SUM SQL Server 2008上的性能非常低

I have a table of likes/dislikes that contains around 5 million rows. When I use the following query to get the data, it completes in 2+ minutes. Is there a better approach to how I store and retrieve the likes/dislikes? Every time someone likes/dislikes a post it adds a column to the db. 0 For dislike and 1 for like. I then need the sum of both columns for each user and then return the users with the most amount of likes versus dislikes. If I take out the SUMs for likes/dislikes, the query returns in 4 seconds. I also have indexes on UserID and everything that I am grouping by. Here is the query:I have a table of likes/dislikes that contains