable in this study corresponded to the profiles of blood glucose monitoring conducted by the T2DM patients, which was measured by four questions regarding the methods and frequency of blood glucose monitoring. The first question was “How many times on average did you test your blood glucose by the professionals in the primary health institution(s) per month?”; The second question was “How many times on average did you test your blood glucose by the pharmacists in the pharmacies per month”; The third question was “How many times on average did you test your blood glucose by the professionals in the secondary or tertiary hospital (s) per month”; The fourth was “How many times on average did you test your blood glucose using glucometers at home per month”. The answers to the four questions were counted based on the frequency of capillary blood glucose monitoring (fasting blood glucose and postprandial blood glucose), ranging zero to 360 times per month. The zero means that patients never tested their blood glucose; and the 360 times was the maximum limit that the patients could not be reached. These four questions were categorized through Latent Profile Analysis (LPA), and then presented as a single variable of the profiles of blood glucose monitoring. In addition, the frequency of twice a month was used to assess the characteristics of blood glucose monitoring in different subgroups based on LPA, according to the Guideline for Blood Glucose Monitoring in China.6

LPA is a type of cluster analysis, which is used to group population into a specific (k) number of unique categories. Based on LPA, patients are most similar to each other within each category, and are most different between the categories.31 LAP estimates two parameters based on maximum-likelihood: 1) class membership probabilities, which represent individuals’ probability of belonging to each class; and 2) item-response probabilities conditional on class membership (conditional response probabilities), which refer to the conditional probability a particular response given the individual is in a certain class. Based on their highest latent class probability, individuals are assigned to one class exclusively.31 In this study, the selected four observed variables used in the LPA were the frequency of blood glucose conducted with four different methods, including by professionals in primary healthcare institutes, by pharmacists in pharmacies, by professionals in secondary or tertiary hospitals and by self-monitoring using glucometers at home. To find out the optimal number of categories, 2–6 class solutions were modelled, and the outputs were assessed; Each solution was then compared using model fit indicators (Akaike information criterion [AIC], Bayesian information criterion [BIC], entropy values),32,33 until an appropriate number of patterns was decided according to empirical and theoretical interpretations regarding simultaneously considering and interpreting multiple indices of fit. Once an optimal class solution was determined, the individual classes were characterized with respect to frequency of blood glucose tests in certain methods that defined each group.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.