Census’s Supplemental Poverty Measure (SPM) nearly doubles the elderly poverty rate compared to the “Official” Poverty Measure (OPM), a result of the SPM subtraction of medical out-of-pocket (MOOP) expenditures from income. Neither the SPM nor OPM counts health benefits or assets as resources. Validation studies suggest that subtracting MOOP from resources worsens a poverty measure’s predictive validity and excluding assets exacerbates this bias, since assets fund MOOP. The SPM is based on a 1995 NAS report that recommended a health-exclusive poverty measure, despite considering it, conceptually, a “second best” to a Health-Inclusive Poverty Measure (HIPM). We analyze the reasons for the NAS recommendation and argue that constructing a HIPM is now feasible if we conceptualize health needs as a need for health insurance, and if plans with non-risk-rated premiums and caps on MOOP are universally available, a condition largely met by the Affordable Care Act and Medicare Advantage Plans. We describe four HIPM variants and present analyses that suggest the SPM treatment of MOOP results in a less valid measure of elderly poverty and an overstatement of the elderly poverty rate (by up to 5.5 percentage points or 50 percent). Many elderly classified as poor by the SPM’s unlimited MOOP deduction are not poorly insured persons with incomes near the poverty line, but well-insured persons with incomes well above the poverty line.