Recruitment and retention
Between November 2021 and March 2022, the four selected care homes (see Table 1) ran the CONTACT programme 24/7 for 2months. Despite ending as planned, the feasibility study did not meet its pre-determined progression criteria for a full RCT.
Of 156 screened residents (see Fig. 2), 105 consented (either personally or through a nominee) to wear a device, with 102 (97%) wearing them at the start of the 2-month intervention. Of the 225 staff deemed eligible, 82% (n = 178) agreed to participate, but 20 dropped out before the intervention started.
CONTACT feasibility study CONSORT diagram 28th June 2023
Ineligibility amongst residents was solely due to staff concerns that wearing the device could pose a risk of harm. Of the residents who declined to wear the devices, 14 did not give a reason, two were disinterested, four did not receive consent from their nominees, and two passed away before they could return their consent forms.
Of staff, 17 opted not to participate, with eight outright declining, seven not providing a reason, one objecting to wearing the device, and one simply expressing a lack of enthusiasm. Contextual factors for non-participating staff included six leaving the care home, five with imminent maternity leave, and seven categorised by managers as “rarely present” (sic.) bank staff.
The demographic profiles of the homes were female and white. Most residents had been in the homes for an extended period, and both staff and residents had been vaccinated against COVID-19. More than a third of residents lived with a dementia diagnosis (see Table 4).
Acceptability and feasibility of intervention delivery
Ease of administering devices to residents, staff, and external visitors
Getting devices to participants was moderately successful, with 70% of screened residents and 87% of staff receiving BLE wearables. But participation in CONTACT was burdensome and added to regular work. Staff highlighted screening processes, obtaining consent, and registering participants as particularly laborious. COVID-19 restrictions meant homes conducted recruitment themselves; limited digital and data infrastructure meant screening was manual and time-consuming. Larger homes bore a heavier burden; despite this, Homes 1–3 managed to complete screening on time.
Recruiting residents lacking mental capacity [31] to make decisions for themselves, and thus provide consent, meant contacting designated consultees, which further added to home workload. In some instances, the homes found the workload associated with the study outweighed the perceived benefits.
I find I have to shuffle things around to make it work. When things were heavier, I would usually finish at 5, but during the screening and consent time I had to stay late at night to contact the families. It was hard it fit it into an already hard day (Home 1, study champion).
The study’s research governance requirements contributed to CONTACT’s complexity. Every BLE wearable device’s unique number (used by the study team) needed to be cross-referenced against a “master log” in each home for the home to identify the wearer. Communications involving identifiable data were carried out via a secure file transfer system. However, university secure databases for registering participants and reporting COVID-19 cases encountered technical issues, adding further delays.
Homes 1–3 successfully dispensed devices within a month from consent and before the feasibility start date. Conversely, home four managed to issue only 66% of their BLE wearables after the study start date, with a median delay of 52.5 days (range 31, 60). Because of Home 4, the median time from consent to issuing resident devices was 33 days (range 20, 60). Several reasons were given for the 10 resident withdrawals, including residents not wanting to wear a device or feeling distressed or confused by them.
Issuing staff devices was efficient. Homes distributed them in a median of 32 days (range 12, 60). Home 4 again took longer, with a median of 35 days (range 12, 60). Reasons for staff withdrawals included no longer wanting to wear the device and finding the device irritating or inconvenient.
An original study objective was assessing the feasibility of BLE wearables for tracking visitors’ (relatives and community professionals) movements within the homes. All the homes conveyed that implementing the necessary procedures for this was not possible due to staffing constraints. Homes one and two did not have permanent reception staff, and the other homes simply judged procedures as too burdensome. Consequently, tracing visitors was dropped from study procedures.
We successfully appointed study champions in each home. Each home was informed in advance, and as part of their participation requirements, that there would be study tasks to be accommodated as work in the home. But it was clear that they struggled to absorb CONTACT-related work into non-research day to day work. Consequently, it was deprioritised by homes:
It was the time element. I don’t have an administrator or anyone else to help me with my tasks; it’s just me. CONTACT wasn’t at the top of the list by far. We said we would try our best with it, but we couldn’t (Home 3, manager and champion).
Home managers scored aspects of CONTACT familiarity, and current and future chances of “normalisation” using NoMaD (see Table 4). Managers from Homes 3 and 4 (compared to Home 2) felt more familiarity with CONTACT and that it was a more normal part of work by the end of the intervention (Home 2’s use-based familiarity diminished or stayed the same). Whilst the manager of Home 1 believed CONTACT could become part of normal work, they left before completing their post-implementation scoring.
Device loss and damage were noteworthy. Eleven percent of resident devices (n = 12) and 7% of staff devices (n = 7) were lost. Almost half (47%, n = 9) of lost or damaged devices were replaced. Fewer staff devices were lost (3%, n = 5) or damaged (4%, n = 7). Just 8% (n = 1) were renewed.
Fob wearables required frequent battery changes: 15% (n = 38) in Homes 3 and 4. These were supposed to be done by the homes, but Home 4’s delays meant a research team member undertook these over two visits. Card wearables in Homes 1 and 2 required no battery changes.
Acceptability and feasibility of structured CONTACT feedback
Home (1, 2, and 4) managers provided assessments of the (i) understandability; (ii) influence on IPC thought, and (iii) likelihood of changes based on the report (Fig. 3).
Managers’* assessed understandability, IPC influence, and change likelihood — structured reporting. *Home 3 did not provide post-scheduled report data
No clear patterns were evident in the assessment of report sections (see Fig. 3) by home managers. Only two homes (2 and 4) provided a judgement on structured report Sects. 5 and 6, and only Home 4 provided an assessment of report Sect. 6. For Homes 1, 2, and 4, the reports were unlikely to lead to change. Only one home (Home 1) was ambivalent (neutral) towards Sects. 1 and 2. And all 3 completing homes viewed Sect. 3 as most unlikely to induce any change.
The quantified assessment of CONTACT’s inability to induce change was evident in qualitative findings. CONTACT’s research study context, and delivery alongside competing pressures such as maintaining staffing and pre-existing infection prevention and control (IPC) requirements, diminished the perceived value of the study’s information, contributing to an overall perception that the study was of limited value:
The triggered report covered mostly what we knew already. The scheduled report identified which residents are most at risk, but what can you really do with that information? We can make people isolate but then you lose staff. The staff do a lateral flow test before work every morning, that’s the protection we already have without losing too many staff (Home 4, study champion).
…it could work, preventing us having to close because we’ve got 2 cases out of 80 for any infection. We can easily isolate pockets of people if we needed to and staff as well. So, I can see if we didn’t have the national guidelines in place, where it would give me research-based information to make risk assessment decisions…. In the guidelines, it does say that registered managers are accountable for decisions. Outside of a trial, it would have given me the confidence to say this is what the infection is doing, and we can safely isolate that and carry on doing what we are doing with the other residents, so the residents don’t suffer from lack of visitors (Home 4, manager).
A significant barrier to feasibility, reducing trust and study compliance was staff concern at “being tracked”. As a result, scheduled reports were not shared by Home 4’s management with other staff. Reports were disseminated in the other homes. The follow-up support call from researchers after each report was perceived as highly beneficial by managers and champions.
Delivering CONTACT required training for study champions and home staff. Of the 34 individuals invited to attend virtual training across 9 sessions, almost two-thirds (65%) participated (Table 5).
Acceptability and feasibility of study design/implementation processes
Despite securing the necessary ethical and research governance approvals, we were unable to link residents in the homes to NHS (National Health Service) data. Dialogue with NHS Digital began a year before the intervention period, but linkage proved impossible in the timeframe. DSHC infection and mortality data for the homes was eventually secured, after the intervention period.
Data capture
Only around 29% (n = 70) of devices functioned as expected, with only minor differences between resident (29%) and staff (28%) devices. Differences between (Fig. 4) homes were evident: more day-to-day variability in Homes 1 and 3; relatively stable adoption in Homes 3 and 4; a visible dis-adoption trend in Home 3. Within Home 2 (Fig. 5), resident data was relatively complete and stable, but staff data was partial, variable, and notably absent for a short time early in the implementation period. Apparent device malfunction could be due to battery failure, inappropriate device placement, or staff not updating weekly logs for active devices — a crucial element for correctly processing data. Data transmission from our commercial partner to the university’s secure database experienced no issues.
Proportion of active devices correctly recording per day by home
Proportion of active devices correctly recording for residents and staff — Home 2
During the feasibility period, 33 (32%) of 102 residents and 53 (34%) of 158 staff reported COVID-19 infections, suggesting self-reported COVID-19 was a feasible primary outcome. However, the single reported case of staff gastroenteritis suggests “other infections” was a less feasible outcome. Although all homes provided reported deaths (n = 7, 7%) during the intervention, only two homes (3 and 4) shared data regarding whether the deaths were COVID-19 related and the months from registration or device issue to death. Despite 86 infection notifications, only 52 (60%) contact reports were requested by the homes.
Progress against predefined criteria
The study did not meet any of our quantitative criteria for progression to a definitive RCT. Additionally, qualitative data from the homes indicated study demands were too burdensome and excessive. Projected compliance and participation rates were too low to justify a definitive trial (Table 6).
Add Comment