Maybe skipping too much sleep this week is making me a little groggy, but let me make sure I understand the scenario:
In you example system, you've used the continuation method to place planet in the baseline orbit# (whose location is forced by the preexisting world and its climate). So you have an orbital location, based on whatever temperature range is appropriate for the pre-existing world, and a computed (Step 1) baseline number for it, so there is one less then the baseline number of orbit# to allocate inside this mainworld orbit#. The Spread for this star would be as determined by the Spread formula on page 48, potentially adjusted by the Maximum Spread formula on page 48.
Now are you asking if this Spread number applies to other stars in the system or whether it needs to be recalculated for each Secondary star (or Secondary and Companion pair)?
Well, it depends on whether you want to use the Optional Rule on page 49 (the first column optional rule, still under Step 5.) If not, then the Spread you've calculated for the Primary star applies to every Secondary star or secondary star and companion pair.
If you do want to use the optional rule, then your assumption that you would need to go back to previous steps is correct. But you'd want to step further back to each star or star pair and create a separate baseline number - this can get hairy, because you could start getting into situations where, for instance, the baseline number for this secondary set of orbits is, so low that kicks out all the orbits into unallowable regions. That's why the equation at the top of page 49 is there - Maximum Secondary Spread = ... This will force those orbits into the space available. Which is why I'd be careful recomputing this (plus, what is your definition of 'total worlds' now for the purpose of the DM? - I could argue that one both ways, but it is safer to use the system's total world, not the secondary's).
So that optional rule to recompute multiple baseline numbers and then spreads is hard to handle programmatically (especially in a spreadsheet). Best to stick with keeping the computed primary spread and modifying it down to a maximum secondary spread if necessary. The optional rule is fine for bespoke, hand crafted systems, but it introduces too many contingencies.
Short answer: Use the same Spread for all stars or star pairs, limited by the Maximum Secondary Spread formula if necessary.
As for the question on whether the continuation method forces the planet to be in orbit around the primary star, it doesn't, but you're right it isn't clearly indicated anywhere and the procedures are focused on placing the existing mainworld in relation to the habitable zone of the primary star Orbit#. Doing otherwise makes it more art than science, meaning a heavy use of Rule 0 might be necessary. First, you have to make sure the orbit supporting the proper climate fits for that star. (this is also true for the primary star if a secondary star disrupts the habitable zone). If not, you need to move stars around to make it work the way you want to. There is a risk, especially in multi-star systems that no star can support a habitable zone planet, but if you start with such a beast, it has to go somewhere, so something has to shift.
The baseline number is really meant to to determine the spread for the primary star and in many instances, the primary star is going to have the bulk of the planets. If the secondary is at a far location, you can treat primary and secondary fairly independently, divide the worlds as indicated in Step 1, and then move through the steps 2+ for each star's worlds. That's essentially using that optional rule, going back to Step 1 for each star/star pair.
I allowed myself an out with a statement on page 45: In the continuation method, if that placement is unacceptable because of a known
mainworld environment, the Referee may need to move stars to different Orbits#.
That's a lot of typing and maybe not exactly what you're asking, so feel free to continue to as clarifying questions. And I'll try to think and communicate clearly.
