Blood Pressure Observations

Search:

Description

This phenotype is used to determine the "primary" blood pressure for a given minute from the monitor data, as well as to cache derived metadata about each blood pressure.

Limitation

Eliminating artifact values

Solving time misalignment (e.g. systolic values is offset from diastolic values by a few seconds)

Picking one set of BPs when multiple sources are available (e.g. arterial vs cuff)

Data quality is limited to individual site documentation accuracy, completeness, and variable mapping, which is performed by each site, independently. MPOG, as the coordinating center, reviews data quality at each data upload and monthly via Data Diagnostics and Case Validation tools. This review, while thorough, may not contain all data elements relevant to this phenotype and thus should be carefully reviewed by the user.

Value Type

Numeric

Enumeration

Value	Value Code	Definition

Return Columns

Column Name	Data Type

Granularity

Logic

Step 1: Pull all values with the following concepts that are within the time bounds of 4 hours before Anesthesia Start and PACU End Time.:

3011 BP Sys Invasive Unspecified Site 1
3012 BP DIas Invasive Unspecified Site 1
3013 BP Mean Invasive Unspecified Site 1
3015 BP Sys Non-invasive
3020 BP Dias Non-invasive
3025 BP Mean non-invasive
3026 BP Sys Invasive Unspecified Site 4
3027 BP Dias Invasive Unspecified Site 4
3028 BP Mean Invasive Unspecified Site 4
3030 BP Sys Aterial Line (Invasive, Peripheral)
3035 BP Dias Arterial Line (Invasive, Peripheral)
3040 BP Mean Arterial Line (Invasive, Peripheral)
3041 BP Sys Invasive Unspecified Site 2
3042 BP Dias Invasive Unspecified Site 2
3043 BP Mean Invasive Unspecified Site 2
3046 BP Sys Invasive Unspecified Site 3
3047 BP Dias Invasive Unspecified Site 3
3048 BP Mean Invasive Unspecified Site 3
3475 BP Sys Invasive Unspecified Site 5
3476 BP Dias Invasive Unspecified Site 5
3477 BP Mean Invasive Unspecified Site 5

*Aortic BP is not used because it does not have the same clinical meaning

Step 2: Truncate seconds from the BP timestamp

Step 3: If multiple values are present for a given minute and concept, take the highest value

Step 4: Combine components of the same BP category into a single row for each minute. See example below:

**BEFORE**
BP Time	BP Concept	Value
9:01 AM	BP Sys Non-invasive	120
9:01 AM	BP Dias Non-invasive	80
9:01 AM	BP Mean Non-invasive	93
9:01 AM	BP Sys Arterial Line (Invasive, Peripheral)	125
9:01 AM	BP Dias Arterial Line (Invasive, Peripheral)	100
9:01 AM	BP Mean Arterial Line (Invasive, Peripheral)	108

**AFTER**
BP Time	BP Category	Systolic	Diastolic	Recorded Mean
9:01 AM	BP Non-Invasive	120	80	93
9:01 AM	BP Arterial Line (Invasive, Peripheral)	125	100	108

Step 5: For each blood pressure, compute an approximate BP mean from the systolic and diastolic to account for the absence of a recorded mean. [Computed Mean = (SYS + DIAS + DIAS) / 3 ]

Step 6: Apply the following artifact rules to each value

Rule	Artifact Code
Any of the original component sys, dias, or mean values were marked as artifact in the original system	1
Sys > 150 AND (Sys - Dias) < 30	2
Sys >= 100 AND Sys <= 150 AND (Sys - Dias) < 15	3
Sys < 100 AND (Sys - Dias) < 10	4
Sys > 200 AND (Sys - Dias) < 50	5
Sys <= 10 OR Dias <= 10	6
Sys = Dias = Mean	7
Recorded or Calculated Mean <= 10	8
Mean >= 150	9
Single Spike in Dias	11
Sys > 350 OR Dias > 250	12
Recorded Mean > Sys	13
Dias > Recorded Mean	14
Dias > Sys	15
Dias is Missing OR Sys is Missing	16

Step 7: Find single value spikes by comparing each BP dias to the BP values of the same category one minute before and after. If the previous value and next value are within 10% of each other and the middle value is twice that of the previous value, mark the middle value with artifact code 11.

Step 8: For each minute within a case that has a BP value, rank all BP values within that minute to determine the "primary" BP for that minute

Non-artifact values are ranked over artifact values
Values with both a systolic and a diastolic are ranked over those that don't (e.g. BP mean only was recorded)
Values with a recorded BP mean are ranked over those that don't
Of any remaining BP values, the one with the highest systolic is chosen
BP components (sys/dias/mean) are grouped together based on source (invasive vs. non-invasive)

Step 9: For every "primary" BP that is not artifact, compute the number of minutes until the next available BP value

Output: This phenotype will result in 4 columns:

BP_Time - Time of the BP Value
BP_Sys, BP_Dias, BP_Mean_Recorded, BP_Mean_Computed- These columns lists all of the BP values for each minute that a BP value was captured. Including the ones that the MPOG algorihm above determined to be Artifact.
MPOG_Marked_Artifact_CD - This column lists the artifact code that the value falls into based on the rule table above.
BP_Sys_Cleaned, BP_Dias_Cleaned, BP_Mean_Recorded_Cleaned, BP_Mean_Computed_Cleaned - These columns return only the values determined to not be artifact based on the rule table above. (Note, artifact values are removed)

Example Case:

BP_Time	BP_Sys	BP_Dias	BP_Mean_Recorded	BP_Mean_Computed	MPOG_Marked_Artifact_CD	BP_Sys_Cleaned	BP_Dias_Cleaned	BP_Mean_Recorded_Cleaned	BP_Mean_Computed_Cleaned
3/29/24 8:56	21	-8	-6	2	1
3/29/24 8:57	161	71	93	101	1
3/29/24 8:58	240	3	75	82	6
3/29/24 8:59	342	286	11	305	12
3/29/24 9:00	179	76	115	110	0	179	76	115	110
3/29/24 9:01	200	144	160	163	9
3/29/24 9:02	214	99	144	137	0	214	99	144	137
3/29/24 9:03	229	153	245	178	1
3/29/24 9:04	205	93	139	130	0	205	93	139	130
3/29/24 9:05	304	301	304	302	2
3/29/24 9:06	218	181	194	193	5
3/29/24 9:07	138	63	249	88	1
3/29/24 9:08	104	95	93	98	3
3/29/24 9:09	132	64	91	87	0	132	64	91	87
3/29/24 9:10	126	62	86	83	0	126	62	86	83
3/29/24 9:11	120	58	246	79	1
3/29/24 9:12	111	57	77	75	0	111	57	77	75
3/29/24 9:13	106	54	73	71	0	106	54	73	71
3/29/24 9:14	92	51	92	65	7
3/29/24 9:15	144	74	99	97	0	144	74	99	97
3/29/24 9:16	128	65	89	86	0	128	65	89	86
3/29/24 9:17	116	59	250	78	1
3/29/24 9:18	115	60	250	78	1
3/29/24 9:19	110	57	78	75	0	110	57	78	75
3/29/24 9:20	104	55	73	71	0	104	55	73	71
3/29/24 9:21	104	54	72	71	0	104	54	72	71
3/29/24 9:22	157	82	250	107	1
3/29/24 9:23	156	81	109	106	0	156	81	109	106
3/29/24 9:24	141	70	96	94	0	141	70	96	94
3/29/24 9:25	204	139	167	161	9
3/29/24 9:26	114	58	79	77	0	114	58	79	77
3/29/24 9:27	105	55	74	72	0	105	55	74	72
3/29/24 9:28	95	50	67	65	0	95	50	67	65
3/29/24 9:29	129	68	90	88	0	129	68	90	88
3/29/24 9:30	130	69	91	89	0	130	69	91	89
3/29/24 9:31	110				8
3/29/24 9:32	120	62	83	81	0	120	62	83	81
3/29/24 9:33	131	66	90	88	0	131	66	90	88
3/29/24 9:34	133	67	91	89	0	133	67	91	89
3/29/24 9:35	60	57	58	58	4
3/29/24 9:36	136	68	93	91	0	136	68	93	91
3/29/24 9:37	137	68	93	91	0	137	68	93	91
3/29/24 9:38	110	57	168	75	13
3/29/24 9:39	114	50	74	71	0	114	50	74	71
3/29/24 9:40	65	55	52	58	14
3/29/24 9:41	109	51	72	70	0	109	51	72	70
3/29/24 9:42	138	67	93	91	0	138	67	93	91
3/29/24 9:43	153	115	130	128	11
3/29/24 9:44	123	55	80	78	0	123	55	80	78