There are lots of policies out there in the world that are not optimal. Every policy change involves significant legislative, grass roots political, and bureaucratic mobilization. So, changing everything at once isn't possible (in addition to being highly unpredictable). How do you prioritize?
One way to do it is to focus on issues where the status quo is clearly wrong to anyone who really understands it. These issues are ones where with fortitude, it is possible to build bipartisan coalitions, and where the results of the policy change, while probably impossible to measure at a macro level, are almost surely good.
Solutions to clear problems that are uncontroversial make up a large part of the legisltive agenda. But, even when the status quo is clearly out of whack, for example, in the case of disparate sentencing for crack and powder cocaine, or the double taxation of corporate distributed earnings of C corporations, if the issue is politically hot, you can be tilting at windmills. There is near total academic consensus that these are both bad ideas and has been for decades, but in best case scenarios they are mitigated rather than resolved on the floors of Congress.
Another approach is to dig into the often muddy issues that don't have clear solutions not because the current approach is clearly wrong, but because the consequences of even minor tweaks in the policy can matter a lot because it is an issue that has a big impact on a great many people. In these issues, the lack of a clear consensus solution to the problem, or even a clear consensus recognition that there is a problem, can prevent policy makers of any political persuasion from getting too righteous and entrenched about a particular approach, so there is room for policy movement. But, the stakes are so high no matter what change is made, that securing more than incremental change can be difficult. Health care reform fits more closely in this category.