A Method to Estimate Causally Related Cost Savings of a Preventive Intervention: An Analysis Using the Wellness-Star? Database
Author(s)
Akihiro Kakinuma, BA1, Masaru Kinugawa, M.S.1, Yuuri Miyamori, M.Eng1, Kenji Sato, M.S.1, Sayuri Yamamoto, B.Ag.1, Kunihiko Tanno, Dip.IT2, Kyousuke Nakamura, B.S.2, Ataru Igarashi, PhD3, Naoki Ikegami, PhD4, Kosuke Iwasaki, MBA5, Tomomi Takeshima, PhD5, AYANO CHIDA, High School Diploma5.
1Nippon Life Insurance Company, Tokyo, Japan, 2Nissay Information Technology Co., Ltd., Tokyo, Japan, 3Tokyo Univ. Facul. of Pharm. Dept. of Health Economics & Outcomes Research, Tokyo, Japan, 4Keio University, Tokyo, Japan, 5Milliman, Inc., Tokyo, Japan.
1Nippon Life Insurance Company, Tokyo, Japan, 2Nissay Information Technology Co., Ltd., Tokyo, Japan, 3Tokyo Univ. Facul. of Pharm. Dept. of Health Economics & Outcomes Research, Tokyo, Japan, 4Keio University, Tokyo, Japan, 5Milliman, Inc., Tokyo, Japan.
OBJECTIVES: When estimating the cost-saving effects of preventive interventions, only the portion of the cost difference between the health status groups is causally related. However, extracting causality from data has been challenging. Using disease-specific cost decomposition logic patented by Nissay Information Technology, the difference for each disease could justify the causality.
METHODS: We focused on weight loss for individuals with obesity and improvement of sleep quality for individuals with poor sleep. “Wellness-Star☆” database provided by Nippon Life Insurance Company, including claims and checkups data of about 200 health insurers, was used. We compared total and disease-specific costs Per Member Per Year (PMPY) between the obese group (BMI > 30) and the normal-weight group (18.5 < BMI < 25) for obesity, and between the well-slept group and the poor-slept group, based on surveys, for sleep quality. Diseases were classified using the first three digits of the ICD-10 codes.
RESULTS: PMPY for the normal-weight group (N=652,553) and obese group (N=64,132) was €708.01 and €1,289.83. The 3 diseases having the largest differences were type 2 diabetes (€9.45), essential hypertension (€7.85), and unspecified diabetes (€6.67), all of which have obesity as a risk factor. PMPY for the well-slept group (N=577,318) and poor-slept group (N=319,975) were €793.15 and €915.32. The 3 diseases having the largest differences between the two groups were chronic kidney disease (€0.66), sleep disorders (€0.65), and type 2 diabetes (€0.64), all of which seem not results but causes of the poor-slept condition.
CONCLUSIONS: When estimating the cost-saving effects of preventive interventions, a method was proposed that decomposes correlations by disease and determines causality for each condition individually. However, this approach is not necessarily applicable to all types of preventive interventions. While weight control may be feasible, improving sleep quality presents greater challenges.
METHODS: We focused on weight loss for individuals with obesity and improvement of sleep quality for individuals with poor sleep. “Wellness-Star☆” database provided by Nippon Life Insurance Company, including claims and checkups data of about 200 health insurers, was used. We compared total and disease-specific costs Per Member Per Year (PMPY) between the obese group (BMI > 30) and the normal-weight group (18.5 < BMI < 25) for obesity, and between the well-slept group and the poor-slept group, based on surveys, for sleep quality. Diseases were classified using the first three digits of the ICD-10 codes.
RESULTS: PMPY for the normal-weight group (N=652,553) and obese group (N=64,132) was €708.01 and €1,289.83. The 3 diseases having the largest differences were type 2 diabetes (€9.45), essential hypertension (€7.85), and unspecified diabetes (€6.67), all of which have obesity as a risk factor. PMPY for the well-slept group (N=577,318) and poor-slept group (N=319,975) were €793.15 and €915.32. The 3 diseases having the largest differences between the two groups were chronic kidney disease (€0.66), sleep disorders (€0.65), and type 2 diabetes (€0.64), all of which seem not results but causes of the poor-slept condition.
CONCLUSIONS: When estimating the cost-saving effects of preventive interventions, a method was proposed that decomposes correlations by disease and determines causality for each condition individually. However, this approach is not necessarily applicable to all types of preventive interventions. While weight control may be feasible, improving sleep quality presents greater challenges.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EE20
Topic
Economic Evaluation, Methodological & Statistical Research, Real World Data & Information Systems
Topic Subcategory
Cost/Cost of Illness/Resource Use Studies
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity), Mental Health (including addition)