Appendix CSelection criteria and validity scores for included and excluded studies for the systematic review comparing planned home birth and planned hospital birth and the systematic review comparing planned standalone midwife-led unit and obstetric unit birth

Publication Details

Method used for this review

The difficulty of conducting a randomised controlled trial (RCT) to evaluate effectiveness and safety of planned home birth and planned standalone midwife-led unit birth, compared with planned obstetric unit birth, is evident from the literature. The paucity of good-quality evidence necessitated the inclusion of studies using a range of methodologies as described in Chapter 1 (methodology section). The details of the search strategies employed are provided on the accompanying CD-ROM.

Inclusion/exclusion criteria

The best study design to address the effectiveness of an intervention is an RCT. However, there is a higher incidence/prevalence of benefits than adverse events, especially serious outcomes, in many clinical contexts, and therefore an RCT will not necessarily be the best method to use to demonstrate safety of an intervention.

Studies that employed an adequate randomised design were regarded as having the highest validity [++]. Any study that considered planned birth populations with an additional adequate study design which controlled for background medical and/or obstetric risks between different places of birth and/or reported relevant outcomes was also included and assigned as having acceptable internal validity [+]. Any study that did not report relevant outcomes or that did not meet the criteria above was considered invalid and excluded ([−]). Use of regression analysis, matched control design, and/or any other means to control the risks of these two groups was regarded as relevant.

Women who planned birth at a place outside hospital settings (e.g. home birth and/or standalone midwife-led unit birth) but had adverse outcomes were more likely to have been transferred to hospital before birth and therefore be considered as an obstetric unit birth. To compensate for this required that any observational study comparing clinical outcomes between births outside and within obstetric units should consider women who planned birth outside an obstetric unit with those who planned hospital birth. Controlling for risk factors of these two groups is critical.

Transfer rates were obtained from any study where the validity was regarded as [++] or [+].

Applicability to UK setting

Any study conducted in the UK since 1980 was regarded as having the highest applicability to the current UK setting. Clinical practice in the UK was considered to have been significantly different before 1980. Any study conducted in high income countries since 1980 was also considered valid, and therefore included if there was no UK study available or when the included UK studies could not provide enough information to make a conclusion.

Outcome measures

Another important factor is the availability of appropriate outcomes.

For effectiveness, relevant outcomes included mode of birth, incidence of obstetric interventions and any other relevant clinical outcomes as defined in the guideline questions.

Primary outcomes for safety were defined as intrapartum-related perinatal mortality (IPPM), and maternal mortality.

IPPM, defined as death from intrapartum ‘asphyxia’, ‘anoxia’ or ‘trauma’, (Wigglesworth classification 3),600 was considered to be the most important outcome to assess safety of place of birth. IPPM includes stillbirths and death in the first week but excludes deaths of low birthweight infants or as a result of multiple abnormalities. If there was no relevant single study that reported IPPM, perinatal mortality was used. Similarly, maternal mortality was considered the most important outcome to assess safety of place of birth for mothers (women). If no relevant single study reported this outcome, other important maternal morbidities such as incidence of postpartum haemorrhage (PPH) were reported.

Where there is no single study reported IPPM and/or maternal mortality, perinatal mortality, neonatal morbidities and maternal morbidities were considered as proxy and hence reported.

Note: For further details of included studies, please refer to the evidence tables in the accompanying CD-ROM.

Planned home versus hospital birth

Table C.1. Included studies (planned home birth versus planned hospital birth).

Table C.1

Included studies (planned home birth versus planned hospital birth).

Table C.2. Excluded studies (planned home birth versus planned hospital birth).

Table C.2

Excluded studies (planned home birth versus planned hospital birth).

Planned standalone midwife-led unit versus obstetric unit birth

Table C.3. Included studies (planned standalone midwife-led versus obstetric unit birth).

Table C.3

Included studies (planned standalone midwife-led versus obstetric unit birth).

Table C.4. Excluded studies (planned standalone midwife-led versus obstetric unit birth).

Table C.4

Excluded studies (planned standalone midwife-led versus obstetric unit birth).