@article{10.11648/j.ijde.20251004.13,
  author = {Argwings Kodhek and Alice Achieng’ Ojwang and Francis Oguya and Fredrick Otieno and Isaac Okeyo},
  title = {Development of a Composite Performance Score Tool for Diabetes Mellitus Management: Features and Framework},
  journal = {International Journal of Diabetes and Endocrinology},
  volume = {10},
  number = {4},
  pages = {98-106},
  doi = {10.11648/j.ijde.20251004.13},
  url = {https://doi.org/10.11648/j.ijde.20251004.13},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ijde.20251004.13},
  abstract = {Background: Type 2 diabetes mellitus (T2DM) remains a critical public health challenge in low- and middle-income countries (LMICs), where existing assessment tools often fail to integrate the full spectrum of clinical, behavioral, and contextual determinants of care. This study developed and validated the Management Performance Composite Scoring Tool (MPCST), a multidimensional instrument designed for primary healthcare (PHC) settings, grounded in the Chronic Care Model (CCM). Methods: An exploratory sequential mixed-methods design was conducted over four months across five public PHC facilities in Nairobi, Kenya. The qualitative phase involved six focus group discussions (n=52) and ten key informant interviews to identify domains relevant to T2DM management. In the quantitative phase, 181 adults with stable T2DM were selected via systematic random sampling. Data collection employed validated instruments alongside standardized clinical and biochemical assessments. A composite scoring algorithm was developed using statistical machine learning and deployed in a Streamlit-based application for real-time clinical utility. Psychometric validation included content validity (Lawshe’s method), generalizability theory, and convergent and concurrent validity analysis. Results: Among 181 participants (63.5% female; mean age 55.9 years), glycemic control was poor (mean HbA1c: 9.39 mmol/L), and lifestyle performance was suboptimal. Tool development incorporated clinical (e.g., HbA1c, BMI), behavioral (diet, activity, distress), and contextual (social determinants, healthcare access) domains. Ordinal regression yielded excellent model fit (pseudo R² = 0.949, AIC = 69.821). Geometric mean aggregation improved behavioral sensitivity, while minimax strategies optimized clinical indicator selection. The MPCST demonstrated high reliability (G = 0.8351), strong convergent (r = 0.74) and concurrent (r = 0.565) validity. Conclusion: The MPCST is a rigorously validated, contextually relevant, and scalable tool for evaluating T2DM care quality in LMIC PHC systems. Its dual-format (manual and app-based) design supports routine use, longitudinal tracking, and integration into health systems, thereby enhancing clinical decision-making and patient-centered care.},
 year = {2025}
}

Copy | Download

TY  - JOUR
T1  - Development of a Composite Performance Score Tool for Diabetes Mellitus Management: Features and Framework
AU  - Argwings Kodhek
AU  - Alice Achieng’ Ojwang
AU  - Francis Oguya
AU  - Fredrick Otieno
AU  - Isaac Okeyo
Y1  - 2025/12/09
PY  - 2025
N1  - https://doi.org/10.11648/j.ijde.20251004.13
DO  - 10.11648/j.ijde.20251004.13
T2  - International Journal of Diabetes and Endocrinology
JF  - International Journal of Diabetes and Endocrinology
JO  - International Journal of Diabetes and Endocrinology
SP  - 98
EP  - 106
PB  - Science Publishing Group
SN  - 2640-1371
UR  - https://doi.org/10.11648/j.ijde.20251004.13
AB  - Background: Type 2 diabetes mellitus (T2DM) remains a critical public health challenge in low- and middle-income countries (LMICs), where existing assessment tools often fail to integrate the full spectrum of clinical, behavioral, and contextual determinants of care. This study developed and validated the Management Performance Composite Scoring Tool (MPCST), a multidimensional instrument designed for primary healthcare (PHC) settings, grounded in the Chronic Care Model (CCM). Methods: An exploratory sequential mixed-methods design was conducted over four months across five public PHC facilities in Nairobi, Kenya. The qualitative phase involved six focus group discussions (n=52) and ten key informant interviews to identify domains relevant to T2DM management. In the quantitative phase, 181 adults with stable T2DM were selected via systematic random sampling. Data collection employed validated instruments alongside standardized clinical and biochemical assessments. A composite scoring algorithm was developed using statistical machine learning and deployed in a Streamlit-based application for real-time clinical utility. Psychometric validation included content validity (Lawshe’s method), generalizability theory, and convergent and concurrent validity analysis. Results: Among 181 participants (63.5% female; mean age 55.9 years), glycemic control was poor (mean HbA1c: 9.39 mmol/L), and lifestyle performance was suboptimal. Tool development incorporated clinical (e.g., HbA1c, BMI), behavioral (diet, activity, distress), and contextual (social determinants, healthcare access) domains. Ordinal regression yielded excellent model fit (pseudo R² = 0.949, AIC = 69.821). Geometric mean aggregation improved behavioral sensitivity, while minimax strategies optimized clinical indicator selection. The MPCST demonstrated high reliability (G = 0.8351), strong convergent (r = 0.74) and concurrent (r = 0.565) validity. Conclusion: The MPCST is a rigorously validated, contextually relevant, and scalable tool for evaluating T2DM care quality in LMIC PHC systems. Its dual-format (manual and app-based) design supports routine use, longitudinal tracking, and integration into health systems, thereby enhancing clinical decision-making and patient-centered care.
VL  - 10
IS  - 4
ER  - 

Copy | Download