Skip Navigation


Age and Ageing Advance Access originally published online on December 18, 2006
Age and Ageing 2007 36(2):184-189; doi:10.1093/ageing/afl146
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
36/2/184    most recent
afl146v1
Right arrow E-Letters: Submit a response to the article
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Olai, L.
Right arrow Articles by Svärdsudd, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Olai, L.
Right arrow Articles by Svärdsudd, K.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Copyright © The Author 2006. Published by Oxford University Press on behalf of the British Geriatrics Society.

Prognosis assessment in stroke patients at discharge from hospital

Lena Olai1,2,, Marianne Omne-Pontén2, Lars Borgquist3 and Kurt Svärdsudd1

1 Uppsala University, Department of Public Health and Caring Sciences, Family Medicine and Clinical Epidemiology Section, Uppsala, Sweden
2 Centre for Clinical Research, Dalarna, Sweden
3 Linköping University, Department of Health and Society, General, Practice, Linköping, Sweden

Address correspondence to: L. Olai. Tel: +462318316 Fax: +462318375. Email: lena.olai{at}Ltdalarna.se


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
 References
 
Background: accurate prognostic assessments of need of help, health and dwelling situation in stroke patients are important for patient management, rehabilitation, discharge planning, and for providing reliable information to patients and their relatives.

Objective: to analyse factors affecting the accuracy of discharge prognosis assessments.

Design: prospective study of stroke patients discharged from hospital.

Setting: two cities in central Sweden.

Subjects: three hundred and ninety stroke patients, 65 years or older, living in their own homes and having no dementia diagnosis prior to hospital admission.

Methods: at discharge, physicians, nurses, occupational therapists and physiotherapists in the Departments of Internal Medicine and Geriatrics were asked to make an individual prognosis assessment regarding patients’ need for help, health and dwelling situation at 3 and 12 months after admission to hospital.

Results: the prognosis assessments were on average accurate in 68.4% (3 months) and 61.5% (12 months), far better than chance (33.3%). There were no significant differences between staff categories. The accuracy was influenced by a number of patient linked factors, such as activity degree, household situation before admission, and Mini Mental State Examination level and need of help measured 1 week after discharge, and ranged from 22 to 89%, depending on factor combinations.

Conclusions: prognosis assessments based on clinical judgement were on average quite accurate but the accuracy varied markedly with patient linked factors. Feedback of outcome might be one way to further improve the accuracy of prognosis assessment.

Keywords: prognosis assessment, outcome, functional capacity, health situation, dwelling, stroke, elderly


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
 References
 
On a global level, stroke is the second largest cause of death and it remains a serious health problem for survivors after discharge from hospital [1]. Accurate assessments of the expected health and functional development are needed for clinical decision-making [2], to guide patient management and improve rehabilitation [3], for discharge planning [2, 3], and to provide appropriate prognostic information to patients and next of kin [2–4]. A number of models have been developed to improve prognosis assessments [5], some complex, others easy to use. However, in medical practice clinical judgement is most often used.

Problems inherent in clinical prognosis assessment in stroke patients is the heterogeneity of the patient group [6] and the many professional categories involved in the care of patients, possibly causing limited dissemination of important information among staff or a limited knowledge of the patient needs.

We therefore made a study of the accuracy of prognostic assessments made by four staff categories in a prospective longitudinal study of stroke patients followed from hospital discharge until 1 year after admission, and of possible accuracy determinants.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
 References
 
Setting
The study was performed in the cities of Falun (population 55,000) and Borlänge (population 47,000), central Sweden, with similar age and sex distributions as the national Swedish population [7]. The two cities are served by Falun General Hospital, where stroke patients were cared for in the Department of Internal Medicine, either in a stroke unit or a general internal medicine ward and afterwards possibly at a geriatric ward. Since hospital admission is free of charge for the patient virtually all patients with clinical signs and symptoms indicating stroke are admitted to hospital [8].

Subjects
The inclusion criteria were 65 years or older, discharged alive from the Department of Internal Medicine during the period 1 September 1999 to 31 May 2001 after an acute stroke, including intracerebral haemorrhage, brain infarction, or stroke of undetermined pathological type (ICD10 codes I61, I63, I64) [9], living in own home before admission, and no pre-admission dementia diagnosis.

Out of 432 potential participants 42 died in the Department of Internal Medicine. The 390 survivors constitute the study population (Appendix 1 available online at http://www.ageing.oxfordjournals.org). One week after discharge 351 patients were interviewed by a registered nurse. A second (n = 324) and a third (n = 297) interview were performed 3 and 12 months after admission. Approximately 95% of the eligible subjects participated in each interview. Reasons for non-participation are given in Appendix 1. For some non-participants, data on dwelling and functional capacity could be obtained from other sources. Forty interviews were performed in duplicate with excellent agreement ({kappa} ≥ 0.95). In addition, the nearest next of kin, indicated by the patient, received a questionnaire at the time of the first interview, of whom 261 (75.4%) of the eligible responded.

Data collection
The following variables were used as measures of pre-morbid state. Data on independence regarding transfer, personal hygiene and dressing immediately prior to the stroke incident were obtained for 365 subjects from the national stroke register [10]. Information on community home help during the year before admission was obtained by scrutinizing the Social Service records, and on help from next of kin by patient interviews and nearest next of kin questionnaire. Information on appointments and home calls with primary health care centre (PHCC) staff, hospital admissions, out-patient appointments, and co-morbidity, was obtained by scrutinising PHCC and hospital patient records. At the first patient interview activity degree prior to admission was measured with the Activity Score subscale of the Gothenburg Quality of Life instrument [11], which includes a list of 34 specified activities with response alternatives ‘performed often or regularly’ (=2), ‘occasionally’ (=1) or ‘never’ (=0). The scores were then summed to an overall total score. In addition, information on education, marital status and co-habitation was sought.

Remaining functional ability was estimated from a Katz ADL [12] assessment and a Mini Mental State Examination (MMSE) [13] at the first patient interview.

At each transfer between the stroke unit, internal medicine or geriatric wards and at discharge from the hospital, the patient’s physician, nurse, occupational therapist and physiotherapist were asked to assess the prognosis regarding ‘need of help’, ‘health situation’, and ‘dwelling’ at three and 12 months after admission to hospital, based on their own clinical judgement, or to leave items blank in case of inability to make an assessment. Altogether more than 50 assessors participated, the exact number unknown because of granted anonymity.

The outcomes to be forecasted were given in three levels based on data from the second or third patient interview or register data. The staff was asked to indicate for each of three outcomes at what level they thought the patient would be at 3 and 12 months. For the outcome variable ‘need of help’ the three levels were ‘no need of help’ (Katz ADL degree A), ‘need of some help’ (Katz B–E), and ‘completely dependent’ (Katz F–G). For the outcome variable ‘health situation’ the levels were ‘completely recovered’, ‘unchanged situation or better but no full recovery’, and ‘deteriorated or deceased’. For the outcome variable ‘dwelling’ the levels were ‘living in own apartment or house’, ‘living in special housing’, and ‘living in nursing home’.

Finally, the basis for the assessments was sought. Possible responses were medical, nursing, or rehabilitation state, tests, discharge planning information, expected support from next of kin, experience, intuition, other (open-ended alternative), or combinations.

Altogether 1470 assessment forms (1385 completed) for 292 subjects were obtained, of whom 117 had assessments made on two occasions. In 85 patients no assessments were made because of staff inability to make an assessment, in most cases due to few days in the ward, and in 13 patients because of death in the geriatric ward. The average number of staff category assessments was 3.6 per patient and assessment occasion. The inter-staff agreement (Spearman’s correlation coefficients pair-wise across all combinations) at 3 and 12 months was 0.60–0.63 for ‘need of help’, 0.36–0.42 for ‘health situation’ and 0.70–0.76 for ‘dwelling’. All subjects gave oral informed consent and the Research Ethics Committee of Uppsala University approved the study.

Statistical considerations
Data were analysed with the SAS programme package [14]. Assessment accuracy was measured as number of accurate assessments in relation to total number of assessments done and expressed as a percentage, 33.3% representing chance given the assumption of no knowledge among the staff of the outcome distribution in this patient category.

The analysis of factors influencing the prognosis assessment accuracy was performed with univariate logistic regression with accuracy of individual assessments (yes or no) as the dependent variable and potentially influencing factors as independent variables. Factors affecting at least one outcome were then entered in final multivariate analyses.

Optimism and pessimism were measured as the proportion of inaccurate assessments more favourable or less favourable than the real outcome. Only two-tailed tests were used. To account for multiple testing P-values <0.005 were regarded as significant.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
 References
 
Characteristics of the study population
Mean age was 78.7 years, and 224 (57.4%) were females, Table 1. More than half had only compulsory schooling, half of the men were married and half of the women were widows. Before admission to hospital half of the women and one-third of the men were living alone in their own homes. Pre-morbid state is shown in Table 1. The majority could dress, maintain their personal hygiene, transfer without assistance, had some form of help and had contacts with PHCC staff during the past 12 months before admission to hospital. More than one-fifth had contacts with the hospital as in- or outpatients. One-third had a previous cerebro-vascular event. A history of hypertension, heart disease or diabetes mellitus was common.


View this table:
[in this window]
[in a new window]

  		Table 1. Socio-economic data and pre-morbid state during 12 months before admission to hospital

 
Forty-four percent of the women and 51% of the men were cared for in the stroke unit, the remaining in general internal medicine wards. The average time spent in these wards was 7 days, inter-quartile range 3–9, 43% were discharged from the hospital and the remaining patients were transferred to a geriatric ward where they stayed on average for another 30 days, inter-quartile range 14–41. The dominant diagnosis was cerebral infarction, 88%, equal for men and women. In 97%, a computerised tomography scan was performed. Remaining functional ability measured 1 week after discharge showed that slightly more than half were in Katz grade A, one-fourth were classified as grade F or G, and the remaining were in grades B–E. Mean MMSE-level was 23.3.

Outcome and prognosis assessments
Outcomes at 3 and 12 months after admission to hospital are shown in Table 2. More than half of the patients had no need of help, an unchanged health situation or no full recovery, and more than 75% lived in their own homes. The overall accuracy of the assessments in 3 months was on average 68.4% and for 12 months 61.5%. For 3 months the accuracy was 64.5% for ‘need of help’, 62.4% for ‘health situation’, and 78.4% for ‘dwelling’, and the corresponding proportions for 12 months were 60.1, 56.2 and 68.2%, all highly significantly better than chance (P < 0.0001). The inaccurate assessments were dominated by overly optimistic assessments regarding the ‘health situation’ and overly pessimistic assesments regarding ‘dwelling’.


View this table:
[in this window]
[in a new window]

  		Table 2. Outcome 3 and 12 months after admission to hospital, based on interviews and register data

 
In univariate logistic regression analyses patient age, co-habitation, pre-morbid state (activity degree, hypertension history), remaining functional ability, staff occupational experience, ward speciality, and assessment made by a physician all affected the assessment accuracy of one or more outcomes. No influence was found for patient sex, co-morbidity other than hypertension, first or recurrent stroke event, stroke type, time lag between assessment and outcome per assessment period, and staff categories other than physician.

In a set of multivariate logistic regression analyses, with assessment accuracy entered as dependent variable and all the variables univariately affecting outcome accuracy entered as independent variables the effect size and direction of each factor on outcome was fairly consistent in the 3 and 12 months’ perspective, although some were significant and others were not (Table 3). The accuracy of ‘need of help’ assessments was affected negatively, i.e. made the assessment more difficult, by the factors living alone before admission and need of help at discharge and positively, i.e. facilitated the assessment, by activity degree before admission and hypertension history. The accuracy of ‘health situation’ assessments was affected negatively by ward speciality (stroke unit), and positively by pre-morbid state (activity degree), and remaining functional ability. The accuracy of ‘dwelling’ assessments was affected negatively by patient age, living alone before admission, and need of help at discharge, and positively by activity degree before admission and MMSE-level at discharge.


View this table:
[in this window]
[in a new window]

  		Table 3. Adjusted odds ratios for factors influencing the accuracy of prognosis assessment. Significant ratios are in bold text

 
Combinations of determinants affected the assessment accuracy markedly. In the example shown in Appendix 2 (available online at www.ageing.oxfordjournals.org) the accuracy of the assessments of ‘health situation’ at 3 months ranged from 22% for independent subjects with low MMSE score, to 89% for dependent subjects with high MMSE score. The corresponding range for 12 months was 23 to 77%.

The overall most common bases for the assessments were rehabilitation status (78.2%), medical status (74.5%), nursing status (60.7%), experience (38.0%), intuition (23.0%), and expected support from next of kin (17.9%).


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Conclusions
 Key points
 Declaration of conflicts of...
 References
 
Prognosis assessments at discharge from hospital regarding the course of events three and 12 months after admission to hospital were thus far more accurate than expected by chance. The accuracy was influenced by a number of patient linked factors, with a somewhat different set up for each outcome. Combinations of these factors affected assessment accuracy markedly.

The study was performed as a prospective longitudinal cohort study of consecutive patients surviving the internal medicine or stroke unit phase of hospital care. There were no differences in age, sex and diagnosis distribution between the study group and national stroke patient data [15]. Some of the interviews were recorded in duplicate with excellent agreement. All forms were validated or tested with satisfactory results before the data collection period began. We have therefore no reason to believe that the results are affected by selection or other bias.

A large number of prognostic models involving a variable number of predictors have been proposed. Most of the 83 models reviewed by Counsell and Dennis [5] had potentially serious deficiencies in internal and statistical validity, many had limited generalisability, and none had been adequately validated. Few studies have addressed the issue of clinical judgement versus formal decision support. Counsell et al. found no model markedly better than clinical judgement [16]. They reported 65% accurate prognosis assessments in inpatients, approximately the same level as our 56–78%. The 3 and 12-month results in the present study were fairly consistent.

Several of the measures used in this study such as living alone, pre-morbid state, event severity and the course during the hospital stay, and remaining functional ability are probably used consciously or unconsciously in clinical judgement. After adjustments, there were no significant differences in assessment accuracy among the staff categories. Benner [17] pointed out the importance of experience, but in this study the factor experience was offset by other, more significant factors. Several studies [16, 18], including the present one, showed that various staff categories tend to be overly optimistic in their assessments of the patients’ health situation and long-term need of help, and in our study also overly pessimistic about dwelling, perhaps because of low knowledge of the ability among home help and PHCC staff to manage severe illness. The ‘health situation’ assessments performed by the stroke unit staff were less accurate than assessments performed in other wards, which may be surprising. However, their assessments were done fairly early in the hospital period. The patients' condition may have changed in the later course. In addition, the patients were referred back to their general practitioners, limiting the information feedback on outcome to the assessing staff. One way to further improve the accuracy of prognosis assessment might be to provide more information feedback to the assessing staff.

The study results are limited by the inclusion of only subjects 65 years or older, admitted from their own homes and with no documented dementia prior to admission. However, by this restriction a number of problems are avoided, such as dementia limiting the interview data reliability, and for subjects under 65 years support from social insurance and work rehabilitation authorities would make the study population more homogeneous. The strengths of the study are the large study population, the large number of assessors, the high data measurement precision and the prospective design.


   Conclusions
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
Key points
 Declaration of conflicts of...
 References
 
The accuracy of prognosis assessments at discharge from hospital regarding the course of events 3 and 12 months after admission to hospital was far better than expected by chance. The accuracy was influenced by patient linked factors, such as activity and household situation before admission, and MMSE-level and need of help at discharge.


   Key points
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
 Key points
Declaration of conflicts of...
 References
 

  • Prognostic assessments regarding ’need of help’, ’health situation’, and ’dwelling’ in elderly stroke patients performed by four staff categories were far more accurate than expected by chance.
  • There was a considerable variation in the accuracy of assessments, 22 to 89%, due to individual factor combinations.
  • Feedback might be one way to further improve the accuracy of prognosis assessment.


   Acknowledgements
 
This study was supported by grants from the Vardal Foundation, Centre for Clinical Research Dalarna, the Dalarna County Council and Uppsala University.


   Declaration of conflicts of interests
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
References
 
None of the authors has any conflict of interest regarding the content of this report.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusions
 Key points
 Declaration of conflicts of...
 References
 

  1. Ingall T. (2004) Stroke–incidence, mortality, morbidity and risk. J Insur Med 36 143–52.[Medline]
  2. Ostbye T, Steenhuis R, Wolfson C, Walton R, Hill G. (1999) Predictors of five-year mortality in older Canadians: the Canadian Study of Health and Aging. J Am Geriatr Soc 47 1249–54.[Web of Science][Medline]
  3. Appelros P. (2002) Stroke Severity and Outcome: In Search of Predictors Using a Population-based Strategy. Stockholm Karolinska Institute Library.
  4. Collins TC, Petersen NJ, Menke TJ, Souchek J, Foster W, Ashton CM. (2003) Short-term, intermediate-term, and long-term mortality in patients hospitalized for stroke. J Clin Epidemiol 56 81–7.[CrossRef][Web of Science][Medline]
  5. Counsell C and Dennis M. (2001) Systematic review of prognostic models in patients with acute stroke. Cerebrovasc Dis 12 159–70.[CrossRef][Web of Science][Medline]
  6. Kwakkel G, Wagenaar RC, Kollen BJ, Lankhorst GJ. (1996) Predicting disability in stroke–a critical review of the literature. Age Ageing 25 479–89.[Abstract/Free Full Text]
  7. . Statistics Sweden. (2002) Population Statistics. Stockholm Available from:http://www.scb.se/. Accessed October, 1.
  8. Asplund K, Bonita R, Kuulasmaa K, et al. (1995) Multinational comparisons of stroke epidemiology. Evaluation of case ascertainment in the WHO MONICA Stroke Study. World Health Organization Monitoring Trends and Determinants in Cardiovascular Disease. Stroke 26 355–60.[Abstract/Free Full Text]
  9. . International Classification of Diseases and Related Health Problems. (1997) Tenth Revision, Swedish version (ICD10). Stockholm National Board of Health and Welfare.
  10. Asplund K, Hulter Asberg K, Norrving B, Stegmayr B, Terent A, Wester PO. (2003) Riks-stroke - a Swedish national quality register for stroke care. Cerebrovasc Dis 15Suppl. 1, 5–7.
  11. Tibblin G, Tibblin B, Peciva S, Kullman S, Svardsudd K. (1990) "The Goteborg quality of life instrument"–an assessment of well-being and symptoms among men born 1913 and 1923. Methods and validity. Scand J Prim Health Care Suppl 1 33–8.[Medline]
  12. Katz S, Ford A, Moskowitz R, Jackson B, Jaffe M. (1963) Studies of illness in the aged. The index of ADL: a standardized measure of biological and psychological function. JAMA 914–9.
  13. Folstein M, Folstein S, McHugh P. (1975) "Mini-Mental State": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12 189–98.[CrossRef][Web of Science][Medline]
  14. Statistical Analysis System. (1995) Cary, NC SAS Institute Inc. SAS Campus Drive.
  15. Glader EL, Stegmayr B, Norrving B, et al. (2003) Sex differences in management and outcome after stroke: a Swedish national perspective. Stroke 34 1970–5.[Abstract/Free Full Text]
  16. Counsell C, Dennis M, McDowall M. (2004) Predicting functional outcome in acute stroke: comparison of a simple six variable model with other predictive systems and informal clinical prediction. J Neurol Neurosurg Psychiatry 75 401–5.[Abstract/Free Full Text]
  17. Benner P. (1982) From novice to expert. Am J Nurs 82 402–7.[CrossRef][Web of Science][Medline]
  18. Korner-Bitensky N, Mayo N, Cabot R, Becker R, Coopersmith H. (1989) Motor and functional recovery after stroke: accuracy of physical therapists’ predictions. Arch Phys Med Rehabil 70 95–9.[Web of Science][Medline]
Received 5 June 2006; accepted in revised form 24 October 2006.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 StrokeHome page
L. Olai, M. Omne-Ponten, L. Borgquist, and K. Svardsudd
Survival, Hazard Function for a New Event, and Healthcare Utilization Among Stroke Patients >=65 Years
Stroke, November 1, 2009; 40(11): 3585 - 3590.
[Abstract] [Full Text] [PDF]

Home page
 NeurologyHome page
A. E. Baird
Improving stroke prognosis
Neurology, October 6, 2009; 73(14): 1084 - 1085.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
36/2/184    most recent
afl146v1
Right arrow E-Letters: Submit a response to the article
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Olai, L.
Right arrow Articles by Svärdsudd, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Olai, L.
Right arrow Articles by Svärdsudd, K.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?