93 / 290
P-221
Profile of re-hospitalization in a private health provider that caters
exclusively geriatric patients in Brazil, in 2015
S.A.D. Silveira, Jr, E.F. Parrillo, A.L.A. Nascimento, D. Cabral.
Prevent
Senior, Brasil
Objective:
Evaluate the profile of readmissions in a particular health
care provider that caters exclusively geriatric patients (Prevent Senior)
in Brazil during the year 2015.
Method:
We conducted a prospective study between January 01 to
December 31, 2015, where evaluated all readmissions that occurred
during that period.
Results:
Were analyzed in 2,622 (8.01%) readmissions, of a total
population of 32.700 admissions. 45.53% of these re-admitted in up to
seven days post discharge; 44.41% readmitted for the same reason
discharge. 66% of readmissions were due to infectious aetiology,
12.9% from cardiovascular causes, 9% for palliative care, 6.6% for renal
insufiencia and 5.5% by other causes. The overall hospital mortality
was 9.28%, and the average length of stay was 4.42 days.
Conclusion:
The results presented show a low rate of re-hospitaliza-
tion, low death rate and low time of permanency for the population
studied. The main reason for the readmission of this population were
the infectious causes, with pneumonia being the most responsible.
Such results show indirectly an efficient in-hospital management
mechanism of the different conditions that affected the population
studied.
P-222
HIV as potential risk factor for falls and risk factors for falls in
older treated HIV-infected
O.S. Smeekes, K. Kooij, F. Wit, P. Reiss, N. van der Velde.
AMC
Department of Geriatrics & the AIGHD
Introduction:
Recent studies have shown a high prevalence of falls in
the middle age HIV-infected. This study aimed to explore if HIV
positivity is associated with increased fall risk and to identify risk
factors for falls in older HIV-infected persons.
Methods:
Data from the AGEhIV cohort was analyzed cross-sectionally
to assess the association between HIV positivity and fall history
(recurrent falls, any falls) and to assess potential fall risk factors within
the HIV positive group with multivariable logistic regression. Patients
were excluded if fall registration was missing. Variables included in
analyses were HIV characteristics, demographics and classical fall risk
factors.
Results:
In total, 535 HIV-infected and 522 HIV uninfected partici-
pants were included, with a median age of 52 (interquartile range
[IQR] 48
–
59 vs. 47
–
58). The HIV-infected fell recurrently in 11.6% of
cases vs. 8.9% of the HIV uninfected (P = 0.136). HIV was not sig-
nificantly associated with fall history. An independent interaction was
found between age and HIV for both outcome measures, suggesting
that HIV related falls occur at younger age (P = 0.073 vs. P = 0.025).
Significant independent factors associated with recurrent falls in the
HIV positive group were male gender OR 0.3, 95%CI (0.1
–
1.0), BMI OR
0.8, 95%CI (0.7
–
1.0), fear of falling, dizziness OR 0.9, 95%CI (1.2
–
38.0)
and anti-depressants OR 3.2, 95%CI(1.1
–
9.6).
Key conclusions:
In this large cohort study, fall prevalence in
HIV infected persons as well as fall risk factors are comparable to
general population. Data suggests that in middle-age HIV positivity
is associated with fall risk, potentially due to the underlying
complications.
P-223
STOPP/START version 2: development of software applications:
easier said than done?
P. Anrys
1
, B. Boland
2,3
, J.M. Degryse
3,4
, J. De Lepeleire
5
, M. Petrovic,
S. Marien
1
, O. Dalleur
1
, G. Strauven, V. Foulon, A. Spinewine
1
.
1
Université catholique de Louvain, Louvain Drug Research Institute,
Clinical Pharmacy Research Group,
2
Cliniques Universitaires Saint-Luc,
Geriatric Medicine,
3
Université catholique de Louvain, Institute of Health
and Society,
4
Department of Public Health and Primary Care, KU Leuven,
Leuven,
5
Departement Public Hea, KU Leuven
Introduction:
Explicit criteria, such as the STOPP/START criteria, are
used for both clinical practice and research to identify potentially
inappropriate prescribing (PIP). There has been growing interest in the
development of software applications to automatically detect PIP.
Methods:
In the context of the COME-ON study [1], a software
application was developed to detect PIP instances from the research
database. The detection should be as sensitive and specific as possible
as there is no subsequent evaluation by a clinician. During this process,
some difficulties arose for which decisions had to be taken by the
research team.
Results:
We encountered four kinds of issues: 1. Some criteria are
not as explicit as they should be: e.g. the list of anticholinergic drugs
has to be established; some terms are not (precisely) defined:
e.g.
“
long-term aspirin
”
. 2. Specific information that is not easily
available is sometimes required: e.g. information about lack of efficacy,
contraindication. 3. There is no universal coding system for medica-
tions that meets the requirements to apply all criteria: The ATC codes
do not distinguish between medications with different routes of
administration or formulation. For such criteria, coding by the national
identification code is required. 4. In order to improve specificity,
several criteria would benefit from additional rules:
“
unless
”
rules
could be added to decreased the risk of false positives.
Key conclusions:
The next version of the STOPP/START criteria could
be enriched to make them more directly transferable to algorithms,
to minimize variations between research teams and to enhance
specificity.
Reference
[1] Anrys P.,
et al.
, Collaborative approach to Optimise MEdication use
for Older people in Nursing homes (COME-ON): study protocol of a
cluster controlled trial.
Implementation Science
, 2016. 11(1).
P-224
Predictors of sedentary status in overweight and obese patients
with multiple chronic conditions, a cohort study
M. Joven
1
, I. Croghan
1
, D. Schroeder
2
, S. Quigg
1
, J. Ebbert
1
, P. Takahashi
1
.
1
Department of Internal Medicine, Mayo Clinic,
2
Department of Health
Science Research, Mayo Clinic, Rochester, MN
Objective:
Obese patients with multiple chronic conditions often
require walking to improve their health. However, these patients may
have barriers to walking. We sought to determine the risk factors at
baseline that impacted sedentary status (<5,000 steps a day) after four
months of some pedometer use.
Methods:
We conducted a secondary analysis using a cohort design.
Patients over 18 years of age were enrolled with a BMI >25 kg/m
2
and
>7 chronic conditions. Primary outcome was <5,000 steps a day on a
pedometer after 4 months. Potential predictors included demograph-
ics, biometrics, comorbid health conditions, self-rated health, and
length of pedometer use. We compared the predictors to sedentary/
non-sedentary status using Pearson chi square or logistic regression.
We created a final multivariable model.
Results:
We enrolled 130 patients with an average age of 63.6 years
(+/
−
15). 72% were women and 98% were white. At 4 months, 55% were
sedentary. We observed that increased age, cumulative comorbid
health, BMI, waist circumference, and sedentary baseline status
predicted sedentary status. Reduced self-rated physical activity,
physical quality of life, baseline step count and gait speed all predicted
sedentary status at 4 months as well. In the multivariable model, BMI
and physical QoL were significantly associated with being sedentary
(p values of 0.014 and 0.03 respectively).
Conclusions:
We observed that lower physical QoL and higher BMI
were associated with being sedentary after pedometer use. This study
outlines potential barriers to future activity with pedometers.
Poster presentations / European Geriatric Medicine 7S1 (2016) S29
–
S259
S87