Skip to main content
. 2022 Feb 22;12(2):e055773. doi: 10.1136/bmjopen-2021-055773

Table 4.

Numbers of pregnancy episodes with recorded outcome missing which were consistent with applied criteria for each scenario*

Scenario Description N of pregnancy episodes with evidence of this scenario (% of total episodes with missing outcome) N of pregnancy episodes with evidence of this scenario only (% of total episodes with missing outcome) N of pregnancy episodes with evidence of an outcome in linked HES (% of linkage eligible episodes with recorded outcome missing†) N of episodes during current registration and UTS follow-up (% of total episodes with missing outcome)*
Denominator 932 604 932 604 424 375† 932 604
Problem 1: The women was pregnant at the time of the database record, but the outcome was not captured in CPRD primary care data.
Scenario 1a The pregnancy outcome occurred in hospital or elsewhere and information wasn’t fed back to the practice. 211 070 (22.6) 1934 (0.2) 211 070 (49.7) 139 084 (14.9)
Scenario 1b The outcome of the pregnancy is recorded in the primary care data but has no event date associated with it. 1595 (0.2) 48 (0.0) 523 (0.1) 475 (0.1)
Scenario 1c The pregnancy occurred before the patient was registered at the practice or before UTS. 415 807 (44.6) 204 176 (21.9) 60 423 (14.2) 0 (0.0)
Problem 2: The women was pregnant at the time of the database record, but the pregnancy was still ongoing at the end of available follow-up in the database.
Scenario 2a The patient transferred out or died before the putative end of pregnancy. 177 557 (19.0) 40 191 (4.3) 71 012 (16.7) 117 571 (12.6)
Scenario 2b The last collection date of the practice was before the putative end of pregnancy. 65 141 (7.0) 22 039 (2.4) 24 091 (5.7) 58 698 (6.3)
Problem 3: The patient was not pregnant at the time of the database record.
Scenario 3a Episode is derived from historical pregnancy information recorded in the first few months after the patient joined the practice. 10 235 (1.1) 588 (0.1) 3058 (0.7) 3875 (0.4)
Scenario 3b Patient asks for advice while planning a pregnancy. 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Problem 4: The pregnancy record belongs to another pregnancy episode in the Register.
Scenario 4a Delay in recording the outcome of a pregnancy, algorithm calculates the last menstrual period date (LMP) too late and uncovers records at the beginning of pregnancy creating this pregnancy with recorded outcome missing. 61 662 (6.6) 9299 (1.0) 23 099 (5.4) 35 255 (3.8)
Scenario 4b The LMP is derived from the data and is wrong resulting in early codes being uncovered creating this episode. 29 057 (3.1) 4022 (0.4) 11 304 (2.7) 17 110 (1.8)
Scenario 4c The LMP has been shifted earlier in time uncovering records at the end of the pregnancy. 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Scenario 4d A code recorded relating to the patient’s delivery history is incorrectly identified by the algorithm as a delivery uncovering records at the end. 349 874 (37.5) 113 688 (12.2) 90 274 (21.3) 219 505 (23.5)
Scenario 4e The outcome of the pregnancy episode has been misclassified as antenatal 38 848 (4.2) 8000 (0.9) 6611 (1.6) 18 222 (2.0)
None These pregnancy episodes did not meet the criteria for any identified scenarios. 106 458 (11.4) 94 769 (10.2)

*A version of this table restricted to episodes which occurred during practice UTS follow-up and patient’s current registration is given in the appendices (online supplemental appendix 16).

†Denominator=pregnancy episodes which had at least 1-day overlap with the available HES follow-up period and where the woman was eligible for linkage.

CPRD, Clinical Practice Research Datalink; HES, Hospital Episode Statistics; UTS, up-to-standard.