Body Mass Index (BMI)

The ReadMe file has been updated with more detailed directions for network implementation.

Hi there,

In readme, there is a file EM_BMI_covariate_algorithm_<date>. I can't find it in the PheKB. Can you update PheKB? Thanks.

Cong

In the section of the script copied below, I think that I may be overlooking something. Can you please help me understand where data for #inclusion_events are coming from? I see the CREATE statement for #inclusion_events, but I am not sure where it gets populated with data. Thank you! Chad

create table #inclusion_events (inclusion_rule_id bigint,
person_id bigint,
event_id bigint
);

with cteIncludedEvents(event_id, person_id, start_date, end_date, op_start_date, op_end_date, ordinal) as
(
  SELECT event_id, person_id, start_date, end_date, op_start_date, op_end_date, row_number() over (partition by person_id order by start_date ASC) as ordinal
  from
  (
    select Q.event_id, Q.person_id, Q.start_date, Q.end_date, Q.op_start_date, Q.op_end_date, SUM(coalesce(POWER(cast(2 as bigint), I.inclusion_rule_id), 0)) as inclusion_rule_mask
    from Phen_BMI_qualified_events Q
    LEFT JOIN #inclusion_events I on I.person_id = Q.person_id and I.event_id = Q.event_id
    GROUP BY Q.event_id, Q.person_id, Q.start_date, Q.end_date, Q.op_start_date, Q.op_end_date
  ) MG -- matching groups

)  --end with cteIncludedEvents

select event_id, person_id, start_date, end_date, op_start_date, op_end_date
into Phen_BMI_included_events
FROM cteIncludedEvents Results
WHERE Results.ordinal = 1
;

The ReadMe file has been updated to note that if the OMOP person_id field is not the eMERGE subjid, then the identifier should be changed before uploading data to PheKB.

We are trying to implement this code in PostgresSQL on which our OMOP is based. So I need to convert some functions into PostgresSQL .

May I know why we are adding 0 to this which makes no difference ? :

1. DATEADD(day,0,OP.OBSERVATION_PERIOD_START_DATE) <= E.START_DATE AND DATEADD(day,0,E.START_DATE) <= OP.OBSERVATION_PERIOD_END_DATE

Why are we adding this function to add "0"?

Can't we just use :

OP.OBSERVATION_PERIOD_START_DATE <= E.START_DATE AND E.START_DATE <= OP.OBSERVATION_PERIOD_END_DATE

2. DATEADD(day,-1 * 0, event_date)  as end_date 

This is again adding "0"

Can we just use event_date as end_date? Why add 0?