High-quality perinatal care is of central importance to public health and the broader agenda for global health (ten Hoope-Bender et al, 2014). In antenatal care, the focal point has shifted from maternal physical health and fetal wellbeing to an increased awareness of psychosocial factors (Hildingsson et al, 2014). In Western nations such as the UK, New Zealand, the Netherlands, Ireland and Australia, various models of maternity care, including midwife-led, medical-led and shared care (Hatem et al, 2008) have been developed. In Victoria, Australia, women mostly receive maternity care through the public health system (Forster et al, 2016). Public maternity models such as shared-care (where a woman sees her GP for most appointments, and hospital doctors for key check-ups during pregnancy) or midwife-led care attempt to provide continuity of care and information provision, increased involvement of women in decision-making and normalisation of the birth process. Despite efforts to provide more holistic models of care, research still identifies shortcomings in service delivery, such as a lack of timely information and emotional support (Hildingsson et al, 2013).
As in many Western countries, new mothers in Victoria remain in hospital for up to 2 days after the birth of their baby (Forster et al, 2016; Goodwin et al, 2018). Postnatal hospital care has long been negatively critiqued (Rudman and Waldenström, 2007; Beake et al, 2010) as the ‘Cinderella’ of maternity services (Senate Community Affairs References Committee, 1999; Morrow et al, 2013). In a study exploring views and expectations of postnatal care (Forster et al, 2008), women expressed a desire for constant professional support to gain confidence in caring for their infant and feel supported in parenting.
Inadequate communication in the woman/caregiver relationship has been found to be a central aspect in unsatisfactory care (Redshaw et al, 2007; Hunter et al, 2008). Relationships are integral to clinical care, influencing a woman's decision to seek further information and support (Featherstone and Broadhurst, 2003), which in turn has the potential to affect morbidities associated with the childbearing period. Furthermore, effective communication between maternal and child health nurses, GPs and community services is required, although a disconnect between these parties is often reported (Department of Health and Ageing, 2010; Prime Minister's Science Engineering and Innovation Council, 2008).
In addition to the provision of quality collaborative maternity services, preparation for motherhood can influence adjustment (Lederman and Lederman, 1987; Serçekus and Mete, 2010). While there are opportunities to enhance maternal health literacy via maternity models of care (Batterham et al, 2016), antenatal education is the main vehicle for the provision of information about childbirth and parenthood. Despite this, there is no robust evidence that education in the antenatal period reliably attains its most frequently stated objectives of decreasing distress during birth, or improving parenting (Gagnon and Sandall, 2007). Given such criticisms, attempts have been made to develop more comprehensive models of care, one example being the ‘CenteringPregnancy’ care model. CenteringPregnancy was implemented as a mode of antenatal care blending assessment, education and support in group settings (Ickovics et al, 2007; Teate et al, 2011). While this model succeeded in improving pregnancy outcomes, it still mostly only supports women in the antenatal period (Novick et al, 2013), hence the need for more maternity-service provision addressing complex bio-psychosocial needs across the entire perinatal period.
On clinical outcome indices such as mortality rates, maternity care may appear to indicate steady positive trends (Kildea et al, 2008; Knight et al, 2014; 2016); however, mortality rates alone do not convey the complete picture. Issues such as obstetric intervention rates in low-risk pregnancies (Dahlen et al, 2012), breastfeeding duration (Thulier and Mercer, 2009), and the rising number of women developing antenatal and postnatal depression (Leigh and Milgrom, 2008) are all of concern. This study aimed to address common maternal morbidities and limitations identified in models of care and education across the perinatal period.
Methods
Study aim and hypothesis
The aim of this study was to identify the potential efficacy of a multifaceted, group-based, psycho-educational programme for primiparous women and their partners. It was hoped that programme participants would demonstrate the following outcomes, compared to a control group:
Design
This alternate-allocation pilot trial compared two groups: one who received an intervention programme (routine care with the addition of a weekly telephone call) and a comparison group, who received routine care. Recruitment of potential participants ceased at 23 (Figure 1) because, as a pilot study, groups of 6–10 post-attrition were preferred. While alternate-allocation is not a traditional random method of allocation, it has been described as ‘a method that can abolish selection bias equally well if applied strictly’ (Kerr et al, 2004: 81). Pilot studies play a significant role in health research, as they can provide evidence for the development and validation of higher-quality experimental studies (Anderson and Prentice, 1999). As a small pilot study, this trial was exploratory and not powered to reliably detect smaller significant differences or associations.
Development of the intervention programme
The content of the pilot programme was developed with the involvement of the multidisciplinary team, who later facilitated sessions. Programme content was guided by preliminary research (Buultjens et al, 2013a; 2013b; 2016; 2017a; 2017b), and previous studies (Currid, 2004; Hoddinott et al, 2012), and based on a proactive, rather than reactive, approach to perinatal care. A full copy of the evidence-based literature that informed the programme is available from the corresponding author.
The intervention was a group-based, psycho-educational programme, beginning in the third trimester. The programme was conducted weekly for 2-hour sessions until approximately the eighth week after birth, as literature on depression suggests that it is by this time that higher initial levels of stress of being a parent have been experienced, and many of the life changes associated with having a baby have become established (Gavin et al, 2005). This is also when parents master new tasks, such as breastfeeding and fostering a secure mother-infant attachment. Of the antenatal sessions, one session was included in the intervention programme that focused on increasing couples' awareness of each other's psychosocial concerns in the transition to parenthood. The group process was thought to allow women to normalise their experiences in the transition to parenthood, such as physical changes, hormonal and emotional changes, and changes to couple relationships (Duncan and Bardacke, 2010). A copy of the full curriculum is available from the corresponding author.
Setting
The study was conducted in Melbourne, Australia following ethics approval from La Trobe University (FHEC09/215). Eligible participants were women above 18 years of age who were experiencing their first pregnancy. They (and their partners) were all able to speak and comprehend English because English-language instruments were used in the research.
Recruitment
A self-selected sample was recruited through various forms of advertisements, including an online pregnancy, baby and parenting website, print media, and flyers posted in GP clinics. Each source advertised the study and invited interested participants to contact the research team. Details of group allocation and study participation are illustrated in Figure 1.
When women contacted the principal researcher, they were informed that they would be randomly assigned in the order they returned their consent forms. Therefore, group assignment was determined via sequential number formula, of which odd numbers represented the comparison group (telephone support) and even numbers represented the intervention group (programme). Potential participants were not aware of the alternate allocation sequence and to further ensure allocation concealment (Schulz and Grimes, 2002), participants did not have knowledge of group assignment, until they were all notified on the same day.
Measures
A number of outcome measures were used over five time periods, at baseline (T1: 34–36 weeks gestation); the conclusion of the final antenatal session (T2: 38–40 weeks gestation); following the first postnatal session (T3: 2-5 weeks postnatal); the conclusion of the final postnatal session (T4: 5-8 weeks postnatal) and at follow up (T5: 12-14 weeks postnatal). Each measure was given to members of both the programme and comparison groups. The scales used (EDPS, MSPSS, SDAS, MPAQ and PSCS) broadly addressed four psychosocial areas: maternal mood, marital adjustment, social support and mother-infant attachment. Breastfeeding outcomes were also assessed for comparison purposes.
Data analysis
All statistical analyses were performed using SPSS 19.0. Descriptive statistics comprising frequencies, means, standard deviations and percentages were produced. Inferential statistics included: Levene's test of homogeneity of variance, the independent-samples t-test of means for continuous variables, and Fisher's exact tests for categorical variables. To assess the effect of intervention over time as compared with baseline, repeated measures analysis of variance was conducted (Fowler, 2013). Differences at specific time points between the intervention and comparison group were examined using independent sample t-tests, with the important Cohen's d computed at T4 and T5 (follow-up) to measure the effect size involved. Particular analysis of time-points 4 and 5 were chosen to measure and compare the size of the intervention effect; that is, the magnitude of the difference between mean scores of the intervention versus the comparison group (Thalheimer and Cook, 2002).
Using the intention-to-treat principle, the pilot study reported outcome measures irrespective of the number of sessions participants attended. As noted in previous research, using this principle avoids over-optimistic assessments of the efficacy of an intervention arising from the exclusion of non-completers (Heritier et al, 2003).
Findings
A total of 18 women expecting their first baby (10 in the intervention arm and 8 in the comparison arm) participated in this research. No significant differences were found between women in the intervention and comparison groups (P values all >0.05) (Table 1). Women in the study were aged 24–39 years, and most were well educated, with 83.3% (n=15) having a tertiary level of education. In total, 88.9% (n=16) had a combined family income greater than AUD $50 000 per annum.
| Intervention (n=10) | Comparison (n=8) | P-value | |
|---|---|---|---|
| Maternal age (years) mean (SD) | 32.56 (3.25) | 31.86 (5.30) | 0.503a |
| Country of birth n (%) | |||
| Australia | 9 (90) | 5 (62.5) | 0.163b |
| Other | 1 (10) | 3 (37.5) | |
| Marital status n (%) | |||
| Engaged | 0 (0) | 2 (25) | 0.94 |
| Married | 9 (90) | 3 (37.5) | |
| De facto | 1 (10) | 3 (37.5) | |
| Education level n (%) | |||
| Certificate | 0 (0) | 1 (12.5) | 0.568b |
| Diploma | 2 (20) | 0 (0) | |
| Undergraduate | 3 (30) | 4 (50) | |
| Postgraduate | 5 (50) | 3 (37.5) | |
| Household income in AUD n (%) | |||
| <$50,000 | 0 (0) | 2 (25) | 0.315b |
| $50 000–$99 999 | 3 (30) | 3 (37.5) | |
| >$100 000 | 7 (70) | 3 (37.5) | |
| Type of health cover n (%) | |||
| Private | 5 (50) | 3 (37.5) | 0.596b |
| Public | 5 (50) | 5 (62.5) | |
| History of mental health difficulties n (%) | |||
| Yes | 3 (30) | 2 (25) | 0.813b |
| No | 7 (70) | 6 (75) | |
Independent sample t-test;
Fisher's exact test (Fisher's exact test was used given the number of participants was <20)
Infant feeding outcomes
With regards to infant feeding (exclusive breastfeeding in comparison to formula, or a combination of breastfeeding and formula), there were no differences at T3 (P=0.069). However as illustrated in Table 2, at T5 (follow-up), there were distinct differences between groups, with the intervention group reporting greater exclusive breastfeeding than the comparison group (P=0.01).
| Intervention (n=10) (%) | Comparison (n=8) (%) | P-value* | |
|---|---|---|---|
| Time 3 | 0.059 | ||
| Breastfeeding | 9 (90) | 4 (50) | |
| Formula or combination of breastfeeding and formula | 1 (10) | 4 (50) | |
| Time 5 | 0.003 | ||
| Breastfeeding | 10 (100) | 3 (37.5) | |
| Formula or combination of breastfeeding and formula | 0 (0) | 5 (62.5) | |
Standardised psychosocial measures
Table 3 presents the scores on the tests, and reports on any between-group differences. The two later measurement points (T4 and follow-up (T5)) were of particular interest (as they were post-intervention) and effect size calculations were computed to measure the relative magnitude of the impact of the intervention programme. For example, a P-value indicates only the probability, or otherwise, of a real difference or association when the null hypothesis is true. Reporting the effect size therefore facilitates the interpretation of the result. Without an estimate of the effect size, the practical implications of the differences cannot be fully understood.
| Scale | Intervention (n=10) (%) | Comparison (n=8) (%) | P-value | Cohen's d |
|---|---|---|---|---|
| Edinburgh Postnatal Depression Scale (EPDS) | ||||
| Time 1 | 5.70 (3.59) | 6.88 (3.64) | 0.503 | -0.35 |
| Time 2 | 3.78 (2.86) | 8.63 (3.02) | 0.004 | -1.76 |
| Time 3 | 4.13 (1.73) | 8.13 (3.98) | 0.027 | -1.45 |
| Time 4 | 3.38 (1.60) | 9.25 (5.23) | 0.015 | -1.70 |
| Time 5 (Follow-up) | 4.5 (2.68) | 9.25 (4.77) | 0.017 | -1.35 |
| Multidimensional Scale of Perceived Social Support (MSPSS) | ||||
| Time 1 | 60.80 (19.49) | 64.62 (20.16) | 0.689 | -0.20 |
| Time 4 | 66.87 (16.56) | 60.62 (20.16) | 0.489 | 0.36 |
| Time 5 (Follow-up) | 62.80 (17.73) | 59.75 (19.29) | 0.732 | 0.18 |
| Spanier Dyadic Adjustment Scale (SDAS) | ||||
| Time 1 | 123.20 (7.66) | 123.5 (14.27) | 0.955 | -0.03 |
| Time 2 | 125.89 (9.67) | 123.75 (11.46) | 0.682 | 0.22 |
| Time 3 | 126.62 (7.59) | 122.62 (9.04) | 0.354 | 0.51 |
| Time 4 | 124.87 (7.59) | 109.12 (27.53) | 0.157 | 0.88 |
| Time 5 (Follow-up) | 128.00 (7.60) | 106.62 (36.39) | 0.144 | 0.92 |
| Maternal Postnatal Attachment Questionnaire (MPAQ) | ||||
| Time 3 | 82.70 (8.00) | 74.17 (8.53) | 0.058 | 1.10 |
| Time 4 | 85.09 (5.25) | 73.40 (13.76) | 0.051 | 1.25 |
| Time 5 (Follow-up) | 84.72 (6.83) | 77.41 (8.78) | 0.064 | 1.00 |
| Parenting Sense of Competence Scale (PSCS) | ||||
| Satisfaction subscale | ||||
| Time 3 | 37.62 (4.31) | 36.25 (6.34) | 0.620 | 0.27 |
| Time 4 | 39.12 (3.27) | 36.12 (6.68) | 0.227 | 0.63 |
| Time 5 (Follow-up) | 39.7 (4.08) | 38.75 (6.30) | 0.703 | 0.19 |
| Efficacy subscale | ||||
| Time 3 | 30.87 (2.36) | 30.50 (4.11) | 0.813 | 0.12 |
| Time 4 | 34.00 (1.51) | 31.50 (4.04) | 0.120 | 0.91 |
| Time 5 (Follow-up) | 34.4 (2.17) | 32.62 (4.34) | 0.272 | 0.57 |
| Total score subscale | ||||
| Time 3 | 68.5 (6.35) | 66.75 (10.40) | 0.665 | 0.22 |
| Time 4 | 73.12 (4.42) | 67.62 (10.56) | 0.153 | 0.75 |
| Time 5 (Follow-up) | 74.10 (5.68) | 71.37 (10.47) | 0.489 | 0.36 |
Time 1 = 34–36 weeks gestation; Time 2 = 38–40 weeks gestation; Time 3 = 2–5 weeks postnatal; Time 4 = 5–8 weeks postnatal; Time 5 = Follow-up, 12–14 weeks postnatal
From these tests of psychosocial functioning, significant differences were observed on the EPDS. All measures (EDPS, MSPSS, SDAS, MPAQ and PSCS) showed moderate-to-large effect sizes at T4 and T5.
Table 4 extends these between-group analyses by presenting additionally the results of a 2x2 ANOVA test to identify the interaction between the programme and the passage of time.
| Scale | Group effect | Time effect | Group x time effect | ||||||
|---|---|---|---|---|---|---|---|---|---|
| F-stat (df) | P-value | η2 | F-stat (df) | P-value | η2 | F-stat (df) | P-value | η2 | |
| EDPS | 10.275 |
0.006 | 0.423 | 0.066 |
0.801 | 0.005 | 3.809 |
0.071 | 0.214 |
| MSPSS | 0.230 |
0.639 | 0.016 | 1.881 |
0.192 | 0.118 | 5.997 |
0.028 | 0.300 |
| SDAS | 1.367 |
0.262 | 0.089 | 2.132 |
0.166 | 0.132 | 5.265 |
0.038 | 0.273 |
| MPAQ | 6.575 |
0.022 | 0.320 | 7.256 |
0.017 | 0.341 | 0.296 |
0.595 | 0.021 |
| PSCS |
1.116 |
0.321 | 0.070 | 0.309 | 0.074 | 16.576 |
0.001 | 0.542 | 1.057 |
| PSCS |
1.351 |
0.265 | 0.088 | 66.207 |
<0.001 | 0.825 | 7.356 |
0.017 | 0.344 |
| PSCS |
0.940 |
0.349 | 0.063 | 6.036 |
0.028 | 0.301 | 0.168 (1,14) | 0.688 | 0.012 |
EPDS: Edinburgh Postnatal Depression Scale; MSPSS: Multidimensional Scale of Perceived Social Support; SDAS: Spanier Dyadic Adjustment Scale; MPAQ: Maternal Postnatal Attachment Questionnaire; PSCS: Parenting Sense of Competence Scale
When comparing scores of the intervention group with these of the comparison group across the five measures (Table 4), intervention group scores were more positive on all indices. Positive trends in the data were further noted at T4 and T5 (follow-up), arguably the two most important points in time.
From the between-within ANOVA results of Table 4, it can be observed that there were both group and time effects. As was perhaps expected, time made a difference to scores on both depression and self-efficacy. On average, across both groups, EPDS scores improved over time, and there was a tendency for self-efficacy scores to improve among all participants, supporting the self-correcting role of ‘experience’. Group effects when observed explained some relatively large amounts of variance (see eta-squared (h2) value of 0.42 for the EPDS).
There is insufficient space for a detailed explication (scale by scale) of the results obtained. Most notable trends were detected in the EPDS (an important index of maternal mood) and the MPAQ (measuring maternal perceived attachment to infant, see Tables 3 and 4).
Discussion
The aim of this health-promoting programme was to facilitate the transition to parenthood for first-time parents by means of a multifaceted group psychosocial and educational intervention. An ambitious set of specific hypotheses was proposed in anticipation of positive effects of the intervention programme on psychosocial constructs as well as increased breastfeeding and reduced intrapartum intervention.
The research was implemented across both the antenatal and postnatal periods, which is consistent with earlier evidence suggesting that interventions are best administered during both periods to create sustained beneficial effect (Vieten and Astin, 2008). This study applied a range of validated outcome measures, and it was therefore anticipated that study findings would be pertinent to both clinical practice and future research. The intervention model was able to target a number of women through the group format, not only increasing a woman's social network and health literacy, but also addressing numerous clinical care components, i.e, biopsychosocial factors and morbidities associated with perinatal health. This framework supports women in a proactive, rather than reactive manner, and the positive preliminary results highlight a research priority.
While this trial was exploratory and not adequately powered to reliably detect many observed differences, the results suggest that the stable positive mood among programme participants, when compared with the comparison group, was likely to be an intervention effect. The between-group differences indicated for the comparison group an increase of approximately 35% in mean EPDS from baseline to follow-up, whereas the intervention group showed a decrease of approximately 21% from baseline to follow-up. Significant P-values were reported at all postnatal time-points (T3=0.027, T4=0.015 and T5=0.017). These results are compatible with findings of Dennis et al (2009), who confirmed that women who received peer support had a significantly minimised chance of a diagnosis of postpartum depression at 12 weeks than those in the comparison group. Minimising the risk of emotional distress in mothers has high cost-benefit value, given that only around 25% of postnatal depression cases are reported and 63% may be dealt with as late as 1 year postpartum (Rice et al, 2001). Furthermore, early detection and treatment of postnatal depressive symptoms is thought to increase overall functional status of mothers (Fathi et al, 2017).
Effective parenting is key to fostering healthy infant development. A Cochrane review (Bryanton and Beck, 2013) evaluated the effects of structured postnatal education delivered by an educator to an individual or group on infant general health and parent-infant relationships. The review conclusion suggested that the benefits of such programmes to women and their infants remain unclear, and that larger, better-designed studies were recommended. The present study went one step further than what is offered in most parent education sessions, including mother-baby practical activities, designed to build and strengthen the bond reinforced by positive associations with one's infant. The findings of the MPAQ suggested that the intervention had a positive impact on the maternal bond. The results of the intervention programme illustrated a consistent increase from T3 to follow-up. In particular, large effect sizes were observed at T4 and T5 (Cohen's d of 1.25 and 1.00, respectively). While many parents do not have social support, role models to foster positive parenting behaviours (Hanna et al, 2002) or opportunities and resources to facilitate positive mother-infant interactions, this highlights the importance of considering incorporating such a component in future education and service provision. This is further supported by evidence that demonstrates that satisfaction with emotional support from maternity care professionals can potentially mitigate maternal psychological strain during the postnatal period (Razurel et al, 2017).
The preliminary results of this study show that a brief couples' session addressing topics such as communication, role responsibilities and problem solving, may positively influence couples' adjustment. Mean SDAS scores showed substantial differences between the intervention and comparison group, which was maintained at follow-up, suggesting positive intervention effects. Furthermore, large effect sizes were noted at T4 (0.88) and T5 (0.92). The group x time effect showed significance (P<0.05), including a large effect size (η2=0.273). This study found that increasing a partner's awareness of the postpartum period and helping them to anticipate any difficulties from communication, roles and responsibilities had encouraging effects, although this was only a small sample comprised of participants who chose to attend the sessions. While the notion of including men into more sessions may be attractive for women (Stamp et al, 1995; Mattern et al, 2017), it may not be appealing to men (Matthey et al, 2004).
This study hypothesised that the group programme participants would report greater social support. While mean scores reported did not reach statistical significance, trends noted in the results were in a positive direction, and the interaction effect of group and time was statistically significant (P<0.05) in improving the mother's scores as measured by the MSPSS. Evidence suggests that maternal support is significantly associated with maternal self-efficacy in the postnatal period (Haslam et al, 2006). The findings of this study indicate a similar correlation, as programme participants also had consistently encouraging results on the efficacy subscale as measured by the PSCS. This result is consistent with the tenets of Bandura's (1997) theory of self-efficacy, wherein social influence in the form of informal support can constructively influence a parent's confidence in their ability to perform key tasks.
Reported breastfeeding was a planned index of the success of the intervention. As anticipated, the proportion of women breastfeeding in the intervention group (90%) was higher at T3, compared with those in the comparison group (50%). Results were maintained at follow-up with 100% of the women in the intervention exclusively breastfeeding, versus only 37.5% in the comparison group. These results are a step in the right direction towards meeting the World Health Organization and UNICEF recommendations of exclusive breastfeeding for first 6 months (World Health Organization, 2003). As identified in previous research, women value being shown, rather than told, how to breastfeed (Memmott and Bonuck, 2006), and the practicality of the study intervention seemingly facilitated timely breastfeeding instruction, reinforcement and support.
In summary, while high statistical significance was limited to the EPDS scores, consistent patterns and positive trends were noted when contrasting the intervention and comparison groups on all other measures (Table 4). Furthermore, large effect sizes (as defined by Cohen (1988)) were observed for the EPDS, SDAS, MPAQ and PSCS (satisfaction and efficacy sub-score) at follow-up. One obvious explanation of many of the inconsistent results at the 5% significance level is the small sample size. Consequently, further research should aim at obtaining a much larger sample, which will improve power to detect differences between groups.
No cost-effectiveness analysis was undertaken as it was outside the scope of this study. However, it has been reported that health care expenditure increases for mothers with depression (Kendall-Tackett, 2005), and as such, improving women's emotional wellbeing alone may yield economic benefit. Although this study was limited by the small sample, it illustrated a promising model of care that has the potential to reduce waiting times to access specialist perinatal services, increase continuity of care, and overall improve perinatal outcomes. Capitalising on the shared learning environment, the programme further provided active learning opportunities and timely skills training throughout the transition to parenthood.
Although beyond the direct scope of this research, midwifery workforce issues are of international concern, and thus highlight the need to develop sustainable models of midwifery practice that not only foster healthy workload infrastructure, but are also aligned with midwifery core values. As illustrated in the present study, a group model focusing on the antenatal and postnatal episodes of care can directly facilitate labour and birth outcomes by increasing women's health literacy and self-efficacy through education, preparing the partner for birth and beyond, and more broadly, building parental capacity to enable informed decisions. Existing literature further supports that psycho-education during pregnancy can be a significant factor in improving women's birth confidence (Toohill et al, 2014). In addition, a model such as this encourages multidisciplinary work to provide childbearing women timely information and referral. In many instances, specialist collaboration does not go beyond referral to other services, and it is unusual for information to be fed back in a timely manner to the referring practitioner or service (Kruske et al, 2006). This study demonstrated a collaborative model, enabling information to be commonly available to all perinatal health professionals involved in the woman's care. In practice, this is a step towards reducing conflicting information, and although not evaluated here, has the potential to improve a woman's quality of reproductive life, reduce health care costs, and improve maternal and child health outcomes.
Limitations of the study
While randomisation was not used in this pilot, it is not a major limitation of the study, as the alternation allocation method used obviated possible selection bias. More important is the generalisability of the findings, which is constrained firstly by the small sample size and secondly by the characteristics of the sample. Participants in the sample were predominately white, Australian-born mothers in a stable relationship, and in general were relatively well educated, middle-class women. While partners were included in some aspects of the intervention, the outcome measures were limited to the mothers; therefore, no analysis of intervention effects on the partner was possible. Moreover, this study did not target women who had a diagnosis of clinical depression or anxiety, or those at-risk for postnatal depression as in previous research (Zlotnick et al, 2006). Finally, because of the large number of tests applied and the 0.05 alpha (a) level used, the possibility of a type-1 error must be acknowledged. All the significant results found need to be supported with data from a larger number of participants and with fewer primary outcome measures specified.
Conclusion
This article reported on a pilot perinatal health study, developed within a health-promoting framework. The emphasis on prevention and facilitation of multiple areas and constructs amid the transition to parenthood is the potential strength of this intervention model. However, as a professionally-led intervention, it may incur high costs, thus an economic analyses needs to be conducted with consideration of the existing perinatal healthcare expenditure. Limitations notwithstanding, the pilot programme was associated with many positive results, including improved emotional wellbeing, mother-infant attachment, and high rates of breastfeeding. Replication of this programme in a different setting with a larger, more diverse sample of women would allow for greater generalisability of results. Furthermore, this programme could complement existing models of hospital maternity care, with the potential to decrease the number of hospital antenatal visits, and increase educational opportunities and social capital. This novel approach is now ready for more rigorous testing (including measurement of clinical outcomes, such as preterm birth) with a no-intervention control in a large randomised controlled trial.