Author, year, country | Study sample selection bias | Bias in measurement of exposure | Bias in measurement of outcomes | Attrition bias | Analysis method reported and appropriate | Closure independent of other changes over time | Potential confounders adjusted for and listed |
UK Studies | |||||||
Fleming/East Lancashire study | |||||||
Fleming15, 2013, UK (abstract) | LOW All births in East Lancashire Maternity Services catchment | LOW All births in catchment area affected by the closure No of OUs closed reported | LOW Objective outcome (CS) | UNCLEAR Not reported | UNCLEAR None reported | UNCLEAR None reported | UNCLEAR None reported |
East Lancashire Hospitals NHS Trust 2017, UK (unpublished data) | UNCLEAR Unpublished data, no details | UNCLEAR Unpublished data, no details | LOW Objective outcomes (BBA, SB, NM) | UNCLEAR Unpublished data, no details | UNCLEAR Unpublished data, no details | UNCLEAR Unpublished data, no details | UNCLEAR Unpublished data, no details |
Mackie et al
16, 2014, UK (abstract & unpublished) | LOW Data from Maternity Information System | LOW All births in catchment area affected by the closure. No of OUs closed reported | LOW Objective outcomes (SB, NM, BBA, MM, ICU admission, perineal tears) | UNCLEAR Not reported | UNCLEAR None reported | UNCLEAR None reported | UNCLEAR None reported |
Other European Studies | |||||||
Blondel et al 5, 2000, France* | LOW Data from birth certificates | LOW No of OUs closed reported | LOW Objective outcome (BBA) | LOW 11% excluded due to missing data | LOW Analysis method was described and appropriate, a multi-level model analysis | UNCLEAR None reported | LOW Adjusted for maternal age, occupational category and rurality |
Hemminki et al 18, 2011,m Finland | LOW Data from Finnish medical birth register | LOW All births in catchment area affected by the closure No of OUs closed reported | LOW Objective outcome (BBA) | LOW Births with missing information excluded (<0.05%) | LOW Analysis method was described and appropriate, a regression model with adjusted analysis | UNCLEAR None reported | LOW Adjusted for Parity, plurality, age, socioeconomic status and smoking |
Engjom et al 7, 2014, Norway | LOW Data from Medical Birth Registry of Norway | LOW All births in Norway affected by the closure No of OUs closed reported | LOW Objective outcome (BBA) | LOW All units report to Medical Birth Registry | LOW Analysis method appropriate, a logistic regression model, crude and adjusted results given | UNCLEAR None reported | LOW Adjusted for maternal age, parity, education and partner status |
Grytten et al 17, 2014, Norway | LOW Data from Medical Birth Registry of Norway | LOW All births in Norway No of OUs closed reported | LOW Objective outcomes (NM, Infant mortality) | LOW All maternity units report to Medical Birth Registry | Unclear Difference-in-difference statistical method used, but reporting of findings were unclear | LOW Adjusted for trend in infant outcomes based on local hospitals that were not closed. | LOW Maternal age, immigrant status, level of education, marital status, predisposing medical factors and characteristics of the birth |
Canadian Studies | |||||||
Le Coutour et al 21, 1990, Canada | LOW Data from MED-ECHO - regional data collection system | LOW All births in catchment area No of OUs closed reported | LOW Objective outcome (CS) | UNCLEAR No information | UNCLEAR No details on data analysed method | UNCLEAR None reported | HIGH No adjustment |
Allen et al, 22 2004, Canada | LOW Data from Nova Scotia Atlee Perinatal Database | LOW All births in catchment area affected by the closure No of OUs closed reported | LOW Objective outcomes (SB, NM) | UNCLEAR Population based dataset but no information about missing data | LOW Analysis method appropriate and data from logistic regression models were reported | UNCLEAR None reported | LOW Maternal age, smoking and maternal diseases |
Hutcheon et al 20, 2017, Canada | LOW Data from British Columbia Perinatal registry | LOW All births in catchment area affect by the closure No of OUs closed reported | LOW Except for third/4th degree tears. Objective outcomes (BBA, PM, NM, ICU admission, CS, MM, blood transfusion, Maternal admission to ICU) | LOW >99% complete | LOW Used a within-community fixed-effects design and Poisson regression | LOW Using difference in difference analysis which separates the effect of the closure from underlying time trends of reported outcomes | HIGH No adjustment |
*Blondel et al 2011 included in travel distance and OU closure.
BBA, born before arrival; CS, caesarean section; ICU, intensive care unit; MM, maternal mortality; NHS, National Health Service; NM, neonatal mortality; PM, perinatal mortality; SB, stillbirth.