Myasthenia Gravis Healthcare Resource Use and Cost of Therapy in Australia
Author(s)
Hansoo Kim, BSc, MSc, PhD, Leonard Lee, MN, Jeremy Welton, BSc, PhD, MEpi, Josh Byrnes, BEc, PhD.
Griffith University, Gold Coast, Australia.
Griffith University, Gold Coast, Australia.
Presentation Documents
OBJECTIVES: The identification of myasthenia gravis (MG) patients in administrative health databases is challenging due to the disease rarity, presentation heterogeneity, and limitations of available diagnostic codes. This study aims to to establish a retrospective cohort of patients with MG from a national database to estimate healthcare resource utilisation (HCRU) and associated costs.
METHODS: This study utilised the National Integrated Health Services Information (NIHSI), a comprehensive Australian dataset inclusive of outpatient visits and drug use, inpatient admissions, and emergency department presentations and mortality. Complicating identification, MG diagnosis is often delayed due to similar presentations with other, confounding neurological conditions, with this delay often resulting in an underestimation of HCRU and costs. In addition to diagnostic codes being limited to the inpatient components of NIHSI, they are also insensitive, for instance, with ocular MG where electrodiagnostic testing sensitivity is low. To address these challenges, a novel algorithm was developed in collaboration with MG specialists and a pharmacoepidemiologist to improve the accuracy and sensitivity of MG case identification in the NIHSI dataset.
RESULTS: The algorithm incorporates diagnostic codes, treatment patterns, diagnostic testing, and procedural data. Five inclusion criteria were developed to improve case identification while maintaining specificity and accounting for similarly presenting, confounding neurological conditions. These were (1) extended use of pyridostigmine with exclusion of confounding medications, (2) record of therapeutic thymectomy, (3) pyridostigmine use in conjunction with AChR antibody testing and follow-up treatment, (4) two or more inpatient or ED claims with MG diagnosis codes, and (5) a combination of diagnosis codes with supporting treatment or test data.
CONCLUSIONS: This refined identification algorithm aims to overcome the limitations of relying on diagnostic or treatment codes, increasing the accuracy of HCRU and disease burden in MG. This study offers a reproducible framework for characterising rare diseases in large administrative health databases.
METHODS: This study utilised the National Integrated Health Services Information (NIHSI), a comprehensive Australian dataset inclusive of outpatient visits and drug use, inpatient admissions, and emergency department presentations and mortality. Complicating identification, MG diagnosis is often delayed due to similar presentations with other, confounding neurological conditions, with this delay often resulting in an underestimation of HCRU and costs. In addition to diagnostic codes being limited to the inpatient components of NIHSI, they are also insensitive, for instance, with ocular MG where electrodiagnostic testing sensitivity is low. To address these challenges, a novel algorithm was developed in collaboration with MG specialists and a pharmacoepidemiologist to improve the accuracy and sensitivity of MG case identification in the NIHSI dataset.
RESULTS: The algorithm incorporates diagnostic codes, treatment patterns, diagnostic testing, and procedural data. Five inclusion criteria were developed to improve case identification while maintaining specificity and accounting for similarly presenting, confounding neurological conditions. These were (1) extended use of pyridostigmine with exclusion of confounding medications, (2) record of therapeutic thymectomy, (3) pyridostigmine use in conjunction with AChR antibody testing and follow-up treatment, (4) two or more inpatient or ED claims with MG diagnosis codes, and (5) a combination of diagnosis codes with supporting treatment or test data.
CONCLUSIONS: This refined identification algorithm aims to overcome the limitations of relying on diagnostic or treatment codes, increasing the accuracy of HCRU and disease burden in MG. This study offers a reproducible framework for characterising rare diseases in large administrative health databases.
Conference/Value in Health Info
2025-09, ISPOR Real-World Evidence Summit 2025, Tokyo, Japan
Value in Health Regional, Volume 49S (September 2025)
Code
RWD261
Topic Subcategory
Health & Insurance Records Systems
Disease
SDC: Musculoskeletal Disorders (Arthritis, Bone Disorders, Osteoporosis, Other Musculoskeletal)