|
Current Projects
Factors determining
differential host
susceptibility to
malaria
Client:
Walter Reed Army
Institute of Research,
Silversprings, Maryland,
USA
BGStats Consulting
provides Statistical
Consulting services in
designing a series of
epidemiological studies
planned to be conducted
in West Kenya aimed to
understand the
pathogenesis of severe
malaria and to develop
better methods to treat
or prevent these
complications.
The rationale for the
planned studies is to
investigate and
understand the
association between
genotypes of a
complement receptor
(CR1) and decreased
susceptibility to severe
malaria. CR1 is a
key multifunctional
molecule that regulates
complement activation,
enhances the removal of
immune complexes from
circulation
and
is involved in the
development of
antigen-specific
antibody responses. It
also has been implicated
in the pathogenesis of
severe malaria.
Lipoprotein(a) and
cardiovascular risk -
Meta analysis of
prospective studies
Clients:
Pfizer GmbH, Karlsruhe, Germany.
Division of
Nephrology, University
of Würzburg,
Germany;
Synlab - Medizinisches
Versorgungszentrum für
Labordiagnostik,
Heidelberg, Germany
BGStats Consulting
has been
contracted to conduct a
systematic review and
meta analyses to
investigate the
association between
Lipoprotein(a) and
different clinical
manifestations of
cardiovascular diseases.
Feasibility and safety
of autologous bone
marrow mononuclear cell
transplantation in
patients with advanced
chronic liver disease
Client:
Gastro-Hepatology
Unit, Federal University
of Bahia, Salvador,
Bahia, Brazil;
Monte Tabor Foundation
Milano, Italy
BGStats Consulting
has been
contracted
to provide statistical
consulting support for a
clinical study
evaluating the safety
and feasibility of bone
marrow cell (BMC)
transplantation in
patients with chronic
liver disease on the
waiting list for liver
transplantation.
Effect of statins on low
density lipoprotein
cholesterol and
cardiovascular risk in
the presence and absence
of diabetes mellitus -
Systematic review and
meta analysis
Client:
Pfizer GmbH, Karlsruhe, Germany.
Division of
Nephrology, University
of Würzburg,
Germany;
Synlab - Medizinisches
Versorgungszentrum für
Labordiagnostik,
Heidelberg, Germany
BGStats Consulting
has been
contracted to conduct a
systematic review and
meta analyses to
investigate the
association between
statin induced
reductions in LDL
cholesterol and
different clinical
manifestations of
cardiovascular disease
in the presence and
absence of diabetes
mellitus.
A dynamic recurrent
events analysis of a
randomized controlled
trial on the efficacy of
Atorvastatin in patients
with type 2 diabetes on
hemodialysis
Client:
Pfizer GmbH, Karlsruhe, Germany.
Division of
Nephrology, University
of Würzburg,
Germany.
Clinical
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
BGStats Consulting is
contracted to conduct an
advanced data analysis
of a randomized
controlled trial
undertaken to estimate
the efficacy of an
Atorvastatin
intervention in patients
with type 2 diabetes
(the 4D study).
In contrast to a
previously conducted
time-to-first-event
analysis BGStats will
implement a novel
analysis approach
addressing several
dynamic data features
that are common in
modern longitudinal
studies. The analysis
applied allows to
consider the multiple
and/or recurrent
cardiovascular events
recorded on the same
patient, and to control
for confounding or
effect modification of
time-varying variables
(e.g. age, time since
last event etc.).
Selection of
Previous Projects
Effect of statins on low
density lipoprotein
cholesterol and
C-reactive protein: a
meta-analysis
Client:
Pfizer GmbH, Karlsruhe,
Germany.
Clinical
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
BGStats Consulting is
contracted to conduct a
meta-analysis with the
objective to identify
factors affecting
reductions in low
density lipoprotein
cholesterol (LDL-C) and
C‑reactive protein (CRP)
during statin therapy
and to elucidate the
relationship between the
effects on both. A
systematic literature
review was conducted
retrieving 58
publications with 82
randomized statin
treatment arms and 17641
patients. Summary
estimates for the
average changes in LDL-C
and CRP from baseline
were determined and
meta-regression analyses
were conducted to model
effect heterogeneity
among studies due to
dose, type of statin and
baseline level. A
weighted partial
correlation analysis was
performed to quantify
the relationship between
the statin induced
reductions in LDL-C and
CRP.
The association between
low density lipoprotein
cholesterol, statins and
cardiovascular events: a
meta-analysis
Client:
Pfizer GmbH, Karlsruhe, Germany.
Clinical
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
BGStats Consulting is
contracted to
undertake a
meta-analysis with the
aim to to determine
whether the magnitude of
the risk reduction
according to different
cardiovascular endpoints
correlates with the
magnitude of the statin
induced reduction of low
density lipoprotein
cholesterol (LDL-C). A
systematic literature
review was conducted
retrieving 18 randomised
studies with a total of
97,861 participants.
Differences in average
LDL-C reductions between
the intervention and
control groups during
the follow-up and
relative risks according
to different clinical
endpoints (total and
coronary mortality, MI,
stroke,
revascularizations,
etc.) were extracted
from the original
publications and
meta-regression analyses
were taken out to
analyze the relationship
between the statin
induced
LDL-C
reduction and the
reduction in
cardiovascular risk.
Statistical Analysis of
the
LURIC study - a large
prospective cohort study
with the aim to
investigate functional
genomics,
pharmacogenomics and
long term prognosis of
cardiovascular disease.
Client:
LURIC GmbH, University
of Freiburg, Germany.
LURIC is an ongoing
prospective study of
currently more than 3300
individuals in whom the
cardiovascular and
metabolic phenotypes
CAD, MI, dyslipidaemia,
hypertension, metabolic
syndrome and diabetes
mellitus have been
defined in all study
partipiciants.
BGStats has been
subcontracted to perform
statistical analyses of
the data collected in
this study with the aim
to identify genetic and
environmental risk
factors for
cardiovascular and
metabolic diseases, e.g.
-
Statistical Analysis
of the genetic data:
Inference of
haplotypes from SNP
data by applying
mathematical
simulation algorithms.
-
Association analyses:
risk factors -
clinical outcomes
-
Analysis of the follow
up data:
Identification of risk
factors for incidence
rates according to
different clinical
events observed during
the observational
period (Multiple event
-, Competing risk
analysis)
A
prognostic score for
metastatic breast cancer
patients.
Client:
Department of Clinical
Oncology, Medical
University of Graz,
Austria
The
aim of this study is to
find a prognostic score
to estimate the risk of
metastatic breast cancer
patients at the time of
first occurrence of
metastasis. A historical
cohort of 343 metastatic
breast cancer patients
will be analyzed using
different multivariate
statistical methods (Cox
Regression,
Classification Trees,
etc.) to assess the
impact of established
prognostic factors (e.g.
hormonal receptor status,
grading, disease free
interval, metastatic
site, Her2 receptor) on
the survival time. The
estimated relative risk
obtained by the
statistical models
and expert know-how of
clinicians
will
be used to calculate
a
prognostic score. In
clinical practice the
score will be very
useful
for classifying
breast cancer patients
in different risk groups.
The score will be
validated
using resampling
methods (bootstrap,
jackknife) and validation
samples.
Selection of
Previous Projects
Apolipoprotein
E genotypes and
haplotypes and their
association with
coronary artery disease
Client:
Clinical
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
In
this study we
investigate the
association of Apolipoprotein (Apo E)
polymorphisms and
haplotypes for ApoE
levels, lipid levels and
cardiovascular risk in
the large Ludwigshafen
Risk and Cardiovascular
Health (LURIC) study,
a cohort study which
includes 3316
patients hospitalized
for coronary angiography
between June 1997 and
May 2001. BGStats
is
analyzing the genetic
data collected in this
study.
Since the phase of the
SNPs (sequence on the
chromosom) usually is
unknow, mathematical simulation
algorithms
(e.g. Clarks Algorithm,
EM approaches, Stephens
Algorithm, etc.) will be
used to infer the ApoE
haplotypes.
The
reconstructed haplotype
frequencies will be
analyzed in order to
evaluate the association with
different clinical
outcomes, e.g.
coronary artery
disease,
myocardial infarction,
etc.
Analysis of the
Brazilian BCG-REVAC
Trial - a large cluster
randomization trial to
estimate the efficacy of
a BCG revaccination in
school children.
Client:
London School of Hygiene
& Tropical Medicine,
London, UK and
Department of Public
Health, Federal
University of Bahia,
Salvador, Brazil
BGStats
analyzed data from
a
large
cluster randomized trial
with the
objective of estimating
the efficacy
of a
second
dose
of
BCG
vaccination
against tuberculosis
given
to school children.
The
study population
consisted of more
than 200.000
children aged 7
to 14
years,
enrolled in more
than 700
state
schools from the cities
of Salvador and Manaus,
Brazil. Since
schools and not
students
were
the unit of
randomization in
this trial, special
statistical methods for
cluster randomization
trials had to be used to
infer unbiased parameter
estimations.
Evaluation
of treatment sequences
in metastatic breast
cancer
using
a multi-state model.
Client:
Department of Clinical
Oncology, Medical
University of Graz,
Austria
Partner:
Institute of Medical
Biometry, Charité,
Humboldt University of
Berlin, Germany
Standard
therapy
for metastatic
breast cancer
involves a
sequence of cytotoxic
and/or hormonal
treatments. If
recurrence occurs after
an initially positive
response
to treatment,
the treatment is usually
changed. BGStats
analyzed a dataset
from a cohort of 270
metastatic breast cancer
patients treated at the Department
of Clinical Oncology,
Medical University of
Graz. Using a
multi-state approach the
course of the patients
during the different
clinical response or
progression states was
modelled as a stochastic
process. A set of common
prognostic factors and
the different applied
treatments were included
as covariates in the
statistical model.
A
”partial” Markovian
structure of the
observed stochastic
process
was assumed
to
estimate the
parameters
of interest using
the extended Cox-model.
The results of the
multi-state models
showed that the
influence of prognostic
factors is
constant
during the entire
metastatic phase
and
also
independent
of the
number, type and success
of the
applied treatment
sequences.
Determination
of Cyclosporine A in
whole blood: comparison
of a chromatographic
method with three
different immunological
methods.
Client:
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
The
monitoring of
cyclosporine A (CsA)
levels in whole blood
serves to optimize
immunosuppressive
therapy after
transplantation.
Currently,
chromatographic methods
(HPLC) and immunological
methods
are used to
determine cyclosporine
concentration in blood. It is well
known that
discrepant results occur between these
two assay systems,
resulting from different
cross-reactivity
of monoclonal antibodies
with metabolites. In
this study we evaluated
immunological CsA assays(
AxSYM, EMIT, Dimension)
with HPLC, taking into
account the
cross-reactivity with
CsA metabolites.
BGStats fitted a
linear mixed-effects
model
using a patient specific
random effect
and the metabolites as
covariates to
quantify
cross-reactivity between
the various
immunological methods
and the gold standard (HPLC).
Health
economic aspects of
diagnosis and therapy of
congenital
Human Cytomegalovirus Infection
(HCMV).
Client:
Institute of Medical and
Chemical Laboratory
Diagnostics, Medical
University of Graz,
Austria.
BGStats
performed a cost
effectiveness analysis
to
evaluate the impact
of different screening
and therapy strategies
for
congenital HCMV
infection. The
parameters needed for
the multivariate model,
e.g. HCMV seroprevalence,
transmission rate,
treatment effect, etc.
were estimated applying
a meta analysis approach.
Furthermore a
multivariate sensitivity
analysis was performed
to compare
the cost
benefit,
i.e.
cost caused by
congenital HCMV
infections – cost of
screening and therapy)
of different simulated
scenarios.
|
