COMPANIONSHIP by a lay labour supporter to modify the clinical birth environment : long-term effects on MOTHER AND CHILD

Navorsingsontwerp Tweehonderd-en-sestig primigravidas wat in gestabiliseerde baring was met geen verloskundige komplikasies, en wat geen ondersteuner by haar gehad het nie, het toestemming verleen om aan die studie deel te neem. Hulle was ewekansig toegewys om addisionele ondersteuning deur 'n “doula” gedurende baring te ontvang, of die gewone standaard prosedure van verpleegsorg te ontvang. Die deelnemers en hul babas is een jaar na die bevalling opgevolg.


Introduction
The birth of a child and adaptation to motherhood is generally accepted to be a stressful life event.During this time women may experience significant physi ological, phycological and emotional changes in their lives.
One of the most frequent complications of childbirth is postpartum depression (1).Perinatal stress, and in particularly post partum depression have been associated with negative effects on children's cog nitive and behavioural development (2).The arrival of a child frequently disrupts established marital and sexual relation ships (3,4).Maternal adaptation to child birth and maternal behaviour during the early postnatal period may influence mother-infant interactions and subse quent growth and development of the child (5).
The way in which women adapt to these changes may be influenced by the sup port they receive during the uniquely vul nerable time of labour.Continuous sup port is an essential component of child birth in many cultures especially where home deliveries take place.
Labouring women are often supported by their mothers or a family member.With the shift of the place of birth from home to the hospital, many practices have changed.Although fathers or relatives have in recent years been allowed into labour rooms, a considerable number of women still undergo labour without con tinuous support (6).Hodnett (7) has reviewed 11 randomized c o n tro lle d tria ls of s u p p o rt from caregivers (social or professional) during childbirth.The trials were conducted in G u atem ala , F in la n d , S outh A frica , Canada, France, Belgium, Greece, and the United States, under widely diverse circumstances.
She concludes that the continuous pres ence of a support person during labour has positive short term effects on mater nal and neonatal outcomes.
We conducted a randomized controlled trial to evaluate the effect of supportive com panionship on labour and various aspects of adaptation to parenthood (8).We found that the support group were more likely to report that they felt that they had coped well during labour (60 vs 24%, p < 0.00001).
At six weeks women in the support group were more likely to be breastfeeding ex clusively (51 vs 29%, p < 0.01).Higher self-esteem scores (74.5 vs 58.8, p <0.0001), and lower postpartum depres sion (10.4 vs 23.3, p < 0.0001) ratings were attained at six weeks postpartum by women who received support (9).
To investigate whether the above posi tive short term features may have self-re inforcing effects that could contribute to long-term psychosocial outcom es we have followed the women up, one year after birth.Clinical, psychological and biochemical results of the original study have been reported previously (8,9,10,11,12,13,14,15).
Because of expected loss to follow-up, enrollment was continued after conclu sion of the original study, to increase the sample size from 189 to 262 women.

Research objectives
The objective of this study was to exam ine whether supportive companionship given by a "doula" (lay labour supporter) to nulliparous women would have long term psychosocial effects on the mother.
It was postulated that the positive short term features experienced by women who received supportive companionship by a doula during labour, may have self reinforcing effects on feeding practices, the mothers' perception of their infants, parenting practices and the mothers' re lationship with their partners one year after delivery.

Sample and m ethods
A randomized controlled trial was con ducted at an academ ic state hospital serving a low income, multi-cultural, multi racial urban population in the Western Suburbs of Johannesburg, running over a period of 56 months, from February 1989 through October 1993.Low risk nuliiparous women in estab lished labour with cervical dilatation of less than six cm and who had no sup portive companion with them were asked to participate in the study.The trial was verbally explained to the subjects and written consent was obtained.The sub jects were informed that they could with draw from the study at any time and that they had only have a 50% chance of re ceiving a labour companion.
A short baseline questionnaire was com pleted and baseline blood samples for the m e a su re m e n t o f c o rtis o l, catecholamines and endorphines were taken.An external fetal heart rate trans ducer and uterine contraction monitor were applied if not already in use.
Participants were then allocated to an ex perimental or control group by means of randomly ordered cards in numbered, sealed opaque envelopes.After the en velope was opened the patient was in formed whether she would receive sup port from a "doula" or not.If the patient was in the control group, she was looked after by the midwifery and medical staff according to normal hospital routine.
If the patient was in the study group she received similar care, and in addition a "doula" was introduced to her.This ad ditional su pp ort usually com m enced within ten minutes after randomization.
For the purpose of this part of the study, structured interviews were conducted one year after delivery.
The interviewers were blind to the group allocation except for very few interviews in which the mother voluntarily told the interviewer that she had been allocated a "doula" during labour.
The s u b je c ts w ere c o n ta c te d at 11 months after delivery to make an appoint ment for an interview at their home (sup port 49, control 50) or at the hospital (s9, c5).
If it was not possible for the subject to come to the hospital or for the researcher to go to her home, the questionnaire was completed by means of a telephonic in terview (s1, c1) or an adapted postal questionnaire (s5, c11).Three midwives were employed and trained to conduct the one year questionnaire.
Each interviewer was allocated a group of subjects to follow-up, irrespective of randomized allocation.The interview took about 40 minutes to complete.A verbal introduction was given and sub jects were assured of confidentiality.
The questionnaire assessed the follow ing principle outcomes: feeding prac tices, mothers' perception of their infants, parenting practices, mothers' relationship with their partners, self-esteem, and post partum depression.
The Coopersmith Self-esteem Inventory was used to evaluate self-esteem .It measures the attitudes towards the self in domains such as social, occupational, family and personal areas of experience (17).
Postpartum depression scores at six weeks were initially measured using the Pitt Depression Inventory (18).Pitt does not describe a "cut off" point for depres sion.Depression is monitored in terms of changes in the individual's scores from one point of assessment to another and not as absolute values.For the purpose of this analyses we have divided the scores into two categories.
During the course of the study it was de cided that the Edinburgh Postnatal De pression Scale (EPDS) (19), which had been developed specifically to assist pri mary health care professionals to detect mothers suffering from postnatal depres sion, would be more appropriate.Moth ers who score above a threshold of 12 are more likely to be suffering from a de pressive illness (Table 3).

Intervention
The hospital routine at the time of the study did not encourage a companion to be present during labour or delivery, mainly due to lack of space and privacy in the delivery rooms.Artificial rupture of membranes during the active phase of labour was used frequently.Epidural analgesia was not available, but Pethi dine HCL and Hydroxyzine HCL were commonly used for analgesia.The new born was usually taken to the nursery shortly after birth.
The mothers were moved from the labour ward directly after third stage or suturing of the perineum to the postnatal ward.The infants were taken to the mother af ter they had been examined, bathed and given a dextrose water feed.
Mothers who chose to breast feed and who appeared to have problems were offered help by the nursing staff.Sup plementary feeds were commonly pre scribed by doctors.
Mothers and infants went home 24 -48 hours after delivery.They were followed up by the district midwife for another two days, or more if needed.Participants in the support group received the same care as the control group, the only difference being the support of the doula during la bour and delivery.
Continuous social support in terms of this study meant the continuous presence of a " Doula" during labour.The three doulas were mature women who had no medi cal, nursing nor traditional midwifery ex perience, but all had experienced child birth themselves.The work was volun tary, though a small expense allowance was paid (R150 per month).
The doulas were drawn from the same community as the patients and although they were not known personally to the participants, they shared common val ues.The doulas were asked to concen trate on emotional support, comforting, encouraging, reassuring and praising the participant.
Tangible support included physical touch or back rubbing, assistance with chang ing of position, giving fluid to drink or calling the nursing staff.The doulas at no stage gave any informational support such as advice on positioning, guidance with pushing or breastfeeding advice.The doulas were introduced to the par tic ip a n ts fo r the firs t tim e after randomization and stayed with them con tinuously during labour and in most cases during delivery.

Lim itations of study
The low follow-up rate (50%) in this study remains a matter of concern although we have no reason to believe that bias was introduced, as similar proportions of sub jects were followed up (s51% vs c 49%), and the baseline variables of the groups followed up remained comparable.Use of two different measures of depression also limits the power of the study to de tect differences.

Results
The 262 women who agreed to take part in the study were randomly allocated to a support group (s126) and to a control group (c136).Although extensive and va riou s m eth od s of fo llo w up were persued, only a 50% follow up rate (s64/ 126 vs C67/136) was reached.Maternal services were not free at the time of the study and many women gave false ad dresses.
The population also belonged to a politi cally and socially disadvantage commu nity which was very mobile.
Biographical data showed no significant differences between the two groups.The mean ages for the two groups were the same (s21 vs c21) years.
Slightly more m others in the support group were unemployed (s56% vs 48%), and similar proportions were unmarried (s71% vs c69%).Most women lived with their family (s65% vs c67%).The infants in the two groups were similar with re gard to gender (s54% vs c49% male), w eight (s10kg vs c10kg) and length (s74cm vs c73cm) at one year.
Nearly half of the mothers in both groups were using bottle feeding only at one year.Mean d uration of exclusively breast-feeding were sim ilar (s8 vs c 7wks).The most common breast-feeding problems experienced were insuffi cient milk production and nipple prob lems.
The m ost frequent reasons given by mothers for introducing bottle feeding were insufficient milk and returning to work (Table 1).
Fewer mothers in the support group felt that their relationship with their partner was worse since the birth of the child and significantly fewer did not resume inter course within three months of the birth (Table 2).The Coopersmith Self-esteem score did not any longer distinguish sig nificantly between the support and con trol group at one year.There were no differences in postpartum depression scores between the two groups at one year (Table 3).

Discussion
The 52% overall rate of breastfeeding at one year was somewhat higher than that found in a previous survey (39%) of a s im ila r p o p u la tio n by R ou ssou w & Jansen (20).In the initial phase of the current study, breastfeeding at six weeks was significantly more frequent in the support group than the control group (8).
The beneficial effect of labour support on early breastfeeding did not persist at one year.
The p oint at w hich m others resum e sexual relationships with their partners depends on many physical, psychologi cal and social factors including self-esteem and postpartum depression.Re search has shown that the majority of women resume intercourse before twelve weeks, most before seven weeks post partum ( 21).
In our study, most of the women resumed intercourse o nly after tw elve weeks.Fewer in the support group reported that their relationship with their partner had deteriated after the birth of the child, and significantly fewer had not resumed in tercourse within three months.Support during labour may enhance the emotional readiness for the mother to resume a sexual relationship and may enhance the relationship with her partner.
It is estimated that as many as 20% of women will suffer mild to moderate de pression following childbirth (22).At one year a high percentage of women (30%) still suffered some depression.
In general, the clear-cut psychosocial benefits of labour companionship found at six weeks postpartum, such as greater self-esteem and less postpartum depres sion, were no longer discernable at one year.
This may in part have been due to the lim ited follow -up and the difficulty of measuring psychosocial parameters ac curately.However, it is more likely that the benefits of labour companionship are with time diluted by the multiple other experiences and influences which may affect a mother and her child over the first year of life.
There is some evidence that ongoing support is associated with a decreased
The failure to demonstrate clear effects persisting for one year does not detract from the fact that there are clear short term benefits and no known risks asso ciated with social support during labour and delivery.

Hofmeyr GJ, Nikodem VC, Wolman W- L, Chalmers BE & Kramer TR.
So cial and professional su pp ort during childbirth.In: Chalmers I, Enkin M & Keirse MJNC, eds.Effective care in preg nancy and childbirth.Oxford: Oxford Uni versity Press, 1989;807 Com panionship to modify the clinical birth en vironment: effect on progress and per ceptions of labour, and breastfeeding.British Journal o f Obstetrics and Gynae cology 1991;98:756-764.