Development of a Core Outcome Set (COS) for treatment of depression during or after pregnancy (antenatal and postpartum depression)

Summary

Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) has enabled relevant interested parties (particularly patients, researchers, and healthcare personnel), from Sweden and other countries, to agree on which outcomes should be included in a Core Outcome Set (COS) for future research studies in the treatment of antenatal and postpartum depression. In total, the COS included nine outcomes (Table 1).

Table 1. outcomes included in the core outcome set
* Outcomes which are included in the COS, based on the results of survey 2 the remaining outcomes were included after discussions during the consensus meeting.
Self-assessed symptoms of depression, assessed with a scale that captures differences in sleep*
Diagnosis of depression as assessed by a clinician should include a structured interview
Parent to infant bonding
Self-assessed symptoms of anxiety
Quality of life
Satisfaction with the study intervention
Suicidal thoughts, attempts or completed suicide*
Thoughts of harming the baby, including thoughts of extended suicide*
Adverse events including spontaneous or induced abortion, miscarriage, fetal death and neonatal death

Background

A core outcome set (COS) is an agreed standardised set of outcomes that should be assessed and reported, as a minimum, in all clinical trials in specific areas of health or health care. The outcomes that are to be included in different COS are selected by a consensus process, in which healthcare personnel, researchers, and patients should be included. By developing and implementing COS, the aim is to enable the results from various studies to be more readily comparable and synthesised, and that the basis for decisions, for patients and healthcare personnel, will therefore be strengthened (Figure 1).

Figure 1 A schematic picture of the intended use of Core Outcome Sets

Schematic illustration of 3 research studies that explores 6–8 outomes each, whereas 4 are mutual (Core Outcome Set)

Aim

Development of a COS for future research studies in the treatment of antenatal and postpartum depression.

The primary target groups for this report are researchers and research funders. The project has been commissioned by the Swedish government as part of its efforts for the promotion of women’s health. This is the first time SBU has facilitated the development of a COS.

The area of antenatal and postpartum depression was chosen based on the results from an overview of existing and ongoing COS within the field of maternal health [1].

Method

The project was registered in the COMET initiative registry [2] and is developed according to theCOS-STAD Recommendations [3]. The project followed an a priori established protocol (Appendix 1).

The study consisted of three parts: (1) a systematic overview of the outcomes reported in systematic reviews and randomised controlled trials (RCTs) published 2018-2019, as well as in ongoing RCT on the topic, (2) a two-round Delphi survey with relevant stakeholders to prioritise the outcomes used in the studies, and (3) a consensus meeting where the final COS is decided (Figure 2). The project does not address the questions of which method should be applied to assess the outcomes, or when the outcomes should be assessed.

Figure 2 Flow chart of the project.

Flow chart of the project: Suggestions from the project and working groups and Systematic overview of outcomes used in research to List of extracted outcomes to Delphi process by online survey x 2 To top ranked outcomes from survey = Core outcome set to Consensus meeting

Identifying outcomes in research

In order to identify outcomes used in current research, a literature search for relevant randomised controlled trials (RCT) and systematic reviews published 2018–2019 as well as study protocols for RCT (no time limit) was undertaken (search strategy presented in Appendix 2).

Inclusion criteria

Population

  • Pregnant women, or their partners suffering from depression
  • New mothers or their partners suffering from depressions

Intervention

Any intervention given for depression

Control

Any type of control was accepted

Outcome

Outcomes relating to the effect of the treatment given

Study design

Randomised controlled trials or systematic reviews. Both protocols and published studies were included.

Limitations

Conference abstracts, studies investigating prevention of depression and studies investigating treatments given to persons suffering from a combination of depression and other chronic conditions such as HIV infection, were excluded.

Language

No language restriction applied to the search strategy, but the final inclusion was only for studies in English or Scandinavian languages

Search period

From 2018 to 2019 (clinicaltrials.gov no time restriction). Final search September-October 2019 (full search strategy presented in Appendix 2)

Databases searched

Medline (OVID), Cochrane library (Wiley), PsycINFO (EBSCO), Cinahl (EBSCO) and ClinicalTrials.gov

Client/patient involvement

Yes

All identified studies were screened for relevance independently by two members of the study management. Data were extracted by one review author and checked by another review author.

Following data were extracted from the studies:

  • Reference
  • Study design
  • Year for publication
  • Intervention/s used
  • Outcomes, with notification if outcomes are stated as primary or secondary
  • Toll for outcome measure
  • Time for outcome measure

After extraction, the identified outcomes were analysed and all outcomes that are unique were listed. Some of very similar outcomes in that list were brought together, for example outcomes measuring different hormonal or pharmaceutical levels were combined as under biological parameters.

Development of a Core Outcome Set

This process comprises a two-round Delphi survey and a consensus meeting. Relevant stakeholders (the working group) participated in the entire process, giving their perspectives on which outcomes are so important that they should be included in the COS. To allow international participation, all material was translated into English and information about the project was disseminated both within Sweden and internationally. In all, 222 people expressed an interest in participating, of whom 24 were international participants (Table 3). All participants belonged to one of the following four stakeholder group:

  • Patient- or relative
  • Researcher within the field
  • Health professional
  • HTA-agencies policy makers and others

In the Delphi study, all participants independently ranked the various outcomes in terms of importance and relevance, in two consecutive surveys. The information sent to the participants are presented in Appendix 3. Before they answered the second survey, they were shown how the different stakeholder groups, had ranked the outcomes in the first survey. In the first survey there was also an opportunity to suggest additional outcomes which were not identified in the actual research. The results of the surveys were analysed according to previously determined criteria, as to whether the outcome should be included in the COS. One exception from the pre-set protocol was made after the Delphi process. This to enable all outcomes included in the consensus meeting to be properly discussed. In total 23 outcomes were passed forward to the consensus meeting. Based on the results from the second survey three outcomes (marked with * in Table 1) were classified as consensus in, and thereby included in the COS. In addition, outcomes fulfilling the below specified criteria where discussed during the meeting:

  • Any outcome for which at least 70% of participants in either one of the stakeholder group scored as critically important (7–9)
  • The ten outcomes given the highest scores by each stakeholder group

The entire process concluded with a meeting at which a selection of the participants who had been in the Delphi study reached consensus about which further outcomes, over and above those which had already met the inclusion criteria, should be included in the COS. After the consensus meeting, all those who had completed the second questionnaire were given access to, and the option of making comments on, the final COS.

Results

Identifying outcomes in research

A total of 165 studies (Figure 3) were included and are described in further detail in Appendix 4. Excluded studies with reason for exclusion are presented in Appendix 5. On the basis of the included studies, a list was made of 93 different outcomes which were used in the research. As well, a further 13 outcomes were added by the project participants. Thus, in all, 105 outcomes were included in the consensus process. On average, the RCT: s contained six outcomes and the systematic reviews four outcomes (Table 2). The three most common outcomes in the included studies were self-assessed symptoms of depression, clinical diagnosis of depression and self-assessed symptoms of anxiety.

Table 2 Summary of results from the systematic overview of outcomes.

  Number of studies Antenatal depression Postpartum depression Both antenatal samt postpartum depression Number of outcomes in the studies, Median; Mean (min–max)
RCT-protocol 104 33 60 11 5; 6 (1–24)
RCT-published
(2018–2019)
27 12 12 3 5; 6 (1–19)
Systematic reviews published
(2018–2019)
34 8 16 10 2; 3 (1–13)
Total 165 53 88 24 5; 6 (1–24

 

Figure 3 Flow Chart.

Of 1772 records identified through database searching there were 165 studies included: 34 systematic reviews (2018–2019), 27 published RCTs (2018–2019) and 104 RCT protocols. Flödesschema

Development of a Core Outcome Set

A flow chart describing the different steps of the development of the COS are presented in Figure 4.

Figure 4 Flow chart of the different steps in the development of the Core outcome set.

From 945 outcomes who was identified in the studies became 9 outcomes in COS (Core Outcome Set) after a consensus meeting. Illustration

Out of the 222 people who expressed an interest to partake in the project, 151 completed survey 1 and 123 completed survey 2 (Table 3). All outcomes included in the survey and a summary of the results are presented in Table 4. A more detailed presentation of the results from survey 1 and survey 2 are available in Appendix 6 and Appendix 7).

 

Table 3 Characteristics of participants in the Delphi online survey.

* Brazil (n=1), Norway (n=1), United kingdom (n=4), Ireland (n=7), Finland (n=1), Germany (n=2), China (n=2), Nigeria (n=1), France (n=1), Holland (n=1), Portugal (n=1), India (n=1), Israel (n=1).
  Signed up for participation Completed survey 1 Completed survey 2 Participated in consensus-meeting
Researcher within the field 41 30 31 3
HTA-agencies, policy makers and others 13 11 9 2
Patient- or relative 69 43 33 4
Health professional 99 67 50 4
Total (from countries other than Sweden) 222 (24*) 151 (20) 123 (18) 13 (0)

 

After analysing the result of both surveys, three outcomes met the criteria for inclusion in the COS (marked * in Table 1). As well as these three outcomes, a further 20 were discussed during the consensus meeting.

 

Table 4 Summary of results from survey 1 and 2. Highlighted in Bold are outcomes discussed during the consensus meeting.

a Indicates that the outcome has been listed as primary in at least one study.
b The outcome has only been identified in studies on antenatal depression.
c The outcome has only been identified in studies on postpartum depression.
d The outcome has been added by the project management.
e The outcome has been added by participants in the delphi survey.
  Survey 1 Survey 2 Number of studies including this outcome (of 165)
Outcome Number of replys Median (IQR1-IQR3) Number of replys Median (IQR1-IQR3)
Depressive symptoms 149 7 (6–9) 123 8 (7–9) 133a
Depression in partner 144 6 (4–7) 123 5 (4–6,5) 1
Psychiatric diagnosis of depression 147 7 (6–8) 119 8 (6–9) 57a
Recovery from depression (remission) 146 7 (6–8) 121 7 (5–8) 17a
Depression relapse 144 7 (6–8) 121 7 (6–8) 4a
Time to remission from depression 143 6 (5–8) 122 6 (5–7) 1a,c
Was depression treatment initiated 145 7 (6–9) 122 7 (6–9) 6a
Change in antidepressant medication 141 6 (4–7) 120 6 (4–7) 3
Expert rating of overall disease burden 137 6 (4–8) 118 6 (4–7) 17a
Anxiety symptoms 141 7 (6–8) 119 7 (5–8) 51a
Fear of childbirth 142 6 (4,25–8) 119 6 (4,5–8) 0d
Psychiatric diagnosis of anxiety 141 7 (5–8) 118 6 (5–7) 8a
Other psychiatric or neuropsychiatric diagnoses 138 6 (5–7) 117 6 (5–8) 2b
Mood 139 6 (5–8) 119 6 (5–8) 5
Self-harm 141 7 (6–8) 118 7 (6–9) 0d
Suicidal thoughts or attempts  or completed suicide 140 9 (7–9) 114 9 (8–9) 10a
Cognitive ability 143 6 (4–7) 117 5 (4–7) 4
Eating disorder symptoms 138 6 (4–7) 118 5 (4–6) 2
Visit to clinic for fear of childbirth 141 6 (4–8) 118 5 (4–7) 0d
Post-traumatic stress symptoms 141 7 (6–8) 119 7 (4,5–8) 1b
Psychiatric diagnosis of postpartum stress syndrome in relation with childbirth 141 7 (5–9) 117 7 (5–8) 0d
Psychiatric diagnosis of postpartum psychosis 138 7 (6–9) 115 7 (6–9) 0d
Domestic/ intimate partner violence 136 8 (5–9) 118 7 (6–9) 1c
Family functioning 137 6 (5–8) 119 6 (5–7) 4
Quality of relationship 139 5 (4–7) 119 5 (4–6) 9a
Parent to infant bonding 140 7 (6–9) 119 7 (6–8) 18
Attachment 139 8 (6–9) 118 7 (5,25–8) 2a,c
Parenting Sense of Competency 139 6 (5–8) 119 6 (5–8) 13a
Parental stress 139 6 (5–7) 118 6 (5–7) 15a
General stress 139 6 (4–7) 118 6 (4–6) 5
Parent-infant interaction 138 7 (6–9) 116 7 (5–8) 11a
Attitudes about parenting 137 6 (4–7) 117 5 (3–6) 4
Family planning 141 5 (4–7) 118 4 (3–6) 1b
Daily functioning level/ activities of daily living 137 6 (5–8) 119 6 (5–7) 23
Sleep 138 7 (6–9) 119 7 (5,5–8) 8a
Fatigue 137 6 (5–8) 119 6 (5–8) 2c
Quality of life 140 6 (5–8) 117 6 (5–8) 26a
Sexual desire/function 135 5 (3–6) 119 4 (3–6) 1c
Stigma 138 6 (4–8) 117 6 (4–7) 2b
Social relationships 138 6 (5–7) 119 6 (4–7) 9
Perceived ability to cope with life challenges 136 6 (4,75–7) 117 6 (4–7) 1c
Effort-based decision-making 135 6 (4–7) 118 5 (4–6) 4b
Self-compassion 135 5 (4–7) 117 5 (3–6) 1b
Lifestyle 134 6 (4–7) 116 5 (4–6) 1b
Verbal fluency 122 3 (2–4) 108 3 (2–5) 1
Personality 128 4 (3–6) 112 3 (3–5) 2
Perceived social support / Help-Seeking 137 7 (5–8) 117 6 (5–7) 12
Socio-emotional competence 131 6 (4–7) 117 5 (4–6) 4
Confidence to self-help 133 6 (4–7) 116 5 (3–7) 3
Mentalizing, reflective functioning and level of mindfulness 133 5 (4–7) 116 5 (3–6,25) 5a,c
Self-esteem 135 6 (4–8) 117 6 (4–7) 3
Adverse event 133 7 (5–8) 116 6 (5–8) 29a
Spontaneous abortion or stillbirth 127 7 (6–9) 112 8 (5–9) 1b
Perinatal mortality 126 8 (6–9) 111 8 (6–9) 1b
Infections during pregnancy 119 5 (3–6) 108 4 (3–6) 1b
Residual physical problems after delivery 133 7 (6–9) 115 7 (5–8) 4a,b
Birth injury 130 6 (5–9) 115 7 (5–8) 5
Pain 131 6 (4–7) 115 6 (4–8) 1b
Gestational hypertension or preeclampsia 124 6 (4–7) 103 5 (4–7) 3b
Gestational age at delivery 130 6 (5–8) 111 6 (4–7) 5b
Gestational diabetes 121 5 (3–7) 107 5 (3–7) 1b
Estimated blood loss 121 6 (4–7) 105 5 (3–7) 2b
Induction of labour 126 6 (4–7) 110 5 (4–7) 1b
Mode of delivery 132 6 (5–8) 114 6 (4–8) 6a
Fetal growth assessment by ultrasound 124 5 (4–7) 108 5 (3–6) 1b
Complications associated with the placenta 117 5 (3–7) 108 5 (3–7) 1b
Vaginal and/or faecal microbiota 107 3 (3–6) 102 3 (3–5) 1b
Transition of care from pregnancy to postpartum 133 7 (6–9) 115 7 (5–9) 1a,b
Breastfeeding 131 6 (4–8) 114 6 (4–7) 8
Child attention 126 5 (4–7) 109 5 (4–6) 3c
Duration of crying episodes 127 6 (4–7) 109 5 (3–6) 1b
Child development 133 7 (5–8) 112 6 (5–8) 19a
Neurodevelopmental and neurobehavioral outcomes 128 7 (5,75–8) 111 7 (5–8) 2a,c
Infant sleep 131 6 (5–7) 112 6 (4–8) 1b
Vaccinations received/ Immunization status 123 4 (2–6) 109 3 (2–6) 4
Infant or child’s temperament 132 5 (4–7) 112 4 (3–6) 6a
Biological parameters in the infant 110 5 (3–7) 96 4 (3–6) 6
Well-baby check-up status 132 6 (4–7) 111 5 (3–7) 2
Congenital malformations 125 6 (4–8) 104 6 (4–7) 1b
Infant antibiotic use 117 4 (3–6) 100 3 (3–5,25) 1b
Birth weight, height and head circumference 126 5,5 (4–6) 106 5 (3–6) 8a
Small for gestational age (SGA) 123 6 (4–7) 105 5 (4–7) 3b
Apgar score 123 5 (4–7) 106 5 (3–7) 4a
Neonatal intensive care unit admission 126 6 (5–8) 106 6 (5–8) 4
Cost-effectiveness 130 6 (4–7) 106 5 (4–7) 9a
Use of health services 131 6 (4–7) 110 5 (4–6) 16
Engagement and retention with health services 133 6 (4–7) 109 5 (4–6) 3
Sick-leave 133 6 (4–7) 113 6 (4–7) 3
Satisfaction with intervention 136 7 (6–8) 114 7 (6–8) 31a
Therapeutic alliance 135 7 (6–9) 113 7 (6–9) 2
Quality of care 136 6 (5–8) 114 7 (5–8) 5
Parent experience 133 7 (6–8) 114 7 (5–8) 2a
Pregnancy/ delivery experience 135 7 (5–9) 114 7 (6–9) 2a,b
Biological parameters 112 5 (4–7) 99 4 (3–6) 15a
Maternal vital signs/ physical examination 116 4 (3–6) 101 4 (3–6) 6a
Knowledge about anti-depressant treatment in pregnancy 128 5 (4–7) 110 5 (4–7) 2b
Partner’s attendance 135 6 (5–7) 114 6 (4–7) 2
Coping strategies among relatives 126 5,5 (4–7) 112 5 (4–6,25) 1b
Separation or divorce - - 112 5 (3–7) 0e
Ability to return to work/ change of work - - 113 5 (4–7) 0e
Obsessive thoughts or intrusive thoughts - - 110 6 (4–7) 0e
Fear of something bad happening to the baby - - 113 6 (4–8) 0e
Thoughts of harming the baby, (including thoughts of extended suicide) - - 113 8 (7–9) 0e
Level of compliance with the given treatment - - 112 7 (6–8) 0e
Mental health literacy - - 113 6 (4–8) 0e

 

Thirteen people participated in the consensus meeting: all four stakeholder groups were represented. During the meeting the group agreed on inclusion of a further six outcomes in the COS and also to amend the formulation of some of the included outcomes. No further comments were received from other participants with respect to the COS which was developed during the consensus meeting. The outcomes which are included in the final COS are described in Table 1.

Discussion

The aim is that the outcomes included in this COS are to be measured in future studies and systematic reviews of treatment of antenatal and postnatal depression (intervention studies). If this is achieved it will result in improved potential to synthesise the results from different studies. In the long term this will lead to a stronger evidence. This can be important for some of the outcomes which occur less frequently, for example suicide, and where a large patient material is necessary in order to discern differences between various treatment methods. Most of the outcomes in the agreed COS are currently used in research (Figure 5). However, only one of the selected outcomes, self-assessed symptoms of depression, is used in more than 50% of the identified studies. The second most commonly used outcome, diagnosis of depression as assessed by a clinician only occurred in 33%.

Figure 5 Frequency of the COS included outcomes in the identified studies.

Number of outcomes who is included in COS: Self-assessed symptoms of depression, Diagnosis of depression as assessed by a clinician, Parent to infant bonding, Self-assessed symptoms of anxiety, Quality of life, Satisfaction with the study intervention, Suicidal thoughts, attempts or completed suicide, Thoughts of harming the baby, including thoughts of extended suicide, Adverse events

In the protocol for the present study the maximum number of outcomes to be included in the COS had already been limited to ten. We believe that a greater number is not practical. After going thru the actual studies in this field, we can state that generally they contain around six outcomes, i.e. fewer than the number included in this COS. Thus, despite our ambition to try to limit the number of outcomes, the nine outcomes included in this COS, may be to many. This might lead to difficulties for researchers to include further outcomes specific to the research question. This project has not addressed how and when these outcomes should be measured, directives that should be included in a comprehensive COS. We intend, however, to pursue these questions in a future project.

This COS included an outcome which has not been identified in any study, namely ”thoughts about harming the child”.  This highlights the need to involve patients, healthcare personnel and researchers in the process of selecting which outcome to include in a COS, and also to encourage the project participants to suggest outcomes over and above those which are already used in research.

SBU would like to thank all those who have been involved in, or assisted in dissemination of information about, the project, especially all those who participated in the development of the COS.

Project group

Experts

  • Maria Jonsson, Associate professor and senior consultant at the Department of Obstetrics and Gynecology, Uppsala university hospital, Sweden.
  • Alkistis Skalkidou, Professor of Obstetrics and Gynecology, Dept. of Women’s and Children’s Health, Uppsala University and senior consultant at the Gynecologic Department, Uppsala university hospital
  • Frida Trönnberg, Patient representative

SBU

  • Christel Hellberg (Project Manager)
  • Marie Österberg (Project Manager)
  • Ann-Kristine Jonsson (Information Specialist)
  • Sara Fundell (Project Administrator)

External reviewers

  • Erik Forsell, PhD, Research and Development-Psychologist
  • Ann Josefsson, Professor, Senior consultant in Obstetrics and Gynecology
  • Sverker Svensjö, PhD, Senior Consultant, Surgery

References

  1. SBU. Core outcome sets inom förlossningsvård. Sammanställning och analys av studier. Stockholm: Statens beredning för medicinsk och social utvärdering (SBU); 2020. SBU-rapport nr 309. ISBN 978-91-88437-51-8.
  2. COMET Initiative. COMET Database. Treatment of perinatal depression: protocol for a systematic review of outcomes in the literature and identification of a core outcome set using a Delphi survey. [cited 2020 April 1]. Available from: http://www.comet-initiative.org/Studies/Details/1421.
  3. Kirkham JJ, Davis K, Altman DG, Blazeby JM, Clarke M, Tunis S, et al. Core Outcome Set-STAndards for Development: The COS-STAD recommendations. PLOS Medicine 2017;14:e1002447.
Download summary

SBU Policy support identifies and presents available scientific evidence to support policy and decision making, including the development of national guidelines, at other government agencies. In consultation with professional experts, SBU staff generates supporting documentation to address the various questions that have been posed.

Published: 6/8/2020
Report no: 314
Registration no: SBU 2019/290
https://www.sbu.se/314e