Group singing for postnatal depression (Music and Motherhood)

1. Introduction

Version followed: Breathe Melodies for Mums (M4M), the 10-week mother-and-baby community singing programme run by Breathe Arts Health Research. The delivery format follows the SHAPER-PND trial protocol (Estevao et al., 2021). Session content follows the process evaluation of the original Music and Motherhood trial (Fancourt & Perkins, Royal College of Music, 2015–16). Where the two differ, the difference is named in the text. [source: Estevao et al. 2021 protocol; Fancourt & Perkins process evaluation; Breathe research page]

What it is. This is a community arts-in-health programme. It is not psychotherapy, and it is not music therapy. Groups of mothers with postnatal depressive symptoms come with their babies to weekly one-hour singing workshops in children's centres or other community venues, or on Zoom. Professional workshop leaders trained by Breathe run the sessions. Mothers learn songs from around the world, lullabies and simple vocal and instrumental activities, and they write songs about their babies and motherhood. Nobody interprets the music, processes emotions clinically or gives a diagnosis. [source: Estevao et al. 2021; Fancourt & Perkins process evaluation]

How change is thought to happen. In the 2018 trial, mothers with moderate–severe symptoms (EPDS 13 or above) improved faster with singing than with usual care. A creative-play group run by the same leaders in the same format did not differ from usual care. The authors read this as a sign that features specific to singing mattered, beyond social support alone. [source: Fancourt & Perkins 2018 BJPsych accepted version] Mothers' own accounts named five features of the singing groups:

• a social, authentic and multicultural creative experience;
• a way to calm their babies;
• immersive "me time";
• a sense of achievement and identity;
• a stronger bond with the baby.

[source: Perkins, Yorke & Fancourt 2018, abstract] The SHAPER-PND protocol also proposes biological pathways (stress-hormone/HPA axis, inflammation, oxytocin) and psychosocial ones (mother–infant interaction, social support, loneliness). It describes the oxytocin evidence as conflicting. [source: Estevao et al. 2021]

Evidence status, as the sources give it.

• 2018 trial (n=134 analysed). At the EPDS≥10 level there was no significant group difference. In the EPDS≥13 sensitivity analysis, singing improved faster than usual care by week 6, but it did not differ significantly from play, and the gap had narrowed by week 10. The authors call for replication in larger samples. [source: Fancourt & Perkins 2018]
• SHAPER-PND RCT (199 mothers, published October 2025). Both arms improved by week 10. Only the singing arm showed significant reductions at weeks 20 and 36. Dropout at week 10 was 23% with singing and 43% in the signposted control. The cost was described as within NICE thresholds. [source: KCL news release; Breathe research page] The full 2025 paper is not drawn on, so effect sizes and the detailed findings on implementation and adverse events are [uncertain] here.
• How the programme presents itself. Breathe describes the programme as "an additional option" for women already getting help, not a replacement for psychological or medical treatment. [source: KCL news release]

Who delivers it, and what training is required. Leaders are professional workshop leaders with experience in community, hospital and school settings, mostly in music or mixed arts. Breathe trains them. In the original trial, training covered an introduction to the research, an overview of PND, guidance on sensitive situations, and safeguarding training from a perinatal health visitor. All staff had enhanced background checks (DBS) and signed a protection policy for vulnerable adults and children. [source: Fancourt & Perkins process evaluation; Estevao et al. 2021] No regulated professional title or registration that this leader role requires is identified [uncertain]. Delivering the programme under its name, with its research-backed format, means going through Breathe, whose research page lists a commissioning route. [source: Breathe research and news pages] UK music therapy is a separate, HCPC-regulated profession, and the British Association for Music Therapy is its professional body [synthesis]. A psychotherapist should not present this group as music therapy, and should not run it as a therapy group. The psychotherapist's role is to refer, to keep holding clinical care and risk, and to work alongside the group. [synthesis]

2. Framework

This is a fixed-length programme, not a staged therapy. The outline below goes from referral to ending.

Stage A — Referral and screening (before week 1).

• Goal: find mothers with postnatal depressive symptoms and screen out women who need more urgent care.
• SHAPER-PND procedure:
  • The mother completes an online prescreen that includes the EPDS.
  • A score of 10 or above leads to a screening visit at home or on Zoom. Below 10, she is told she is not eligible now and is signposted to community support such as talking therapies and mother–baby groups.
  • Eligibility: mother aged 18 or over, baby aged 0–9 months, satisfactory English, internet access, able to consent.
• 2018 trial differences: babies up to 40 weeks; EPDS≥10; women scoring 8–9 were offered one rescreen.
• Risk procedure (2018): for a high EPDS score or any response on the self-harm item, staff spoke with the mother, reminded her of services and wrote to her GP.
• Not found: neither source describes formal exclusion of psychosis, severe depression or active suicidality.

[source: Estevao et al. 2021; Fancourt & Perkins process evaluation] If you refer in, add your own perinatal screening first (Section 4.1). [synthesis]

Stage B — The 10-week block.

• Format:
  • 10 weekly 60-minute sessions.
  • 8–12 mothers with their babies (SHAPER-PND), or 10–13 (2018).
  • The same group and the same leader for the whole block.
  • Mothers sit in a circle on the floor on cushions and mats, with babies within reach.
  • Sessions run in person or on Zoom.
  • In 2018 the time slots were early afternoon on weekdays.
• Staffing: a lead facilitator plus one or two assistants. A separate coordinator tracks attendance, phones mothers who miss a session, checks on wellbeing, monitors risk and reports to health professionals when needed.

[source: Estevao et al. 2021; Fancourt & Perkins process evaluation]

How the block progresses.

• The weekly structure stays largely the same: a welcome song, then a mix of relaxing and energetic singing and instrument activities.
• Longer songs are learned gradually over the weeks.
• Leaders deliberately step back over time so that mothers take ownership.
• Content evolves as the babies grow.
• Co-written songs and recordings build up over the block.

[source: Fancourt & Perkins process evaluation] The sources give no fixed week-by-week curriculum [uncertain].

Stage C — Ending and afterwards.

• Group recordings go home with the mothers for listening.
• In 2018, the process evaluation recommended adding social coffee time and more formal take-home resources.
• SHAPER-PND followed mothers up at weeks 20 and 36.
• Breathe now offers ten-week programmes both in person and online.

[source: Fancourt & Perkins process evaluation; Estevao et al. 2021; Breathe research page]

Criteria for moving on. The sources set no clinical criteria for moving between stages. Completing the 10 weeks is the endpoint. The decision rules below are [synthesis]:

• Pause or step up: if a mother's risk rises (for example new suicidal thoughts, psychotic symptoms, or concern for the baby), escalate her to clinical care. She may continue singing alongside that care, but the group is never the safety plan.
• After the block: if EPDS stays at or above the referral level, review her clinical care. Do not simply repeat the block.

3. Guiding principles

• Free, local, non-clinical access. Sessions are free and held in familiar community venues. Practical meaning: refer women who would not attend a clinic. Present the group as a class, not a treatment session. [source: Estevao et al. 2021; process evaluation]
• Mother and baby together. The programme works with mother and baby as a pair. Practical meaning: babies are always in the room and part of every activity, and the programme aims to improve mother–infant interaction as well as mood. [source: Estevao et al. 2021]
• Singing itself is the active ingredient. Practical meaning: keep the session about singing and making music. Do not let it drift into a general coffee group or craft play, which did not outperform usual care. [source: Fancourt & Perkins 2018]
• Authentic, social, multicultural experience. Practical meaning: use songs from around the world and real adult singing, not only nursery rhymes. [source: Perkins et al. 2018 abstract; process evaluation]
• Calm the baby, give the mother time for herself. Practical meaning: choose activities that soothe infants, so the mother can be absorbed in the music for a while. [source: Perkins et al. 2018 abstract]
• Achievement and identity. Practical meaning: songs build over the weeks and mothers co-write their own, so they leave with something they made. [source: Perkins et al. 2018 abstract; process evaluation]
• Continuity. Practical meaning: keep the same leader, the same group, the same slot and a recurring welcome song. Fidelity reporting tracked substitute cover when a leader was ill. [source: Estevao et al. 2021; process evaluation]
• Gradual handover of ownership. Practical meaning: the leader does less over the block and the mothers lead more. [source: process evaluation]
• Separate roles; risk sits with the coordinator and clinicians. Practical meaning: keep the leader, assistant and coordinator roles distinct. The coordinator monitors risk and liaises with health professionals, and the leader is not the clinician. [source: process evaluation]
• Care for the facilitators. Practical meaning: leaders reported the emotional load of working with struggling mothers. Hold a debrief after every session. [source: process evaluation]
• Comfortable, spacious venue. Practical meaning: a cramped room with prams left outside and no space to step out of the circle is a barrier. Choose rooms that have space and staff on hand. [source: process evaluation]
• No judgement of voice or skill. Practical meaning: musical ability is never assessed, no one sings alone unless she chooses to, and you sing along yourself. [synthesis]
• Arts activity, not therapy. Practical meaning: leaders do not interpret what mothers disclose or process trauma in the group. Disclosures go to the coordinator and the mother's clinicians. [synthesis, consistent with Breathe's "additional option" framing]

4. Intervention inventory

The programme sources describe the session activities in 4.3–4.13 and the staffing and support procedures in 4.14–4.17. The perinatal safety steps (4.1) and the psychotherapist's parallel role (4.18) combine the sources with NICE CG192 and are marked accordingly. Example language is illustrative wording written for this guide throughout. No scripts are published.

4.1 Perinatal safety screening and referral into the group

• Purpose / when. Before a mother joins, and again whenever her state changes. The singing group suits persistent subthreshold or mild–moderate postnatal depressive symptoms. It is not a route for severe illness or acute risk. [source: Estevao et al. 2021 for eligibility; NICE CG192 for the risk steps; the combination is synthesis]
• Steps.
  1. Measure symptoms. Use the EPDS or PHQ-9, as NICE recommends for assessment and for monitoring through the first postnatal year. The programme's own entry threshold is EPDS 10 or above. Describe licensed scales by name only and do not reproduce their items. [source: NICE CG192 1.5.5, 1.5.8; Estevao et al. 2021]
  2. Screen for severe illness. Ask about past or present severe mental illness (bipolar disorder, psychosis, severe depression), past specialist or inpatient care, and severe perinatal illness in a first-degree relative. A current or past severe illness, or a suspicion of one, means referral to secondary services, preferably the specialist perinatal mental health team, with the GP informed. The group can at most be an addition to that care. [source: NICE CG192 1.5.9–1.5.10]
  3. Screen for postpartum psychosis. Sudden-onset symptoms that suggest postpartum psychosis need referral for immediate assessment, within 4 hours, by secondary or specialist perinatal services. Do not wait for a group start date. For women at higher risk (a personal history of severe illness or a family history of severe perinatal illness), monitor closely in the first 2 weeks after birth. [source: NICE CG192 1.5.11–1.5.12]
  4. Assess suicide and self-harm risk. Any self-harm item response on the EPDS, or any disclosure, calls for a full risk assessment. Cover suicidal thoughts and intent, self-neglect, self-harm, risks to others including the baby, substance use, smoking and domestic abuse, and involve the partner or family if she agrees. Match the help to the level of risk, inform all relevant professionals including the GP, and agree what she and her family should do if things get worse. Where lethal-means counselling is needed, limit it to safe storage and removal. [source: NICE CG192 1.6.3, 1.6.5; synthesis for means counselling] The programme's own minimum when risk shows up at screening was to talk with the mother, remind her of services and write to her GP. [source: process evaluation]
  5. Assess the baby's safety and the relationship. Ask about the mother–baby relationship, domestic violence and abuse, and other caring responsibilities. If you suspect maltreatment, follow local safeguarding procedures. [source: NICE CG192 1.6.1, 1.6.4]
  6. Name the fear of disclosure. Women may hold back because of stigma or fear that the baby will be taken into care, so raise this directly. [source: NICE CG192 1.5.1]
  7. Agree consent to share. With her agreement, give the group coordinator only what they need: the name of a contact clinician, and how to reach you or the GP if concerns come up. [synthesis]
  8. Keep clinical care running. For moderate or severe depression, NICE points to high-intensity psychological therapy and/or medication. Present the group as an addition to these. [source: NICE CG192 1.8.3; KCL release]
• Example language (illustrative). "This is a singing class for mums and babies, not a therapy group. It runs alongside what we're doing here. Before I refer you, I'd like to check a few things to make sure it's the right fit right now." "Some mums have frightening thoughts after birth, about harming themselves, or thoughts that don't feel like their own. It's common to be scared to say this. Have you had anything like that?"
• Watch for.
  • Recent sudden change in mental state, confusion, perplexity, unusual beliefs or not sleeping: treat as urgent. [synthesis; NICE gives the 4-hour rule]
  • Rising EPDS scores, hopelessness, or new thoughts of self-harm.
  • Persistent detachment from or hostility towards the baby.
  • Distressing intrusive thoughts of harm to the baby. These need careful assessment: they are often ego-dystonic, as in perinatal OCD, and are not by themselves a safeguarding issue. [synthesis]

4.2 Room and session set-up (including infant and hearing safety)

• Purpose / when. Every session. A comfortable, spacious room supports engagement, and cramped rooms were a recorded barrier. [source: process evaluation]
• Steps.
  1. Lay out the circle. Put soft cushions and mats on the floor in a circle, keep babies within the mother's reach, leave room for prams inside the space and a clear exit route from the circle. [source: process evaluation; Estevao et al. 2021]
  2. Prepare instruments. Lay out simple instruments: maracas, drums and hand chimes. In SHAPER-PND (pandemic era) instruments were not shared and were disinfected before and after each session. [source: Estevao et al. 2021]
  3. Infant safety. [synthesis]
     • Check every instrument for small or loose parts and cracked shakers that could spill fill; anything that fits in a baby's mouth is a choking risk.
     • Keep hot drinks out of the circle.
     • A baby who falls asleep should be moved to a flat, firm, clear surface on their back, or held. They should not be left to sleep in soft cushions.
     • Keep the floor clear for crawlers.
     • Feeding and nappy changes can happen at any time.
     • The assistant knows where the first-aid kit is. In 2018 the coordinator had enhanced infant first-aid training. [source for the training: process evaluation]
  4. Hearing safety. [synthesis]
     • Keep drums and percussion soft. Do not hold loud instruments, amplified sound or shouting close to infants' ears.
     • Play guitar or ukulele unamplified.
     • In online sessions, ask mothers to keep the device volume moderate and the speaker away from the baby's head.
  5. Safeguarding basics. Staff hold enhanced background checks, a signed protection policy, a known safeguarding lead, and a coordinator who can be reached during sessions. [source: process evaluation]
• Example language (illustrative). "Come in whenever you arrive. Feed, change or walk about whenever your baby needs to; nothing here is a performance."
• Watch for.
  • Mothers hovering at the edge or leaving prams at the door.
  • Overheating rooms.
  • Babies left unattended on raised surfaces.

4.3 Welcome song

• Purpose / when. Opens every session. The same song each week introduces the mothers and babies to one another and marks the start. [source: process evaluation; Estevao et al. 2021]
• Steps.
  1. Teach a short, simple welcome song with a slot for each pair's names.
  2. Go round the circle, singing to each mother and baby by name.
  3. Repeat it every week so it becomes familiar.
  4. Over the weeks, invite mothers to lead it. [details synthesis]
• Example language (illustrative). "We'll sing hello to each baby and mum. Just join in with the tune; the words will come."
• Watch for.
  • Mothers who avoid being named.
  • Whether mothers start singing the song unprompted, a sign of familiarity.
  • A mother who is visibly flat or absent during the opening; the coordinator may follow up. [synthesis]

4.4 Songs from around the world

• Purpose / when. The core repertoire. It gives an authentic, social and multicultural creative experience that goes beyond nursery songs. [source: process evaluation; Perkins et al. 2018 abstract]
• Steps.
  1. Choose short songs from different cultures and languages, with simple, repetitive melodies.
  2. The leader sings the song first while mothers listen. In 2018, mothers listened to songs sung by the leader before learning and singing them with their babies. [source: Fancourt & Perkins 2018]
  3. Teach it line by line by call and response, then sing it together. Teaching by ear is assumed here; the sources do not specify the method [synthesis].
  4. Invite mothers to bring songs from their own languages and families.
  5. Check the meaning of a song's text before using it. [synthesis for steps 4–5]
• Example language (illustrative). "This is a lullaby from Ghana. I'll sing it once, then you echo me a line at a time." "Does anyone have a song from home they'd like to teach us?"
• Watch for.
  • Musical trauma triggers. A song tied to a bereavement, an abusive caregiver, a difficult birth or a faith a mother has left can bring up strong distress.
  • Signs of distress: going quiet, tears, or leaving the circle.
  • If a song triggers distress, drop it without singling her out. [synthesis]

4.5 Vocal play with "motherese"-style sounds

• Purpose / when. Short vocal exercises using infant-directed noises and sound effects. They warm up voices and engage babies. [source: process evaluation]
• Steps.
  1. Lead short sound games: sliding pitches, animal or vehicle sounds, exaggerated vowels.
  2. Ask mothers to make the sounds face to face with their baby.
  3. Pause so the baby can respond. [steps synthesis]
• Example language (illustrative). "Let's slide our voices up like a lift… and down. Try it looking right at your baby."
• Watch for.
  • Babies orienting to the mother's face and the mother noticing it, a sign of engagement. [synthesis]
  • Self-consciousness about "silly" sounds. Lead by doing it yourself.

4.6 Sound bath (sustained-note relaxation)

• Purpose / when. A relaxation technique in which the mothers hold a sustained note together. [source: process evaluation] Use it in the calmer part of a session. [synthesis]
• Steps.
  1. Settle the group, with babies held or lying close.
  2. Give a comfortable starting pitch.
  3. Everyone hums or sings a long note. Each mother breathes when she needs to and comes back in, so the sound keeps going.
  4. Optionally shift to a second note to make a simple chord.
  5. Fade out together and sit in the quiet for a moment. [steps synthesis]
• Example language (illustrative). "Find a hum that feels easy. Breathe whenever you need to and slip back in; the group keeps the sound going."
• Watch for.
  • Breath-focused or eyes-closed work can bring up panic or dissociation in some trauma survivors. Make eyes-open and simply listening explicit options. [synthesis]
  • Signs it is working: settled babies and softened posture.

4.7 Lullabies in simple harmony or rounds, with touch

• Purpose / when. Calm songs sung in basic harmonies or rounds, with the mother hugging or stroking her baby. They calm babies and give mothers absorbed time and a sense of achievement. [source: process evaluation; Perkins et al. 2018 abstract]
• Steps.
  1. Teach the melody in unison first.
  2. Add a simple second part or split the group into a round.
  3. Invite mothers to rock, hug or stroke their babies as they sing.
  4. Close softly. [steps 1–2, 4 synthesis] Touch here is between mother and her own baby only. Leaders do not handle babies without the mother's explicit invitation. [synthesis]
• Example language (illustrative). "Hold your baby however feels cosy, and sway as we sing. Half of you start; the other half come in when I nod."
• Watch for.
  • Lullabies can bring back a mother's memories of her own childhood and caregivers, both warm and painful. [synthesis]
  • A mother who holds her baby stiffly or avoids touch: note it gently and share with clinicians if concerning. Do not interpret it in the group. [synthesis]

4.8 Energetic songs with movement (standing and bouncing)

• Purpose / when. Livelier songs in which mothers stand and bounce their babies. They balance the relaxing activities and lift energy. [source: process evaluation]
• Steps.
  1. Alternate these songs with calm ones within the session.
  2. Invite mothers to stand if they wish, holding their baby securely.
  3. Move gently in time with the music.
  4. Offer a seated option. [steps 1, 4 synthesis]
  5. Infant safety [synthesis]:
     • Support the head and neck of young infants.
     • Bounce gently, never shake.
     • Mind the space between standing pairs.
     • Mothers recovering from a caesarean or birth injury may stay seated.
• Example language (illustrative). "Stand up if you'd like, or stay where you are. Little bounces to the beat, keeping baby snug against you."
• Watch for.
  • Laughter, eye contact and babies' excitement are good signs.
  • Mothers wincing or avoiding standing may be in physical pain.
  • Overstimulated babies who start crying need a calmer song next. [synthesis]

4.9 Longer songs learned over several weeks

• Purpose / when. Some songs are built up gradually across the block. This gives continuity and a sense of achievement. [source: process evaluation; Perkins et al. 2018 abstract]
• Steps.
  1. Introduce one section per week.
  2. Recap the earlier sections each time.
  3. Add harmony or actions once the melody is secure.
  4. Aim to perform or record the whole song near the end. [steps synthesis]
• Example language (illustrative). "Last week we learned the first verse. Let's sing it again, then add the next part."
• Watch for.
  • Mothers who missed sessions may feel lost. Give a quick recap and recordings (4.12).
  • Growing confidence and volume across weeks. [synthesis]

4.10 Simple instruments with babies

• Purpose / when. Mothers and babies play maracas, drums or hand chimes together. The leader accompanies some songs on guitar or ukulele. [source: Estevao et al. 2021; process evaluation]
• Steps.
  1. Offer an instrument to each pair.
  2. Show how to help the baby hold it or tap it.
  3. Play along to a familiar song.
  4. Collect the instruments afterwards, and clean them if hygiene requires. [steps 1–2 synthesis]
• Example language (illustrative). "Pop the shaker in baby's hand and shake together, or shake it gently near them so they can watch."
• Watch for.
  • Mouthing, loose parts and volume (see 4.2).
  • Comparison between mothers about whose baby "does" more. Praise each pair's own response. [synthesis]

4.11 Co-writing songs about babies and motherhood

• Purpose / when. Mothers co-write lyrics about their babies or motherhood and set them to simple melodies. In 2018 this was described as creating new songs together about motherhood. It builds achievement, identity and ownership. [source: Fancourt & Perkins 2018; process evaluation; Perkins et al. 2018 abstract]
• Steps.
  1. Choose a familiar tune or a simple new melody.
  2. Gather words and phrases from the group: the baby's name, habits, or what nights or mornings are like.
  3. Shape them into a verse together.
  4. Sing it and refine it over later weeks.
  5. Include it in the recordings. [steps 1–4 synthesis; recordings: source]
• Example language (illustrative). "Give me one word for what 3am feels like right now… and one for your baby's smile. Let's make a line out of those."
• Watch for.
  • Lyrics can surface raw feelings such as exhaustion, ambivalence or grief. Accept them warmly, keep the group's focus on the song, and let the coordinator check in afterwards. Do not explore them therapeutically in the room. [synthesis]
  • Hopeless or self-harm-related content in lyrics or comments: pass it to the coordinator the same day for follow-up. [synthesis]

4.12 Recordings for home listening

• Purpose / when. Group recordings are shared on private online platforms or on CDs, so mothers can keep singing at home. [source: process evaluation; Estevao et al. 2021] The 2018 evaluation recommended more formal take-home resources if funding allows. [source: process evaluation]
• Steps.
  1. Get consent to record from every mother before recording.
  2. Record audio only unless all agree to video.
  3. Share only through a private channel.
  4. Do not include babies' full names or identifying details. [steps synthesis]
• Example language (illustrative). "If everyone's happy, I'll record today's songs so you can sing them at bedtime. Say no at any point and we won't."
• Watch for.
  • A mother who declines recording, for example because of domestic abuse or a confidentiality worry. Respect this without asking for reasons. [synthesis]

4.13 Online (Zoom) delivery

• Purpose / when. The format runs in person or online. SHAPER-PND moved sessions to Zoom when in-person meetings were not possible, and Breathe now offers both. An online study (SHAPER-PNDO) also ran. [source: Estevao et al. 2021; Breathe research page] The SHAPER-PNDO paper is not drawn on, so online-specific procedures are [uncertain].
• Steps [synthesis]:
  1. Keep the same 60-minute structure.
  2. Mute everyone during group singing to avoid lag, and have the leader sing in time with a backing track.
  3. Unmute for hellos and chat.
  4. Agree beforehand how to contact the mother's emergency contact and clinician if she becomes distressed on camera.
• Example language (illustrative). "You'll be on mute while we sing, so sing out as loud as you like. Only your baby can hear you."
• Watch for.
  • Mothers with cameras off for many weeks.
  • The isolation of solo online participation.
  • Unsafe settings behind the screen.

4.14 Gradual handover of ownership

• Purpose / when. Leaders deliberately step back over the block so mothers lead more. [source: process evaluation]
• Steps [synthesis]:
  1. From mid-block, invite volunteers to start the welcome song.
  2. Let mothers choose songs from the repertoire.
  3. Have mothers teach a song from home.
  4. In the final weeks, run sections that the mothers lead.
• Example language (illustrative). "Who'd like to start us off this week?"
• Watch for.
  • Never pressure a quieter mother to lead.
  • Mothers leading or suggesting songs unprompted is a sign of ownership.

4.15 Coordinator attendance and wellbeing follow-up

• Purpose / when. The coordinator tracks attendance, phones mothers who missed a session, checks on wellbeing and reports to health professionals when needed. In SHAPER-PND the control arm also received texts and calls encouraging attendance. [source: process evaluation; Estevao et al. 2021]
• Steps.
  1. Take the register every session.
  2. Phone a mother who misses a session within a day or two.
  3. Ask about practical barriers, such as nap times or transport, and how she is.
  4. Escalate risk through the agreed route: her GP, her clinician, or the perinatal team; urgent services if there is immediate risk. [steps 2–4 timing and detail synthesis]
• Example language (illustrative). "We missed you and your little one today. How are things? Anything making it hard to get there?"
• Watch for.
  • In 2018, dropouts were mostly logistical, and nap timing was a barrier for some mothers. [source: process evaluation]
  • Repeated absence together with withdrawal on the phone is a clinical warning sign. [synthesis]

4.16 Facilitator debriefs

• Purpose / when. A formal debrief for leaders after each session. Leaders reported the emotional burden of the work. [source: process evaluation]
• Steps [synthesis]:
  1. Leader, assistant and coordinator meet for 10–15 minutes after the session.
  2. Review any concerns about specific mothers or babies, and decide who acts and by when.
  3. Note what worked musically.
  4. Check how each staff member is.
• Example language (illustrative). "Before we pack up: anyone we're worried about today, and who is following up? And how are each of us doing after that session?" [synthesis]
• Watch for.
  • Leaders taking on a counsellor role.
  • Unrecorded safeguarding worries.

4.17 Optional social coffee time after sessions

• Purpose / when. Not part of the trialled sessions. Mothers, leaders and the coordinator all suggested it, and the process evaluation recommended it. Mothers described friendships forming and a sense of joining a "tribe of mothers." [source: process evaluation]
• Steps [synthesis]:
  1. Offer 15–30 minutes after the singing, in the same or an adjoining space.
  2. Keep hot drinks on a table away from babies and crawlers.
  3. Keep it informal and unstructured.
• Example language (illustrative). "There's tea and coffee next door if you'd like to stay and chat. No pressure, head off whenever your baby needs to." [synthesis]
• Watch for.
  • Coffee time should not replace singing time, since singing is the active ingredient. [synthesis based on Fancourt & Perkins 2018]

4.18 The psychotherapist's parallel work

• Purpose / when. For clients you refer: link the group with your own therapy without turning it into therapy. [synthesis throughout]
• Steps.
  1. Complete 4.1 first.
  2. Help the client find a local programme, through Breathe or through local health visitors, children's centres or social prescribers.
  3. Address practical barriers: nap timing, transport, getting out of the house.
  4. In therapy, ask what she noticed in the group: moments of closeness with her baby, enjoyment, connection with other mothers. Use these as behavioural-activation or attachment-focused material within your own model.
  5. Keep monitoring with the EPDS or PHQ-9.
  6. Plan the ending and what continues after week 10: singing at home with the recordings, and other mother–baby groups.
• Example language (illustrative). "When you sang that lullaby and she settled, what was that like for you?"
• Watch for.
  • The client dropping therapy because she is "singing now" while still symptomatic.
  • Shame about missing sessions.

5. Handling problems

The sources document only a few delivery problems: cramped venues, nap timing, logistical dropout, leaders' emotional load, and play mothers feeling their babies were too young. These are tagged [source: process evaluation]. Everything else in this section is [synthesis].

Acute risk disclosed in or around a session (suicidal thoughts, intent, or thoughts of harming the baby) [synthesis; risk steps from NICE CG192 1.6.3–1.6.5]

• The leader does not assess. Quietly bring the mother and baby out of the circle with the assistant or coordinator, and keep her company.
• The coordinator follows the agreed escalation route:
  • If there is immediate danger, call emergency services or take her to urgent care. Make sure the baby is with a safe adult.
  • Otherwise, contact her GP, clinician or perinatal team the same day, with her knowledge.
• If you are her therapist, carry out a full risk assessment, update the safety plan, and inform the relevant professionals.
• The group continues only once clinical care is in place.

Possible postpartum psychosis (sudden confusion, unusual beliefs, agitation, not sleeping, or behaviour out of character)

• Treat this as an emergency. NICE requires referral for assessment by secondary or specialist perinatal services within 4 hours. [source: NICE CG192 1.5.12]
• Do not leave her alone with the baby while help is arranged. [synthesis]

Safeguarding concern about the baby

• Concerns include rough handling, neglect, or the mother saying she cannot keep the baby safe.
• Follow local safeguarding procedures through the named safeguarding lead. Record what you saw and heard, not your interpretation. [source: NICE CG192 1.6.4; synthesis for the recording advice]
• Remember that fear of the baby being removed stops many women from disclosing. Where it is safe, be open with her about what you are doing and why. [source: NICE CG192 1.5.1; synthesis]

Tearfulness or overwhelm during singing

• This is common and is not in itself a risk sign.
• The leader offers a tissue, makes it normal to step out or just listen, and keeps the music going so she is not the centre of attention.
• The assistant can sit with her or hold the baby, if she invites it.
• The coordinator checks in afterwards.
• Do not explore the material in the group.

Musical trauma triggers

• A song, language or genre may be linked to abuse, loss (including baby loss or a NICU stay), a traumatic birth or religious coercion.
• Drop or swap the song without asking why.
• Let mothers flag songs to avoid privately at the start.
• Avoid songs whose lyrics are about lost or dying children or similar themes.

"I can't sing" or self-consciousness

• Normalise this. The leader sings imperfectly and cheerfully, and no one sings alone.
• Humming, playing an instrument or just listening all count as taking part.
• Singing quietly to the baby is the main thing.

Crying, feeding or sleeping babies

• Carry on singing. Many crying babies settle with the music.
• Mothers feed, change or walk about whenever they need to.
• A sleeping baby goes onto a safe sleep surface or stays held (see 4.2).
• Never judge a mother by her baby's mood.

Non-attendance and dropout

• In the original trial, dropout was mostly logistical, nap timing helped some mothers and hindered others, and dropouts demoralised the group. [source: process evaluation]
• Make follow-up calls (4.15) and offer recordings for missed weeks.
• Choose a session time that suits the babies' age.
• Acknowledge absent members briefly in the group without speculation. [synthesis]

Comparison and judgement between mothers

• Avoid developmental milestone talk in the circle. (The play arm opened by sharing babies' recent development, and some play mothers felt their babies were too young.) [source: process evaluation for the play arm]
• Praise each pair's own way of joining in. [synthesis]

A mother who stays flat or worsens over the block

• Re-measure with the EPDS or PHQ-9.
• Review her clinical care: high-intensity therapy and/or medication for moderate–severe depression. [source: NICE CG192 1.8.3]
• Do not treat repeating the singing block as the fix.

Physical safety incidents

• Possible incidents include a choking risk from an instrument, a fall, or a scald.
• Stop the activity, give first aid, record the incident, and report it through the venue and organisation procedures.
• Review the equipment and the room.

Language barriers and exclusion

• SHAPER-PND required satisfactory English. [source: Estevao et al. 2021]
• In practice, songs from mothers' own languages can include those with limited English, but you still need interpreter access for screening and risk conversations.

Fathers, co-parents and other carers

• The programme as researched is for mothers. Whether to include others is a local decision outside the evidence. [uncertain]

Leader burnout or boundary drift (the leader becoming a counsellor)

• Use regular debriefs (4.16). [source: process evaluation]
• Keep the leader, assistant and coordinator roles separate. [source: process evaluation]
• Leaders need a clear route to hand concerns to clinicians.

Ruptures with the referring therapist

• A client may feel "sent away" to a singing class. Present the group as an addition and keep sessions going.
• Repair by asking what the referral meant to her.

Ending

• Some mothers feel a loss when the group finishes.
• Plan with them in week 8–9:
  • keep in touch with the group, if all agree;
  • sing at home with the recordings;
  • find other mother–baby groups and community singing;
  • keep up their clinical follow-up.

Sources

Sources:

• Fancourt & Perkins — The effect of singing interventions on symptoms of postnatal depression: three-arm RCT (author accepted version, RCM Research Online)
• Fancourt & Perkins — Creative interventions for symptoms of postnatal depression: a process evaluation of implementation (author version, UCL Discovery) — journal and year not shown in the copy opened [uncertain]
• Estevao et al. (2021) SHAPER-PND trial: clinical effectiveness protocol of a community singing intervention for postnatal depression, BMJ Open 11:e052133 (UCL Discovery copy)
• Perkins, Yorke & Fancourt (2018) How group singing facilitates recovery from the symptoms of postnatal depression: a comparative qualitative study, BMC Psychology 6 (DOAJ record, abstract only)
• RCM Research Online record for Perkins, Yorke & Fancourt 2018
• Breathe Arts Health Research — Breathe Melodies for Mums research page
• Breathe Arts Health Research — Newly published research on Breathe Melodies for Mums
• King's College London — Group singing reduces symptoms of postnatal depression for up to six months
• NICE CG192 — Antenatal and postnatal mental health: clinical management and service guidance

Further reading (not drawn on) [uncertain for details]:

• Bind R, Pariante CM, et al. (2025). Clinical effectiveness, implementation effectiveness and cost-effectiveness of a community singing intervention for postnatal depressive symptoms: SHAPER-PND randomised controlled trial. British Journal of Psychiatry. DOI 10.1192/bjp.2025.10377 (details from the KCL release).
• SHAPER-PNDO online feasibility study (2023), Pilot and Feasibility Studies, and a 2025 qualitative implementation study in Frontiers in Health Services, both listed on the Breathe research page.
• Breathe Melodies for Mums programme and registration page (breatheahr.org/programmes/melodies-for-mums/), not retrievable.