Diabetes causes complications and collateral diseases, reducing quality of life and increasing medical costs. The Japanese government has promoted measures for the prevention of diabetes aggravation. Although glycemic control is reported to prevent the development of complications, assessment of the effects on overall medical cost is insufficient. We examined the medical cost by the analysis of hemoglobin A1c (HbA1c) level.

A Japanese employee-based health insurance claims database with annual medical check-up data was analyzed. Excess medical cost was calculated as the difference between medical cost and standard medical cost (defined as the average medical cost for individuals of same age and sex). Percentage of excess medical cost was calculated by dividing excess medical cost by standard cost, and compared between individuals with or without treatment for diabetes.

Of 4,307,184 individuals with HbA1c data, four percent of them received treatment for diabetes. For treatment of 6.5 percent of HbA1c, excess medical cost increased to 124 percent. The medical cost increased by an additional 20.4 percent (95% CI: 17.1–23.8) when the HbA1c level increased one percent. Treatment for less than six percent of HbA1c caused an increase consistent with the HbA1c level. The relative risk of iron deficiency anemia, unspecified of those with less than six percent of HbA1c against those with seven to eight percent was the highest, 2.15.

An increase of medical cost for individuals with treatment for high HbA1c is likely associated with diabetic complication. The raise for individuals with lower HbA1c level may be related to anemia. Despite the younger age and healthier life of the analyzed individuals, since they were insured by employee-based health insurance who took regular medical check-ups, more expensive medical cost was observed for those having higher HbA1c level.

Self-reported outcomes are considered to be useful to understand physical and mental conditions in daily life. Sleep quality is an important factor related to healthy lifestyle and work productivity, as well as to diseases. We examined the relationship of sleep condition with lifestyle and diseases based on self-reported sleep.

A Japanese employee-based health insurance claims database with annual medical check-up data was used. Individuals were questioned about sleep quality as: “Do you get enough rest by sleeping?” during the medical check-up. The prevalence of diseases and medical check-up data were compared between those who answered “Yes” or “No”.

Among 1,310,157 individuals who answered about sleep quality, 540,564 (41.3 percent) answered “No”. The female ratio was around 38 percent for both answers, and the average age was lower for those who answered “No” (45.3) than “Yes” (47.3). Matched individuals for same examination year, sex, and age were 536,218 in each group. Individuals diagnosed with sleep disorder were 8.7 percent of those who answered “No”, representing the highest relative risk (RR=1.64), followed by other anxiety disorders (RR = 1.47), and depressive episode (RR = 1.45), with statistical significance. Other diseases diagnosed in more than 200,000 patients, and which had significantly higher RRs in patients who answered “No”, included vasomotor and allergic rhinitis (RR = 1.09), disorders of refraction and accommodation (RR =1.02), acute upper respiratory infections (RR = 1.11), gastritis and duodenitis (RR = 1.17), and acute bronchitis (RR = 1.13). The RR of other diseases of the liver (RR = 1.13), diabetes (RR = 1.12), hypertension (RR = 1.08), and disorders of lipoprotein metabolism and other lipidaemias (RR = 1.06) were also significantly higher for those who answered “No”.

Sleep quality is suggested to be associated with various chronic diseases as well as mental disorders. Therefore, self-reported outcomes should be a useful tool to understand health-condition, prevent the onset and progression of diseases, and evaluate patient-centered care.

Mr. Shinjiro Koizumi and some younger members of Japan's National Diet suggested a new policy, “Health Gold License” which would introduce financial incentives to encourage population health management, with people receiving medical checkups receiving a reduction in coinsurance from the current 30 percent to 20 percent. In this research, to evaluate the policy, we adjusted confounding factors of those insured who receive medical checkups (Medical-Checkup Group) and those who do not (Non-Medical-Checkup Group) using claims data, and estimated the effect of medical checkups on medical costs.

We analyzed Japanese employee-based claims data provided by the Japan Medical Data Center Co. Ltd. for the 3 million insured from January 2005 to December 2015. Two regression models were developed. Under model A, explanatory variables were year, age, dummy variables for various hierarchical condition categories and for medical checkups. Under model B, explanatory variables were estimated medical costs per patient per month (PMPM) in 2012 and a dummy variable for medical checkups. We also simulated the financial impact if Japan introduced Health Gold License for all insured.

The coefficients of medical checkups in model A and in model B were -JPY4,816 PMPM and -JPY8,735 PMPM, respectively. The gap of medical costs between the Medical-Checkup Group and Non-Medical-Checkup Group was JPY4,588 PMPM, without any adjustment. If all of those insured received medical checkups, the breakeven coinsurance would be 27.2 percent.

The Medical-Checkup Group is less expensive than Non-Medical-Checkup Group by at least 30%, therefore, the break-even coinsurance for them would be 0 percent. However, because most of those insured have already gone to medical check-ups every year, if the coinsurance were reduced from 30 percent to 20 percent for all insured, the finance would be largely negative. The break-even as 27.2 percent, we believe, would not incentivize the Non-Medical-Checkup Group to receive medical checkups. Therefore, the coinsurance reduction proposed under Health Gold License is not fully justified financially.