Group by using left join and sum() = duplicate aggregates

[Nibinaear]

I have a query here involving LEFT JOIN which works fine but I need to add cats c (categories) and subcats s (subcategories).

I'll post the background to the original problem here:

Fixes has p_id col which corresponds to problems pk (I've grouped these).

Problem p            Fixes f
p_id                p_id    f_id    fix_count
1                    1    1    40
2                    1    2    20
3                    1    3    40 

So I wanted to group by fixes.p_id and sum the fix_count to get

Query
p.p_id            p.name        sum(fix_count)
1                Name info...    65 

There is not always a link to fixes. So I need a LEFT join.

Working a treat now but no cats or subcats:

SELECT IF(sum(f.num_fixed) IS NULL,0,sum(f.num_fixed)) AS num_fixed, p.*
FROM problem p
LEFT JOIN fixes f ON p.p_id = f.p_id
GROUP BY f.p_id
ORDER BY p.date_added DESC; 

Results:

num_fixed    p_id        name        cat_id
100            47        thing        1
0            48        rktjhrkjh        1 

When I add the new joins I have this:

SELECT IF(sum(f.num_fixed) IS NULL,0,sum(f.num_fixed)) AS num_fixed, p.*
FROM problem p
LEFT JOIN fixes f ON (p.p_id = f.p_id)
JOIN cats c ON c.cat_id=p.cat_id
JOIN subcats s ON s.cat_id = c.cat_id
GROUP BY s.cat_id,f.p_id
ORDER BY p.date_added DESC; 

num_fixed    p_id        name        cat_id
300            47        thing        1
0            48        rktjhrkjh        1 

There is one category (hardware) and 3 subcats linked so it groups the three rows and ends up summing three copies of num_fixed. I end up with 300 instead of 100. So I need to know how to group the cols and end up with 100 again with the new joins.

(This was stiched together from 2 examples so some values may not add up right.)