The overall health of a population can be viewed as an indicator of social welfare. Yet, individual health itself is complex and multidimensional, influenced by endogenous choices, as well as exogenous environmental and genetic factors. Moreover, defining a mapping from individual health to social welfare can involve onerous assumptions. This paper adopts a nonparametric approach to ranking individual health as a function of several biomarkers--Body Mass Index (BMI), glycohemoglobin (HbA1c), total cholesterol, alanine aminotransferase (ALT), serum creatinine, white blood cell counts (WBC), etc. With this ranking in hand, we use a nonparametric approach to map individual health into social welfare using minimal assumptions (e.g., monotonicity and concavity). Results show that the distribution of wellbeing became worse-off from 1988 to 2018, although there has been a slight rebound since 2009. Moreover, the distribution has widened: those prone to a higher health status have become better-off while those prone to poorer health have become worse-off which, thereby raising inequality and here policy implications need to be focused on. Finally, we construct counterfactual distributions of wellbeing to explore if the change in the distribution is attributed to socio-demographic factors. Findings show that age, gender and race/ethnicity cells combined with education can explain very little of the negative shift than the attributes without education while leaving a substantial portion unexplained.