The currently accepted answer by bhamby is certainly correct, but it’s worth checking if using several correlated subqueries is much slower than a single group by (hint: it most likely is):
SELECT
A.ItemID,
MAX(CASE WHEN A.Item = 'Meeting' THEN Value END) AS Meeting,
MAX(CASE WHEN A.Item = 'Advise' THEN Value END) AS Advise,
MAX(CASE WHEN A.Item = 'NoAdvise' THEN Value END) AS NoAdvise
FROM A
GROUP BY A.ItemID
It’s also a bit simpler in my opinion
SQLFiddle (in PostgreSQL, but works on DB2 LUW as well)