select o1.ids
      ,sum(case when o1.status=3 then datediff(day,o2.date,o1.date) else 0 end) as d3
      ,sum(case when o1.status=2 then datediff(day,o2.date,o1.date) else 0 end) as d2
      ,sum(case when o1.status=1 then datediff(day,o2.date,o1.date) else 0 end) as d1
  from table_o as o1
       left outer join table_o as o2
                    on o2.ids=o1.ids
                   and o2.status=o1.status+1
 where o1.status between 1 and 3
 group by o1.ids
 order by o1.ids

declare @t table
(
Ids varchar(5),
[Status] int,
[Date] date
)
insert @t
values 
('ID1',	4,	'02-May-13'),
('ID1',	3,	'10-May-13'),
('ID1',	2,	'16-May-13'),
('ID1',	1,	'20-May-13'),
('ID2',	3,	'08-May-13'),
('ID2',	2,	'10-May-13'),
('ID2',	1,	'19-May-13')


SELECT IDs,
DATEDIFF(dd,MAX(CASE WHEN Status=4 THEN date END),MAX(CASE WHEN Status=3 THEN date END)) AS [4-3],
DATEDIFF(dd,MAX(CASE WHEN Status=3 THEN date END),MAX(CASE WHEN Status=2 THEN date END)) AS [3-2],
DATEDIFF(dd,MAX(CASE WHEN Status=2 THEN date END),MAX(CASE WHEN Status=1 THEN date END)) AS [2-1]
FROM @t
CROSS JOIN (VALUES (0),(-1),(-2),(-3))t1(n)
GROUP BY IDs--,[Status]+n


output
-----------------------------------
IDs	4-3	3-2	2-1
-----------------------------------
ID1	8	6	4
ID2	NULL	2	9