Abstract
Aims:
To explore healthcare professionals' views about the training and support needed to rollout closed-loop technology to pregnant women with type 1 diabetes.
Methods:
We interviewed (n = 19) healthcare professionals who supported pregnant women using CamAPS FX closed-loop during the Automated insulin Delivery Amongst Pregnant women with Type 1 diabetes (AiDAPT) trial. Data were analyzed descriptively. An online workshop involving (n = 15) trial team members was used to inform recommendations. Ethics approvals were obtained in conjunction with those for the wider trial.
Results:
Interviewees expressed enthusiasm for a national rollout of closed-loop, but anticipated various challenges, some specific to use during pregnancy. These included variations in insulin pump and continuous glucose monitoring expertise and difficulties embedding and retaining key skills, due to the relatively small numbers of pregnant women using closed-loop. Inexperienced staff also highlighted difficulties interpreting data downloads. To support rollout, interviewees recommended providing expert initial advice training, delivered by device manufacturers together with online training resources and specific checklists for different systems. They also highlighted a need for 24 h technical support, especially when supporting technology naive women after first transitioning onto closed-loop in early pregnancy. They further recommended providing case-based meetings and mentorship for inexperienced colleagues, including support interpreting data downloads. Interviewees were optimistic that if healthcare professionals received training and support, their long-term workloads could be reduced because closed-loop lessened women's need for glycemic management input, especially in later pregnancy.
Conclusions:
Interviewees identified challenges and opportunities to rolling-out closed-loop and provided practical suggestions to upskill inexperienced staff supporting pregnant women using closed-loop. A key priority will be to determine how best to develop mentorship services to support inexperienced staff delivering closed-loop.
Clinical Trials Registration: NCT04938557.
Introduction
Pregnant women with type 1 diabetes (T1D) are advised to aim for target glucose levels of 3.5–7.8 mmol/L for at least 70% of the day. 1 Most women experience difficulties attaining these targets. 2,3 Alongside challenges calculating carbohydrates and determining insulin dosages, physiological changes result in periods of insulin sensitivity in early pregnancy followed by increasing insulin resistance as pregnancy progresses. 4 To help optimize glycemic management in pregnancy, guidelines recommend that women receive input and support from joint diabetes/antenatal teams every 1–2 weeks. 5
Closed-loop systems have begun to transform T1D management. 6 These systems combine a real-time continuous glucose monitor (CGM) with an insulin pump and an algorithm that automatically adjusts insulin delivery based on sensor glucose levels. To date, four hybrid closed-loop (HCL) systems are used commercially with regulatory approval in the United States and Europe, 7 with further systems under development. 8 The CamAPS FX platform, which includes customizable glucose targets suitable for optimal pregnancy glucose targets, is the only system currently licensed for use in T1D pregnancy. 9
Healthcare professionals play pivotal roles in mediating people's expectations, experiences, and use of diabetes technologies. 10 –15 Hence, to support pregnant women using HCL, it is important that their perspectives are considered. To date, interview studies have focused on healthcare professionals' experiences of supporting adolescents using HCL, 14,16 or have consulted those with limited knowledge and first-hand experience of using HCL in T1D pregnancy. 12,17
To address this gap in the literature and provide bespoke recommendations for supporting HCL use in pregnancy, we report findings from interviews with healthcare professionals who participated in a UK-based, multicenter, randomized controlled trial—Automated insulin Delivery Amongst Pregnant women with Type 1 diabetes (AiDAPT). The trial, which began in September 2019 and closed out in November 2022, explored the biomedical and psychosocial effects of using HCL (CamAPS FX) compared with CGM in pregnant women (aged 18–45 years) receiving routine clinical care. 18 For details of the HCL used, see Box 1. Women randomized to HCL used the CamAPS FX system from early pregnancy (∼12 weeks) until delivery (∼36–38 weeks).
The Closed-Loop System Used During the Trial
CGM, continuous glucose monitor; GMI, Glucose Management Indicator; HbA1c, glycated hemoglobin; HCL, hybrid closed-loop.
Healthcare professionals at local sites, working alongside a research educator, trained women to use the system, provided study-related support, and were responsible for participants' routine clinical care. Due to the Covid-19 pandemic, research visits and device training were available virtually via video call or telephone, and most participants' diabetes appointments were also offered virtually, mirroring practices in routine clinical care. For further details regarding the AiDAPT trial and training provided, see Box 2.
Details About the Automated Insulin Delivery Amongst Pregnant Women with Type 1 Diabetes Trial and Training Provided
AiDAPT, Automated insulin Delivery Amongst Pregnant women with Type 1 diabetes; MDI, multiple daily injections; T1D, type 1 diabetes.
Elsewhere, we have described how healthcare professionals favored an inclusive approach to HCL use and rollout due to their perception and understanding that virtually all pregnant women can experience clinical and/or quality-of-life benefits from using a HCL. 19 Here, we explore healthcare professionals' views about challenges and opportunities that a national rollout could present, including their training and resource needs, to support HCL use in pregnant women in routine clinical care.
Methods
We used an inductive, semistructured interview design, informed by topic guides to ensure that the discussion remained relevant to addressing the study aims, while allowing flexibility for participants to raise issues they perceived as salient. The study involved concurrent data collection and analysis, which allowed issues identified in early interviews to inform questions asked in later ones. Research ethics (Cambridge Central Research Ethics Committee: 18/EE/0084) and governance approvals were obtained in conjunction with approvals for the wider trial.
Recruitment
Healthcare professionals were recruited from eight participating trial centers across the UK, after they had ≥6 months' experience of supporting women using HCL. Recruitment continued until there was good representation of different grades of staff from across the eight sites and no new findings were identified in new data collected.
Data collection and analysis
Telephone interviews were conducted by D.R., an experienced (nonclinical) qualitative researcher between June 2021 and April 2022. Topic guides were developed in light of earlier investigations of HCL use, 14,16 and inputs from clinical coinvestigators, were revised in response to emerging findings and used flexibly according to healthcare professionals' roles and involvement in the trial. Key topics relevant to the reporting in this article are described in Box 3. Interviews that lasted 1–2 h were digitally recorded and transcribed in full.
Key Areas Explored in the Interviews
We undertook qualitative descriptive analysis, which focuses on identification and description of minimally theorized explanatory themes and is suited to understanding and illuminating issues relevant to policy and practice. 21,22 To promote rigor, four experienced (nonclinical) qualitative researchers (D.R., J.L., R.I.H., and B.K.) undertook independent analyses, which involved repeatedly reading and cross-comparing individuals' interview transcripts to identify cross-cutting themes. Each researcher wrote a separate report before meeting to discuss their interpretations and agree on a coding frame that captured areas of relevance to addressing the study aims. Coded datasets were then subjected to further analyses to allow more granular insights to be developed. The qualitative analysis software package NVivo20 (QSR International, Doncaster, Australia) facilitated data coding and retrieval.
Analytical workshop
Preliminary findings were discussed at an online workshop, held in September 2022, involving principal investigators, healthcare professionals who participated in the trial, other trial staff, and members of the qualitative research team. A ‘What? So What? Now What?’ approach was used to develop realistic and practical recommendations relevant to a range of diabetes professionals and clinical settings. 23 Individuals who were unable to attend the meeting contributed to the generation of recommendations via email correspondence.
Results
The sample comprised 19 healthcare professionals: 11 doctors, six diabetes specialist nurses (four nurses, two nurse consultants), one dietitian, and one diabetes midwife. See Table 1 for further details. Below, we report interviewees' accounts of the workforce and skill deficits likely to affect wider rollout of HCL in maternity clinics, their suggestions for how these might be addressed, and their views about how investing time and resources to upskill staff upfront could lead to reductions in future workloads. As responses did not differ according to professionals' roles, individual characteristics are not separated out in the reporting below. Key illustrative quotations are included below; for additional quotations, see Table 2.
Participant Characteristics
Percentages do not sum to 100% due to rounding.
AiDAPT, Automated insulin Delivery Amongst Pregnant women with Type 1 diabetes; HCL, hybrid closed-loop.
What challenges need to be addressed to support pregnant HCL users in routine clinical care?
Skills shortages
As several interviewees observed, after setting up and linking components, training and supporting women to use the HCL app/algorithm was “pretty straightforward” (HP-003) because “there aren't an enormous number of things to change… it seems a fairly intuitive interface” (HP-012). Given this, interviewees suggested that the greatest challenge to rolling out HCL technology to pregnant women would reside with upskilling centers and colleagues with limited experience interpreting CGM data and/or supporting insulin pump users (Table 2).
Additional Participant Quotations
T1D, type 1 diabetes.
Several noted that it would be particularly challenging to upskill healthcare professionals in smaller sites with a fluid workforce, because of staff shortages and difficulties in appointing new members who are able to support technology use. It was also noted that, in some smaller and less experienced sites in particular, existing staff lacked time or motivation to take on new training, an issue heightened by the Covid-19 pandemic (Table 2). Many also emphasized that even in experienced centers, a wider group of staff would need to be upskilled, to avoid reliance on a “subset” (HP-009) with relevant technology expertise (Table 2).
Moreover, interviewees indicated the importance of upskilling a wider corpus of hospital staff (e.g., in Accident and Emergency, Maternity Assessment and Delivery units) who provide support to women presenting acutely (e.g., with diabetic ketoacidosis): “so when patients are admitted… wider healthcare professionals aren't doing things they shouldn't be doing” (HP-015).
Several expressed additional concerns that, without widespread upskilling, existing regional disparities in healthcare professionals' skills and expertise might result in women in certain locations, especially those living far away from centers with HCL technology experience, having inequitable access (Table 2).
Challenges specific to T1D pregnancy
As interviewees further observed, while similar challenges would apply to upskilling an existing workforce to support HCL more generally, these were amplified in maternity clinics, where even large centers supported relatively few T1D pregnant women (e.g., ∼20 per year). This, as some noted, made it difficult for staff to embed knowledge and skills due to the low volume of T1D pregnant women using HCL: “you have two or three months break [gap], and you go: ‘oh my God. What was it, that thing we said we must remember for this time?’” (HP-004). Some also suggested that this challenge would be compounded if several HCL systems “which are all slightly different” (HP-016) in terms of the training and support required, were available (Table 2).
Healthcare professionals' training, resource, and support needs to rollout HCL
Training resources
To address the above challenges, interviewees proposed several solutions. To help ensure staff trained to support HCL use retained key skills and competencies, interviewees suggested making training resources, videos, and webinars available online, which healthcare professionals could (re)visit at a time of their choosing. They also recommended providing succinct checklists to limit “having to read a hundred and fifty-page user guide” (HP-003) (Table 2).
In keeping with the approach some described having adopted during the initial rollout of insulin pumps, some interviewees also discussed whether centers might begin by supporting pregnant women using one HCL system only (Table 2).
Infrastructure to support clinical teams to develop and consolidate local HCL expertise
When describing their trial experiences, interviewees praised the involvement of a research educator with CamAPS FX expertise who supported professionals at local sites when training pregnant women to use HCL (Table 2). Given their positive experiences, interviewees suggested that a similar approach be used to facilitate rollout, where individuals transitioning onto HCL received initial training from device manufacturers, replicating existing models to initiate insulin pump therapy (Table 2).
As an alternative, a minority proposed a hub-and-spoke model, “led from a central point” (HP-006), where women would receive training and glycemic support outsourced to “centers that are a bit more experienced supporting local teams … at least until [local] centers are more familiar with using the technologies” (HP-014).
Technical support for healthcare professionals
As with pump initiation in early pregnancy, healthcare professionals emphasized the need to invest time and resources during the first 2–4 weeks of HCL use. This included supporting women unfamiliar with component devices (e.g., assisting with set and sensor changes) or who “felt a bit nervous because they're having to trust the machine” (HP-009). All interviewees described how the ability to access technical, telephone support, which was “available on hand, 24/7” (HP-001) from the trial's research educator had been invaluable, especially when supporting women during this initial transition period (Table 2).
Hence, they suggested that if HCL technology was rolled out to pregnant women, staff would need access to a helpline, ideally on a 24 h basis, which offered similar technical support to that provided by the research educator, and which mirrored provisioning by pump and CGM manufacturers (Table 2).
Glycemic management support for healthcare professionals
Interviewees with limited or no prior experience of HCL use expressed a need, initially at least, for support to review and interpret women's HCL data, especially during the initial period after a woman had transitioned onto the system “to help people get the best out of it [HCL]” (HP-008). Specifically, interviewees noted that: “there was a requirement to have … support around trying to interpret patterns, thinking about carb counting, thinking about where hypos come in… as the system begins to change its algorithm and zero in on particular insulin doses. And that, probably for all of these women, was a bit more intensive for the first two or three weeks” (HP-012).
As some interviewees observed, it took time and practice to develop skills to interpret data and determine which system settings to adjust: “you really only get the sense of the data … once you do it lots” (HP-013). Hence, some interviewees described benefiting from consulting and learning from experienced colleagues during the trial (Table 2):
Interviewees suggested potential ways to provide these kinds of support in routine clinical care. Reflecting on their use of the HCL's data sharing/remote access capabilities to obtain support from central trial staff, several, for instance, suggested that this functionality could enable inexperienced colleagues to seek mentorship/peer support from experienced staff at their own site or a regional expert: “I think just having a mentor until you've got going… someone to ring up… who you share downloads with and say: this is what I did. What do you think? You know, I think that would be valuable.” (HP-015)
Some also suggested that this kind of support could be provided using an online portal where “you could drop the image of what you're unsure about, and then somebody who has expertise could perhaps either call you, if that's necessary, or could give feedback based on what they can see.” (HP-006). Several also described benefiting from intersite case-based meetings arranged to share learning and discuss challenging cases during the trial: “to hear about other people's experiences of what had gone wrong and what had gone well” (HP-002). To facilitate wider rollout, they suggested establishing closed-loop pregnancy “masterclass” (HP-013) meetings or workshops to help develop competence and confidence reviewing and interpreting data (Table 2).
Short-term investment for long-term gains
While interviewees emphasized that wider rollout of HCL would require an initial investment of time and effort to train up staff to support technology use, most expressed optimism that, in the long-term, their workloads could lessen. As several explained, after optimizing settings, improvements in glycemic management when using the HCL could result in women requiring less support in the later stages of pregnancy: “if women are getting [more] time in range… through the technology doing it, we're probably going to have less input into micromanaging them” (HP-007). Others observed that women using HCL experienced fewer episodes of hypoglycemia, hence, “later support is hopefully less burdensome, if indeed there are fewer hypos and fewer overnight hypos” (HP-012).
Furthermore, healthcare professionals described how it was “a lot less onerous” (HP-004) to support women using HCL, because the automated regulation of basal insulin rates meant their focus was “almost all-around mealtime boluses and I don't have to think about the basal” (HP-006): “if we could put the type 1's on the automatic pump, that would do at least some of the work for us… It's the CGM data that's the issue, looking at the CGM data. And that's why closed-loop is good because you're only looking at … the post meal bit. Whereas when you're looking at the CGM data for open loop and MDI you have to look at everything.” (HP-009)
Indeed, some speculated that, after making initial adjustments to settings/insulin-to-carbohydrate ratios, women using HCL might need less frequent or intensive glycemic reviews, although there would still be a need for someone to “eyeball the data” (HP-010). To this end, interviewees highlighted the potential value of using a system (similar to the CamAPS FX reports), which provided summary glycemic metrics for HCL users at each site/clinic. As interviewees explained, color-coded data enabled them to “interpret… a mountain of data within like five minutes” (HP-008) and quickly establish: “what I need to know, across a clinic or a population base level. So it'll instantly highlight women who are above or below target, anyone who's having a ton of alarms, anything that's kind of out of the ordinary.” (HP-014)
As several further noted, using these kinds of reports had enabled them to “triage” individuals and identify who required support (Table 2).
Discussion
This study reports healthcare professionals' views about the challenges and opportunities that could arise from rolling-out HCL technology to pregnant women with T1D. Our findings, alongside recommendations developed in our online workshop, offer practical ways to upskill professionals and potential ways to promote widespread and equitable access to this technology (see Box 4 for a summary of these recommendations).
Summary of Recommendations
Interviewees emphasized that healthcare professionals providing training and support to pregnant HCL users would need to be proficient in the use of its constituent (CGM and insulin pump) technologies. They also raised concerns that existing disparities in insulin pump and CGM expertise 24 –27 might generate inequities in access to HCL technology. This is a valid concern given that others have identified professionals' lack of skills and competencies as being barriers to diabetes technology use more generally. 10,15,16,28
Moreover, interviewees noted how infrequent exposure to small numbers of pregnant users might limit opportunities for staff to apply their skills and retain expertise, meaning that this issue could be more pronounced when supporting HCL use in T1D pregnancy. Interviewees, however, proposed a range of “low cost” solutions, such as making webinars, videos, training resources, and succinct checklists available online to help staff refresh their skills. These recommendations were endorsed by workshop participants who noted that several of these resources are already available for healthcare professionals supporting CamAPS FX users. Such training resources would need to be updated at regular intervals to keep abreast of rapidly changing HCL technology.
In keeping with interviewees' accounts, workshop participants also recommended that general hospital staff be given guidance to support HCL users who present acutely at emergency departments and maternity units. They suggested that existing mandatory training modules be updated to raise staffs' basic awareness and understanding of HCL technology and alert them to situations when they should contact diabetes specialists to seek clinical input (for further recommendations, see Box 4). In response to some interviewees' suggestion that only one HCL system initially be supported, workshop participants noted that this is not a feasible option because women are increasingly using different HCL systems before pregnancy. Participants noted that it would be unreasonable to ask existing users to switch systems and, in keeping with Sherr et al.'s recommendations, 29 emphasized the need to respect users' choices and preferences.
To ensure equitable access via a national rollout, our findings suggest that it will be important to provide intensive, targeted initial support to inexperienced sites and staff. There was broad consensus among workshop participants that training to use the HCL's constituent components should be provided by device manufacturers; a similar recommendation has also recently been made by Sherr et al. 29 Workshop participants also suggested that a central hub-and-spoke model (as used in the trial) would be unsuitable for real-world implementation. They expressed a strong preference for initial glycemic management after HCL initiation to be provided by local teams, with whom women had existing, trusting therapeutic relationships. With regard to interpreting HCL users' data and providing women with glycemic advice/support, interviewees valued opportunities to consult experienced colleagues, in particular (but not exclusively) during the initial weeks of HCL use.
To this end, workshop participants reinforced interviewees' suggestions by recommending local teams or individuals who lack experience reviewing and interpreting HCL data be offered mentorship/peer support from more experienced colleagues. They also recommended implementing interviewees' suggestions to develop collaborative models of support by establishing intra- and/or intersite case-based learning sessions or HCL pregnancy “masterclasses” led by experienced colleagues. However, workshop participants suggested that further consideration should be given to whether such support is provided locally, regionally, or nationally and by whom.
Moreover, they emphasized that funding may be needed to ensure that experienced staff have capacity to support less experienced colleagues/centers. Interviewees noted that such support could be provided virtually (by sharing HCL users' data with experienced colleagues); a model made logistically feasible by healthcare professionals' ability to access women's data and support glycemic management remotely. 30,31 If a “masterclass” model is implemented, healthcare professionals would need regular refresher sessions to take account of rapid developments in HCL technology.
In line with interviewees' suggestions, workshop participants recommended that provision of HCL technology in routine care should include technical support for healthcare professionals, available via a 24 h telephone helpline, and similar to the support provided by insulin pump and CGM companies. Telephone support is already available for healthcare professionals supporting CamAPS FX users and as other HCL systems become available, 8 similar technical support helplines will be needed. As workshop participants noted, these helplines would need to be able to offer bespoke recommendations for HCL use in T1D pregnancy.
While there may be some initial costs associated with developing the above training and support strategies, interviewees described how, in the long-term, workloads (and, hence, potentially costs) might be reduced because women using HCL might require less glycemic management input, especially during later pregnancy, than those using insulin pump or multiple daily injections regimens. This perspective mirrors that of healthcare professionals supporting adolescent HCL users 16 and parents of young HCL users, 32,33 who both described less need for healthcare professional input once initial adjustments to the system had been made.
Moreover, healthcare professionals in our study observed how email reports containing summary glycemic metrics had helped them to identify and triage individuals requiring additional support. Indeed, our findings suggest that using such metrics to triage caseloads could help address concerns raised in earlier studies that more time would be required during consultations to process and interpret large volumes of HCL data. 17
Strengths and limitations
This is the first study to consult healthcare professionals about the training and support needed to support a national rollout of HCL to pregnant women with T1D. The topicality and relevance of our findings have been enhanced by using a workshop at the end of the study to generate contemporary practical, implementable, and clinically relevant recommendations. A potential limitation is that our study took place within the context of a UK-based trial. Many participating centers had experience supporting insulin pump and CGM users, and some had HCL experience as a result of being involved in previous trials.
Hence, some interviewees may have been technology enthusiasts, with accordingly distinctive views about HCL rollout in T1D pregnancy, specific to a UK healthcare context, where there is universal CGM access. To offer a wider perspective on how a national rollout of HCL in T1D pregnancy could be achieved, future research could explore the perspectives of healthcare professionals working in less experienced centers or who have limited or no experience of supporting HCL users. Future research could also explore the experiences of healthcare professionals who support HCL users in countries that have different healthcare structures to those in the UK.
Conclusions
Our recommendations offer important solutions and suggestions to address the challenges involved in rolling out HCL systems to pregnant women and are likely to have relevance to supporting rollout to other patient groups. A key priority will be to determine how best to develop mentorship services to support inexperienced staff delivering closed-loop therapy.
Footnotes
Acknowledgments
We are deeply grateful to the professionals who made time to take part in this research and shared their experiences with us. We acknowledge the contribution of Prof Fiona Denison, who was a coinvestigator on the AiDAPT study and supported initial delivery of the qualitative research. Prof Denison died in January 2022, and we miss her deeply.
Authors' Contributions
J.L. conceived and designed the interview study with input from H.R.M. D.R. collected the data, which was then analyzed by J.L., D.R., R.I.H., and B.K. J.L. conceived the concept for this article. D.R. and J.L. coproduced the first draft, with further input from R.I.H. and B.K. All authors reviewed, edited, and approved the final version of the article.
Disclaimer
The views expressed in this publication are those of the author(s) and not necessarily those of the MRC, NIHR, or the Department of Health and Social Care.
Author Disclosure Statement
A.B. reports: receiving speaker fees from Novo Nordisk, Sanofi, Dexcom, and Abbott Diabetes Care. C.B. reports: receiving speaker fees from ABCD Diabetes Technology Network. S.H. reports: receiving speaker and advisory board fees from Dexcom, Medtronic, Sanofi, and Ypsomed; and being Director of ASK Diabetes Ltd., he receives consulting/training fees from CamDiab Ltd. E.M.S. reports: receiving honoraria for talks given to Abbott Diabetes Care and Eli-Lilly. R.H. reports: receiving speaker honoraria from Eli Lilly, Dexcom, and Novo Nordisk; receiving license and/or consultancy fees from B. Braun and Abbott Diabetes Care; patents related to closed-loop; and being Director at CamDiab. H.R.M. reports: sitting on the Medtronic European Scientific Advisory Board and receiving speaker honoraria from Dexcom, Abbott Diabetes Care, Medtronic, and Novo Nordisk. D.R., J.L., R.I.H., B.K., K.B.-K., C.C., A.R.D., K.F.H., T.T.M.L., R.S.L., D.R.M., A.M., G.R., R.M.R., and S.L.W. have no conflicts of interest to declare.
Funding Information
This project (NIHR 16/35/01) is funded by the Efficacy and Mechanism Evaluation (EME) Programme, an MRC and NIHR partnership. Additional support for the artificial pancreas work has been provided by the National Institute for Health Research Cambridge Biomedical Research Centre, JDRF, and Wellcome Strategic Award (100574/Z/12/Z). Dexcom supplied discounted continuous glucose monitoring devices.
For the purpose of open access, the author has applied a CC-BY public copyright licence to any Author Accepted Manuscript version arising from this submission.
Supplementary Material
Supplementary Data
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
