Agar 2017.
| Study characteristics | ||
| Methods |
Aim: to assess the effects of facilitated case conferencing (FCC) versus usual care in improving EoL care for people with advanced dementia living in nursing homes Study design: cluster‐RCT; 2 arms (FCC intervention; usual care) Unit of randomisation: nursing home (stratified by organisational affiliation) Consumer involvement: none explicitly stated Funding source: Australian Department of Health. Authors declared they have no conflicts of interest |
|
| Participants |
Participants: people with advanced dementia and their carers Setting and geographic location: residential nursing homes in Sydney and Brisbane (Australia) Methods of recruitment: Sites: identified from Australian government list (websites) and approached in alphabetical order to participate (to minimise selection bias) 20 sites in major cities Family members provided consent for resident participation in trial Selection criteria for participation in study: Inclusion:
Diagnosis of person approaching EoL: as above for diagnostic markers; defined so as to identify people with FAST stage 7 and functional dependency, which in turn identifies people with an average survival of less than 6 months Target of intervention: Resident (person with dementia) Family member or friend who knows the person well (prior to dementia diagnosis) involved in making decisions on the patient’s behalf Protocol stated: visits the resident at least once/fortnight, knew resident prior to dementia diagnosis, willing to be involved in decisions about care, English proficiency at a level to allow completion of outcome assessments Age: patients: intervention 84.7 (SD 7.9), UC 85.8 (SD 8.2) Gender: intervention 61% female, UC 58% female Ethnicity/culture/language: born in Australia intervention 70%, UC 52% (significantly different) Other PROGRESS aspects: Focus was on advanced dementia/nursing homes with at least 50% patients with dementia, larger sized residential homes. Urban populations No exclusions mentioned re: literacy level, comorbidity, etc. but results may not be applicable to smaller/non‐specialist nursing home settings and populations Numbers of participants: see Additional Table 1 |
|
| Interventions |
Intervention: facilitated case conferencing (FCC) and patient‐centred palliative care training Aim of intervention: to improve care at the EoL (primary) Considerations included train‐the‐trainer, evidence‐based organisational culture change and expected advanced dementia trajectory Comparison: usual care No education, training or support provided over usual practice. No restriction on service‐provider training/education where this was usual practice Further details of UC not reported. Authors note however that the difference between intervention and UC sites may have been narrowed by the use of means other than CC to share decisions with residents/family members (e.g. 1:1 conversations). These may have influenced EoL care and satisfaction in the UC group (i.e. narrowed the gap between intervention and UC) Delivered by: nurse, trained as Palliative Care Planning Co‐ordinator (PCPC), worked 2 days/week or equivalent to:
Co‐intervention(s): not reported (other than palliative care training) Setting: residential nursing home; no further details Materials, procedures, content: Nurse, trained as Palliative Care Planning Co‐ordinator (PCPC), worked 2 days/week or equivalent to:
Content (key features): predefined specific clinical triggers for CC; shared agenda setting model (resident, family, multidisciplinary team) could identify areas for discussion; required attendance by resident and/or family/decision‐makers; facilitate by PCPC to ensure optimal participation by attendees; followed by communication strategy to summarise actions and plan from CC. In‐person meeting When and how much: author response indicates further data on these factors being analysed. Also reported median duration was 48 minutes (IQR 30 to 60). Sessions were always conducted at a single session (but some residents had more than 1 over time) Tailoring: discussion topics in FCC meeting tailored to what was important to the resident. Could include care planning, current and future treatment decision‐making, information sharing, meeting residents’ needs or preferences, ACP Author response indicated that topics were identified/put on the agenda for the CC meeting via the PCPC seeking "advice on what to include on the agenda from the resident’s family, GP and nursing staff beforehand (including people who were unable to attend). There was also usually a trigger for the CCs and any related issues were by default on the agenda (e.g. return from hospital, declining health, family concern)" Modified during study: no. Fidelity monitoring was modified to account for variations in level of detail of reporting across nursing homes Fidelity assessed: fidelity was assessed (‘dose at resident level’) for per protocol analyses at resident level. Not assessed as planned as many units did not collect information to make the assessment. Reverted to a simpler measure of intervention dose: resident received a CC or not Dose at nursing home level assessed extent to which PCPCs could: work 2 days/week; supported by managers; fulfilled expectations of training; diffused their role among other staff Theoretical base: factors considered when developing the intervention, and purpose, seem reasonable and logical. No further specific theoretical constructs |
|
| Outcomes |
Primary outcomes None reported Primary outcomes ‐ adverse events None reported Secondary outcomes Family‐rated EoL care: 3 subscales [Quality of EoL care, family‐rated]
Method: face‐to‐face or telephone interview with research team Timing: 4 to 6 weeks after death Scale and scoring: End of Life in Dementia Scales
Resident CAD‐EOLD [Quality of EoL care, nurse‐rated] Method: nurse‐rated. Face‐to‐face or telephone interview with research team Timing: as soon as possible after death of patient Scale and scoring: CAD‐EOLD; higher scores = better (more comfort) Resident SM‐EOLD [Quality of EoL care, nurse‐rated] Method: nurse‐rated. Face‐to‐face or telephone interview with research team Timing: as soon as possible after death of patient Scale and scoring: SM‐EOLD; higher scores = better (lower symptom frequency) Quality of life [Quality of life] Method: nurse‐rated Timing: 3‐monthly Scale and scoring: Quality of life in Late‐stage Dementia (QUALID); 11‐item scale Staff attitudes to, knowledge of and confidence in providing palliative/EoL care [Health practitioner evaluation of preparedness to communicate] Method and timing: assessed for PCPCs before and after training (by research staff) Assessed for other facility staff before and after training (by PCPC) Scale and scoring: Palliative Care for Advanced Dementia tool, 35 items (qPAD) Costs [Costs of subsequent care] Method: training, CC and routine healthcare costs to be considered Cost utility (benefit estimated as QALYs). QoL for economic analyses to be assessed by nurse‐rated EQ‐5D‐5L Person‐centred approach to care [Ratings of concordance with patient preferences for EoL care] Method: rated by observation, resident and family reports and documentation Timing: unclear Scale and scoring: Care and Activities and Interpersonal Relationships and Interactions domain of the Person‐Centred Environment and Care Assessment Tool (PCECAT). 18 items, each rated 0 (not at all) to 3 (all of the time) |
|
| Notes | Protocol prospectively registered 2012, updated 2017 (Improving dementia end of life care at local aged care facilities; www.anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12612001164886) Cluster‐RCT; ICC (0.050 used to calculate required sample size (not met because of unexpectedly low mortality rate amongst sample). Analyses described adjusted for ICC values (reported). Individual results presented. Seems likely that results were appropriately adjusted. ICCs scores reported for different scales and were reported as variable (some lower, some higher than predicted 0.05 level) (e.g. see page 6) This trial reported data related to care received in the last month of life. We did not extract these data as we judged them clinical, rather than fitting with the focus of this review on communication Data were not available/analysable for the following outcomes: quality of life (QUALID), staff attitudes to and knowledge of providing palliative/EoL care (qPAD), person‐centred care (PCECAT), or costs. Duration of EoL discussions was available only for the intervention arm (non‐comparative) and was therefore not included in the review |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Computer‐generated allocation sequence; block randomisation Allocation (unit of randomisation) by nursing home (stratified by 2 factors: organisational affiliation (part of organisation or not); dementia‐specific unit or not) |
| Allocation concealment (selection bias) | Low risk | No details in trial report Protocol states that statisticians responsible for allocating sites were blinded to allocation (page 21) |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Authors state that staff, family members and residents were blinded to the aim of the study (but those in the intervention sites may have noticed changes in practice) |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | Research staff were blinded to study aim; only collected data from sites in 1 arm to minimise the chance they would identify differences in practice between intervention and usual care sites |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Large proportion of data (participants randomised) missing. However, these were comparable for the 2 study groups, and was due to participants not dying during the study period (outcomes assessed for this study were focused on those around death) Withdrawal rates for other reasons were low and comparable across groups |
| Selective reporting (reporting bias) | Low risk | All primary outcomes set out in protocol are reported QUALID scores (nurse‐assessed) were planned for assessment 2‐weekly; in trial report this was 3‐monthly Other outcomes related to patient‐centred care, attitudes, knowledge, etc. were not reported but stated in protocol. These are unpublished and data were provided by the trial authors (not used in the review) |
| Other bias | Low risk | Baseline differences between groups: higher staff knowledge levels intervention group at baseline (16 (4) versus 14 (3)) in usual care group Less frequent visitors in intervention group (19% had daily visitors versus 34% usual care group) Unclear whether these represent important sources of bias Authors took measures to avoid contamination between sites Selective recruitment of cluster participants: authors approached nursing homes on Australian registry/list in alphabetical order to minimise selection bias |