EXISTS with Group By Causes execssive run time

In query that are constructed where an EXISTS statement is used causes increased query times due to the sorting and grouping performed within the sub-query. Example:
UPDATE QUERY_GLOBAL_TEMP
SET panel_count = -1
WHERE QUERY_GLOBAL_TEMP.panel_count = 4
        AND EXISTS
        (
                SELECT 1
                FROM
                        (
                                SELECT
                                        /*+ index(observation_fact fact_cnpt_pat_enct_idx) */
                                        f.patient_num
                                FROM observation_fact f
                                WHERE f.concept_cd IN
                                        (
                                                SELECT concept_cd
                                                FROM concept_dimension
                                                WHERE concept_path LIKE '\i2b2\DIAG\ICD10\ku5d\8s2b\nigp\%'
                                        )
                                GROUP BY f.patient_num
                        )
                        t
                WHERE QUERY_GLOBAL_TEMP.patient_num = t.patient_num
        )

In oracle, excessive time is spent performing the grouping and sort that causes a lot of inefficiency in the queries. The first occurrence of the patient_num should be enough to satisfy the EXISTS without the need to group the sub-query results which requires a complete and repetitive scan of the index.

Additional logic within the query construction should be added to not append the group by when an EXISTS is used.