Age and Ageing Advance Access originally published online on June 1, 2007
Age and Ageing 2007 36(5):589-592; doi:10.1093/ageing/afm057
Post-acute transfer of older people to intermediate care services: the sooner the better?
SIR—We recently reported a randomised controlled trial investigating locality-based community hospital (CH) care for older people compared to elderly care wards in a district general hospital (DGH)[1]. The intention-to-treat analysis, appropriate to the pragmatic study, demonstrated improved independence outcomes at 6 months in favour of CH care. Outcome was assessed by change in the Nottingham extended activities of daily living scale (NEADL) [2]. The adjusted mean (95% confidence interval) between-group difference in outcome was 5.30 (0.64–9.96) points. The study protocol, based on existing local clinical practice, included the expectation that patients randomised to CH care would transfer within 2 days of randomisation. We anticipated that clinical service exigencies might intrude causing delays in transfer. To investigate the effect of delayed CH transfer on outcomes, we undertook pre-specified additional analyses, the results of which we report here.The trial methods are described in full elsewhere [1]. Briefly, patients who had been admitted acutely to a care of the elderly department in a DGH were eligible for the study if they were registered with a general practitioner in the primary care trust served by the study CH and were considered by the responsible geriatrician to be medically stable and in need of post-acute rehabilitation care. Patients were then approached for recruitment to the study by a research nurse who was independent of the ward team. The trial was approved by the local research ethics committee and patients and/or relatives gave informed consent. After recruitment, patients were randomised (in a ratio of 2:1 CH:DGH) to receive rehabilitation in the study CH or to remain for rehabilitation in the DGH elderly care wards. Patients in the CH group were assessed by the consultant-led multidisciplinary team and received an individual care plan. Patients in the control group received ongoing specialist elderly care department multidisciplinary care involving consultants, nurses, therapists, dieticians and pharmacists and in accordance with good clinical practice as described in the National Service Framework for Older People. Each patient had an individualised care programme as determined by their multidisciplinary needs assessment. Transfer to the CH was arranged independently of the research team by a member of a case management team employed by the primary care trust within which the CH was located and the established local protocol was to transfer patients within two working days. We collected data contemporaneously on the reasons for delay in transfer to the CH; or, where this information was not available, we inspected the medical records.
Patients were assessed at baseline, at 1 week after hospital discharge, and at 3 and 6 months after recruitment. Our primary outcome measure was the NEADL [2], a valid and reliable measure [2], [3] of independence in four areas of daily life: mobility, kitchen, domestic and leisure activities. The score range is 0–66; higher scores are associated with greater independence. The effects of delays in CH transfer were investigated in exploratory analyses. Baseline characteristics and outcome were compared for three sub-groups of patients who had received the intended treatment allocation: those patients who had transferred to the CH within the specified 2 days (‘early transfer’ group); those who had transferred to the CH after 2 days (‘late transfer’ group) and the controls (‘no transfer’ group) who remained as allocated in the DGH.
The study was funded by The Health Foundation. The funders played no role in the design, execution, analysis and interpretation of data, or writing the study report.
Statistical analysis
We used a scatter plot and Pearson's correlation coefficient to examine the relationship between time to transfer and outcome for patients in the CH group. We carried out two adjusted comparisons of the changes in scores on the primary outcome measure (NEADL score) from baseline to 6 months using analysis of covariance to adjust for the baseline variables of age, sex, institutional care, and baseline Barthel index (BI) score [4], [5]: one analysis comparing ‘early transfer’, ‘late transfer’ and control groups (with post hoc investigation of pairwise differences using a Bonferroni adjustment); and the other analysis comparing CH and control groups (all patients treated as allocated) but including time to transfer as a covariate.
Two hundred twenty patients were randomised: 141 to locality-based CH care (intervention group) and 79 to usual care (control group). The randomised group allocation was not achieved for 26 patients: 16 patients remained inappropriately in the DGH, ten were transferred inappropriately to the CH. The main reasons for allocation deviation were CH bed closure through infective gastro-enteritis (8/16) and extreme DGH bed pressures (8/10). When the 16 patients who were unable to transfer to CH care are excluded, there were 73 patients who transferred or died within 2 days of randomisation (72 transferred, 1 died), and 49 patients who transferred after more than 2 days. Three patients died in hospital after more than 2 days without transfer (Figure 1).
|
Fifty-four of the 73 patients who were transferred to the CH in 2 days or less were assessed at 6 months (17 died, 2 withdrew); 37 of the 49 patients who were transferred more than 2 days after randomisation were assessed at 6 months (12 died) and 50 of the 69 patients who remained in the DGH group were assessed at 6 months (18 died, 1 withdrew).
Comparisons of baseline measures showed that the three groups were similar (Table 1). Patients in the three groups who were assessed at 6 months (‘early transfer’ group n = 54; ‘late transfer’ group n = 37; ‘no transfer’ group n = 50) were also similar at baseline (data not shown). The range of times from randomisation to transfer was 0–12 days. The median time from randomisation to transfer for the ‘late’ transfer group was 4 days. The median (interquartile) length of stay from randomisation to hospital discharge in the ‘early’ transfer CH group was 15 (9–24) days compared to 20 (13–30) days in the ‘late’ transfer group (Table 1). Reasons for delay in transfer were: 31 (63%) patients: organisational and/or administrative time taken to organise transfer; seven (14%) patients were not able to transfer promptly as the CH was temporarily closed (the main reason for closure was because of infective gastro-enteritis); seven (14%) patients became unwell after randomisation but before transfer; one (2%) patient initially refused to transfer then changed her mind; one patient (2%) was awaiting a test and result; two (4%) patients had no reason recorded why there was a delay in transfer.
|
The scatter plot suggested a tendency towards worse outcome following late transfer and this was supported by a significant negative correlation (Pearson correlation coefficient r = –0.239;P = 0.023) (please see the Appendix 1 in the supplementary data on the journal website http://www.ageing.oxfordjournals.org/). There was a significant difference between the ‘early’ transfer (n = 54), ‘late’ transfer (n = 37) and control (n = 50) groups for change in NEADL scores from baseline to 6 months using analysis of covariance to adjust for baseline variables (ANCOVA P = 0.030). Pairwise comparisons showed a significant difference only between the ‘early’ transfer and control groups. Adding time to transfer as a covariate improved the fit in the adjusted comparison of changes in scores on the NEADL from baseline to 6 months between the groups of patients managed as randomised (‘early’ plus ‘late’ transfer groups, n = 91; control group n = 50) and showed a significant difference in favour of the CH group (adjusted mean difference 9.88; 95% CI 3.83–15.93; ANCOVA P = 0.002).
The effects of timing of post-acute transfer to intermediate care have not been previously investigated. We hypothesised that, if intermediate care confers benefit over usual care, then earlier transfer might be associated with better outcomes than delayed transfer. We have explored this hypothesis using outcome data previously obtained in a randomised trial evaluation of CH care for older people [1].
Some caution is required in the interpretation of the results as the analyses are exploratory, are based on small patient numbers and may be subject to bias as only those patients treated according to randomised allocation and assessed at 6 months (141/220 = 64% of the study group) have been included. In addition, the scatter plot showed some outliers (large increase in NEADL or long transfer time) which influenced the analyses and strengthened the relationship between outcome and early transfer. Nonetheless, we have found provisional evidence that time to CH transfer was an important factor that was associated with improved independence at 6 months. The evidence emerges from three separate analyses. Firstly, there is a significant inverse correlation between NEADL independence outcomes and time to transfer. Secondly, the improved independence outcomes we observed in the trial appear to be confined to the early transfer sub-group of patients. The NEADL score changes for the usual care (DGH) and the delayed transfer groups were very similar. Lastly, the inclusion of the factor ‘time to transfer’ is associated with an additional improvement in the analysis of covariance statistical model.
It is tempting to conclude that patients benefit more from early transfer, but the association could also be explained by ill-health causing delay in transfer. For seven patients, the reason given for delay suggests this was the case. However, for the majority (31 patients; 63%), the reason for delay was organisational and/or administrative rather than patient related.
These provisional findings need clarifying in future prospective studies but, if confirmed, would be important for practitioners, patients and researchers. The importance to practitioners and patients is self-evident—that systems of care should ensure the transfer for post-acute rehabilitation should be as soon as possible after medical stability has been achieved. Researchers would need to consider transfer delays in the design of future intermediate care studies. Indeed, delays in transfer may be an important reason for apparent lack of differences in outcomes in comparison studies investigating intermediate care services [6].
- Transfer to a community hospital is associated with improved independence at 6 months for older people requiring post-acute rehabilitation care.
- The time to transfer may be important: in an exploratory analysis, late transfer to a community hospital was associated with a worse outcome at 6 months as measured by the Nottingham extended activities of daily living scale.
None
Supplementary data for this article are available online at http://ageing.oxfordjournals.org.
1 Academic Unit of Elderly Care and Rehabilitation, University of Leeds, St Luke's Hospital, Little Horton Lane, Bradford, West Yorkshire BD5 0 NA, UK
2 Aysgarth Statistics, Aysgarth Cottage, 3 Mynchen Road, Knotty Green, Beaconsfield, Buckinghamshire HP9 2AS, UK
*To whom correspondence should be addressed Email: john.young{at}bradfordhospitals.nhs.uk
References
- Green J, Young J, Forster A. A randomised controlled trial evaluation of locality-based community hospital care for older people. Br Med J (2005) 331:317–22. originally published online 1 Jul 2005; DOI: 10.1136/bmj.38498.387569.8F.
[Abstract/Free Full Text] - Nouri FM, Lincoln NB. An extended activities of daily living scale for stroke patients. Clin Rehabil (1987) 1:301–5.
[Abstract/Free Full Text] - Harwood RH, Ebrahim S. The validity, reliability and responsiveness of the Nottingham extended activities of daily living scale in patients undergoing total hip replacement. Disabil Rehabil (2002) 24:371–7.[CrossRef][Web of Science][Medline]
- Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J (1965) 14:61–5.[Medline]
- Collin C, Wade DT, Davies S. The Barthel ADL Index: a reliability study. Int Disabil Stud (1988) 10:61–3.[Medline]
- Young JB, Robinson M, Chell S. A whole system study of intermediate care services for older people. Age Ageing (2005) 34:577–83.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||