We used the Good Reporting of a Mixed Methods Study (GRAMMS) checklist to guide the reporting in this article (21).
Practice Characteristics
Table 3 provides an overview of the baseline characteristics of all practices. Overall, 35% of available GPs and 80% of available PNs from the participating practices agreed to participate. There was no difference in the proportion of GPs and PNs participating by allocation.
Characteristics of the providers
Table 4 provides self-reported baseline provider characteristics. There was an even gender mix among GPs but PNs were predominantly female (92%) and predominantly younger with 54% being aged between 20-34-years. Among the PNs, 65% worked part time whereas 85% of GPs worked full time. Compared to GPs, PNs had worked fewer years in general practice (mean 4 years) and less years in their current practice (mean 2 years). Almost half of the GPS and PNs reported that they never or rarely used apps or websites with their patients. PNs were more likely than GPs to indicate they were moderately/very confident to show patients how to use apps (51% versus 22%).
Organisational Readiness domains
Table 5 displays assessment of the organisational readiness domains for each intervention practice. The one NSW practice that did not recruit any patients has not been included in the discussion of intervention specific capacity.
a) Motivation
Due to a targeted practice recruitment of those interested in prevention and research, we can assume some level of motivation on the part of practices to participate and implement the HeLP GP intervention. Incentives were also provided in the form of one-off monetary practice payment for participation (AUD $1,000), completion of PN health checks (AUD $40 per patient) and follow up (AUD $20 per patient). Continuing Professional Development (CPD) activities were also provided for GPs and PNs which contributed towards meeting professional educational criteria.
Analysis of the recorded observations from the ROs and facilitators working with the practices revealed that four practices displayed strong commitment to providing good preventive care and/or a belief that the practice staff should contribute to research. Increasing numbers of overweight patients within practices was also a common reason why practices wanted to instigate a weight management intervention with their patients. For one GP, a recent close working relationship with the organising centre (UNSW) and a strong interest in obesity drove the decision to participate.
All except one practice indicated at the outset that the intervention would be achievable, although the time commitment required by PNs raised concerns for some. In one smaller Sydney practice where the GP worked across two sites, and the PN worked part-time, reluctance was observed on the part of the PN because the GP had agreed to participate while she was on leave. Despite generally good motivation levels initially, these were observed to decline in some practices over time. This was not a short intervention and a 12 -14-month commitment was required by practices for set up, training, patient recruitment, intervention delivery and data collection. RO’s and facilitators described a steady loss of motivation over time due to various factors including work pressures, inadequate or changes in staffing, organisational barriers, and competing clinical priorities such as general practice accreditation and a busy influenza season. For one rapidly expanding practice, motivation decreased as the practice became busier and pre-occupied with mounting organisational issues.
b) General Organisational Capacity
Climate/Culture
Multiple and varied practice structures were observed among the participating practices. Of note were reported cases where hierarchical, haphazard, or disorganised internal systems reduced staff participation in decision making, particularly around involvement in the trial. Most practices displayed a ‘top-down approach’ where decisions were made by the GP/s resulting in situations where other practice staff felt they had limited opportunity to contribute to decision making, and hence limited opportunity to initiate or influence changes within the practice. Only in one SA practice were the GP, PN and Practice Manager (PM) observed to equally influence the decision about participation. Here, the PN was pro-active and clearly supported by the GP, indicating mutual trust and respect within the relationship. Conversely, reception staff at this practice were not engaged in decision making despite being tasked with the distribution of trial materials to potentially eligible patients.
Despite a general tendency for PNs in this study to lack decision making capacity within the practice, they seemed consistently autonomous in their day-to-day clinical work. GPs were not reported to micro-manage or monitor the PNs delivery of, or the involvement of other staff in the HeLP GP intervention. ROs and facilitators reported that GPs were frequently unaware as to whether the reception staff or PNs were actively undertaking the trial tasks or completing them according to trial protocols. Coordinating roles were sometimes undertaken by PMs. In other cases, PNs coordinated the reception staff and the GP to complete tasks, but this was specific to a few practices only.
Staffing and Resources
No initial difficulties relating to practice space, equipment issues or staffing quotas were identified. Practice eligibility required at least one PN to implement the intervention, but this was sometimes shared across two positions. Consistent and widespread PN turnover however significantly impacted the capacity to implement the Help GP intervention. As PNs infrequently provided prior notification to the trial that they were leaving, we were unable to elicit whether this level of turnover represented dissatisfaction with their employment, or some other work related or personal trigger. Both study groups experienced substantial turnover, the intervention group recording slightly higher rates than the control group (11 versus 8 PNs). PN turnover, however, affected half of all intervention practices at some point and was a persistent disruptive element within the trial. At each occurrence, RO’s had to re-commence orientation and training, and facilitators had to repeat support visits. It was also challenging for the new PNs to pick up the role part way through the trial; they did not always feel confident with the research tasks, did not have existing rapport with the patients, and were sometimes unhappy with the unexpected workload or role. Consequently, the level of engagement of these ‘replacement’ nurses varied, and extra work on the part of the researchers was essential to achieve completion in the practices with consistent turnover. In addition, general workload was a major factor particularly for those PNs who were undertaking other research activities (1 practice) and those working reduced hours or sharing roles.
Leadership and Communication
The degree of practice leadership observed was variable and often linked to the attributes of individuals rather than a practice-wide norm. The level of leadership within each practice was also observed to be influenced by many factors, notably staffing arrangements and the consistency or fluctuation of staffing levels.
Difficulties establishing good lines of communication was observed by the ROs and facilitators. The primary contact in a practice could be the GP, the PM, the PN, or a combination of these. In some cases, contact with the reception staff could be done directly, whereas in others, it was only via one of the other staff members. Particularly difficult were those circumstances where the primary contact could not be reached easily (e.g., a message had to be left for the GP), where the primary contact worked part-time/weekends, and in practices where inter office communication was disorganised. It was crucial to this research to identify the main contact person at each practice and to build individual relationships and develop tailored methods of interacting. Identifying this person, however, took time, and required considerable patience and flexibility. In some practices the reception staff were consistent, in others, there were multiple receptionists working different shifts, or more than one receptionist working at a time. Lines of responsibility and accountability were frequently vague or hard to discern.
a) Intervention specific capacity
The PN Model – skills knowledge and ability to deliver the HeLP GP Intervention
The HeLP GP intervention had three components that PNs could offer to their patients. Feedback and observation indicated that PNs were comfortable with the clinical content of the health check as this largely aligned with their normal work. Despite this, ROs and facilitators experienced significant ambivalence on the part of the PNs to conduct the health checks, noting a mix of disinterest, reluctance, and a lack of confidence by PNs to engage with their patients throughout the intervention. On-line training appeared well received, providing obesity education and tools and instruction to implement the intervention. PNs however, demonstrated variable success with delivering the health checks to consenting patients. One Sydney practice failed to successfully recruit any patients and hence no health checks were conducted at this site. In total 84/120 (70%) consenting patients underwent a health check and completed the baseline assessment. At the point of health check, 16 patients withdrew, 14 patients could not be contacted after three attempts by the PN, and 3 patients were found to be ineligible. The health check was not completed for 2 patients who received insufficient follow up by the PN and 1 patient who was too busy to attend. Similarly, PNs struggled to complete the patient follow-up with 58/84 (69%) patients being seen for the 6-week review (Figure 2). The mean number of days between the health check and the 6-week follow up was 64.2 days (range 42-199 days) indicating that this was frequently provided outside the designated timeframe (42 days).
The remaining components of the intervention (referral to Get Healthy and (mysnapp) required initiation of the health check and were therefore impacted when the health check was not conducted. Referral to Get Healthy could be initiated by phone, fax, or email and did not itself present a barrier for PNs. However, PNs demonstrated great variability in their skills and confidence relating to the introduction and set up of mysnapp. At baseline, almost half of the PNs had reported that they never or rarely used apps or websites with their patients, although PNs were more likely than GPs to say they were moderately or very confident showing patients how to use apps (51% versus 22%). In the individual cases where PN confidence was identified as a factor, additional facilitation was provided to them to encourage maximum uptake.
Within one Sydney practice, the PN model was supplemented by the employment of casual nurses when practice accreditation was prioritised over the intervention, and the PN was removed to that role. To avoid delays, the trial employed two casual PNs specifically to complete the patient health checks. We subsequently also offered this alternative to other practices who were struggling with the completion of health checks, but the offer was not taken up.
Practice factors affecting the implementation of the HeLP GP intervention
General support for the trial and the intervention ranged from welcoming, helpful, and interested to disinterested and in some cases oppositional. ‘Push back’ by reception staff was not widespread but was identified as a prominent issue in some practices and proved challenging to manage. As RO’s were ‘outside’ the practice, they were unable to influence these dynamics and relied on goodwill and rapport building to negate these problems. Some receptionists resented the extra workload that resulted from printing documents, handing these to patients, and managing DCP. In some instances, the primary contact at the practice had no supervisory role over reception staff and, consequently, resistive behaviour remained throughout the trial.
Our experience was that practices uniformly required a great deal more RO support to carry out the required tasks than anticipated. Fortnightly visits were made to each practice through the patient recruitment and intervention period, as well as interim email and telephone support to try and maintain practice staff interest and engagement. Many practices required constant reminders and prompting about tasks and reinforcement about the reasoning for these tasks. Lack of time and workload were frequently cited as reasons for why trial tasks went undone. We did not identify a clear program ‘champion’ at any of the practices (i.e., someone who actively supported the intervention and provided continuous leadership). Although we identified some individuals who had a stronger interest in the intervention and responded in a timelier way, we did not identify clear standout people (or roles) within the practices who were prepared to advocate or promote significant changes, either to facilitate the intervention or because of the intervention. In at least half of the practices, the intervention was generally supported by the practice management (usually the GP/s) however this did not appear to translate to the PNs feeling supported to undertake the intervention. We identified 6/10 practices where PNs expressed that they felt insufficiently supported from within the practice to undertake the intervention with their patients.
Uptake and appreciation for trial systems and resources
An extensive list of resources was developed for the trial and provided in paper, digital, and on-line modes (Table 6). It was not possible to monitor the uptake and use of all resources but the provider’s website containing trial information to support processes and mechanisms for feedback, was generally poorly utilised. While some PNs were enthusiastic about the training, the completion of all three modules proved onerous for some PNs despite being in a format that could be done at a time that was convenient, and which could be used to apply for professional and educational recognition. PNs were not paid to undertake the training.
Over the trial period, most practices experienced some difficulties with the DCP software (used to identify eligible patients and monitor intervention progress). The nature of these issues was widespread and were ongoing for some practices. We cannot therefore discount the negative impact of this disruption, or the negative sentiment generated because of this disruption. Some practices managed their software systems in-house, but many used external IT management companies which added to the complexity and time involved to identify and address system issues. External IT support services sometimes felt uncomfortable about having the trial software installed or were reluctant to support technical issues such as firewalls. Also notable were issues with internet connection and speed which proved particularly problematic for PNs when trying to set up mysnapp for patients. We had anticipated difficulties with the app set-up and had provided troubleshooting documents and a reference video, but we don’t know the extent to which these were utilised.
Practice nurses appreciated the guidance and support provided through the facilitation visits. We envisaged a total of three one-hour facilitation visits to each practice to address requirements and to answer questions. On average 4.7 visits were required to each practice by the facilitator (4.5 visits NSW; 5.3 visits SA). Facilitation number two (the visit where the PNs were briefed around the intervention), took an average of 98 minutes. This extra time was a due to a combination of higher PN need (unfamiliarity with apps, uploading health check templates and training to use trial software) and repeat visits due to PN turnover. Distance of some practices from the research centre resulted in just over 5hrs per practice of travel time (Table 7).