This difference would have the greatest impact on treatment for cancer, in which a treatment protocol is based on genes deleted, amplified, altered in the tumor vs the reference genome for that patient.
In terms of health risks based on SNP genotypes, the data are far from complete. Sure, some level of risk can be assigned to a variant (SNP), say at certain markers within the FTO gene and risk of obesity. However, the complete list of risk alleles for this and numerous other complex traits is not fully described. Furthermore, the impact of environmental factors (EF), or lifestyle choices, on those genetic variants is only beginning to gain wider attention. For example, a risk allele may not show itself as risk until the EF, such as amount of physical activity or percent energy from dietary polyunsaturated fat, passes a certain threshold. Such gene-environment interactions are thought to contribute to the variance in traits (phenotypes), but to what degree is not known.
Added in edit 15 May 2012: Epistasis, or gene-gene interactions, also are important but far from being cataloged for humans. A situation could arise where one allele elevates risk for a certain condition, but compensatory alleles elsewhere in the genome decrease that risk. We don't know the full extent of epistasis in humans.
Thus, in terms of personal healthcare outside of something like cancer and monogenetic disorders, there may be little that is gained from genotyping or complete genome sequencing, little that is compared to other advice you already know. I carry risk alleles for certain conditions, but there is not really any advice one can give me that is specific for those alleles that I have, where that advice is different than or goes beyond general advice health care providers already have dispensed.