C.R. Burton et al.
23. Eriksen HM, Iversen BG, Aavitsland P. Prevalence of nosocomial infections and use of antibiotics in long-term care
facilities in Norway, 2002 and 2003. J Hosp Infect 2004; 57:
316–20.
24. Büla CJ, Ghilardi G, Wietlisbach V, Petignat C, Francioli P.
Infections and functional impairment in nursing home residents: a reciprocal relationship. J Am Geriatr Soc 2004; 52:
700–6.
25. Eckford SD, Keane DP, Lamond E, Jackson SR, Abrams P.
Hydration monitoring in the prevention of recurrent idiopathic
urinary tract infections in pre-menopausal women. Br J Urol
1995; 76: 90–3.
26. Beetz R. Mild dehydration: a risk factor of urinary tract
infection? Eur J Clin Nutr 2003; 57: 52–8.

Age and Ageing 2010; 39: 554–559
doi: 10.1093/ageing/afq077
Published electronically 20 July 2010

27. Newman DK. Urinary incontinence, catheters, and urinary
tract infections: an overview of CMS Tag F 315. Ostomy
Wound Manage 2006; 52: 34–44.
28. Rutchik SD, Resnick MI. The epidemiology of incontinence in
the elderly. Br J Urol 1998; 82: 1–4.
29. Bliss DZ, Savik K, Harms S, Fan Q, Wyman JF. Prevalence
and correlates of perineal dermatitis in nursing home residents.
Nurs Res 2006; 55: 243–51.
30. Berg RW, Milligan MC, Sarbough FC. Association of skin wetness and pH with diaper dermatitis. Pediatr Dermatol 2008;
11: 18–20.
Received 7 July 2009; accepted in revised form 14 May 2010

© The Author 2010. Published by Oxford University Press on behalf of the British Geriatrics Society.
All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

The palliative care needs of acute stroke patients:
a prospective study of hospital admissions
CHRISTOPHER R. BURTON1, SHEILA PAYNE2, JULIA ADDINGTON-HALL3, AMANDA JONES4
1

Centre for Health Related Research, Bangor University, Bangor, UK
Division of Health Research, Lancaster University, Lancaster, UK
3
School of Health Sciences, University of Southampton, Southampton, UK
4
Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK
2

Address correspondence to: C. R. Burton. Tel: +44 (0) 1248 382556. Email: c.burton@bangor.ac.uk

Abstract
Background: despite a mortality rate of approximately 30% in acute stroke, little is known about the palliative care needs of
this group of patients.
Design: prospective study of 191 acute stroke patients admitted to hospital in England. Biographical, medical and strokerelated data were collected. Participants completed the Sheffield Profile for Assessment and Referral to Care (SPARC), a
screening tool for referral to specialist palliative care.
Findings: over 50% reported moderate to significant fatigue-related problems. Approximately 50% reported symptom-related
problems (e.g. pain) or psychological distress (e.g. anxiety). Approximately 25% had concerns about death or dying, and 66% had
concerns about dependence and disability. Over 50% were worried about the impact of stroke on family members. There were
significant main effects of dependence (Barthel Index) (F1,123 = 12.640 P = 0.001) and age (F4,123 = 3.022 P = 0.020), and a
significant three-factor interaction between dependence, age and co-morbidities (F9,123 = 2.199 P = 0.026) in predicting total
SPARC scores.
Conclusions: acute stroke patients have a high prevalence of palliative care needs. Acute stroke services should use the
SPARC for needs assessment. Priority for assessment should be given to patients with a score of <15/20 on the Barthel
Index, a tool already used in most stroke services.
Keywords: stroke, palliative care, needs assessment, elderly

554

The palliative care needs of acute stroke patients

Introduction
Over 130,000 people in the United Kingdom have a stroke
each year, three-quarters of whom are aged over 65. Stroke
is the third most common cause of death in the United
Kingdom and the single most common cause of severe disability. The ‘acute’ stroke phase is def ined as the f irst
30 days after stroke, during which nearly one-third of patients will die [1].
The UK stroke service model [2] reflects international
principles of good practice including rapid admission to specialist, hospital-based stroke units, early access to specialist
rehabilitation services, the transfer of ongoing inpatient rehabilitation into community settings and a range of longterm health and social care services that support life with
stroke.
The National Clinical Guidelines for Stroke [3] and National Stroke Strategy [2] recommend that all stroke patients
should have access to specialist palliative care expertise and
that staff providing this care receive appropriate training.
Within cancer, palliative care is defined as the alleviation
of pain and discomfort when a cure is not possible [4],
and can be an important component of a range of health
care services, whether patients are terminally ill or not,
and in diseases other than cancer [5, 6]. As the stroke disease trajectory is variable, the challenge is to identify models
of practice that support the timely integration of palliative
care.
The end of life care tools to guide practice recommended
in the End of Life Care Strategy [7] include the Liverpool
Care Pathway (LCP) [8] providing generic recommendations
for terminal care. The applicability of the LCP to acute
stroke has not been systematically evaluated, although it
may improve prescribing practice and communication about
prognosis [9]. The Macmillan Gold Standards Framework
[10] supports the strategic evaluation of palliative care in
community services, and the Preferred Priorities of Care
tool [11] supports advance care planning. Whether these
strategies address the clinical challenges of embedding palliative care in acute stroke services, such as difficulties in
accurate prognostication, has yet to be tested.
Prospective information on the palliative care needs of
stroke patients is lacking. A critical review of the international literature [12] identif ied seven studies, four of
which were completed within the United Kingdom. Many
patients did not receive optimal symptom control. Patients
reported insufficient help to overcome psychological morbidity, and informal caregivers identified difficulty accessing
information. The caring experience was distressful and unrewarding, with high reports of insufficient help. The review
concluded that palliative care may have a role in the care of
stroke patients and should be systematically provided on the
basis of need.
The National Clinical Guidelines for Stroke make no distinction between those patients who die in the acute and
later stages of the disease pathway, although these groups
may require different palliative care strategies and interven-

tions. In non-acute stroke, patients have some time to make
plans for death, and professionals have opportunity to implement appropriate interventions. This study focuses on
the palliative care needs of acute stroke patients about which
little is known.

Methods
Aim

The aim of this study was to identify the palliative care needs
in a consecutive cohort of acute stroke patients. The local
ethics committee approved the study.
Study population and sampling

Within one UK city, all acute stroke patients admitted to either
of two acute hospitals were eligible for inclusion in this study.
Stroke was confirmed by CT scan. Children and patients with
subarachnoid haemorrhage were excluded. Patients with a
serious psychiatric illness or severe dementia were excluded
if no family member was available to provide assent.
The study benefitted from the use of a centralised stroke
register. After admission and initial confirmation of diagnosis, all patients were seen by a stroke specialist nurse who
commenced the integrated stroke pathway, consisting of
protocols for all aspects of multi-disciplinary acute, rehabilitation and transfer of care. Stroke Specialist Nurses provided
patients or family members with a study information sheet.
Twice weekly, they informed the research team of those patients introduced to the study for follow-up and recruitment.
A researcher then visited or telephoned the patient (or their
family member) and arranged a suitable time to discuss the
study and take consent. If it was not possible to approach
the patient directly (for example, if the patient was unconscious or had cognitive difficulties), a family member was
approached. Witness consent was available where it was
not possible for the patient to give written consent.
Recruitment ran from October 2006 to June 2007, during which attempts were made to recruit all acute stroke
admissions. Recruitment was regularly reviewed to ensure
the study sample was as consecutive as possible.
Data collection

Biographical data, medical history and stroke-related data
were extracted from patients’ medical records on recruitment.
The Barthel Index [13] assessing functional dependence was
completed by stroke unit staff and cut-offs (1–14; 15–20)
used to identify levels of disability.
Palliative care needs were assessed using the Sheffield
Profile for Assessment and Referral to Care (SPARC), developed to screen patients with advanced illness, regardless of
diagnosis, for specialist palliative care referral [14, 15]. The
SPARC supports clinical decision-making rather than providing a quantitative indicator of whether referral is
warranted. Tool development was underpinned by an extensive literature review and interviews with patients and

555

C.R. Burton et al.
refined through testing in patients referred to specialist palliative care services [15]. It is composed of five domains:

Table 1. Study participants

•
•
•
•
•

............................................................

physical symptoms (21 items)
psychological issues (9 items)
religious and spiritual issues (2 items)
independence and activity issues (3 items)
family and social issues (4 items).

Each item is scored on a four-point scale according to how
participants have been affected by each issue in the previous
month. Two items relating to problems with mobility and fear
of falling were added. The mean length of time from admission to hospital to completion of the SPARC was 7.77 days
(SD 7.55 days). The SPARC was completed by patients as a
questionnaire or by interview with a member of the research
team. Where patients were unable to complete the SPARC, a
close family member was asked to complete on their behalf.
Fifty-nine (31%) assessments were completed in this way.
Due to the complexity of stroke-related impairments, using
carers as proxy raters is a feature of applied stroke research,
although there is a tendency for carers to rate functional problems higher than patients [16]. In palliative care research,
carers can be a vital source of non-clinical data. Whilst agreement in subjective aspects of the patient’s experience can be
problematic, carers can provide reliable information on service quality and observable symptoms [17].
Data analysis

Descriptive statistics and graphical displays provided a
prevalence assessment of palliative care needs in patients
with acute stroke. Predictors of need were identified using
multiple regression (Enter method) and explored using General Linear Modelling techniques.

Results
Study sample

A total of 290 patients were introduced to the study, of
whom 191 stroke cases were recruited (Table 1). The age
(25.7% aged <65 years) and gender profile of participants
is typical of the acute stroke population. Reasons for nonrecruitment were lost to follow-up (13, 4%), declined (28,
9%), died before consent could be taken (24, 8%), cognitive
impairment with no family member (15, 5%), incorrect diagnosis (12, 4%) and rapid discharge (7, 2%).
Profile of needs and problems

The proportions of participants for each level of severity on
SPARC items within each domain are summarised graphically in the Supplementary data available in Age and Ageing
online.
The SPARC physical domain appeared to be sensitive:
although a considerable prevalence of problems and needs
was evident, only approximately 10% report any other ‘un-

556

N (%)

Gender

Male
Female
Age groups
<55
56–65
66–75
76–85
>85
Lives with
Alone
Partner/other family member
Sheltered housing/care home
Co-morbidities
None
1
2
3
4
5 or more
Stroke hemisphere
Left
Right
Unassigned
Stroke history
None recorded
Previous TIA
Previous stroke
Barthel Index score on admission 15 or more
<15

100 (52.4)
91 (47.6)
25 (13.1)
24 (12.6)
32 (16.8)
65 (34.0)
45 (23.6)
91 (47.6)
92 (48.2)
8 (4.2)
3 (1.6)
35 (18.3)
52 (27.2)
48 (25.1)
32 (16.8)
21 (11.0)
76 (39.8)
106 (55.5)
9 (4.7)
131 (68.6)
21 (11.0)
39 (20.4)
93 (48.7)
98 (51.3)

TIA, transient ischaemic attack.

controlled symptoms’. Nearly 80% experienced some form
of communication problem. Fatigue-related problems were
prevalent, with over 50% reporting moderate to significant
problems with ‘feeling weak’, ‘feeling tired’ or ‘being sleepy
during the day’. Approximately 50% reported problems with
‘pain’, ‘memory loss’, ‘headache’, ‘restlessness’ or ‘bladder
problems’. The additional two items indicated that mobility
problems were highly signif icant, which combined with
speech problems, represents the clinical picture. Forty percent had moderate to significant ‘fear of falling’. Mirroring
the prevalence of fatigue-related concerns, 70% reported
problems with ‘feeling everything’s an effort’.
Approximately half experienced some form of psychological distress such as ‘anxiety’, ‘low mood’, ‘confusion’,
‘poor concentration’ and ‘loneliness’. Whilst the prevalence
of religious or spiritual concerns was low, approximately
25% had concerns about death or dying. Approximately
two-thirds had concerns relating to dependence and disability.
Over 50% were worried about the effects of their stroke
on others within their support network. Twenty five percent
felt that they needed more help than their family could
provide.
Predictors of need

Patient dependence was a significant predictor of scores on
physical, psychological, and religious and spiritual SPARC
domains, and with female gender and age, on the dependence and disability domain. Higher levels of dependence
and female gender were associated with greater concerns

The palliative care needs of acute stroke patients
about dependence and disability. Those patients under
55 years of age and those over 75 years of age also had
a higher prevalence of these problems. Patient dependence,
female gender and the number of co-morbidities were significant predictors of scores on the SPARC family and
social domain.
A model of the data was constructed using the four factors
shown to be significant within the regression. This confirmed
the significant main effects of the Barthel Index score (F1,123 =
12.640 P = 0.001) and Age (F4,123 = 3.022 P = 0.020) in predicting total SPARC scores. The model also indicated a
significant three-factor interaction between Barthel Index,
Age and Co-morbidities (F9,123 = 2.199 P = 0.026).

Discussion
We aimed to prospectively investigate the palliative care
needs of a cohort of patients admitted with acute stroke.
Our pilot study highlighted that many stroke service staff
equate palliative and terminal care. Nurses were reluctant
to refer patients to the study until they appeared virtually
moribund. We excluded ‘palliative care’ from subsequent
study material, with the approval of the ethics committee.
We succeeded in assessing a sizeable cohort of patients
within 1 week of admission, which we are confident is
comprehensive and unbiased, and typical of UK acute
stroke services.
Our study sample was representative of age and gender.
Our findings may be different for religious and socio-cultural
groups: we recruited few participants from ethnic minority
groups within this study. A total of 47.6% of the cohort lived
alone. Current demographic shifts in the age of the UK
population, alongside increased mobility and cultural diversity, are leading to structurally and geographically dispersed
families. Consequently, single-person households are increasingly common; older women are more likely than older men
to live alone [18].
As stroke policy emphasises rapid access to specialist services [2], more patients with severe stroke are likely to be in
hospital requiring palliative and end of life care. Using a validated tool to identify palliative care needs, we identified a
wide range of needs, with significant proportions of patients
experiencing them.
Nearly 80% experienced fatigue-related symptoms. This
challenges the shift towards more intensive and early mobilisation after stroke [19] and early supported discharge [2].
Similar proportions experienced some form of low mood,
anxiety, confusion or poor concentration. In addition, nearly
one in five patients expressed some suicidal feelings. In
comparison, 20–30% of patients with advanced cancer have
a formal psychiatric diagnosis, the most common being depression [20]. This suggests the importance of providing
opportunities for patients to express their distress and access
to specialist mental health services.
Studies have demonstrated patients’ fears about feeling a
burden to others, including family members [21, 22]. Whilst

80% felt that family support was sufficient, it should be
noted that our sample was hospitalised. The impact of the
challenges faced by patients and families during transfer of
care and renegotiating life with stroke in community settings
may not have begun to be felt.
Lower physical independence is associated with problems identified by the SPARC. Female gender and age
(both young strokes and those aged over 75 years) are also
associated with concerns about dependence and disability.
Female gender and the presence of co-morbidities are associated with higher levels of family and social concerns. Our
data suggest that the influence of co-morbidities is nonlinear:
patients with no or multiple co-morbidities experienced
greater problems. This possibly reflects the development
of coping strategies by those with multiple co-morbidities
and differs from the results of previous studies [23].
Whether we can extrapolate the problems identified in
this study to a need for specialist palliative care requires careful consideration. Evidence and policy now recognises the
need for a range of health and social care services to ameliorate the full range of stroke sequelae [2, 3]. Research has
identified the need for the integration of palliative care within a range of conditions with high morbidity including heart
disease [24, 25], chronic obstructive pulmonary disease [26],
end-stage renal disease [27] and chronic illness [28]. Many of
the problems identified within this study are part of both the
palliative care and stroke literatures. For example, poststroke fatigue is a well-documented problem, associated
with a high case fatality [29]. In the absence of models to
explain the mechanisms of post-stroke fatigue, both palliative care and rehabilitation perspectives may reduce its
impact and warrant further investigation.
The problems identified by patients or proxies in this
study do not, of course, imply that these are palliative care
needs, or that a generalist model of palliative care within the
acute stroke service, or referral to peripheral specialist palliative care services is required. The SPARC does not provide a
threshold score or problem profile which is indicative of the
need for specialist referral. This may be a useful line of enquiry in further research, requiring follow-up of patients and
care processes over time, as needs change.
Lack of capacity within specialist palliative care [30] underpins the aim of the NHS End of Life Care Strategy [7],
enhancing the quality of generalist palliative care. Embedding a generalist model of palliative care in non-cancer
services requires a debate about the capacity and capability
of the generalist service to address the needs and concerns
of patients and families, clarity about the timing of engagement with specialist services and the development of
practice tools to support model implementation.

Conclusion
Acute stroke patients have a high prevalence of problems
which may be met by the integration of principles of palliative care into stroke services. The incorporation of a targeted

557

C.R. Burton et al.
needs assessment is a useful first step in the refinement of a
generalist model of palliative care for acute stroke. We recommend the SPARC as the basis for needs assessment for
those patients with a score of less than 15 on the Barthel
Index, a routine assessment already used in most stroke services. This will serve as a prompt for staff to consider
palliative care issues, facilitate dialogue between staff, patients and families and, importantly, opens up possibilities
for discussion of sensitive issues including fears about death
and dying. Whilst some of the patient needs identified within this study are also recognised within the stroke literature,
for example post-stroke fatigue, there is a need to investigate
a combined approach to management, integrating principles
of rehabilitation and palliative care.

Key points
• The palliative care needs of acute stroke patients are
poorly understood.
• A screening tool for referral to specialist palliative care
indicated a high prevalence of needs in this patient
group.
• Priority for assessment should be given to patients
scoring <15/20 on the Barthel Index.

Supplementary data
Supplementary data mentioned in the text is available to
subscribers in Age and Ageing online.

Acknowledgements
We wish to thank all the patients, families and staff who
supported this research and to the Stroke Association for
funding this study.

References
1. Roberts SE, Goldacre MJ. Case fatality rates after admission to
hospital with stroke: linked database study. BMJ 2003; 326:
193–4.
2. Department of Health. National Stroke Strategy. London: Department of Health, 2007..
3. Intercollegiate Working Party for Stroke. National Clinical
Guidelines for Stroke. London: Royal College of Physicians,
2008.
4. National Institute for Clinical Excellence. Improving Supportive and Palliative Care for Adults with Cancer. London:
National Institute for Clinical Excellence, 2004..
5. Dunlop R. Specialist palliative care and non-malignant diseases.
In: Addington-Hall JM, Higginson IJ, eds. Palliative Care for
Non-Cancer Patients. Oxford: Oxford University Press, 2001.

558

6. Sepulveda C, Marlin A, Yoshida T, Ullrich A. Palliative care:
the World Health Organization’s global perspective. J Pain
Symptom Manage 2002; 24: 91–6.
7. Department of Health. End of Life Care Strategy. Promoting
High Quality Care for All Adults at the End of Life. London:
Department of Health, 2008.
8. Ellershaw J, Smith C, Overill S, Walker SE, Aldridge J. Care of
the dying: setting standards for symptom control in the last 48
hours of life. J Pain Symptom Manage 2001; 21: 12–7.
9. Jack C, Jones L, Jack BA, Gambles M, Murphy D, Ellershaw
JE. Towards a good death: the impact of the care of the
dying pathway in an acute stroke unit. Age Ageing 2004;
33: 625–6.
10. Thomas K. The Gold Standards Framework. A Programme
for Community Palliative Care. London: Macmillan Cancer
Relief, 2004.
11. Pemberton C, Storey L, Howard A. The Preferred Place of
Care document: an opportunity for communication. Int J Palliat Nurs 2003; 9: 439–41.
12. Stevens T, Payne SA, Burton C, Addington-Hall J, Jones A.
Palliative care in stroke: a critical review of the literature. Palliat
Med 2007; 21: 323–31.
13. Wade DT. Measurement in Neurological Rehabilitation. Oxford: Oxford University Press, 1992.
14. Ahmedzai SH, Payne SA, Bestall JC, Ahmed N, Clark D,
Noble B. Improving Access to Specialist Palliative Care. Developing a Screening Measure to Assess the Distress Caused
by Advance Illness that may Require Referral to Specialist Palliative Care. Sheffield: University of Sheffield. Academic
Palliative Medicine Unit, Royal Hallamshire Hospital and Trent
Palliative Care Centre.
15. Ahmed N, Bestall JC, Payne SA, Noble B, Ahmedzai SH. The
use of cognitive interviewing methodology in the design and
testing of a screening tool for supportive and palliative care
needs. Support Care Cancer 2009; 17: 665–73.
16. Knapp P, Hewison J. Disagreement in patient and carer
assessment of functional abilities after stroke. Stroke 1999;
30: 934–8.
17. McPherson C, Addington-Hall JM. Judging the quality of care
at the end of life: can proxies provide reliable information? Soc
Sci Med 2003; 56: 95–109.
18. Del Bono E, Sala E, Hancock R, Gunnell C, Parisi L. Older
People: A Gendered Review and Secondary Analysis of the
Data. Institute for Social and Economic Research. University
of Essex. Equal Opportunities Commission. Working Paper
Series No. 62.
19. Indredavik B, Bakke F, Slordahl S, Rokseth R, Haheim LL.
Treatment in a combined acute and rehabilitation stroke
unit: which aspects are most important? Stroke 1999; 30:
917–23.
20. Hotopf M, Chidgey J, Addington-Hall J, Ly KL. Depression in
advanced disease: a systematic review. Part 1. Prevalence and
case finding. Palliat Med 2002; 16: 81–97.
21. McPherson CJ, Wilson KG, Murray MA. Feeling like a burden
to others: a systematic review focusing on the end of life. Palliat Med 2007; 21: 115–28.
22. Low JTS, Payne S, Roderick P. The impact of stroke on
informal carers: a literature review. Soc Sci Med 1999; 49:
711–25.
23. Gott M, Barnes S, Parker C et al. Dying trajectories in heart
failure. Palliat Med 2007; 21: 95–100.

CRP: a biomarker for early detection of bacterial infection
24. Addington-Hall JM, McCarthy M. Regional study of care for
the dying: methods and sample characteristics. Palliat Med
1995; 9: 27–35.
25. Gibbs JSR, McCoy ASM, Gibbs LME, Rogers AE, Addington-Hall JM. Living with and dying from heart failure: the
role of palliative care. Heart. 2002; 8: 27–35.
26. Gore JM, Brophy CJ, Greenstone MA. How well do we care
for patients with end stage chronic obstructive pulmonary disease (COPD)? A comparison of palliative care and quality of
life in COPD and lung cancer. Thorax 2000; 55: 1000–6.
27. Drennan J, Cleary J. Quality of life of patients on haemodialysis for end stage renal disease. J Adv Nurs 2005; 51:
577–86.

Age and Ageing 2010; 39: 559–565
doi: 10.1093/ageing/afq067
Published electronically 23 June 2010

28. Fitzsimons D, Mullan D, Wilson JS et al. The challenge of patients’ unmet palliative care needs in the final stages of chronic
illness. Palliat Med 2007; 21: 313–22.
29. Glader E, Stegmayr B, Asplund K. Post-stroke fatigue. A 2year follow-up study of stroke patients in Sweden. Stroke
2002; 33: 1327–33.
30. Gomes B, Higginson IJ. Where people die (1974-2030): past
trends, future projections and implications for care. Palliat Med
2008; 22: 33–41.
Received 10 September 2009; accepted in revised form
5 May 2010

© The Author 2010. Published by Oxford University Press on behalf of the British Geriatrics Society.
All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Serum C-reactive protein as a biomarker for
early detection of bacterial infection in the older
patient
ANGELA LIU, TRIET BUI, HUONG VAN NGUYEN, BIN ONG, QING SHEN, DILAN KAMALASENA
Department of Aged Care and Rehabilitation, Bankstown–Lidcombe Hospital, NSW 2200, Sydney, Australia
Address correspondence to: A. Liu. Tel: (+61) 2 9845 6989. E-mail: angelaliu1682000@yahoo.com.au

Abstract
Background: although C-reactive protein (CRP) is widely used in younger populations, its value for diagnosing bacterial
infection in older population is not well established. This study examined the usefulness of serum CRP level in the early
detection of bacterial infection in older patients.
Methods: in a prospective cohort study, consecutive patients aged 70 years or over admitted to Aged Care wards were
recruited. CRP levels were measured within 24 h of presentation, and their significance in predicting bacterial infections
was analysed. The relationship between CRP and other clinical features of diagnosing bacterial infections (e.g. temperature,
white cell count, neutrophil count, oxygen saturation, blood pressure and heart rate) was also examined.
Results: a total of 232 patients were recruited over a period of 3 months. CRP levels were 21.3±36.0 and 150.5±114.1 mg/l
(mean±SD) in the non-infection and infection groups, respectively (P<0.001). We found that the CRP cut-off value of
60 mg/l had the best combination of sensitivity and specificity. At this level, the sensitivity of diagnosing bacterial infection
was 80.7%, specificity 96.0%, positive predictive value 91.9% and negative predictive value 89.8%. CRP and temperature had
higher sensitivity and specificity than white cell count and neutrophil count in the diagnosis of infection. For every 1-mg/l
increment in CRP, the risk of bacterial infection increases by 2.9%.
Conclusion: CRP is a convenient and useful biomarker to predict early bacterial infection in older patients especially when
other markers are atypical or not present.
Keywords: bacterial infection, C-reactive protein (CRP), elderly, older person

559