select distinct
a.status,
BILLING_TELEPHONE_NUMBER,
LINE_OF_BUSINESS_CODE,
ACCOUNT_CUSTOMER_CODE,
CUSTOMER_ACCOUNT_NUMBER,
SERVICE_NAME,
a.SERVICE_ORDER_ID,
STATUS,
ORIGINAL_ORDER_TYPE,
SUBID,
ORDER_NUMBER,
REGION_CODE,
ORDER_BATCHID,
CREATE_DATE,
ORIGINAL_CREATE_DATE,
COMPLETION_DATE,
SENT_TO_BILL_DATE,
ORIGINAL_SENT_TO_BILL_DATE,
DUE_DATE_SUFFIX,
RELATED_ORIGINAL_ORDER_TYPE,
RELATED_ORDER_NUMBER,
BIR,
b.LINE_ITEM_SEQ_NUM,
COMPENSABLE_PRODUCT_ID,
WORKING_TELEPHONE_NUMBER,
PRODUCT_CODE,
ACTION,
QUANTITY,
RATE,
CONTRACT_EFFECTIVE_DATE,
CONTRACT_USOC_TERM,
CONTRACT_EXPIRATION_DATE,
ORIGINAL_PROCESS_FLAG,
LAST_PROCESS_FLAG,
MCC_CD,
ISSUANCE_SENT_TO_BILL_FLAG,
QTNUM, ltrim(SALESID),
ISSUANCE_SHARE_PERCENT
from acct c, service_order a, so_line_item b, so_salesid f
where c.account_id=a.account_id and a.service_order_id=b.service_order_id and f.service_order_id=a.service_order_id
and a.create_date >= '1/1/2011'
and b.change_flag<>'D'
and char(a.cancel_order_flag)<>'Y'
and a.SENT_TO_BILL_DATE is not null
and b.issuance_sent_to_bill_flag<>'S'
and ltrim(rtrim(b.product_code))<>'TNC' with ur

You will probably need to post the table DDL, including indexes, for someone to know how to help you with the performance of the SQL. However, there are a few things to consider:

• Avoid using functions in predicates. DB2 is not Oracle (POS) and DB2 will compare 'Y' and 'Y ' as equals (unlike Oracle). Not sure why you are using the CHAR and LTRIM/RTRIM functions, but they can in some circumstances prevent indexes on these columns from being used (if indexes exist on those columns).
• It may not improve performance in this case, but always use ISO date formats when possible ('2011-01-01'). This avoids confusion if the client is not US territory code.

In Addition. to marcus_a's comments: Do you really need the SELECT DISTINCT? Is it any quicker if you remove it?

If you get duplicates I would check your joins as you shouldn't need 'distinct' in a properly normalised DB.

Have you current statistics on the tables?
Have you indexes on the join columns?
Have you indexes on the columns on which predicates are applied?
Have you a sound configuration of the database system? (memory pool sizes etc.)