Background and Purpose The aim of the study was to observe patients on a stroke unit and to compare their activity with that of patients on conventional hospital wards to identify aspects of rehabilitation practice that might account for differences in outcome.
Methods Stroke patients admitted to the hospital were observed on three 8-hour shifts over 3 consecutive days. An observer recorded, at 10-minute intervals, where patients were, what they were doing, and whether their positioning was as recommended by rehabilitation therapists. Patients on a stroke unit were compared with those on conventional wards.
Results Stroke unit patients spent less time by their beds and more time in other locations on the ward (P<.001). There were significant differences in the frequency of behaviors observed in the two types of ward (P<.001). Stroke unit patients had significantly more interaction with nurses and therapists (P<.001). They were also more often in the recommended position (P<.001).
Conclusions The proportion of time in therapeutic activity was low in all locations, with patients spending many hours sitting and doing nothing. Despite this, stroke unit patients had more therapeutic contact with staff and were more often in the recommended position. These two features may account for some of the differences in outcome.
Stroke patients receive their rehabilitation in various settings while in the hospital. This may be on a specialized stroke unit or on general medical or geriatric healthcare wards. Stroke units are reported to provide more coordinated multidisciplinary care than is available on acute-care hospital wards.1 Garraway et al2 compared patients randomly allocated to the Edinburgh Stroke Unit with those who received rehabilitation on medical and geriatric wards. Outcomes indicated that stroke unit patients were more independent on discharge, spent less time in the hospital, and required less therapy. Similar results were reported by Kalra et al.3
Although individual studies have shown small benefits from specialized stroke units, a recent meta-analysis suggests that overall there is an advantage of rehabilitation on a specialized stroke unit.4 Outcome was measured in terms of survival and good functional outcome and was recorded early (6 to 16 weeks) and late (1 year) after stroke. Results showed that even when potential confounding variables were removed, benefits seen in the stroke unit group were not significantly diminished. Thus, Dennis and Langhorne1 concluded that rehabilitation on a stroke unit was associated with improved survival and functional outcome up to 1 year after a stroke.
Apart from coordinated care, some stroke units are able to provide more intensive therapy. Studies directly concerned with treatment intensity have found improvements in activities of daily living (ADL) with intensive treatment compared with nonintensive treatment.5 Sivenius et al6 in a comparison of intensive and standard physiotherapy found improvements for gross motor function and ADL, but by 1 year only motor function abilities remained significantly better in those patients who had received intensive therapy.
In a retrospective study, Wade et al7 estimated that the amount of therapy received on general medical wards was approximately 45 min/d for both physiotherapy and occupational therapy combined. This meant that 3.4% of the patients’ time was engaged in therapy, leaving the rest of the day unaccounted for.
However, these studies have not directly observed patient activities. Systematic observations, despite the heavy demand on the observer’s time, yield more accurate data than interviews or retrospective questioning,8 although there is the possibility that the observation may affect the activity being observed. Kennedy et al9 observed 20 spinal injury patients in both acute and rehabilitation phases of recovery. Patients were watched by observers on 4 consecutive days. Results indicated that patients in a rehabilitation center spent 15.7% of the day (75 minutes) in face-to-face contact with a therapist; 44.2% of the time was in the ward area, with half this time in solitary behavior such as gazing around the ward. Despite reporting these findings to staff, a follow-up 8 months later revealed no improvement.
Observations of stroke patients have been conducted and have indicated variations in the intensity of treatment provided. Keith and Cowell10 used direct observational techniques and found that stroke patients in three different hospitals spent nearly one third of their time in therapeutic activity. This is an average of 93 min/d in physiotherapy and occupational therapy and an additional 10 min/d in other recreational therapy. Studies in the United Kingdom have found less intensive therapy input. Lincoln et al,11 using the same behavior profiling as Kennedy et al,9 investigated the amount of therapy received on a stroke unit. The study revealed that 35.8% of the patients’ time involved inactive solitary pursuits, and 26% was in therapy areas. Thus, time in therapy areas represents 2 hours and 5 minutes per day, although not all of this would necessarily have been active therapy. Tinson12 studied a rehabilitation unit and found that stroke patients spent 12.9% of their time in therapeutic activity, which was approximately 62 minutes of therapy, of which 53 minutes were spent in physiotherapy and occupational therapy per day. Almost half of the day (44%) was spent in nontherapeutic activity, mainly on the ward. Two thirds of this time was spent watching others or gazing around the ward.
These observations suggest that a small proportion of time is spent in therapeutic activity. However, it is not known whether the intensity of rehabilitation differs between a stroke unit and other wards where stroke patients receive rehabilitation. Although studies have reported therapeutic activities in relation to physiotherapy and occupational therapy, rehabilitation is also carried out by nurses. One feature of a stroke unit is that nurses become familiar with rehabilitation techniques used by therapy staff and are therefore able to provide a continuity of treatment that is not possible in other settings. We reported the outcome of a randomized, controlled trial to evaluate the Nottingham Stroke Unit.13 Patients treated on the stroke unit had better outcome in personal and instrumental ADL and better psychological adjustment and mood than those who received their rehabilitation on conventional wards. The aim of the present study was to compare patient activity and therapy on the stroke unit with that provided on general medical wards and wards for health care of the elderly to ascertain features that may account for the differences in outcome.
Subjects and Methods
Observations were carried out in three settings: a stroke unit, wards for health care of the elderly, and general medical wards in four hospitals within Nottingham Health District.
Stroke patients admitted to Nottingham hospitals who met the following criteria were considered: (1) those with no other medical problems requiring continued treatment on an acute medical or geriatric ward; (2) those who had no planned discharge date within 2 weeks; (3) those able to tolerate rehabilitation (in half-hour sessions) for at least 2 hours each day; (4) those able to do two of the following: drink, eat, and wash their own face; (5) those able to transfer from bed to chair with help from no more than two nurses; and (6) those independent for toileting before their stroke.
Patients who met these criteria were included in a randomized, controlled trial to evaluate the effectiveness of a stroke unit.13 They had been randomly allocated to the stroke unit or to remain on other hospital wards. At randomization patients had been assessed on the Barthel Index, Rivermead ADL Scale, and Rivermead Motor Assessment. Patients were eligible for observation from June 1992 to August 1993. At any time the most recently randomized patient was seen first. If this patient was on the stroke unit, observations were carried out on three patients simultaneously, these three being patients who had most recently been admitted to the stroke unit but not previously observed. If the patient was on a conventional ward and there was another patient on the same ward, the two patients were observed simultaneously. If no other patient was on the same ward, one patient was observed. The next two most recently randomized conventional ward patients were then observed. Once six patients had been observed (three on the stroke unit and three on conventional wards), the procedure was repeated beginning with the most recently randomized patient at that time.
One independent researcher, an assistant psychologist who was not involved in the patients’ rehabilitation, recorded observations. She had worked previously on both the stroke unit and wards for health care of the elderly but had no direct involvement with patient services such that she would favor one setting. Observations were made on three consecutive days for 8 hours per day. Three shifts—an early (6 am to 2 pm), a middle (8:30 am to 4:30 pm), and a late (2 pm to 10 pm) shift—were observed for each patient on a weekday. The order of these shifts was randomly allocated. On each shift 48 observations were made. Observations were made on average once every 10 minutes; the time intervals were randomly selected so that intervals ranged between 1 minute and 19 minutes. Three series of time intervals were predetermined, and the series used was decided by random allocation. The order of shifts and time series were prepared in advance of the study with the use of computer-generated random numbers.
The observer explained the procedure to staff and patients before visiting the ward to ensure their cooperation. The observer walked onto the ward noting the action and position of the patient at the specified time. The observations included the location of the patient, the activity being performed, the other people present, and the patient’s position. Locations were categorized as follows: ward in vicinity of own bed; ward nonbedside (other locations on the hospital ward not included in other categories, eg, day room and corridor, toilet, and bathroom); medical (any area where medical activities were taking place, eg, x-ray, theater); rehabilitation, including areas specifically for physical, occupational, or speech therapy; and social, including areas such as social club, pub, outing, or a hair salon.
We recorded the patients’ activities using the categories of solitary behaviors and interactive behaviors, as defined by Kennedy et al.9 Solitary behaviors included the following: (1) isolated disengagement: little external activity, physical isolation, eg, nonspecific gaze while sitting, sleeping; (2) inactive individual task: activity not related to self-care involving engagement with environment, eg, focused attention, watching television, observing others; (3) active individual tasks: solitary tasks involving gross physical movements, eg, pushing a table away, reading a book; and (4) independent self-maintenance: activity associated with self-care, eg, eating.
Interactive behaviors included the following: (5) individual interaction task: one-on-one communication associated with task activities, eg, helping with washing; (6) individual interaction verbal: one-on-one communication during nontask activity, eg, chatting with relatives; (7) group interaction task: as for category 6 but involving more than two people, eg, two physiotherapists helping a patient to walk; (8) group interaction verbal: as for category 7 but with more than two people, eg, a family visit; and (9) formal meetings, eg, patient education by ward staff.
The position of the patient was also noted and whether it was beneficial to recovery or not. Positions were classified as good, poor, or not possible to say. This classification was based on clinical experience and theoretical background14 without any evidence for their effect on outcome. The criteria for good and poor positioning were that when sitting, patients should be symmetrical, with trunk and head in the midline and knees in the midline and not rolled outward. The feet were to be level on the floor or on footplates and not rolled in or out. Heels and not just toes were to be touching the floor. The arm should be supported on the lap or on a pillow and not hanging over the side of the chair. The patient should be upright and not slipping down or to one side. In lying, patients should not be flat on their backs or in an uncomfortable position. The “good” position was with the patient on the side with pillows positioned to support the arms and legs. Standing was rated as good if the balance seemed controlled and the support provided helped but was not excessive. In transfers, the trunk should be handled and not the arms; a “poor” rating was made if staff were seen to pull on the patients’ arms or lift them under their armpits.
From June 1, 1992, to August 1, 1993, 760 patients were admitted to Nottingham hospitals and were included in a Register of Stroke Patients. Of these, 171 were included in the evaluation of the stroke unit. Of the 589 patients excluded, 86 died before random allocation, 59 were not independent for toileting before their stroke, 112 were not medically fit for transfer to the stroke unit, 105 had a planned discharge date within 2 weeks, 57 were unable to do two basic ADL, 17 were unable to be transferred from bed to chair by two nurses, 181 were not able to tolerate therapy for 2 hours per day by 5 weeks after stroke, and 7 had met the criteria for inclusion but there had been no bed available at the time they met the criteria. Of the 171 patients included in the ward evaluation study at that time, 76 were observed. Of these, 39 were on the stroke unit and 37 on general medical and geriatric wards.
The biographical characteristics of the patients are shown in Table 1⇓. There were no significant differences between patients on the stroke unit and conventional wards in regard to sex or side of stroke, but stroke unit patients were significantly older than those on the conventional wards. There were no significant differences in disability as assessed at randomization on the Barthel Index, Rivermead ADL Scale, or Rivermead Motor Assessment. Patients included in the observation study were compared with those in the randomized control trial but not observed. There were no significant differences in age (U=7976, P=.21), sex (χ2=0.97, P=.33), or side of stroke (χ2=0.96, P=.81).
The location of patients is shown in Table 2⇓. Stroke unit patients spent less time by their beds and more time in rehabilitation and social areas than patients on conventional wards. Stroke unit patients spent less time in medical areas than those on conventional wards. The location was divided according to the time of day. Observations between 8:30 am and 4:30 pm were considered part of the rehabilitation day; those outside these times were part of the nonrehabilitation day. The location of patients differed significantly between the stroke unit and conventional wards during both time periods (rehabilitation day, χ2=253, P=.001; nonrehabilitation day, χ2=132, P<.001). Stroke unit patients spent less time by their beds, less in medical areas, and more in social areas. Patients were not observed in rehabilitation areas during the nonrehabilitation day, but stroke unit patients were more often in rehabilitation areas during the rehabilitation day. The amount of time in rehabilitation areas was 46 min/d on the stroke unit and 21 min/d on conventional wards. Some rehabilitation also took place outside the rehabilitation areas, and therefore this is less than the amount actually received.
The frequency of each observation category for the different types of ward is shown in Table 3⇓. There was a significant difference between the types of ward in the behaviors observed. Stroke unit patients spent less time in isolated disengagement, individual verbal interactions, and group verbal interactions and more time in individual tasks, self-care, and task interactions. However, the magnitude of these differences was small. A comparison of behavior categories according to time of day was calculated. The distribution of the categories was significantly different between the stroke unit and conventional wards during both the rehabilitation day (χ2=303, P<.001) and nonrehabilitation day (χ2=80, P<.001).
The distribution of interactions with staff, visitors, and other patients is shown in Table 4⇓. Stroke unit patients had significantly more interaction with nurses and therapists and less with doctors, visitors, and porters (P<.05). The amount of time in interaction was 31 minutes in an 8-hour day with nurses and 36 minutes with therapists on the stroke unit and 18 minutes in an 8-hour day with nurses and 21 minutes with therapists on conventional wards. There were no significant differences in the proportion of interactions with other patients.
Patient position was compared between wards, and the results are shown in Table 5⇓. Stroke unit patients spent significantly less time lying down and significantly more time sitting and standing. Comparisons between the types of ward were significant for both the rehabilitation day (χ2=428, P<.001) and nonrehabilitation day (χ2=172, P<.001), but the differences were more marked during the nonrehabilitation day. During the nonrehabilitation day, stroke unit patients spent 70% of their time sitting and 23% lying down, whereas those on conventional wards spent 51% of their time sitting and 40% lying down. Differences were less marked during the rehabilitation day when stroke unit patients spent 84% of their time sitting and less than 1% lying down, whereas those on conventional wards spent 76% of their time sitting and 12% lying down.
The distribution of quality of positioning is also shown in Table 5⇑. Stroke unit patients were less often exposed to poor practice than patients on conventional wards. Significant differences occurred both during (χ2=295, P<.001) and outside (χ2=241, P<.001) the rehabilitation day. Most examples of poor positioning were observed on conventional wards during the nonrehabilitation day, and most examples of good positioning were observed during the rehabilitation day on the stroke unit.
The results reported are based on the observations of one researcher. Although she reported no affiliation to the stroke unit, she may have had expectations about the nature of differences between the settings. Given the number of observations, it seems unlikely that significant bias has occurred since the magnitude of differences are small, even though they are statistically significant. Also, some differences are in a direction contrary to that which might have been predicted. There was a significant difference between the groups in age. This is unlikely to have affected the findings, but if it had, then it would be likely to reduce the difference between the groups rather than increase them. Older patients might be expected to receive less treatment, and if this age difference is a factor, then the actual differences would be greater than those reported.
The proportion of time spent in therapeutic activity was low, which is consistent with previous observations on the same unit11 and with other units in the United Kingdom.12 If the therapeutic working day is considered to be 9 am to 4:30 pm, then on the stroke unit only 36 min/d are spent in contact with either a physiotherapist or occupational therapist. Therapeutic activities with nurses covered a longer day but were also relatively infrequent. However, some therapeutic activities may occur outside rehabilitation areas and with no therapist present. In addition, some therapeutic activities were not observed because patients could not be traced when they were absent from the ward, and some therapists refused the observer access to therapy areas. However, even if all missing observations were therapeutic, the proportion of time in contact with therapists would still not be much higher. Since differences in outcome occurred,13 it suggests that a relatively small increase in therapeutic contact might improve the outcome of patients on conventional wards.
Differences between the therapeutic activity on the stroke unit and on other wards were small, despite better therapist staffing levels on the stroke unit. This may be a result of stroke unit therapists having other commitments, such as teaching. However, overall estimates indicate that twice as much time is spent in rehabilitation areas and in contact with nurses and therapists on the stroke unit compared with other hospital wards. This suggests that the better outcome of stroke unit patients13 may be due to more intensive therapy. The absolute amount of therapy is not high. It should therefore require relatively little additional input to stroke patients on conventional wards to produce an improvement in outcome. Whether providing the additional therapy on conventional wards achieves this effect needs to be evaluated prospectively in a randomized, controlled trial.
In all settings observed, patients spent many hours sitting and doing nothing. The people with whom they had the most contact during their rehabilitation were visitors. If effective use is to be made of hospital rehabilitation resources, then strategies that enhance self-directed practice or treatment by visitors would seem desirable. Visitors are rarely used as assistants in therapy. The philosophy of the stroke unit was to allow visitors to be present on the wards at all times providing that it did not interfere with patients’ rehabilitation. Relatives were taught specific skills, such as transfers and dressing, in preparation for home visits and discharge home. Relatives of patients on conventional wards had greater restriction of visiting hours and less interaction with therapists then those on the stroke unit. Visitors therefore provide the most interaction with patients, but these sessions are rarely therapeutic. If visitors were actively encouraged to conduct rehabilitation activities with patients, this might improve rehabilitation outcome.
Despite random allocation, greater “medical” involvement was noted on conventional wards. It is unlikely that this was due to differences in the patient groups since analysis of patient characteristics did not demonstrate any consistent differences between the two groups. The differences could be due to stroke unit staff having more familiarity with and confidence in their ability to manage the medical aspects of stroke. They may therefore seek the advice of the physicians less than those on conventional wards. Physicians carried out a detailed weekly case conference and ward round but were not always present on the ward. Alternatively, the physicians and medical investigations may be more a part of the routine on a general medical or geriatric ward, and therefore medical care may have been provided automatically.
Social activities were more often engaged in by patients on the stroke unit than by those on conventional wards. It is a policy of the stroke unit to incorporate outings, trips, and social activities in the rehabilitation program. This is seen to be therapeutic and part of preparing a patient for discharge into the community. Such social rehabilitation may reduce the handicap imposed by stroke and facilitate psychological adjustment.
Many hours are spent sitting or lying. Lying down was more frequent on conventional wards, possibly because of the routine of these wards. Since other patients on conventional wards needed to rest, it would not have been noticeable that the stroke patients spent more of the time than was necessary in bed. There may also have been less time for nurses on these wards to assist with transfers into chairs. Since much time is spent sitting or lying, the quality of positioning may affect the outcome. Although there is no evidence that positioning reduces spasticity or improves outcome, it seems unlikely that poor posture will assist recovery. Patients on the stroke unit were positioned poorly at times, but this was less frequent and for shorter periods of time than on conventional wards. This could be because staff on conventional wards were unaware of recommended positioning or did not respond when they noticed patients poorly positioned. Alternatively, staff could simply have not been available to correct patients’ posture. The former explanation seems more likely, since poor positioning continued even when nurses were present. This suggests that an educational program on recommended positioning could be evaluated on conventional wards to determine whether this affects observed positions and outcome.
Comparisons of these observations with those on the same unit but a different hospital ward 5 years earlier11 indicate some changes. In the present study the stroke unit patients spent more time in therapy areas than patients studied from 1987 to 1988, but despite this they had fewer interactions with therapists or nurses. This may reflect an increase in administrative or teaching duties. Participation in solitary active tasks and independent self-care shows an increase. This may reflect the introduction of an activities nurse, whose role has been to introduce patients to leisure activities.
The observation of stroke patients in different settings has indicated some factors that may have contributed to the different outcome of patients who received their rehabilitation on the stroke unit compared with those who were treated on conventional wards. These factors include therapeutic activities, task-oriented rather than verbally oriented interactions, contact with nurses and therapists, and maintaining recommended positions. Which of these factors is the most important will need to be determined from further randomized, controlled trials.
This study was supported by Trent Regional Health Authority. We would like to thank S. Lomas, H. Corson, J. Garnett, C. Ashman, and E. Iliffe for assistance with data collection, W.B. Clarke for assistance with data analysis, P. Radley for assistance with preparing the manuscript, and the ward staff for agreeing to the observations.
- Received July 26, 1995.
- Revision received October 2, 1995.
- Accepted October 16, 1995.
- Copyright © 1996 by American Heart Association
Dennis M, Langhorne P. So stroke units save lives: where do we go from here? BMJ.. 1994;309:1273-1277.
Garraway WM, Akhtar AJ, Prescott RJ, Hockey L. Management of acute stroke in the elderly: preliminary results of a controlled trial. BMJ.. 1980;280:1040-1043.
Kalra L, Dale P, Crome P. Improving stroke rehabilitation. Stroke. 1993;24:1462-1467.
Smith DS, Goldenberg E, Ashburn A, Kinsella G, Sheikh K, Brennan PJ, Meade TW, Zutshi DW, Perry JD, Reeback JS. Remedial therapy after stroke: a randomised controlled trial. BMJ.. 1981;282:517-520.
Sivenius J, Pyorala K, Heinonnen OP, Salomen JT, Riekkinen P. The significance of intensity of rehabilitation after stroke: a controlled trial. Stroke. 1985;16:928-931.
Wade DT, Skilbeck CE, Langton Hewer R, Wood VA. Therapy after stroke: amounts, determinants and effects. Int J Rehabil Med. 1984;6:105-110.
Keith RA. Activity patterns of stroke rehabilitation unit. Soc Sci Med. 1980;14:575-580.
Kennedy P, Fisher K, Pearson E. Ecological evaluation of a rehabilitative environment for spinal cord injured people: behavioural mapping and feedback. Br J Clin Psychol. 1988;27:239-246.
Keith RA, Cowell KS. Time use of stroke patients in three rehabilitation hospitals. Soc Sci Med. 1987;24:529-533.
Juby LC, Lincoln NB, Berman P. The effect of stroke unit rehabilitation on functional and psychological outcome: a randomised controlled trial. Cerebrovasc Dis. In press.
Lynch M, Grisogono V. Strokes and Head Injuries: A Guide for Patients, Families, Friends and Carers. London, England: John Murray; 1991.