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

80 / 290

malnutrition, which can contribute to functional decline. Our aimwas
to evaluate the functional status of patients ≥75 years that survived 12
months (12 M) after discharge and to analyse association with OH and
hygiene.
Methods: Prospective longitudinal cohort study of 100 patients.
Comprehensive geriatric assessment, dental examination and OH
questionnaire performed at baseline. Survival and functional status at
12 M assessed by phone contact and hospital record analysis.
Results: 46 patients survived 12 M follow-up. Average Barthel score
(BS) 12 M 59.2 (baseline 75.3). According BS variation two groups
were defined: A- maintenance/improvement of BS (47.8%); B- decline
of BS. A and B were non-homogeneous concerning average age (86.9 vs
81.8) and baseline BS (70.5 vs 80). Patients in group A presented:
higher average number of teeth (7.50 vs 6.21, ns) but higher
prevalence of caries and periodontal disease (27.3% vs 16.7%, ns);
similar usage of oral prosthesis and autonomy in oral hygiene; higher
usage of toothbrush and toothpaste (63.6% vs 45.8%, ns), lower usage
of mouthwash (22.7% vs 25%, ns).
Conclusion: Better OH status and hygiene habits were not statistically
associated with maintenance/improvement of functional status.
Nevertheless, patients with favourable outcome showed a higher
average number of teeth and higher usage of toothbrush and tooth-
paste. The impact of OH in outcome might be underestimated due to
small sample size.
P-172
Influence of age and multimorbidity on time to readmission
O.B. Fernandes 1, S. Lopes1,2, R. Santana1,2. 1 Escola Nacional de Saúde
Pública, Universidade NOVA de Lisboa, 2Centro de Investigação em Saúde
Pública, Escola Nacional de Saúde Pública, Universidade NOVA de Lisboa,
Lisboa, Portugal
Introduction: Ageing populations and the increasing prevalence of
multiple chronic conditions are a challenge for healthcare delivery and
health system design. Readmissions are frequently studied for its
negative impact on patients and providers. This study aims to explore
the association of time to readmission with age and multimorbidity.
Methods: A database including administrative data from 1.679.634
inpatient episodes from years 2002–14 was considered. Chronic
conditions were identified from all diagnoses coded with
International Classification of Diseases – 9th version – Clinical
Modifications codes (1: present). The considered outcome was
thirty-day hospital-wide all-cause unplanned readmissions (1:
readmitted). Episodes were divided into five age groups: 0–19; 20–
44; 45 –64; 65–84; 85+ years. Gender, number of Elixhauser
comorbidities, and treatment in a vertically integrated unit were also
included. We used a Cox regression to determine the association of
time to readmission with selected covariates.
Results: The observed readmission rate was 5.1% and median time to
readmission was 10 days. The risk of readmission increased through-
out age groups, with increasing likelihood of readmission for
individuals aged 65+ [65 –84: 1.237 (1.208–1.266); 85+: 1.739 (1.691–
1.788)]. Individuals with two chronic conditions presented the highest
risk of readmission (1.368; 1.325–1.413), whilst patients with 5+
presented a likelihood of readmission of 1.276 (1.212–1.343). Male
patients, with more comorbidities, and treated outside vertically
integrated units showed an increased risk of readmission.
Key conclusions: Older patients with multimorbidity had an
increased risk for readmission. An awareness of the factors influencing
time to readmission allows the design of interventions aimed at
increased risk groups.
P-173
Differences between older patients admitted with cancer or
diagnosed with cancer during hospital admission: a palliative care
approach
L. García-Cabrera, J. Mateos-Nozal, M.V. Cerdeira, E. Sánchez-García,
A.J. Cruz-Jentoft, L. Rexach-Cano.
Introduction: Very old patients assessed by hospital palliative care
consultation teams are either admitted with a known cancer or
diagnosed during hospital admission. Our aim was to compare
characteristics and prognosis of each group, in order to better tailor
palliative care.
Methods: We included all patients over 79 years old with cancer
(known diagnosis of found during hospitalization) who were assessed
by a palliative care consultation team during one year. Demographic,
clinical and mortality data were collected.
Results: 167 subjects (37.7% diagnosed during hospitalization). Those
diagnosed with cancer during admission were older (87.1 ± 6.0 vs
85.1 ± 3.6 years, p = 0.017), had fewer comorbidities (CIRS-G 2.1 ± 0.4 vs
2.4 ± 0.4, p < 0.001) and lived alone more frequently (23.8% vs 11%,
p = 0.06) than those admitted with cancer. No differences in gender,
polypharmacy, dementia or functional decline before admission
were found. Subjects with new cancer had significantly more focal
neurologic signs (12.7% vs 7.8%), falls (12.7% vs 6.8%) and constitutional
syndrome (14.3% vs 7.8%), although they had less bleeding episodes
(4.8% vs 11.7%, p = 0.003) than the other group. Those with a new
diagnosis were more frequently admitted to the Internal Medicine
and Geriatrics department (60.3% vs 26.9%), than to the Oncology
department (4.8% vs 35.6%, p = 0.001). They had more lung cancer,
liver and biliopancreatic tumours (23.8 vs 7.7%, p = 0.002) and less low
grade disease (11.1% vs 21.2%). Staging was not completed in more
subjects with new cancer (30.2% vs 5.8%, p = 0.001) and treatment was
not as active (surgery (3.2% vs 34.6% p < 0.001), chemotherapy (1.6% vs
28.8%, p < 0.001), radiotherapy (0% vs 13.5%), palliative care (3.2% vs
11.5%, p = 0.001). Length of stay was longer (17.6 ± 8.9 vs 12.4 ± 10.0
days, p = 0.001). There were no differences in after hospital care, total
mortality or use of palliative sedation, but survival was significantly
shorter for those with a new diagnosis (median 2.3 ± 0.4 vs 40.7 ± 5.1
months, p < 0.001).
Conclusions: Older patients with a new diagnosis of cancer during
admission in need of Palliative Care are different to those admitted
with a known cancer and may have different care needs.
P-174
Study of predictive factors (clinical and personal) of hospital
mortality in a Geriatric Service in Zaragoza (Spain)
I.F. Lacarte1, B.G. Huarte1, F.A. Monzón2, M.G. Eizaguirre1, C.D. Pérez 1.
1Geriatric Service, Hospital Nuestra Señora de Gracia, 2Health Deparment,
Goverment of Aragon
Introduction: The aim of the study is to find out which personal and
clinical factors may be associated with mortality in geriatric
hospitalized patients, face to obtain a predictive model that allows
us to identify people at increased risk.
Methods: There were 318 incomes, between 06.10.2014 and
30.11.2014. Variables studied: age, sex, clinical aspects (personal
background, Barthel index (BI), Charlson index (CI), drugs, SPMSQ
and biochemical parameters. Logistic Regression (LR) was perfor-
med to assess the relationship between death and studied variables.
In previous bivariate analysis, Chi square and ANOVA was used
depending on the type of variable analyzed. Was used SPSS v19.
Results: A LR was performed between the dependent variable “Exitus”
and some explanatory variables “age, dementia, renal function, IB, IC,
omeprazole, high prolactin, hemoglobin, albumin, creatinine, urea,
calcium, GOT, lactate deshydrogenase (LDH” (bivariate p < 0.05).
Significant associations were detected: IB (<60) OR 6.101 (2.013–
18.48), Omeprazole OR 0.468 (0.227–0.961) (protector), OR 1.364
Creatinine (1.054–1.766), Albumin OR 0.463 (0.256 –0.836) (protector)
and LDH OR 1.004 (1.002–1.006). The Model was: Prob(exitus) = 1/
(1 + e (2.587–1.808Barthel + 0,76Omeprazole − 0,311 Creatinine +
0.771 Albumin − 0.004LDH)) The multivariate model correctly classi-
fied 86.7% of patients, showing a high specificity (98.5%) but low
sensitivity (19.6%). The discriminatory power of the model, according
to ROC curve, was 87.6% of the maximum possible.
Poster presentations / European Geriatric Medicine 7S1 (2016) S29–S259 S74