The questionnaire set consisted of 104 items, which took around 20-30 minutes to answer. Permission to use instruments was obtained from the original authors unless available for free access. For instruments not available in Bengali, translation-back translation was done and independently reviewed by Bangladesh experts (English lecturer and obstetrician) to check for appropriate meaning and to ensure the equivalence of the two versions.

The data were collected by the first author with help from two trained local research assistants (RA), from May to December, 2017. The data were collected at baseline, postpartum 2 weeks, and postpartum 6 weeks.

1) Baseline (between postpartum day 1-6): Participants returned completed questionnaires in a collection box placed in the ward and the researcher checked for any missing items. Phone number were obtained for the purpose of further contact. The posttest 1 questionnaire was provided along with a stamped envelope and participants were requested to complete the second questionnaire and send it via mail 2 weeks later. The researcher provided a card marked with the code number to show the RA at posttest to avoid unnecessary exposure of personal information.

The NLPPSC was developed following the ADDIE model [27], according to the following phases (Supplementary material).

Analysis phase: Data on postpartum care and interventions from WHO, UNFPA, World Bank, and a literature review was done using PubMed, CINHAL, EMBASE, Cochrane, PsycInfo, and Google Scholars database, for English publications from 2000-2017. Eleven related articles were identified but none were found for Bangladesh. Interviews with three first-time postpartum mothers in Bangladesh, who were purposively selected, explored actual postpartum needs and wishes for aid from maternal nurses or midwives, especially relating to practicing self-care behaviors at home. Participants noted at least one home visit and/or telephone calls could bridge the connection with maternal nurse/midwives and mothers and would be helpful to carry out effective postpartum self-care. Design phase: Program components built on using SE sources [28] of verbal persuasion (face-to-face discussion with interpretation of the participants’ responses, supplemented by a brochure containing postpartum care information). SE enactive mastery and vicarious experiences were designed as providing opportunity for participants to demonstrate skills and role play how to perform. The intervention period was drafted to be provided in the hospital before discharge, as two 30-minute sessions to accommodate maternal attention span to avoid fatigue. The initial program consisted of two phone calls at 2nd and 4th weeks to stimulate mothers to perform self-care actions and to clarify any ambiguities. Development phase: Program content and format were reviewed by an expert panel in Bangladesh (n=5) consisting of 2 obstetricians, 1 head nurse of the postpartum unit, and 2 women’s health nursing faculty. Panel consensus were obtained with criteria of CVI >85% for each item (mean CVI was .94). Timing for the second follow-up was modified to occur at postpartum 6 weeks. Implementation & evaluation phase: The program was provided to 2 postpartum mothers by convenience sampling at DMCH, using the same inclusion/exclusion criteria for the main study. Based on participant suggestions, we determined a larger space was necessary for the small groups and decided to add using a model placenta and uterus and a newborn doll. Thus, the NLPPSC program was finalized.

The experimental group received the finalized NLPPSC (Table 1), face-to-face education in four 30-minute sessions (small group format of 3-5 mothers) over 2 days during hospital stay, and reinforcement by phone calls (duration 3-5 minutes each) at postpartum 2 weeks and 4 weeks. The control group received the hospital’s standard care (5-minute verbal instruction at the time of hospital discharge).

2) Posttest 1 (postpartum 2 weeks): The response rate for the mailed questionnaires was 76.4%.

3) Posttest 2 (postpartum 6 weeks): Mothers were contacted to identify date and location of the EPI center they would visit for infant vaccination. A trained research assistant met the participant and checked the card’s code number, provided the questionnaire, which was completed on site. Upon completion a gift voucher of 320 Bangladeshi taka (BDT; approximately 4 US dollars) was given as a token of appreciation for participation. For the control group, a postpartum self-care brochure was provided at the end of data collection along with gift voucher. The response rate of participants was 83.8%.

The data were analyzed by IBM SPSS statistics version 21 (SPSS Inc., Chicago, IL, USA). Descriptive statistics, bivariate statistics t-test, chi-square was used to examine homogeneity of variables, and paired sample t-test, and independent sample t-test were used to examine differences within and between groups. Due to attrition rate in follow up (23.5% at 2 weeks, 16.1% at 6 weeks) linear mixed model (LMM) was used to adjust for covariates and examine the treatment effects over three time periods. LMM was used for repeated measurements to explore the change in PPF, depressive mood, and maternal functioning between baseline and at 2 weeks; and baseline and 6 weeks while adjusting age, living status, and family income for baseline differences in the two groups. The LMM is a powerful advanced statistics tool that can estimate changes over time despite missing data points of a subsample, and thus, data from all 68 participants were included for analysis.

The average age of the participants was 20.71 years old. Most lived with their husbands without extra family members (61.8%), and did not work out of the home (85.3%). Monthly family income averaged 25,155 BDT (approximately 300 US dollars). The average gestational age at delivery was 37 weeks, and newborn conditions were generally good (APGAR scores of 8 and more, 92.6%). Forty-seven percent had participated in at least one parenting preparation class, and all had at least one antenatal care (ANC) visit, of whom 52.9% had received at least three ANC, the minimum number recommended by WHO.

The experimental group was slightly older (t=2.90; p<.01), and more participants tended to live with their husband, including other family members (χ²=3.98; p<.05). They also showed slightly higher income compared to the control group (t=-2.38; p<.01). There were no differences in delivery-related information or main variables between the experimental and control groups (Table 2).

For baseline PPF, both the experiment (mean±SD, 26.32±3.21) and control groups (24.68±3.18) showed a moderate level and the experiment group showed statistically significantly higher fatigue (t=-2.17, p<.05). Depressive mood (cutoff of ≥13) was noted in both the experimental (33.3%) and control group (29.6%), although there were no statistically significant differences (χ²=-.29; p=.76). The level of postpartum maternal functioning was slightly less than midpoint, and postpartum maternal functioning was not significantly different between the groups.

The process variables, postpartum self-care, postpartum SE, and postpartum care knowledge are conceptually meaningful for effects on the output variables. Within the experimental group between baseline and postpartum 2 weeks, all variables showed significant improvement, while for the control group, only postpartum SE was improved (t=-9.08, p<.01). For between group differences at postpartum 2 weeks, postpartum self-care (M_dif 10.98; t=9.49, p<.01), postpartum SE (M_dif 17.18; t=18.30, p<.01) and postpartum self-care knowledge (M_dif 8.76; t=11.08, p<.01) of the experiment group were significantly higher compared to the control group (Figure 3).

From baseline to postpartum 6 weeks, the experimental group showed significantly greater differences (p<.01) compared to the control group (Figure 3). For within group differences, there were significant mean differences in all three variables for both the experimental and control groups: For example, the experimental group’s postpartum self-care (M_dif 15.28; t=10.02, p<.01), postpartum SE (M_dif 15.04; t=19.72, p<.01) and postpartum care knowledge (M_dif 5.82; t=6.91, p<.01) levels all improved by 6 weeks. The same pattern was found for the control group, which may suggest that by 6 weeks after childbirth, a natural improvement is likely for mothers in general.

For the LMM analysis to test the effectiveness of NLPPSC intervention on PPF, depressive mood, and postpartum maternal functioning, the covariates that showed statistical significance, i.e., maternal age, living status, and monthly income, were adjusted in the equation. In addition, conceptually meaningful variables of this study, i.e., postpartum self-care behavior, postpartum SE, and postpartum self-care knowledge, were also loaded as covariates (Table 3).

Effects on PPF: The mothers who received the NLPPSC intervention showed statistically significant levels of slightly higher fatigue (β= 1.45, SE=.72, t=1.98, p<.05) over time when compared to the control. However, the treatment effects of NLPPSC intervention group at 6 weeks postpartum was statistically significant in decreased fatigue (ß= -6.17, SE=1.81, t= -3.39, p<.01) compared to the control group.

Effects on depressive mood: The estimated treatment effects did not find significant changes related to the NLPPSC intervention for depressive mood compared to the referent group.

Effects on postpartum maternal functioning: For postpartum maternal functioning, the estimated treatment effects showed that living with husband but not with other family members, was statistically significant (ß= 4.44, SE=1.30, t=3.39, p<.01) compared to women who also lived with other family members. Among other variables, more postpartum self-care knowledge was also statistically significant (ß= .37, SE=.18; t=2.03, p<.05) for maternal functioning for the NLPPSC participants. Maternal functioning at both 2 weeks (ß= 12.76, SE=1.55, t=8.20, p<.01) and 6 weeks (ß= 24.54, SE=3.12, t=7.84, p<.01) showed statistically significant improvement over time. Finally, the treatment effects at 6 weeks were statistically significant (ß= 13.72, SE=3.67, t=3.73, p<.01) in increased maternal functioning in the experimental group compared to the control group.

Ancillary analyses: There was no ancillary analysis.

The NLPPSC significantly increased scores in process indicators, postpartum care knowledge, postpartum SE, and postpartum self-care behavior; and postpartum fatigue and maternal functioning was improved among the outcome indicators. As there were no intervention studies found in Bangladesh to directly compare with this study, studies from others countries were compared in the following discussion. Discussion on outcome indicators are presented first, as they were the main study hypotheses.

Fatigue decreased over time in the experimental group compared to the control. Among the time points (within 2 days after delivery, postpartum 2 weeks, and postpartum 6 weeks), the decrease in fatigue at postpartum 6 weeks was statistically significant in the experimental group. This finding is similar with previous studies that used a psycho-educational program with telephone support to manage PPF with a RCT design [2,3]. Both studies encouraged mothers to prioritize their tasks, and plan for and engage in postpartum self-care behaviors. Although this study found statistically significant improvement in SE at postpartum 2 weeks and postpartum 6 weeks in the NLPPSC group, LMM analysis did not find direct influences of SE for fatigue. This suggests indirect effects of SE on fatigue, which may need to be further explored in future studies. This study found a mid-level of fatigue at baseline. Although using a different measurement of fatigue, a prior survey of new mothers in Bangladesh reported a lower mean fatigue score [22]. This may be related to measurement time or differences in sample characteristics, but further studies are needed to clarify.

The NLPPSC intervention was not statistically significant in decreasing depression scores, when treated as a continuous variable. This finding is similar to a RCT study of a nurse-led postpartum discharge education program to reduce postpartum depressive mood, which did not find differences at 6 weeks nor at 3 months [29]. In contrast, Giallo’s study (2014) [3] used a fatigue management booklet, a home visit, and three telephone support sessions led by a professional, and was effective in decreasing postpartum depressive mood after 6 weeks. Differences might be due to the intensity and number of contacts, as the NLPPSC did not include home visits and had one less telephone contact. Another RCT with first-time mothers [30] reported that the treatment group (12.5%) had lower depression than the control group (25%) at 6 months postpartum. This suggests the possibility that NLPPSC participants may have shown effects over a longer period of time, assuming continued practice of self-care.

This study found a high proportion of mothers with postpartum depression (33% in the intervention group) at 2-3 days within delivery. Of these mothers, 82.4% had received an episiotomy, which may be related to their perception of fatigue. At 6 weeks, the control group showed a high proportion of postpartum depression (n=9, 33% with scores ≥ 13), much higher than 9% reported in a previous study of Bangladeshi women at 6-8 weeks postpartum [31]. This may be due to demographic differences in the two studies. The previous study involved 100 multiparous postpartum mothers with mean age 25.5 years, of whom 54% had one child, whereas this study mostly included young first-time mothers, who may have been more sensitive to depressive mood.

When analyzing depression based on cut off categories, the pattern of depression over the three time points (up to 6 weeks) in mothers in the NLPPSC group steadily decreased, resulting in none at ≥13 at 6 weeks, which is in contrast to the pattern of increased depression in the control group over time (33% at ≥13 at 6 weeks). Future studies are needed to employ the NLPPSC program with more contact intensity and may be needed to be followed up for longer periods, to more accurately manage depressive moods of new mothers in Bangladesh.

In maternal functioning, maternal age and family income had a statistically significant association with increased maternal functioning. Women living with their husband only, reported significantly higher maternal functioning than women also living with other family members. As it happens that after childbirth in Bangladesh, other relatives will often come to support the new mother or she will stay at her parents’ home for some time to aid postpartum recovery, the postpartum living status may temporarily change. Thus, usual living status may not offer clear understanding of how it may be related to postpartum maternal functioning. The process indicator postpartum self-care knowledge was statistically significantly associated with increasing maternal functioning over time. The intervention was significantly effective in increased maternal functioning at 2 weeks and also at 6 weeks follow-up. These findings are supported by findings from the literature. For example, a postpartum educational intervention for first-time mothers consisting of four sessions and phone calls found improvement in maternal functioning status and self-confidence after 8 weeks [7]. In addition, another study [23] found that postpartum knowledge and efficacy influenced maternal functioning.

For the process indicators, self-care behavior, SE, and postpartum care knowledge showed statistically significant increases at both postpartum 2 weeks and 6 weeks, compared to baseline levels. This finding is similar to studies such as a nurse-led self-care intervention for primiparas after birth [32] and a prenatal and postnatal self-care educational program focused on perineal care, nutrition, and sexual activities [17], which found improvement in postpartum self-care knowledge and self-care efficacy.

Overall, the participants in this study were young age (average 20.71 years) and had completed primary to secondary level education (83.8%), which may have helped them to understand the NLPPSC content. In addition, as the majority of the mothers (85.3%) did not work out of the house they may have had sufficient time to perform self-care behaviors at home.

As such, this NLPPSC intervention offers implications for nursing practice, theory development, and research. In practice, maternal nurse caring for new mothers in Bangladesh should provide psycho-educational support using strategies for fatigue management before hospital discharge. Although brief, the NLPPSC was influential in improving the process indicators and improving fatigue and maternal functioning. This would contribute to postpartum mothers performing self-care activities for their wellbeing and their newborns. Although the NLPPSC was planned as a brief practical intervention, in reality the allocated time for intervention took more than usual as participants required more clarification. Therefore, duration of interventional time may need to be increased for future implementation. In regards to nursing theory, the conceptual framework used in this study offers a framework for generating nursing knowledge theoretically. Finally, this study identifies recommendations for further studies. For example, exploratory studies on SE and postpartum depression, and developing culturally appropriate postpartum self-care instruments for use in Bangladesh would be beneficial.

Limitations: This study has some limitations. The low values of internal consistency for some instruments may have been a source of measurement error. This may be related to the fact that although the questionnaire was administered face-to-face to facilitate understanding and completion, the timing (immediately after childbirth) may not have been optimal. For example, ‘I am a good mother’, or ‘I hold my newborn comfortably’ may have been sources of measurement and response bias. Another limitation is that participants were recruited from only one public hospital (DMCH), and convenience sampling might be a threat of selection bias. Furthermore, as this study focused on women delivering at the hospital, it cannot be applied to the sizable proportion of Bangladesh women who deliver at home, and future studies are needed for how to pragmatically and effectively reach this needy population.

In conclusion, the NLPPSC was effective in decreasing PPF and increasing postpartum maternal functioning of first-time mothers in Bangladesh. However, it was not found to be effective in decreasing postpartum depressive mood. The NLPPSC was effective in improving SE, self-care behaviors, and postpartum care knowledge at both 2 weeks and 6 weeks. Postpartum care knowledge was effective in improved maternal functioning, the supports implementing the NLPPSC for new mothers following delivery. Although this study focused on hospital-based births, a relatively small proportion of all births in Bangladesh, this nurse-led program was found to be feasible and can be easily provided for first-time mothers who have hospital births. Further studies are needed to expand the NLPPSC in terms of intensity and/or frequency of contact, as well as consider follow up measurement for a longer period. Further exploratory and interventional studies are also needed on postpartum depression, and development of postpartum self-care instruments based on Bangladesh context.