The phenomenon named kodokushi, meaning death alone without the care or company of anyone inside temporary housing, appeared after the Kobe earthquake in Japan in 1995 with some 250 cases. This paper analyzes the evolution of Japanese temporary houses—to attempt to prevent the problem of kodokushi—from the point of view of management, how services and activities are organized, and design. We will use case studies as our methodological tool, analyzing the responses in 1995 Kobe (50,000 THs), 2004 Chūetsu (3000 THs), 2011 Tōhoku (50,000 THs), and 2016 Kumamoto (4000 THs). This article shows how the Japanese THAs follow a single design that has undergone very little variation in the last 25 years, a design which promotes the social isolation of their residents, making recovery—from the psychological perspective—and helping the most vulnerable members of society, more difficult. In small scale disasters (Chūetsu) applying organization and management measures was able to correct the problems caused by design and there were no cases of kodokushi: in large-scale disasters (Tōhoku), however, the difficulties to implement the same measures resulted in the reappearance of new cases at rates similar to Kobe’s. Our main conclusion is that the design of Japanese THAs must be reconsidered and changed to respond to the real needs of the most vulnerable groups.