Real-World Costs of United States Patients With T1 Diabetes, Removing Misdiagnosed Patients, and Compared With Patients With T1D and (T1.5 or T2) Diabetes
Author(s)
Richard A. Brook, MBA, MS1, Stephen A Schramm, MA2, Alek A Drnach, MS2, Eric Michael Rosenberg, MA, MA2, C. Bernie Good, MPH, MD3.
1President (BH-WW) / SVP (NPRT), Better Health Worldwide / NPRT, Newfoundland, NJ, USA, 2Integrated Analytics, Workpartners LLC, Pittsburgh, PA, USA, 3UPMC Health Plan, Insurance Services Division, UPMC Health Plan, Pittsburgh, PA, USA.
1President (BH-WW) / SVP (NPRT), Better Health Worldwide / NPRT, Newfoundland, NJ, USA, 2Integrated Analytics, Workpartners LLC, Pittsburgh, PA, USA, 3UPMC Health Plan, Insurance Services Division, UPMC Health Plan, Pittsburgh, PA, USA.
OBJECTIVES: Diabetes+diabetic complications are leading causes of morbidity&mortality. Diagnostic accuracy essential for data analysis. 2021 ADA/EASD consensus Report highlights frequent misclassification of type 1 diabetes (T1D) as type 2 (T2D), >40% of T1D cases diagnosed (Dx) age>30y initially treated as T2D. Prior research developed an algorithm to improve diagnostic accuracy in patients (pts) with diabetes (PwD). We assessed T1D Dx claim accuracy, and healthcare costs for uncorrected+corrected T1D pts.
METHODS: Retrospective analysis of Workpartners Research Reference Database (RRDb) of self-insured US employers. Clinical review of pts with T1D claims using ICD-codes,index date=initial Dx date. EMPloyee and spouse pts (without 2nd T1D Dx, use of [INS]ulin or with pre-index T2D) Dx OR without 1y pre-/post-index eligibility excluded. Costs from paid MEDical and prescription claims. Race, absence time&payments from employer records for PwD1=(2 T1D Dxs+INS+no[T1.5D or T2D]) were compared with PwD12=(2 T1D Dxs+2 post-index T2D/T.5 Dxs,+INS). Non-overlapping 95% confidence intervals significant (p<0.05).
RESULTS: 27,749 pts identified with T1D Dxs. 320(1.2%) met study criteria for PwD1 and were 22.3y,50% were male,2.8% were EMPs aged 32.4y,62% male[average salary $57,216,race:26%=white,13%=other,62%=unknown]. PwD1 costs for MED were $13,169($10,005—$16,333),prescription $9105($8,376—$9,834), and plan costs $22,274($18,905—$25,643). 376 pts(1.4%) met criteria for PwD12 were 34.3y,49% male,20.3% were EMPs aged 43.7y,76% males(average salary $77,388,race:19%=white,4%=other,76%=unknown). PwD12 costs: MED =$33,683($25,158—$42,209),prescription=$9421($8,251—$10,592), plan =$43,105($34,282—$51,928). PwD1 pts were younger, had lower salaries, and different races than PwD12 pats. Total Rx costs were similar (~97%) and MED and plan costs for PwD1 were 39.1% and 51.7% of the PwD12 pt costs(p<.05).
CONCLUSIONS: Claims data are useful for analyzing the impact of conditions, however staff may enter incorrect ICD-codes. Our reclassification algorithm allows for better understanding of the costs of patients with only T1D diagnoses, and those with diagnoses of T1D and (T1.5D or T2D). Cohorts are better suited for outcomes research, therapy use, and studying diabetes types.
METHODS: Retrospective analysis of Workpartners Research Reference Database (RRDb) of self-insured US employers. Clinical review of pts with T1D claims using ICD-codes,index date=initial Dx date. EMPloyee and spouse pts (without 2nd T1D Dx, use of [INS]ulin or with pre-index T2D) Dx OR without 1y pre-/post-index eligibility excluded. Costs from paid MEDical and prescription claims. Race, absence time&payments from employer records for PwD1=(2 T1D Dxs+INS+no[T1.5D or T2D]) were compared with PwD12=(2 T1D Dxs+2 post-index T2D/T.5 Dxs,+INS). Non-overlapping 95% confidence intervals significant (p<0.05).
RESULTS: 27,749 pts identified with T1D Dxs. 320(1.2%) met study criteria for PwD1 and were 22.3y,50% were male,2.8% were EMPs aged 32.4y,62% male[average salary $57,216,race:26%=white,13%=other,62%=unknown]. PwD1 costs for MED were $13,169($10,005—$16,333),prescription $9105($8,376—$9,834), and plan costs $22,274($18,905—$25,643). 376 pts(1.4%) met criteria for PwD12 were 34.3y,49% male,20.3% were EMPs aged 43.7y,76% males(average salary $77,388,race:19%=white,4%=other,76%=unknown). PwD12 costs: MED =$33,683($25,158—$42,209),prescription=$9421($8,251—$10,592), plan =$43,105($34,282—$51,928). PwD1 pts were younger, had lower salaries, and different races than PwD12 pats. Total Rx costs were similar (~97%) and MED and plan costs for PwD1 were 39.1% and 51.7% of the PwD12 pt costs(p<.05).
CONCLUSIONS: Claims data are useful for analyzing the impact of conditions, however staff may enter incorrect ICD-codes. Our reclassification algorithm allows for better understanding of the costs of patients with only T1D diagnoses, and those with diagnoses of T1D and (T1.5D or T2D). Cohorts are better suited for outcomes research, therapy use, and studying diabetes types.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EPH203
Topic
Economic Evaluation, Epidemiology & Public Health, Methodological & Statistical Research
Topic Subcategory
Disease Classification & Coding
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity), No Additional Disease & Conditions/Specialized Treatment Areas