This chapter presents information on sampling, recruitment and data collection. Data collection took place between March 2015 and May 2016. Figure 3 presents an overview of which data were gathered when. Each of the methods reflected is described in detail below.
FIGURE 3 Data collection overview. (Figure not reproduced here; see the original article.)
Ward sampling and recruitment
Six ward teams were included in the study. The sample size was determined by funding availability and the plan to run the study in at least two hospital organisations, and at least two ward specialties. Our previous research had highlighted the importance of organisational and ward context in determining nurses’ relational capacity, and this informed a sampling strategy aimed at diversity.15 Senior nurses at two English NHS trusts were invited, on behalf of their organisations, to take part in the study and agreed. They were invited to nominate three acute inpatient units each to take part in the study. They were guided to select three adult inpatient wards in each trust with the highest proportion of patients aged ≥ 65 years. In both trusts, wards with the highest proportion of older patients were MOP, so, to aid diversity of the sample, we also specified that one of the wards needed to be a surgical ward with a high proportion of older patients. Senior nurses were also encouraged to include those wards perceived as less effective in their general performance in addition to those perceived as high-performing wards, because there is a need to know if interventions of this kind can also work in contexts in which staff may not recognise the potential benefits of the intervention and/or do not prioritise improving compassionate care.
Critical care units were not eligible for inclusion, as previous research indicates that nurses’ experiences of providing relational care in critical care environments are very different from their experiences of providing care in more general wards.15 Wards were also excluded if the departure of the ward leader was anticipated in the subsequent 6 months of attempted recruitment, as stable ward leadership was theorised to be an important influencing factor on implementation.
Following the circulation of written information about the study, the chief investigator met with the ward leader for each of the nominated wards. They were given a verbal explanation about the study and a chance to ask questions and discuss the implications of taking part. At the close of the discussion, they were invited to put their ward forward for the study or not, or, if they preferred, to take more time to make up their minds.
Once the wards were recruited, all nursing staff, including registered nurses (RNs) and health-care assistants (HCAs) employed to work in the participating ward teams, were eligible to take part in the study and recruitment processes were designed to maximise their inclusion. Meetings with ward leaders were designed to inform staff about the study. Researchers also offered to visit the wards to talk to the teams about the study and this took place on three wards. Written information sheets about the study were also given to ward leaders to distribute to staff. Posters about the study were displayed on the participating wards in public and staff areas.
Addressing the process evaluation and the pilot CRT in turn, the following sections outline sampling, recruitment and data collection.
Process evaluation
The process evaluation aimed to identify and explain the extent to which the planned intervention was implemented into existing nursing practices on the four intervention wards. Data were collected using a variety of methods in order to gather insights into different aspects of implementation from different viewpoints. The data collected were:
- quantitative ward profile data generated by each ward leader at the outset of the intervention
- qualitative interviews with nursing staff and ward leaders from the participating wards, the PDNs leading CLECC implementation, the matrons overseeing the wards and hospital senior nurses
- field notes made by PDNs delivering the intervention
- observations of a sample of CLECC classroom training days and CLECC action-learning sets for ward leaders
- quantitative records of training delivered.
Data were collected between May 2015 and May 2016 to capture the period just before the implementation period, with follow-up for up to 12 months. Three main phases were used to guide the scheduling of data collection in the process evaluation: period 1 (baseline/early CLECC active training; May–August 2015), period 2 (mid-to-late CLECC implementation period; July–October 2015) and period 3 (after implementation period; October 2015–May 2016).
Process evaluation sampling and recruitment
All members of the participating nursing teams, including RNs and HCAs, were eligible to take part in the qualitative interviews and were invited to take part through posters, presentations and e-mails. We purposively sampled from those who volunteered to capture variations in staff grade and ward. Written consent was sought for the staff interviews and staff were given information about not being obliged to take part and their right to withdraw their consent at any time. We offered a payment of £15 shopping vouchers to individual staff who completed an interview. Individuals recruited early in the study were invited to a second and third interview so that variations over time could be tracked. When such individuals could not be contacted or declined further interviews, new individuals were recruited at the same grade and from the same ward to ensure that variation by ward and grade was maintained. All four ward leaders and two PDNs leading the CLECC implementation were also invited to three interviews each. At the final interview round, all three matrons overseeing the participating wards and two further people in more senior nursing roles in the trust were invited to be interviewed.
Process evaluation data collection
Ward profile
Contextual data were gathered on intervention and control wards through the completion of a ward profile questionnaire by the ward leader (or other senior nurse on the team). These ward-level data included physical layout, specialty, bed occupancy, staffing, sickness rates, agency usage, turnover and shift length. These data were gathered at baseline phase and updated at follow-up phase.
Qualitative interviews
One-to-one face-to-face qualitative interviews were undertaken in three phases as outlined in Process evaluation . All but two participants opted to use a hospital meeting room for the interview. Two participants chose to be interviewed away from the hospital site. The interview schedules were designed to capture individual views and experiences, and to reflect the key NPT concepts of coherence, cognitive participation, collective action and reflexive monitoring. Period 3 interviews with intervention ward staff also included questions on resource implications of CLECC for the purposes of the economic evaluation. Schedules reflected the implementation stage at the time of interview. Appendix 4 shows examples of interview schedules. The interviews lasted, on average, 46 minutes (range 17–70 minutes). Interviews were audio-recorded and transcribed verbatim, and transcripts were checked for accuracy by the interviewer.
The two PDNs kept detailed field notes of their experiences of delivering the intervention.
Observations of training activities
A researcher observed a sample of CLECC classroom training days and ward leader action-learning sets. Data were collected using unstructured non-participant observation, using event sampling, that is, recording observations of a set event rather than at regular periods over time.113 The use of observation allowed for verbal and non-verbal interactions to be recorded in the form of field notes. The researcher adopted the role of complete observer and did not participate in the learning activities.114 These observations were intended to complement the interview findings and the quantitative records of training delivered.
Quantitative records of training delivered
We also explored the feasibility of gathering data on the amount of training delivered through a register of attendance at classroom training and action-learning sessions, and a quantitative record of ward cluster discussions, reflective group discussions and cluster discussions.
Pilot cluster randomised trial
There were two main data collection phases for the pilot CRT and data were gathered from patients, staff and visitors. To enable us to gather data on baseline characteristics and assess against the selected outcomes, we used the following methods:
- observations of staff–patient interactions using QuIS42
- patient questionnaire survey comprising PEECH,100 Picker Patient Experience Questionnaire (short-form)100,115 to measure general care quality, EQ-5D-5L93 and participant demographic details100–102
- nursing staff questionnaire survey comprising JSE, MBI, CC, FECC, selected items from the IHOS battery and participant demographic details26,104,107–109
- visitor questionnaire survey comprising CEC and participant demographic details26
- matron questionnaire survey comprising MAQC26
- ward leader ward profile
- qualitative interviews with nursing staff.
Copies of the questionnaires used are available on request.
We also piloted qualitative interviews about relational care on the wards with a small number of patients (n = 12) and visitors (n = 12) to inform a future process evaluation. An overview of the method and findings from the patient and visitor qualitative interviews is reported in Appendix 5 .
Following baseline data collection in March and April 2015, the ward teams were randomised to intervention or control conditions. The CLECC intervention was implemented on four of the six wards from June 2015, starting on each ward with a 4-month implementation period. Follow-up data were collected on all six wards during February and March 2016.
Pilot cluster randomised trial recruitment
The procedures outlined here mirrored our envisaged procedures for a definitive trial. The opportunity was taken during the feasibility study to evaluate these procedures and further develop them when needed. Pre-screen and screening logs were developed to enable assessments of:
- the timeline of the introduction, approach, discussion and consent process
- the number of people assessed for eligibility
- the number of people approached to join the study
- the number of people recruited into the study
- the number of declined offers and the reasons for these decisions
- participation rates of older patients and patients with cognitive impairment
- the achievement of targets.
Patient recruitment to observations
All adult patients on participating wards were assessed for eligibility to be included in observations of care. Patients were excluded if they were unable to communicate their choices about taking part in the research and a consultee (as defined by the Mental Capacity Act 2005 Code of Practice)110 could not be consulted. Patients who indicated either verbally or non-verbally that they did not wish to take part were excluded, as were patients who were unconscious or those for whom there were clinical concerns that may preclude them from being approached. Patients excluded for clinical reasons included people who were critically ill, at the end of life, or isolated because of a high risk of infection.
The patient sample for observations was determined by randomisation. Up to 24 hours in advance of each scheduled observation, all eligible patients on the ward at the time an observation was scheduled were identified. Patients were placed in a list in alphabetical order of names and each patient was then allocated a number in sequence. A random number generator was then used to select an index patient from the pool of eligible patients.
Each index patient was then approached and informed (verbally with accompanying written information and with aids when needed) about the planned observations. Recruiting researchers were trained to be person-centred and patient, allowing sufficient time for successful communication, and to make environmental modifications to optimise communication. If the index patient indicated verbally or non-verbally that they were happy for the observation to proceed, other eligible patients in the researcher’s field of view were approached, informed about the planned observations, and if they indicated that they were happy for the observation to proceed their care was included in the observations. We did not ask for written consent from patients, but consent was instead recorded by the researcher. If the index patient declined to take part, another index patient was randomly selected, and approached and invited as before. The observations proceeded with data being collected on interactions with all patients who had agreed that the observations could proceed. When an assessment was made that a potential participant did not have capacity to make a decision about taking part in the research, advice was sought from a consultee, in line with the requirements of the Mental Capacity Act 2005 Code of Practice.110 Figure 4 outlines these processes.
FIGURE 4 Recruitment process for observations. (Figure not reproduced here; see the original article.)
Patient sample and recruitment to questionnaire survey
An adapted census approach was used with the patient questionnaire survey. Researchers aimed that all eligible patients on the ward on the scheduled day for questionnaire data collection would be invited to complete a questionnaire. Questionnaire data collection days were planned in advance for each ward to ensure researcher availability, but timing of planned data collection during that day meant that patients would be least likely to be involved in other activities, such as washes, meal times or doctors’ rounds. Before distributing questionnaires, researchers ascertained with the nurse-in-charge which patients were able to be directly approached by a researcher to complete the questionnaire. Patients were excluded if they were critically ill, in receipt of palliative care or at high risk of infection. All eligible patients were then approached and assessed further for eligibility, particularly mental capacity. If the researcher was able to confirm eligibility, patients were informed about the research and invited to take part.
Staff sample and recruitment to questionnaire survey
We planned a census approach to the nursing staff survey in that all RNs and HCAs employed to work in the participating ward teams would be invited to complete a nursing questionnaire. Ward leaders were asked to provide a list of RNs and HCAs employed to work on their ward. Nursing questionnaires were then placed into individually named envelopes and given to ward leaders to distribute to staff.
Visitor sample and recruitment to questionnaire survey
An adapted census approach was used with the visitor questionnaire survey. Researchers aimed that all eligible visitors on the ward on the scheduled day for questionnaire data collection would be invited to complete a questionnaire. Questionnaire data collection periods were planned in advance to ensure researcher availability but also to coincide with visiting time on the individual wards. No exclusion criteria were set for visitors, and so any adult visitors on the ward during a data collection period were approached, informed about the research and invited to take part.
Staff recruitment to qualitative interviews
In addition to the qualitative interviews conducted as part of the process evaluation (which focused on the CLECC intervention implementation), intervention and control ward nursing staff members were also interviewed about ward characteristics, specifically relational care, teamwork and leadership at baseline and at follow-up. The purpose was to provide summary baseline data of group characteristics to inform the pilot CRT. Intervention ward staff recruited to interviews as part of the process evaluation were also questioned about these characteristics during their interviews (see Pilot cluster randomised trial recruitment and Qualitative interviews ). A group of control ward staff was also sampled and recruited in the same way, with their interviews focusing solely on relational care, leadership and teamwork. All RNs and HCAs on control wards were eligible to take part in the qualitative interviews and were invited to take part through posters, presentations and e-mails. We purposively sampled from those who volunteered in order to capture variations in staff grade and ward. Written consent was sought for the staff interviews and staff were given information about not being obliged to take part and their right to withdraw their consent at any time. We offered a payment of a £15 shopping voucher to individual staff who completed an interview. Individuals recruited at baseline were invited to a second interview at follow-up. When such individuals could not be contacted or declined further interviews, new individuals were recruited at the same grade and from the same ward to ensure that variation by ward and grade was maintained.
Pilot cluster randomised trial data collection
Observations
All interactions with eligible patients over a 2-hour observation session were directly observed and coded by a trained researcher. Data gathered included the quality, length and frequency of all interactions between participating patients and staff during each observation session. Data collection was guided by a protocol for use of QuIS in acute settings developed in earlier feasibility work.99 Contextual data were also gathered on the session (number of patients on the ward, staffing levels and skill mix), on the patients (age, gender, evidence of cognitive impairment, agitation at outset of interaction) and on individual interactions (including number of staff, staff type and content of interaction). Patients were assessed as having cognitive impairment in a number of ways, although the research team did not have access to patient records at baseline and so could not look for evidence in the records. Clinical staff were asked about cognitive status before patients were approached. In some cases there was an indicator by the bedside. For instance, in one hospital patients with a known dementia diagnosis would have a magnetic flower mounted on the board by their bed to provide clinical staff with a subtle sign of their cognitive status. If such an indicator was present, this was recorded as evidence of cognitive impairment. In other cases, the researcher would detect signs of impairment as they talked to the patient or during the period of observation.
The platform developed and used for data collection was the Quality of Interactions Tool (QI Tool), a tablet-based interface developed during the feasibility study that enables users to enter data in real-time for subsequent wireless upload to an encrypted central database (see Appendix 6 ).
At each assessment period, researchers observed during time periods (10 × 2-hour observation sessions per ward per 3-week assessment period) randomly sampled over a 3-week period from Monday to Friday, 08.00–22.00. Observation sessions were balanced between wards and time of day. Follow-up data collection was conducted by researchers blinded to ward allocation.
Patient questionnaire survey
Questionnaire responses were written on a hard copy of the questionnaire. If patients agreed to complete a questionnaire, the researcher offered help with completing it and, if the patient was willing, the questionnaire could be completed straight away, taking as much time as was needed, or at a later point in time depending on patient preference. The length of time questionnaire completion took was recorded on each survey. Patients were offered help with interview completion to increase response rates, and whether or not this offer was taken up was also recorded.
Staff questionnaire survey
Questionnaire responses were written on a hard copy of the questionnaire. A post-box was placed on each ward for staff to return their completed questionnaires. Research team visits to the ward and e-mail feedback to ward leaders on completion rates were designed to encourage questionnaire completion. In addition, a prize of shopping vouchers was offered to the team in each hospital with the highest completion rate at each assessment period.
Visitor questionnaire survey
Questionnaire responses were written on a hard copy of the questionnaire. Completed questionnaires were gathered by hand by the researcher.
Completed questionnaires were collated and then individual responses were entered onto the SPSS version 13.0 (IBM Corporation, Armonk, NY, USA) database followed by 100% verification of data entered.
Matron questionnaire survey and ward leader ward profile
Matrons overseeing participating wards were e-mailed a survey, invited to complete it and to e-mail or post it back to the research team. Ward leaders were e-mailed a copy of the ward profile for completion and invited to e-mail or post it back when completed. Ward leaders were also offered help with completing the profile from the research team.
Nursing staff qualitative interviews
One-to-one face-to-face qualitative interviews were undertaken with nursing staff from intervention and control wards, as outlined in Staff recruitment to qualitative interviews , with the purpose of gathering qualitative data on baseline group characteristics. All participants opted to use a hospital meeting room for the interview. The interview schedules were designed to capture individual views and experiences as a member of their ward team, and focused on relational care, leadership and teamwork. The interviews lasted, on average, 46 minutes (range 17–70 minutes). Interviews were audio-recorded, transcribed verbatim, and transcripts were checked for accuracy by the interviewer.
Economic evaluation
The economic evaluation drew on two main data sources outlined above. These were the EQ-5D-5L data from patient questionnaires and the qualitative interview data from staff that focused on the resource implications of CLECC. In addition, the cost of providing CLECC as an intervention was explored.
Chapter summary
This chapter has provided information about the key data sources used to address the study objectives, in particular sampling, recruitment and data collection in relation to the process evaluation, the pilot CRT and the economic evaluation. Chapter 6 provides detail on data analyses.