I've been a GTD follower for the past eighteen months and feel this is the most practical, effective and efficient method of managing a work load. However, over time I've found some heretical ways of tweaking the methodology to overcome some difficulties I experienced practicing GTD in the exact manner described in the book.
The most glaring problem I encountered was the enormous amount of next actions I faced, typically an average of 250-350 at any one time. I realize the GTD philosophy emphasizes the importance of having all possible next action options that one is committed to available for moment to moment review. Unfortunately, despite the conceptual validity of this benefit, I couldn't intuitively prioritize and choose what to do next at any moment when faced with a list of 250 items, all necessary tasks that could not be delegated, wisely ignored or judged not true next actions.
My solution was to prioritize projects on a weekly basis as part of the weekly review. Every week, I look over my projects list, all the multiple action outcomes I'm currently committed to, and decide which outcomes I'll focus on that week. I determine my next actions from those weekly priorities. I still consider the outcomes on the projects list that aren't on the weekly priority selection to be items I'm currently committed to, just not things I choose to work on over the next 5-7 days. I realize this violates GTD principles, however, I find that the far smaller next action lists, 40-80 tasks, more than compensate for the inability to have all options in front of me. For me, the system is unworkable without doing this.
The other adjustment I've made is to keep my projects in an outline format, instead of a laundry list. This solves the problem of determining how to list projects with many sub-projects that are not dependent on each other. I find this makes the weekly review easier, improves my ability to see the big picture and my ability to come up with effective next actions.