Effect of the Growth Assessment Protocol on the DEtection of Small for GestatioNal age fetus: process evaluation from the DESiGN cluster randomised trial

In total, 55 qualitative interviews were conducted including 27 interviews with clinical leads (22 GAP leads at implementing sites and 5 service leads at standard care sites) and 28 interviews with frontline staff (implementing sites only). The interviews took place between February 2018 and May 2019. All interviews were conducted following initial implementation at cluster sites. Three interviews were conducted by phone. We were unable to arrange interviews with either lead or frontline sonographers at one implementing site, but interviews were conducted with all the intended professional groups at all other sites.

Findings

Firstly, we report findings from our qualitative inquiry on two key implementation outcomes, ‘acceptability’ and ‘feasibility’, and outline the influence of context on GAP implementation during the DESiGN trial (see Additional files 7 and 8 for more extensive qualitative data). Verbatim quotes are included, but text presented within square brackets has been summarised for brevity or clarity. These qualitative data provide context for the quantitative data on measures of implementation strength, which we then present.

Acceptability: GAP lead and frontline staff perspectives on the potential value and clinical effectiveness of GAP

Within the CICI framework, staff are considered ‘individual implementation agents’ by virtue of being ‘actively involved in…administering or implementing an intervention’ [29]. Staff perspectives about the value and effectiveness of GAP were important throughout the implementation process. We found evidence of differing views amongst frontline staff and GAP leads about whether GAP was beneficial for women or likely to increase detection of SGA. Some felt the GAP approach was promising:

I generally welcomed it, I was excited about it, I thought it was…a nice rigorous way of decision-making.

(SC21, GAP Lead, Site 10)

Others were more negative:

Generally, I think the majority of us don’t really want to [implement GAP]. We don’t really understand why we are doing it.…

(HP12, Frontline staff, Site 11)

I didn't welcome it. I was a bit sceptical of it and maybe this was influenced from speaking to some of my colleagues….

(SC31, GAP Lead, Site 9).

GAP leads had often attended ‘train the trainer’ events held by the Perinatal Institute months before implementation began. They also learned about GAP through clinical networks, conferences and publications. Frontline staff usually learned about GAP through face-to-face or online training provided by GAP leads. However they felt about GAP, respondents usually agreed it was important to do the DESiGN trial to address this question (see Additional file 7).

Acceptability of GAP implementation: GAP lead and frontline staff perspectives

Frontline staff felt that GAP was a useful intervention because plotting foetal growth onto a customised chart was straightforward, and they hypothesised that this would improve detection of SGA, reduce variations in care by standardising practice and possibly reduce routine interventions. Staff felt that customised charts were acceptable to the women they cared for, and that both the work-based ‘face-to-face’ and online (Perinatal Institute e-learning package) trainings were good. However, providing GAP training was problematic for organisations (see ‘Feasibility of GAP implementation: staff perspectives’).

Whilst some GAP leads believed GAP enhanced standardised assessment of foetal growth, others expressed concerns about a variety of issues, including plotting errors (see Additional file 7). A key issue for both frontline staff and leads was that GAP identified potentially large babies without a corresponding care pathway. Staff felt this could increase women’s anxiety and potentially lead to increased interventions.

[GAP leads to identification of more large babies and therefore] lots of intervention that may not be warranted.

(HP3, Frontline staff, Site 7)

I’d say probably half of the women are coming above the line... I just think a lot of women come up quite high on the chart and it can be quite worrying for them

(HP91, Frontline staff, Site 10)

Leads noted that GAP could also create clinical confusion. Examples included a new focus on estimated foetal weight over foetal abdominal circumference, as sonographers had been taught, and there were differences noticed between what was taught in GAP training and recommendations in implementing site protocols (see ‘assessment of fidelity’).

The sonographers were very uncomfortable with not allowing the AC [fetal abdominal circumference] to drive the decision around further scanning….…[they]…felt that they might get blamed if, you know, the EFW [estimated fetal weight] is normal but the AC is slightly dropping and they didn’t act accordingly

(SC17, GAP lead, Site 11)

Overall, frontline staff found the GAP intervention to be acceptable, despite some reservations, and GAP leads agreed there were benefits, but GAP leads were also aware that GAP introduced new clinical complexity. Frontline staff cited access to good quality training, ease of implementation, and benefit to women, all of which contributed to a sense that GAP was potentially a useful approach to improve SGA detection. Some aspects were considered less satisfactory by frontline staff and GAP leads; these included identification of larger babies, measurement errors, clinical uncertainty, and potential to increase interventions. The ‘acceptability’ data are drawn from 43 interviews and 180 extracts, reflecting good saturation of domain codes, with consistency across sites and participant groups (see Additional file 7).

Feasibility of GAP implementation: GAP lead and frontline staff perspectives

Whilst clinical staff generally considered GAP to be acceptable and beneficial, the feasibility of actually implementing GAP often appeared conditional. For example, training was feasible if there were sufficient staff to provide cover; GAP could be implemented if a dedicated ‘champion’ could focus on this. Key feasibility concerns identified by frontline staff and GAP leads included the anticipated and observed increase in ultrasound scans required, which impacted on sonographer breaks, clinics running late, and led to ‘breaches’ of other clinical targets:

We have a lot of patients who come and usually we are full, we are booked completely, and to fit the patient within three working days is very, very difficult. Sometimes we have to scan during our lunchtime which is not ideal at all but then otherwise we breach the time…

(HP41, Frontline staff, Site 9)

GAP leads also reported how shortages of scan slots and sonographers led to decisions with lower concordance between local and GAP guidelines:

Er, their [GAP] BMI [body mass index, referral point] is, er, lower than ours, so we would only refer if they were 35 and over. Just because all of our women…we’d just be referring everyone

(SC20, GAP lead, Site 7)

Frontline staff also reported that using GAP sometimes meant appointments took longer, due to plotting time, having to hunt for charts or missing information or to seek additional advice or a second opinion, and sometimes this meant that clinics over-ran. Frontline staff and leads also reported problems with accessing information technology (IT, lack of printers or computers in hospital or community settings) or equipment.

During implementation, GAP leads were concerned about the feasibility of providing face-to-face training or releasing staff to undergo online training.

…it’s just not feasible for myself and my colleague to train [hundreds of] midwives between the two of us, when we’re not being given any allocation of time…

(SC06, GAP lead, Site 9)

We have had drop-ins where we get people to try and sit and do their online training. And I think that has been the biggest issue, as far as I know we’re still not at the level that we should’ve… had with the online training.

(SC21, GAP lead, Site 10)

Frontline clinicians and GAP leads also reported software duplication and non-alignment between GAP training and site protocols, practice or software:

So we are using [ultrasound generated charts] in conjunction with the GAP charts still…at the moment, they are running alongside each other which at the beginning did generate some problems…

(HP23, Frontline staff, Site 11)

…the [Trust] IT system doesn’t link in with the Perinatal Institute’s GAP GROW, which is possibly the case for a lot of people’s IT systems…So you end up with lots of bits of paper [laughs] because it’s a bit of a hybrid, and probably every trust has to work out their own little system for that.

(SC04/SC07, GAP lead, Site 8)

Despite these feasibility concerns, GAP leads and frontline staff were committed to improving detection of SGA and worked hard to find solutions to the issues they had identified; for example providers increased scanning capacity ready to introduce GAP, leads planned ahead and wrote business cases for additional sonography staff and resources and sought to give staff protected time to do e-learning (see Additional file 7).

Describing the context of implementation and how it interacts with the implementation process

We conducted a further analysis to examine implementation as a chronological process and document the impact of context; we separated this data into micro, meso and macro levels, using the definitions provided in the CICI framework (see Additional file 8) [29]. Contextual factors affected the early stages of implementation (‘exploration and decision to adopt’ and ‘planning/initial implementation’) and continued to impact during ‘full implementation’. After ‘full implementation’, there were no new observations about the influence of context, perhaps because most interview questions were focused on the implementation phases, but the concerns identified appeared likely to have an impact on the longer-term sustainment of GAP within implementing organisations.

How context affected early implementation

During the planning stage, the external ‘macro’ context appeared influential; GAP leads are regularly referred to targeted national campaigns and policies designed to raise awareness about the UK’s relatively high stillbirth rates [15, 36, 37].

There were multiple triggers, some of them being our own local experiences in reviewing cases where there had been adverse outcomes…That was one trigger. Then the growth assessment guidelines from [RCOG]…was another trigger. Then the Saving Babies’ Lives processes also needed us to look at ways of streamlining our care. Those are the kind of things I would say made us choose [to adopt GAP].

(SC12 and SC22, GAP lead, Site 8)

Lead clinicians at ‘standard care’ sites also identified the national policy context as influential. Post-randomisation, clinicians at these sites also needed to identify ways to respond to the same policy guidance without implementing GAP or using customised charts (Additional file 8).

Whilst there was a national consensus that current practice to detect SGA and prevent stillbirth was problematic, clinicians differed in their views of whether it was possible to implement GAP without additional resources. Two provider organisations were randomised to implement GAP but did not implement the intervention, and financial considerations appeared to have influenced these decisions:

…our Trust is under pressure with finances, so they are cutting down everything. So, that is why the new management didn’t want to spend this additional [money] for the GAP programme. It’s not my decision, it’s a management decision.

(SC1, Clinical lead, Site 13)

…the Research & Development department did try their best, but then when they saw there was no funding, they couldn’t see any value in [participating in the trial]…[but] we see the benefit, the benefit of the trial.

(SC14, Clinical lead, Site 12)

During the early implementation stage (see Table 2), interactions between context and implementation occurred mainly at the ‘meso’ (organisational) level (see Additional file 8). Organisations experienced delays and barriers, mainly due to staff shortages, pressures of work and problems identifying GAP leads with sufficient capacity to invest the time needed. Interviewees also identified strategies and contextual factors that had helped with implementation, such as supportive relationships with colleagues and interdisciplinary working, which meant that staff helped each other to understand and implement new protocols.

Int: Did you find that there was anything that made it easier for you, or something that was supporting you to cascade training to your colleagues?

I think the support that we got from [colleague 1] and also [colleague 2] was very, very helpful. And [colleague 1] was very visible to us and …very willing to answer a question…

(HP5, Site 11)

Impact of context on ‘full implementation’

Context had a notable impact on full implementation (see Table 2), mainly at either the ‘micro’ (individual) and ‘meso’ levels (see Additional file 8). At the ‘micro’ level, it became clear how the acceptability issues identified by staff, and discussed earlier, might impede implementation. The impact of changes to usual practice brought by GAP could be seen, as clinicians began to experience dissonance between what they would previously have done and what they should ‘now’ do, according to GAP:

…our protocol has been historically –[for] 50 years, ever since ultrasound assessment has been [used], practice has been [to scan again in] four weeks, so bringing it down to three weeks …is a bit hard

(SC22, GAP lead, Site 8)

…you know, [at] 36 weeks, and you measure 33 centimetres, your mind tells you, I have to scan this woman! [laughs] But the chart tells you, you don’t need to. So for the midwives it’s a bit of …you know, they have to really feel confident that actually yes, it’s working

(SC25, GAP Lead, Site 10)

On the other hand, staff were also motivated to implement GAP to improve care:

…there were babies being missed [before] and the outcomes were not good for those babies, so [GAP] definitely needed to be implemented.

(HP74, Frontline staff, Site 7)

Context could either impede or favour implementation (see Additional file 8); the micro-contextual analysis showed how staff reflected individually on ‘missed’ SGA cases, and this meant they were receptive to an intervention which might improve care. The work undertaken by staff in response to the organisational (‘meso’) context particularly demonstrated the additional time and workload implications as managers and senior leads attempted to resolve day-to-day implementation obstacles. There were also examples where executive boards or directors had approved extra funding for staff or capital expenditure, allowing implementation to proceed, but staff felt these only partially addressed the shortfalls identified.

Measure of implementation strength

Implementation strength was measured using guidelines on screening for SGA which were collected from all five implementing sites, training records received from the GAP provider and a review of 595 maternity records for babies born during December 2018–February 2019. The demographic characteristics of the women whose maternity records were reviewed are summarised in Additional file 9.

Fidelity

All five sites achieved the target of training > 75% staff members from each professional group in face-to-face methods on the GAP intervention; but only one site achieved the e-learning target (Table 3). This may be explained by the acceptability and feasibility findings on training, whereby some members of staff felt that the e-learning training was unnecessary, and GAP leads found it difficult to release staff from clinical duties for additional training.

Table 3 Overall assessment of implementation strength

The assessment of concordance with GAP guidelines identified wide variation. The guidelines from two of the five implementing sites were assessed as having high fidelity to the recommended GAP guideline; one site guideline had low fidelity (Table 3). Low or medium fidelity was usually caused by adaptations made to the local guidance, prioritising women with some risk factors over others and reducing the frequency of (or removing the recommendation entirely for) ultrasound scans offered to those women. This finding may be partly explained by the qualitative data on acceptability and feasibility of implementing the intervention, in particular the finding that maternity services were concerned about a shortage of ultrasound appointments and sonographers. A detailed breakdown of the deviations from the GAP recommended statements is available in Additional file 10.

Maternal risk of SGA, assessed as per GAP guidance, was compared to that allocated by the assessing clinician for agreement. There was agreement achieved in 84.9% (n = 505) cases. Of those women in whom there was disagreement, 19 women (21.1% of those with disagreement by GAP guidelines) had appropriate risk stratification according to local protocols. Results by risk status and by site are detailed in Table 4.

Table 4 Outcome of the assessment of risk stratification, comparing clinician assessment to GAP and local recommendationsReach and dose

With regard to the measures of dose and reach assessed by notes review, the proportion of women in whom the target was achieved is presented for each site in Table 3. For these measures, site 7 was consistently the lowest scoring site, and site 8 was the highest scoring for three of the four measures. There was evidence of a difference in the dose received between nulliparous and multiparous women (38.1% vs 21.9%, p < 0.001, Table 5) across all 5 sites implementing GAP. Overall, implementation reach was generally good, but the rate of dose delivered was low. The low dose delivered may be partly explained by the low or medium strength of fidelity, particularly when this relates to the offer or frequency of foetal growth scans for women at higher risk of SGA.

Table 5 Proportion of low-risk women with at least the minimum expected number of fundal height plots on GROW chartOverall

There was wide variation in the scores for each component achieved by sites. Site 7 consistently scored lowest for the majority of measures. There is no individual site which consistently scored highest for each implementation outcome; however, site 8 scored highest for the most implementation strength components.

留言 (0)

沒有登入
gif