Objective: To examine how the family member’s mental health is affected by the presence of a member who needs a long-term care, after controlling for own sex, own age, own work-status, the familial relationship with the member, household income, spending for nursing care, and socioeconomic characteristics. Background: Long-term care insurance was introduced in Japan in year 2001 mainly to provide home care to support family caregivers, but, due to financial problems, it has gone through major cut backs every three years. In the meantime, family still continues to play key role in the provision of long-term care, and, little is known regarding the effects of LTCI or its reforms on the family burden. We begin by comparing the mental health of each member in a caregiving family with a similar person in a family in the general population (w/o care). Data: We use 2001, 2004, 2007, and 2010 samples of Comprehensive Survey of Living Conditions by the Ministry of Health, Labour and Welfare. Each survey year data contains information of almost 750 thousand individuals in 250 thousand households, on family structure and individual health. In addition, about 5 thousand families providing long-term care to its member are “harvested” in each survey year. Method: We pool all of the samples as an independently pooled cross section data and estimate a logistic regression on the data. Our dependent variable is the binary variable for own mental health status which is equal to 0 for a no stress or no concern and equal to 1 otherwise. Our explanatory variables are the indicator variable for the presence of a family member who needs care, index of the care-need level, family income, expenditure on nursing care, as well as the dummy variables for own age, sex, working status, familial relationship, year, type of house and so on. Following the estimation, we will compute average marginal effects (AMEs) of the variables of our model. Policy implications: Building the huge database is quite a challenge, and no results are yet available. This is the first attempt in Japan focusing on all members’ mental health in a family with or without caregiving burden. Although a number of psychometric scales are available for family caregivers, introducing such a scale in a large scale government survey is not realistic in Japan. If, instead, the stress or SAH can serve as an index of caregiving burden of the family, it would be very useful for policy planning and intervention (e.g. willingness to pay for LTCI).