European Geriatric Medicine - Supplement 1 - Volume 7 - October 2016

49 / 290

predictive value is high. We wished to see how often a negative
dipstick was clinically helpful in the setting of the nonspecifically
unwell older adult who presented to hospital.
Methods: We performed weekly spot audits to evaluate how often a
negative urine dipstick was helpful in the assessment of a non-
specifically unwell older adult (by ruling out UTI)
Results: We reviewed the case records of 216 patients. We found 14
non-specifically unwell older adults. One of these patients had a
negative urine dipstick but was started on antibiotics for a UTI anyway.
Conclusions: Negative urine dipsticks are rarely helpful in the
assessment of nonspecifically unwell older adults. On the one occasion
a urine dipstick was negative, the patient was still treated for a urinary
tract infection.
P-056
Community-acquired pneumonia and prognostic score systems in
the very elderly
A.R. Nogueira 1,2, V. Borba1,2, D. Ferreira 1,2, A. Aragão2, M.T. Veríssimo1,2,
A. Carvalho1,2. 1Faculty of Medicine of the University of Coimbra,
Portugal; 2Coimbra Hospital and Universitary Center, Portugal
Background: Community-acquired pneumonia is an important health
problem affecting the very elderly patients, leading to high rates of
morbidity and mortality. Several prognostic scoring systems were
developed aiming to recognize the critical ill patients and help the
decision making process. However its applicability in the very elderly
patients may be limited.
Methods: We reviewed the hospital stays at the Internal Medicine
Service ward of a 80 years or older population, diagnosed with
community-acquired pneumonia during the month of January
2014. Their demographics were analyzed and their results at CURB,
SOAR and SCAP risk scores were calculated. We also reviewed other
potential prognostic factors, namely the presence of hypoxemia
(PaO2 < 60 mmHg) and laboratory findings (C-reactive protein and
albumin levels) at the admission. The in-hospital mortality was
recorded, as well as the 6-month, 1-year mortality and the hospital
readmission within 30 days after discharge. A 95% confidence interval
was used.
Results: The medium length of hospital stay of the analyzed popu-
lation (95 patients) was 7 days and the in-hospital mortality rate was
29%. We found a correlation between higher results at the CURB score
and the in-hospital, 6-month and 1-year mortality (p < 0,05). No
significant correlation was founded between the results at the SOAR
and SCAP risk score and the patient’s mortality or 30-day readmission.
We were not able to find any correlation between each of the other
variables (hypoxemia, C-reactive protein and albumin levels) and the
mortality or readmission.
Conclusions: In our study CURB risk score was able to predict
patient’ s mortality and hospital readmission. Further adjustments to
the existent risk scores may be needed to improve their sensitivity.
P-057
What is the best biomarker to predict infection in elderly standard
care patients?
A. Nunes Ferreira 1, M. Camacho1, I. Nogueira Fonseca1, M. Marques1,
D. Buendia 1, T. Rodrigues1, A. Pina1, P. Cantiga Duarte1. 1CHLN, Hospital
Santa Maria, Lisbon, Portugal
Introduction: Evaluate the best sepsis biomarker in elderly patients
and determine its sensibility and specificity predicting the risk of
infection.
Methods: Retrospective study, 834 patients admitted to a medical
standard care ward, screened for suffering at least 2 SIRS (Systemic
Inflammatory Response Syndrome) criteria and at least 1 hemoculture
performed. Infection Probability Score (IPS) and sepsis biomarkers
(Leucocytes, C-Reactive Protein (CRP)) were evaluated.
Results: In the final cohort (n = 124), aged 75 ± 15 years, the infection
prevalence was 88%, with 22% bacteremia. The average CRP in infected
patients was higher than in non-infected patients (17.27 mg/dL vs
7.04 mg/dL; p < 0.001), yielding a 0.82 ROC-AUC (CI 95%: 0.72–0.93;
p < 0.001). The CRP optimal cut-off 8.35 mg/dL presented an 80%
sensibility and 72% specificity predicting infection. IPS and leucocytes
were higher in infected patients but performed sub-optimally in
predicting infection, 0.70 ROC-AUC (CI 95%: 0.53–0.86; p = 0.016),
optimal cut-off 14 (73% sensibility, 64% specificity) for IPS and 0.68
ROC-AUC (CI 95%: 0.53–0.83; p = 0.032), optimal cut-off 10705/μ L
(62% sensibility, 71% specificity) for leucocytes. The performance of
CRP and IPS was more significant in patients under 75 years-old [ROC-
AUC 0.87 for CRP and IPS (p = 0.001)], whereas in patients with at least
75 years, only CRP demonstrated to be useful in predicting infection
0.79 ROC-AUC (CI 95%: 0.66–0.93; p = 0.03).
Conclusion: In standard care ward patients, although CRP should be
privileged in evaluating risk of infection, IPS and leucocytes may also
be considered. However, in elderly patients, only CRP demonstrated
to be useful in predicting infection, which may increase the rationale
to take hemocultures or initiate antibiotics.
P-058
Urinary tract infection – a qualitative study exploring the human
factors contributing to misdiagnosis
K. O’Kelly1, K. Phelps2, E. Regen2, D. Kondova1, S. Conroy2. 1University
Hospitals of Leicester, UK, 2University of Leicester, UK
Urinary problems in older people are common and there is general
agreement regarding their assessment with clear guidelines as to the
use of urinary dipstick testing, including knowledge of asymptomatic
bacteriuria. There is significant variability in practice leading to a
misdiagnosis rate of up to 40% in older patients in hospital. A mapping
review conducted by our group indicated that human factors play a
significant contributory role.
To investigate these human factors we undertook a series of interviews
with nursing and medical clinicians responsible for managing older
people with possible urinary tract infection (UTI) in urgent care
settings. Interviews took place over a three month period and each
lasted approximately 30 minutes. Twenty interviews were conducted
in total with each being recorded and transcribed for analysis using
the “Framework” approach. Analysis was facilitated by the use of the
software package NVivo 1.
Our findings indicate a multi-faceted and complex set of human factor
explanations. Themes included: Time pressures, the need for prompt
diagnosis and the need to rule out life-threatening diagnoses such
as sepsis, often drive inappropriate urinary dipstick testing despite
guidelines to the contrary and often in the absence of appropriate
symptoms. Inexperienced clinicians may interpret dipstick results
incorrectly (eg asymptomatic bacteriuria) and are more likely to arrive
at an incorrect diagnosis. Junior clinicians are also more diagnosis-
driven and less likely to adopt a “watch and wait” approach.
An awareness of these human factors will now help tailor an educa-
tional intervention aimed at improving the accuracy of UTI diagnosis.
P-059
Internal Medicine and older patients in Emergency Department
P. Oliveira, B. Frutuoso, R. Veríssimo, A. Oliveira. Internal Medicine
Department, Centro Hospitalar de Vila Nova de Gaia/Espinho
Aging is a problem affecting Emergency Medicine, requiring qualified
staff for special needs of older patients. We intended to understand
what is the paper of internal medicine in the assessment of the older.
We conducted a retrospective, observational study, with a randomized
sample of 14% of total patients admitted during a month in an
Emergency Department (1252/8872 episodes), excluding children
and obstetric patients. We collected demographic data, Manchester
triage, medical specialty involved, the length of stay (LOS) and
destination. Statistical analysis was performed in SPSS ® 23.0 using
qui-square, Mann-Whitney and Kruskal-Wallis tests. In our sample,
53% were female, the average age was 53,9 years, and 32,4% had
≥65 years. Median LOS in the Emergency Department (ED) was 215
minutes. Median LOS in geriatric patients were higher than younger
Poster presentations / European Geriatric Medicine 7S1 (2016) S29–S259 S43