BMF CP135: The Conditioning Effects of Body Mass Index and Insulin Levels in Shaping Diabetes Outcomes

Ni Putu Wulan Purnama Sari

Faculty of Nursing, Widya Mandala Surabaya Catholic University, East Java, Indonesia

“…. the wings must be pretty, the body well-balanced, and the face good-looking.”

In “The Most Beautiful Bird”; Wild Wise Weird (2024)

1. Project description

The risk factors for type 2 Diabetes Mellitus (DM) are broadly categorized as modifiable and nonmodifiable. Among these, Diabetes pedigree function (a measure of family history) and age are nonmodifiable risk factors for DM. This study seeks to identify which modifiable parameters of physical measurement, e.g., body mass index (BMI), and biomedical test results, e.g., insulin levels, matter most in conditioning the effects of nonmodifiable risks on DM outcomes. As age and family history of DM are fixed parameters, professional efforts in regular screening and effective management of modifiable parameters are needed, especially among females.

1.1. Main objectives

The current study is conducted to examine the following research questions:

• How are the nonmodifiable risks, i.e., age and family history, and modifiable parameters, i.e., BMI and insulin levels, associated with DM?
• Is the association between sex and family history of DM conditional on BMI?
• Is the association between sex and family history of DM conditional on insulin levels?

1.2. Materials

The Granular Interaction Thinking Theory (GITT) will be employed for the conceptual development of this study, while the Bayesian Mindsponge Framework (BMF) analytics will be utilized for statistical analysis [1-3]. Employing these frameworks, we explore a systems-thinking orientation to understand how nonmodifiable risks can be modified through modifiable clinical parameters, such as BMI and insulin levels, in developing DM outcomes. The dataset comprises responses from 1,065 females aged 21-86 years in Bangladesh, of whom 840 are diabetics, reflecting a natural imbalance that real-world clinical trends show [4]. Statistical analyses will be conducted using the bayesvl R package, which utilizes the Markov chain Monte Carlo (MCMC) algorithm for estimation.

1.3. Main findings

Results of preliminary analysis showed that only BMI is negatively associated with DM outcomes. Among the rest, family history of DM shows the least reliable positive results, compared to age and insulin levels.

The moderation analysis revealed that the moderating role of BMI in the association between age and family history with DM is unclear, indicating that maintaining a normal BMI may not yield the expected outcome. Conversely, the positive moderating role of insulin levels in the association between family history and age with DM is reliable, with a stronger effect found in the interaction involving family history (see Figure 1).

This study suggests that the expected protective role of maintaining a normal BMI in conditioning the effects of age and family history on DM is not reliable. BMI alone may not buffer genetic or age-related risks. Conversely, insulin levels reliably strengthen the association between family history and age with DM outcomes. The effect is particularly pronounced when family history is involved, suggesting insulin dysregulation amplifies genetic susceptibility. Efforts to regulate insulin levels in DM prevention among females are highly recommended, such as lifestyle interventions, i.e., diet, physical activity, sleep, and stress management, combined with early monitoring of insulin resistance markers.



Figure 1: The estimated posterior distributions with HPDI at 89%

2. Collaboration procedure

Portal users should follow these steps to register to participate in this research project:

1. Create an account on the website (preferably using an institution email).
2. Comment your name, affiliation, and your desired role in the project below this post.
3. Patiently wait for the formal agreement on the project from the AISDL mentor.

If you have further inquiries, please contact us at aisdl_team@mindsponge.info

If you have been invited to join the project by an AISDL member, you are still encouraged to follow the above formal steps.

All the resources for conducting and writing the research manuscript will be distributed upon project participation.

Mentor for this project: Ni Putu Wulan Purnama Sari.

Other members who have joined this project: Minh-Hoang Nguyen and Quan-Hoang Vuong.

The research project strictly adheres to scientific integrity standards, including authorship rights and obligations, without incurring an economic burden at participants’ expenses.

References

[1] Vuong QH, Nguyen MH (2025). Developing Bayesian probabilistic reasoning capacity in HSS disciplines: Qualitative evaluation on bayesvl and BMF analytics for ECRs. https://dipot.ulb.ac.be/dspace/bitstream/2013/397326/3/wp25008.pdf

[2] Vuong QH. (2026). Once the Door Swings Shut. http://books.google.com/books/about?id=Qur0EQAAQBAJ

[3] Vuong QH, Nguyen MH, La VP. (2022). The mindsponge and BMF analytics for innovative thinking in social sciences and humanities. Walter de Gruyter GmbH. https://books.google.com/books?id=EGeEEAAAQBAJ

[4] Bhuiyan MY, Ayon SS, Hossain ME, Miah MSU, Sarower AH, Bappee FK. (2026). A clinical dataset on type-2 diabetes including demographic, anthropometric, and biochemical parameters from Bangladesh. Data in Brief, 112457.