Objectives were to use meta-analytic methods to determine the effects of supplementing rumen-protected choline (RPC) starting prepartum on production and health of dairy cows. The literature was systematically reviewed and 23 experiments, including up to 74 treatment means and 1,938 cows, were used. All 23 experiments had a non-supplemented treatment (0 g choline/d). For supplemented cows, the mean ± SD (range) amount of choline ion fed was 12.7 ± 4.3 g/d (3.2 to 25.2). Data collected included number of cows, parity, days on treatment pre- and postpartum, choline ion fed (g/d), ingredients and nutrient content of prepartum diets, and LSM and respective SEM for DMI, production, and incidence of diseases. The contents of NE_L (Mcal/kg), metabolizable protein (MP) as % of DMI (MPDMI), and metabolizable methionine as % of MP (METMP) were estimated for each prepartum diet using NRC (2001). Mixed models were fitted using the MIXED and GLIMMIX procedures of SAS. Models contained the random effect of experiment and data were weighted by the inverse of the SEM squared. Models included the linear and quadratic effects of choline ion, NE_L, MPDMI, METMP, days on treatment prepartum, and interactions between choline and NE_L, and choline and METMP. Additional meta-analytical statistics were used to estimate the effect size using the METAN and METAREG procedures of STATA. Increasing prepartum intake of choline ion linearly (P < 0.001) increased pre- (β = 0.0187 ± 0.0045; weighted mean difference [WMD] = 0.28 kg/d) and postpartum DMI (β = 0.0438 ± 0.0125; WMD = 0.47 kg/d), ECM (β = 0.1773 ± 0.0219; WMD = 1.61 kg/d), fat yield (β = 0.0063 ± 0.0009; WMD = 0.08 kg/d), and protein yield (β = 0.0048 ± 0.0007; WMD = 0.06 kg/d) in cows. Supplementing RPC reduced (P = 0.05) the risk of retained placenta (relative risk, RR = 0.69; 95% CI = 0.47–1.02) and tended (P = 0.07) to reduce the risk of mastitis (RR = 0.78; 95% CI = 0.58–1.04), but it had no effect on displaced abomasum, ketosis, or metritis. These results support the recommendation of feeding RPC starting prepartum to dairy cows. Responses were linear up to 25 g/d of choline ion.