Skip to content
Snippets Groups Projects
Commit 656b5c81 authored by Josef Hardi's avatar Josef Hardi
Browse files

feature(optum): allow specifying the starting occurrence date

parent d238c8a0
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 83 additions and 56 deletions
phs/data/optum.py
+ 83
− 56
View file @ 656b5c81
......@@ -158,23 +158,9 @@ class Optum:
t1.observation_period_start_date,
t1.observation_period_end_date
FROM (
SELECT
t0.person_id,
t0.gender_source_value,
t0.year_of_birth,
t0.zip,
t0.condition_source_value,
t0.index_date,
t0.observation_period_start_date,
t0.observation_period_end_date,
row_number() OVER (PARTITION BY t0.person_id ORDER BY t0.index_date, t0.condition_source_value) AS ordinal
FROM (
-- START 'Entry Criteria' >>>>
{entry_criteria}
-- <<< END 'Entry Criteria'
) AS t0
WHERE DATE_ADD(t0.observation_period_start_date, interval {obs_prior} DAY) <= t0.index_date
AND DATE_ADD(t0.index_date, interval {obs_subsequent} DAY) <= t0.observation_period_end_date
-- START 'Entry Criteria' >>>>
{entry_criteria}
-- <<< END 'Entry Criteria'
) AS t1
WHERE t1.ordinal = 1
)
......@@ -187,40 +173,54 @@ class Optum:
# SQL template for getting a population with a certain disease
_ENTRY_CRITERIA = """
WITH {person_table}, {person_claims_table}, {condition_occurrence_table}
SELECT DISTINCT
t3.person_id,
t3.gender_source_value,
t3.year_of_birth,
t3.index_date,
t3.observation_period_start_date,
t3.observation_period_end_date,
t3.zip,
t4.claim_id,
t4.condition_source_value
SELECT
t5.person_id,
t5.gender_source_value,
t5.year_of_birth,
t5.zip,
t5.condition_source_value,
t5.index_date,
t5.observation_period_start_date,
t5.observation_period_end_date,
row_number() OVER (PARTITION BY t5.person_id ORDER BY t5.index_date, t5.condition_source_value) AS ordinal
FROM (
SELECT
t1.*,
t2.claim_id,
t2.index_date
SELECT DISTINCT
t3.person_id,
t3.gender_source_value,
t3.year_of_birth,
t3.index_date,
t3.observation_period_start_date,
t3.observation_period_end_date,
t3.zip,
t4.claim_id,
t4.condition_source_value
FROM (
SELECT
t0.*
FROM
person AS t0
WHERE
t0.zip in ({zipcode_valueset})
) AS t1
INNER JOIN
person_claims AS t2
ON t1.person_id = t2.person_id
AND t2.index_date >= t1.observation_period_start_date
AND t2.index_date <= t1.observation_period_end_date
) AS t3
INNER JOIN (
{event_occurrence}
) AS t4
ON t3.person_id = t4.person_id
AND t3.claim_id = t4.claim_id"""
SELECT
t1.*,
t2.claim_id,
t2.index_date
FROM (
SELECT
t0.*
FROM
person AS t0
WHERE
t0.zip in ({zipcode_valueset})
) AS t1
INNER JOIN
person_claims AS t2
ON t1.person_id = t2.person_id
AND t2.index_date >= t1.observation_period_start_date
AND t2.index_date <= t1.observation_period_end_date
) AS t3
INNER JOIN (
{event_occurrence}
) AS t4
ON t3.person_id = t4.person_id
AND t3.claim_id = t4.claim_id
{event_occurrence_filters}
) AS t5
{continuous_observation_filters}"""
def info(self):
"""Returns a brief description about the dataset.
......@@ -301,13 +301,38 @@ class Optum:
"""
event_occurrence_sql = ""
event_occurrence_filters_sql = ""
continuous_observation_filters_sql = ""
for k, v in entry_criteria.items():
if k == 'condition':
condition_valueset=",".join(f'"{x}"' for x in self._CONDITIONS[v]['codes'])
event_occurrence_sql = f'SELECT t0.*\nFROM condition_occurrence AS t0\nWHERE t0.condition_source_value in ({condition_valueset})'
elif k == 'conditionCodes':
condition_valueset=",".join(f'"{x}"' for x in v)
event_occurrence_sql = f'SELECT t0.*\nFROM condition_occurrence AS t0\nWHERE t0.condition_source_value in ({condition_valueset})'
if k == 'event':
event_type = list(v.keys())[0]
if event_type == 'condition':
event_value = v['condition']
if type(event_value) is str:
condition_valueset=",".join(f'"{x}"' for x in self._CONDITIONS[event_value]['codes'])
event_occurrence_sql = f'SELECT t0.*\nFROM condition_occurrence AS t0\nWHERE t0.condition_source_value in ({condition_valueset})'
elif type(event_value) is list:
condition_valueset=",".join(f'"{x}"' for x in event_value)
event_occurrence_sql = f'SELECT t0.*\nFROM condition_occurrence AS t0\nWHERE t0.condition_source_value in ({condition_valueset})'
elif k == 'occurrence':
occurrence_type = list(v.keys())[0]
if occurrence_type == 'dateStartBefore':
occurrence_date = v['dateStartBefore']
event_occurrence_filters_sql = f'WHERE t3.index_date < "{occurrence_date}"'
elif occurrence_type == 'dateStartAfter':
occurrence_date = v['dateStartAfter']
event_occurrence_filters_sql = f'WHERE t3.index_date > "{occurrence_date}"'
elif occurrence_type == 'dateStartAt':
occurrence_date = v['dateStartAt']
event_occurrence_filters_sql = f'WHERE t3.index_date = "{occurrence_date}"'
elif occurrence_type == 'dateStartRange':
occurrence_start_date = v['dateStartRange'].split('/')[0]
occurrence_end_date = v['dateStartRange'].split('/')[1]
event_occurrence_filters_sql = f'WHERE t3.index_date >= "{occurrence_start_date}" AND t3.index_date <= "{occurrence_end_date}"'
elif k == 'continuousObservations':
days_before = v['daysBeforeOccurrence'] if 'daysBeforeOccurrence' in v else 0
days_after = v['daysAfterOccurrence'] if 'daysAfterOccurrence' in v else 0
continuous_observation_filters_sql = f'WHERE DATE_ADD(t5.observation_period_start_date, interval {days_before} DAY) <= t5.index_date AND DATE_ADD(t5.index_date, interval {days_after} DAY) <= t5.observation_period_end_date'
entry_criteria_sql = textwrap.dedent(
self._ENTRY_CRITERIA.format(
......@@ -315,7 +340,9 @@ class Optum:
person_claims_table=self._PERSON_CLAIMS,
condition_occurrence_table=self._CONDITION_OCCURRENCE,
zipcode_valueset=",".join(str(z) for z in zips),
event_occurrence=event_occurrence_sql))
event_occurrence=event_occurrence_sql,
event_occurrence_filters=event_occurrence_filters_sql,
continuous_observation_filters=continuous_observation_filters_sql))
inclusion_criteria_sql = ""
if inclusion_criteria:
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment