This study provides the first comprehensive analysis of the dynamics of labor supply of direct care workers, the lower-skill nursing workers who provide the bulk of long-term care for the elderly in the USA. Our estimates from the 1996 and 2001 panels of the Survey of Income and Program Participation (SIPP) show that the mean (median) duration of employment spells for the same direct care employer is only 9.7 (5.0) months. We find that fewer than one-third of direct care workers leave a job to take another job in the direct care field. There is also little indication of upward mobility in the health sector; direct care workers are approximately equally likely to transition to working as Registered Nurses as they are to working in household service jobs. Additionally, the rate at which spells end in work-limiting disability (5.4%) is very high compared with rates in similar occupations. We estimate duration models of direct care job spell length and find that, after correcting for the endogenous relationship between wages and tenure, wages appear to have a modest effect in preventing turnover; this effect is concentrated among the shortest spells.