Research code: 40200113
Ethics code: IR.TUMS.NIHR.REC.1403.003
Maedeh Gholamazad1 
, Alireza Atashi1 , Fatemeh Hajialiasgari2 
, Shima Mohammadi *3 , Reza Mehrizi4 , Leila Ghamkhar5 , Hossein Effatpanah6
, Alireza Atashi1 , Fatemeh Hajialiasgari2 , Shima Mohammadi *3 , Reza Mehrizi4 , Leila Ghamkhar5 , Hossein Effatpanah6
1- Medical Informatics Department, Breast Cancer Research Center, National Cancer Institute-Iran, ACECR, Tehran, Iran
2- Geriatrics Medicine Department, School of Medicine, Tehran University of Medical Sciences, Iran
3- National Center for Health Insurance Research, Tehran, Iran & Neuroscience Research Center, Shahroud University of Medical Sciences, Shahroud, Iran ,shimamohamadi1365@gmail.com
4- School of Medicine, Tehran University of Medical Sciences, Tehran, Iran
5- National Center for Health Insurance Research, Tehran, Iran
6- Department of School Health, School of Allied Medical Sciences, Asadabad School of Medical Sciences, Iran
2- Geriatrics Medicine Department, School of Medicine, Tehran University of Medical Sciences, Iran
3- National Center for Health Insurance Research, Tehran, Iran & Neuroscience Research Center, Shahroud University of Medical Sciences, Shahroud, Iran ,
4- School of Medicine, Tehran University of Medical Sciences, Tehran, Iran
5- National Center for Health Insurance Research, Tehran, Iran
6- Department of School Health, School of Allied Medical Sciences, Asadabad School of Medical Sciences, Iran
Abstract: (45 Views)
Introduction: Alzheimer’s disease, as the most common cause of dementia in older adults, imposes a substantial burden on the healthcare system. With the growing elderly population in Iran, understanding real‑world prescribing patterns is essential for optimizing treatment and reducing polypharmacy. Association rule mining on insurance registry data provides an opportunity to identify co‑occurrence patterns of medications.
Methods: A descriptive–analytical, non‑experimental study was conducted on registry data from the Iran Health Insurance Organization (2020–2022). From 532,369 prescription records of Alzheimer’s patients, after cleaning, standardization, and removal of rare medications, a binary transactional matrix (presence/absence of drugs) was constructed. The Apriori and FP‑Growth algorithms were applied to extract association rules using Support, Confidence, and Lift.
Results: Donepezil and Memantine dominated with 81% of prescriptions; Rivastigmine and Galantamine had minimal shares, and non‑specific medications were limited. Both algorithms showed similar patterns (FP‑Growth more efficient), yet no meaningful rules met the Support, Confidence, and Lift thresholds; convergence of results indicated high dispersion in real‑world data.
Conclusion: Prescribing patterns in Alzheimer’s disease were consistent with clinical guidelines but lacked stable co‑occurrences for classical association rules. These findings highlight the complexity of clinical data and the need for Sequential Mining and machine learning, providing a foundation for optimizing prescribing and policymaking in Iran.
Methods: A descriptive–analytical, non‑experimental study was conducted on registry data from the Iran Health Insurance Organization (2020–2022). From 532,369 prescription records of Alzheimer’s patients, after cleaning, standardization, and removal of rare medications, a binary transactional matrix (presence/absence of drugs) was constructed. The Apriori and FP‑Growth algorithms were applied to extract association rules using Support, Confidence, and Lift.
Results: Donepezil and Memantine dominated with 81% of prescriptions; Rivastigmine and Galantamine had minimal shares, and non‑specific medications were limited. Both algorithms showed similar patterns (FP‑Growth more efficient), yet no meaningful rules met the Support, Confidence, and Lift thresholds; convergence of results indicated high dispersion in real‑world data.
Conclusion: Prescribing patterns in Alzheimer’s disease were consistent with clinical guidelines but lacked stable co‑occurrences for classical association rules. These findings highlight the complexity of clinical data and the need for Sequential Mining and machine learning, providing a foundation for optimizing prescribing and policymaking in Iran.
Keywords: Alzheimer’s Disease, Drug Prescribing, Data Mining, Association Rules, Apriori, FP Growth, Health Insurance Prescriptions
Type of Study: Research |
Subject:
Special
Received: 2026/04/28 | Revised: 2026/06/6 | Accepted: 2026/05/31
Received: 2026/04/28 | Revised: 2026/06/6 | Accepted: 2026/05/31
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |