Send to

Choose Destination
See comment in PubMed Commons below
Am J Health Promot. 2005 Jan-Feb;19(3 Suppl):249-54.

Association between lifestyle-disease diagnosis or risk status and medical care costs in a Japanese corporation.

Author information

Asia/Pacific Research Center, Stanford University, Stanford, California, USA.



This study examined the differences in medical care costs among (1) individuals who carried the diagnosis of selected lifestyle diseases (diabetes mellitus, hypertension, and hyperlipidemia), (2) individuals whose levels of risks (blood glucose, blood pressure, and total cholesterol) satisfied the diagnosis guidelines yet who did not carry the diagnoses, (3) individuals who had these risks but whose risk levels were not high enough to satisfy the diagnosis guidelines, and (4) individuals without the risks.


A one-time cross-sectional design was used. Health checkup data and medical-claims data obtained from the fiscal year 2000 were examined for correlations.


A total of 3292 employees aged 34 years and older were selected from the entire employee population of 6543 in a Japanese corporation. Employees younger than 34 years old were excluded because their clinical risk data were not available.


On the basis of their absence or presence of diagnoses (obtained from medical claims) and underlying risk levels (obtained from health checkups), employees were categorized into (1) the diagnosed group, (2) the extremely high-risk group, (3) the high-risk group, or (4) the no-risk group. Reimbursement points on medical care claims were summed for each individual during the study period and multiplied by 10 to calculate the total medical care costs, as each point in the reimbursement request form represents 10 yen.


The high-cost case analysis was used. First, the high costs were determined as the costs at or above the 90th percentile. The diagnosis or risk status was examined in its relation to the newly created dichotomous variable (whether the medical costs were at or above the threshold or were lower than the threshold) by using a chi2 test. Furthermore, excluding the diagnosed group, a chi2 test was performed to examine the relationships between the levels of risk and the likelihood of incurring any medical care costs (use vs. nonuse).


Approximately 15% of employees were already diagnosed with at least one of the three diseases (the diagnosed group; n = 490). One-quarter of employees had at least one risk that was high enough to be diagnosed with the corresponding disease if they had sought medical care (the extremely high-risk group; n = 809). There were 1343 employees in the high-risk group and 650 employees in the no-risk group. The diagnosed group had much higher chances of incurring medical care costs at or above the 90th percentile than did any other risk or no-risk group. No difference among the three risk or no-risk groups was found in mean medical care costs or in the likelihood of any use of medical care services after controlling for the effect of diagnosis.


In a Japanese employee population, the diagnosis status of diabetes mellitus, hyperlipidemia, and hypertension was found to be associated with higher medical care costs while risk levels for the diseases were not in a 1-year time period.

[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Loading ...
    Support Center