You'll need to cross-reference this with the actual query execution plan to see if it is reasonable or badly optimised.

This modified query will add the execution plans to the output:

SELECT TOP 100 T.*, qp.query_plan

FROM

(

SELECT

DatabaseName = DB_NAME(qt.dbid)

,ObjectName = OBJECT_SCHEMA_NAME(qt.objectid,qt.dbid) + '.' + OBJECT_NAME(qt.objectid, qt.dbid)

,DiskReads = SUM(qs.total_physical_reads) -- The worst reads, disk reads

,MemoryReads = SUM(qs.total_logical_reads) --Logical Reads are memory reads

,Executions = SUM(qs.execution_count)

,IO_Per_Execution = SUM((qs.total_physical_reads + qs.total_logical_reads) / qs.execution_count)

,CPUTime = SUM(qs.total_worker_time)

,DiskWaitAndCPUTime = SUM(qs.total_elapsed_time)

,MemoryWrites = SUM(qs.max_logical_writes)

,DateLastExecuted = MAX(qs.last_execution_time)

,QueryText = qt.Text

,PlanHandle = qs.plan_handle

FROM sys.dm_exec_query_stats AS qs

CROSS APPLY sys.dm_exec_sql_text(qs.sql_handle) AS qt

GROUP BY DB_NAME(qt.dbid), OBJECT_SCHEMA_NAME(qt.objectid,qt.dbid) + '.' + OBJECT_NAME(qt.objectid, qt.dbid), qt.Text, qs.plan_handle

) T

CROSS APPLY sys.dm_exec_query_plan(T.PlanHandle) as qp

ORDER BY IO_Per_Execution DESC