4.1 | Response rates and participants
All managers of the project municipalities (n = 20) responded to the organisational survey, yielding a response rate of 100%. Data are provided in Table 2.
→ Insert Table 2
The health visitor survey was distributed to 368 health visitors at baseline, and 351 health visitors at follow-up. Figure 2 presents a detailed flow diagram of the inclusion and attrition of participants in the survey waves. The response rates in the control arm and the intervention arm were 82% and 89% at baseline respectively, and 81% in both arms at follow-up, and 242 health visitors completed both baseline and follow-up, of which three changed trial-arm. The results are based on follow-up responses, except for Fig. 6b in which self-reported assessment of breastfeeding support is based on the complete responses in the intervention arm.
Data from health visitor records were obtained from all municipalities (n = 20).
In dialogue meetings, all local representatives participated if possible (n = 13). In the seven meetings analysed, 12–13 local representatives participated in each. A coding tree of the analysis can be seen in Fig. 3.
Sixteen health visitors participated in the three focus groups, with four to seven participants in each. Four out of 10 municipalities participated with only one participant. Participants had between six months and 31 years of experience as a health visitor, two held an International Board Certified Lactation Consultant (IBCLC) degree, and one had not participated in the training programme (data not shown). A coding tree of the analysis of focus groups can be viewed in Fig. 4.
In the interviews with mothers, eight mothers and one father participated. Five mothers were eligible for intensified intervention, and four had accepted the offer. Mothers age ranged from 21–32, and the infants’ age at interview ranged from < 1–9 months (median 4 months). Four had either primary school or vocational education, one had secondary education and three had tertiary education as highest educational attainment. All participants were in a cohabitating relationship or married, three were primiparous and the rest multiparous. Five interviews were conducted face-to-face and three via telephone. A coding tree of the analysis of interviews can be found in Fig. 5.
4.2 | Implementation – Was the intervention delivered as planned – to whom, what elements and were adaptations made?
4.2.1 | The training programme
Of the 134 health visitors in the intervention arm responding at follow-up, 90% (n = 120) responded to have participated in full, while two (1%) had participated in parts of the breastfeeding support training programme (data not shown). Twelve (9%) had not participated at all. Reasons for not participating in the training were ‘not employed at the time’ (n = 7) and ‘leave’ (n = 5) (data not shown). Of the 12 non-participants, 11 reported having received one-to-one education from a colleague at the least. The remaining health visitor had not received any training (data not shown). Of the respondents in the intervention arm, 83% felt that the training had provided them the ability to deliver breastfeeding support in accordance with the intervention (Fig. 6a).
4.2.2 | The breastfeeding support
Most health visitors in the intervention arm reported that it had been their intention to guide and that they had in fact guided in accordance with the intervention, and that they furthermore intended to continue guiding in accordance with the intervention (Fig. 7a), underlining the health visitors’ acceptance of the intervention.
In the focus groups, health visitors expressed that the intervention breastfeeding support was aligned with their previous support, however simpler and therefore clearer. The four key messages, they expressed, helped them to structure the support and remember the most important things.
“We really want to support breastfeeding, so we also thought it was valuable if we could do something to increase the breastfeeding rate. That’s what motivates us […]” (Health visitor 1)
In Fig. 7b, responses to questions regarding specific theory- and evidence-based elements that was part of the intervention is presented. Also here, most health visitors (90–98%) reported that they ‘often’ or ‘almost always’ had delivered support in alignment with the intervention (Fig. 7b). This was confirmed by qualitative data, where the health visitors described that they had tailored the information to fit parents’ needs, and that they had delivered the postcard, the check-the-milk pamphlet and the project website to all the families they encountered, furthermore confirmed by mothers participating in the interviews.
“I think that the simpler it is with such a card (the postcard), the easier it is for them to comprehend. So even if you ask if we support the same way as usually? Yes, to some extent, but it is a good tool because we get to add something illustrative, and I think we lacked that before. And it helps underline the importance of it.” (Health visitor 2)
The exception was the dialogue sheet. During the first dialogue meetings, local representatives had expressed that health visitors found it difficult to implement and use, especially if they had to use it after birth when there were many infant specific tasks to be completed during visits and little time. Therefore, the intervention developers and the local representatives had agreed to only use the dialogue sheet in pregnancy visits. Still, the delivery continued to be incomplete; just a couple of health visitors in focus groups expressed to have found the sheet useful, whereas the remaining participants believed that the sheet just duplicated the obtainment of families’ individual histories in the electronic record system usually done during pregnancy visits. Interview findings also confirmed this, as only two mothers had encountered the sheet.
In municipalities that had not previously had pregnancy visits as part of their usual care, pregnancy visits had been in demand as expressed by local representatives in dialogue meetings and by health visitors in focus groups. This made them easy to implement. Health visitor records showed an increase in most intervention clusters (Table 3), thereby confirming the easy implementation. However, increases were also found in most control clusters (Table 3).
→ Insert Table 3
In interviews, a few mothers told of situations where the delivered breastfeeding support had acted against the intended intervention. For instance, a few mothers had been guided to try to space out feedings to counteract an infant’s rapid weight gain, and another had been told that it was normal to experience pain related to breastfeeding in the beginning.
“When I asked my health visitor, she said that it was normal for breastfeeding to hurt in the beginning, because my body isn’t used to it and things like that. But that it at least should get better. It did when I was almost finished breastfeeding. But in the end, it ran out (the milk). […].” (Mother 3)
4.2.3 | The intensified intervention
Another adaptation found across the qualitative data was the intensified intervention. During dialogue meetings, representatives expressed that health visitors in their teams wished to be able to adapt the delivery mode to include text messages in addition to the planned telephone calls, and that some had found it difficult to find a good way of voicing the reasons behind the offer of the intensified intervention to mothers in the target group without stigmatising them. Mothers confirmed the adaptation of the communication mode and expressed that they had received the planned intervention. Findings from the focus groups expanded these findings in that some health visitors found it easier than others to deliver the offer of the intensified intervention. Finding one’s own way of voicing the offer seemed to be key. Health visitors also talked about situations where they had refrained from offering the intensified intervention. In these situations, they had made a judgment call to prioritise other, more pressing circumstances than offering the intensified intervention, or they anticipated that a specific family would not have the resources to focus that much on breastfeeding.
”I think it depends on how well you ’sell it’ (the intensified intervention), and I think that some considers: ‘In this particular family, there is no point in even introducing it’. At least in our team, when we discussed it yesterday, some of my colleagues had no families in the intensified intervention, because they had reasoned: ‘They are not able to… It will be too much’. And so, the families may not even get the offer”. (Health visitor 1)
Furthermore, health visitors expressed some frustration about the necessity of developing their own unique system to deliver the proactive calls in alignment with the plan and that they would sometimes adapt the scheduled contacts to fit their calendars, for example when a holiday would come up. Some would in such a case offer the contacts before or after the scheduled contact to avoid having colleagues take over, while others would leave it up to a colleague to handle. Health visitor records were highly ambiguous, however had a tendency towards convergence in that the mean number of telephone contacts increased in seven out of the 10 intervention clusters after the implementation of the intervention, whereas this was the case for less than half of the control clusters (Table 3). This also suggests that contamination regarding the intensified intervention across trial arms was unlikely.
4.2.4 | Postponing the introduction of solid food
Implementation of the postponed four-month visit was reported by managers in the organisational survey to have been successful, although sometimes difficult to implement due to barriers related to families’ context to be elaborated below (data not shown).
4.3 | Mechanisms of impact – What were the reactions and interactions with the intervention – among health visitors, among families, and were there any adverse effects related to the intervention?
4.3.1 | The breastfeeding support
Most health visitors (71%) with complete responses, felt that their breastfeeding support had improved since the implementation of the intervention, while some felt that their support had not changed (28%) or even worsened (1%) (Fig. 6b). Health visitors largely believed the intervention was superior to usual breastfeeding support in establishing breastfeeding to ensure the infant’s thriving (76%, Fig. 8a) and in strengthening parents’ belief that they could breastfeeding exclusively for four months (62%, Fig. 8b), while 23% and 37%, respectively, found it equal to usual breastfeeding support. Regarding relationship formation with families and number of inquiries from families since the intervention commenced, health visitors largely reported the intervention to produce similar effects as the usual breastfeeding support (78% and 69%, respectively, Fig. 8c-d).
Confirming these findings, health visitors in focus groups expressed that the intervention fit well within their usual practice and that the training programme had enabled them to apply in-depth knowledge about for example mechanisms for milk production to their existing support.
“I feel that I had a lot of my knowledge confirmed during the training programme and gained some. I think it was good and I now provide the appropriate information depending on the family’s needs.” (Health visitor 8)
Moreover, some health visitors expressed in focus groups to have focussed more on laid-back breastfeeding positioning and on the skin-to-skin approach when supporting families. All expressed satisfaction with the intervention materials, i.e., the postcard, Check-the-Milk pamphlet etc., which they considered appealing and felt had been missing, and further facilitated the provision of support via telephone. The postcard with the four key messages on the front and a website login on the back structured and simplified the breastfeeding support in that it helped health visitors remember to focus on the important things. Confirming this sentiment, in the survey, most health visitors reported that providing breastfeeding support in alignment with the intervention was simpler than before (46%) or the same as before (45%) (Fig. 9). Mothers in interviews were happy with the breastfeeding support they had received, except for the few accounts mentioned above, where the advice and support provided had been in discordance with the intervention. Still, in those cases the dissatisfaction seemed brief. Using the skin-to-skin approach to solve breastfeeding problems was advised in alignment with the intervention and had been successfully applied by a couple of mothers.
4.3.2 | Interactions with the intervention materials
As described in the previous section about implementation, in the qualitative data, health visitors’ reactions to the dialogue sheet were mostly frustration. They generally felt micromanaged to do a task they believed they already covered. A few health visitors disconfirmed this by having found the sheet useful and to facilitate a deeper understanding for the families’ situations, in convergence with accounts from the few mothers who had encountered the sheet.
Health visitors in the focus groups and dialogue meetings found the intervention website useful and expressed that sound, evidence-based knowledge about breastfeeding in Danish had been in demand. They were surprised that to their experiences, families seemed to not have utilized the website much. Website data showed that there had been more than 35,000 visits during a one-year period and confirmed that most logons (86%) had spent up to one minute (Fig. 10), which might cover health visitors’ introduction of the website to families, and the families’ independent logons to give the website a quick browse. However, further 11% of visitors spent between one to 10 minutes on the visit, while the remaining 3% spent 10–30 minutes or beyond (Fig. 10). Thus, even though website data confirmed most visits to be short, a considerable proportion, corresponding to more than 5,000 visits during a one-year period, lasted more than 10 minutes. This suggests that families with interests, utilised the website accordingly.
This heterogenic use of the site was supported by interview data, as most mothers had taken a quick look, but also that some had spent a lot of time on it. The mothers generally expressed that the many videos on the website offered useful support in a timely manner. Viewing videos was expressed as easier than reading under the circumstances of having an infant on the arm, although some mothers with high educational attainment found other available material on the website useful, for example texts and podcasts.
“Yeah, so if there was something we doubted, then he (the father) was the one reading about it and then passing on the information. So, he has used it (the website).” (Mother 6)
Data from the website supported this inclination towards the videos with the ten most visited pages all being videos (Table 4). During dialogue meetings and in focus groups, health visitors were disappointed about the lack of a search option on the website and speculated that a smartphone application would have been used more, however, interview data offered no insights on the matter. Health visitors expressed a wish for continued use of the project website and postcard after the project finished, and interview data confirmed this, as mothers who had found good use of the intervention website deeply wished to have continued access and wanted everyone to gain access.
→ Insert Table 4
4.3.3 | The importance of preparation for breastfeeding
In focus groups, health visitors expressed that a pregnancy visit was an important element in the intervention, providing an occasion to form a relationship with the families and to scope families’ wishes and intentions towards breastfeeding before the arrival of the infant after which challenges and emotions could lead to forgotten intentions. It provided a unique opportunity to tailor breastfeeding support by encouraging and guiding families based on their formerly expressed wishes and their current needs.
“I find it important when we are on pregnancy visits to ask them when you hold the postcard: ‘What do you think? Do you feel like breastfeeding? Did you talk about it?’ […]. So that it is a choice and not something that I tell them to do. I think it is important, because then you have a chat about if they want to and how strongly they feel about it. […]. I think it is so important that you kind of balance it out: ‘What was it that we talked about back then, when they didn’t have those sore breasts and cracks and bleeding and were tired’” (Health visitor 1)
Data from interviews confirmed this, and one mother explicitly expressed how the health visitor’s knowledge of the mother’s expressed wishes from before things got challenging, led to encouragement to try again at the right time, consequently facilitating continued breastfeeding that would have otherwise been discontinued.
4.3.4 | The intensified intervention
The intensified intervention was believed to be time-consuming in both focus groups and dialogue meetings, although conversely, in a few exceptions, health visitors expressed that it preserved time. The health visitors found that the fixed structure of delivering the extra care in the intensified intervention needed development of a systematic approach for it to fit into their individual planning. Health visitors’ reactions to the intensified intervention covered a spectrum of ‘time-consuming waste of time’ to ‘it works really well’. In the former, it was believed that families who experienced needs would reach out themselves, thereby rendering the proactive approach superfluous. In the latter, health visitors had experienced being included in discussions of discontinuation of breastfeeding and believed to have been able to postpone the cessation due to the intensified intervention. They expressed that they were able to prevent discontinuation in multiparous families who ‘usually’ started bottle-feeding at certain time points. Furthermore, in families with concerns about sufficient milk supply due to the sudden emergence of more frequent feedings, they could articulate that infants increase the milk production according to their demand and that this probably was the reason for the changed feeding pattern. If nothing more, they said, they would have at least provided a pat on the back and a sense of security, and thus afforded parents an extra motivation to keep on breastfeeding.
[…] I have been contacted by parents in the intensified interventions: ‘We just want to ask you what you think – would it be okay to bottle-feed tonight?’. […] I have been able to prolong the breastfeeding by informing them: ‘Well, a lot is happening with your child emotionally and that’s why the baby wakes up and cries out for you more – it’s not just about the breastfeeding’. […] Or: ‘The baby eats faster now and gets more milk at a time, and that’s why the feedings are quicker now.’
(Health visitor 11)
Data from interviewed mothers expanded focus group findings with information about how the frequent contacts from the health visitor made them feel that it would be no inconvenience if they themselves needed to contact the health visitor.
Health visitors believed that the proactive telephone calls could potentially preserve time in cases where they would usually provide a needs-based visit ‘just in case’. In the future, they foresaw using the telephone calls to explore needs for a visit instead. They agreed that although the intensified intervention could be warranted in certain cases, the responsibility for identifying which should be up to the health visitors’ professional judgment.
4.3.5 | Postponing the introduction of solid food
The postponement of the four-month visit was discussed in both dialogue meetings and focus groups. In dialogue meetings, the local representatives were worried that a postponement could result in inadvertent consequences by delaying the detection of inadequate motor-skills in infants. Expanding this, health visitors in focus groups expressed that in practice it could be difficult to postpone the visit, because families, especially multiparous-, expected a four-month visit, or because some families, especially primiparous-, were eager to begin introducing solids into the infant’s diet. The organisational survey offered the managers’ view of the uptake, which was overall positive, even with expressed wishes to keep the practice after the intervention ended, and health visitors in the focus group overall found the postponement meaningful. Interview data were silent on this matter.
4.3.6 | Possible adverse effects
During dialogue meetings and in focus groups, concerns were raised about the risk of stigmatising mothers in the target group for intensified intervention by voicing the eligibility criteria. However, this was refuted, as the interviewed mothers did not express to have felt stigmatised. Some had not fully comprehended the motives for the offer. Confirming this, in a focus group one health visitor said that her manager had helped the team overcome the barrier for offering the intensified intervention by telling them to simply leave out reasons for the offer. Another concern raised in focus groups, was that the intensified intervention might inadvertently introduce insecurities in mothers if they interpreted telephone calls as a form of monitoring. Again, interview data disproved this, as mothers said that the close follow-up had stimulated a sense of security.
”Yeah, I thought it was super nice (receiving the proactive calls). I remember being a bit nervous about ‘oh no, will I be judged for being… you know… young’. But she was super nice, and it was quite comforting to know, like… Because she ensures me that she is only trying to help me. […] It felt very safe for me, I think […].” (Mother 2)
4.4 | Context – What contextual factors were important for the intervention to be implemented and delivered as planned – what impeded or facilitated the delivery of the intervention and the mechanisms of change?
4.4.1 | Organisational context
In both trial arms there had been shifts in management and turnover of staff during the project period (Table 2). Despite statistical insignificant differences between groups, proportions were visually higher in the intervention arm with 50% of management and 18% of staff had been replaced since project start, compared to 30% and 12%, respectively. In focus groups, health visitors expressed the importance of management for the implementation, for instance by adapting project tasks such as recruitment to fit the context or by prioritising resources to the project. Participation in the training programme would instil a sense of ownership in the health visitors. Therefore, newly employed health visitors could end up impeding intervention implementation and delivery and some health visitors did indeed point out variation in colleagues’ attitude towards the intervention, they expressed. In the survey among health visitors, only a few had not participated in the training programme as mentioned above and responses suggested that most had felt some extent of managerial support in the intervention (Fig. 11a), thereby disconfirming that lack of managerial support and sense of ownership might have threatened the intervention delivery. Moreover, health visitors largely reported that they had had time to familiarise themselves with the intervention and had read most of the intervention manual (Fig. 11b-c), indicating that the organisations had allocated the resources necessary to implement the intervention.
Two other organisational factors that affected the implementation were workload and financial resources. In focus groups, health visitors expressed how the workload of the intensified intervention would be distributed unequally within teams due to differences in the population composition in each health visitors’ assigned district. Data from the organisational survey showed that almost twice as many municipalities in the control arm had health visitors appointed especially to handle the care of a specific group, for instance young mothers or mothers of low educational attainment (control group: 70% vs. intervention group: 40%), despite the difference being statistically insignificant (Table 2). Financial cutbacks to the health visiting programme put pressure on available resources and on the individual health visitors, thereby impeding the implementation, as it was pointed out by focus group participants.
4.4.2 | The impact of local representatives
Data from the survey among health visitors showed more frequent discussions with colleagues about breastfeeding in the intervention arm than in the control arm (Table 5), which adds information to the findings from the organisational survey, in which managers reported frequency of team meetings (Table 2).
→ Insert Table 5
In dialogue meetings, it became apparent that resourceful local representatives in team meetings and discussions would facilitate implementation and delivery of the intervention by facilitating adaptations to fit practice. Confirming this, health visitors in focus groups pointed to the same aspects being of importance and complemented by expressing that the local representative facilitated a maintained focus on the intervention and gave peptalks when it was needed, and furthermore how the direct link between local representatives and the intervention developers would facilitate a sense of ownership because questions could be raised and quickly answered, or adaptations made continuously.
“Well, I think that the occasional peptalk... And then also that the distance isn’t that long. We have experienced that something was changed, for example: ‘It just doesn’t make sense to call this mother – it feels stupid, because she just breastfeeds […]’. Like, she’ll (the local representative) then ask if it’s okay to just send a text message. And then that was okay. Yeah, then you kind of feel like we have influence.” (Health visitor 16)
Contrary, when local representatives did not provide peptalks or put the intervention on team meeting agendas, the focus on the intervention faded out.
4.4.3 | The project infrastructure
The intervention’s embeddedness in a large trial was frequently discussed during dialogue meetings, as the eligibility criteria for recruiting families to the trial in general and specifically to the intensified intervention placed a double burden on health visitors, because they also had to deliver the intervention. Focus group findings confirmed this but further expanded it informing that the project infrastructure ambiguously acted as a barrier for implementation due to the confusion and frustrations it posed, and as a facilitator because it ensured focus on the intervention and helped health visitors remember, for example, to deliver intervention materials to families.
“[…] Because I think that some of my focus comes when I sit with the family and must present and tell them: ‘We have this breastfeeding project’, and then the postcard and the website are presented, and we talk about enrolling them and what it encompasses. Then we’re already caught up in talking about breastfeeding […]” (Health visitor 13)
4.4.4 | The influence of mothers’ context
In focus groups health visitors expressed that the mothers’ context impacted the intervention. This was confirmed by mothers in interviews. For instance, family members could support or argue against the efforts involved in making breastfeeding work, or a partner could provide a due diligent push to facilitate successful breastfeeding.
”If he hadn’t pushed me to express my milk… Or not push, but support me, then I would have stopped entirely and gone with formula” (Mother 2)
Similarly, members of ‘mother’s groups’ could support each other in breastfeeding, or conversely praise the advantages of infant formula and thereby end up creating a ripple effect. As expressed by a health visitor:
“I have also experienced the impact of the mothers’ group, specifically if they have talked about their experiences with bottle feeding […]. Then when I visit, they tell me: ‘We have started bottle feeding’, and I have not been involvement. I then ask: ‘Where did you get the inspiration from?', ‘We heard from the mothers' group that it was really good’. So, I think that the mothers' group also has an influence on what people choose. And it also differs what kind of mothers' group you are in. Is it a group where everyone breastfeeds? Or is it a mother's group where a large proportion bottle feed? So, it is at least something that I also think we are up against.” (Health visitor 7)
Another contextual factor was parity. Parity could impede the postponement of the four-month visit, as mentioned above. Managers of intervention municipalities confirmed that the families’ context could impede the postponement. Mothers expanded multiparity to act as a facilitating factor for mechanisms in instances where previous experiences provided the right prerequisites to prepare for breastfeeding and seek out as much help as possible, for example from the hospitals’ antenatal care and during pregnancy visits. Conversely, repeated negative experiences with breastfeeding could cause a family to give up.
Additionally, the health care context impacted the intervention. Mothers expressed that overall, the breastfeeding support provided in hospital was consistent with the support provided by the health visitors, however, they also told stories of situations in which the support counteracted the intervention, for example when health professionals handed out nipple shields to alleviate pain or instilled notions about the necessity of supplementing with formula until the milk had come in.
” […] We were at the hospital on day two (after birth), and then a couple of nurses came in, and then she (the infant) was fed some formula, because (they said) it can take a while for your milk to come in properly.” (Mother 1)
A mother’s resources, e.g., psychosocial-, educational- and mental resources, would facilitate the intervention as it enabled her to set the intention about breastfeeding and thereby receive help from her partner, cope with challenges, seek out help from others when needed and critically appraise advice and support received. All in all, having resources enabled a mother to succeed in breastfeeding. However, the intervention was designed to enhance a mother’s self-efficacy, among other things by identifying people in the social surroundings able to support with good breastfeeding accounts and to build action competence to enable the overcoming of obstacles, by for example providing knowledge and evidence-based support. Thus, when mothers with less resources were unable to overcome challenges they faced, it may point to parts of the intervention not being delivered as intended.