I would like to get some feedback from the group concerning the reporting of
modeling results. I have a Pop PK model developed from data arising from 124
pediatric patients, age 1 to 48 months. All of the structural parameters have
been scaled allometrically, with the median body weight used as the reference
value. After accounting for body size, a covariate model was incorporated to
describe maturational changes in CL for young children. The maturation of
clearance was modeled using an exponential model proposed in:
Andersen et al. Population clinical pharmacology of children: modelling
covariate effects. Eur J Pediatr. 2006
Two parameters are estimated as part of this model * the fractional change in
CL for a typical one month old patient (beta - estimated to be 0.76 (0.589,
0.96) for this analysis) and a maturational half-life (TCL - 3.82 (1.57, 6.95)
months). CI’s are from the bootstrap.
The problem that I am running into is how to report the modeling results. It
seems very natural to me to report the model results normalized to median body
weight (L/h/10.4 kg^0.75). One of the study investigators disagrees with me and
would like to report the results on a per kg basis (L/h/kg^0.75). This seems
to be counterintuitive to me, as I tend to think about what represents the
“typical patient.” It also makes no sense to me to represent the CL in a one
kg child. The argument is that reporting in this manner makes more sense to
clinicians and that there is no such thing as a typical child.
So in an attempt to appease the investigator, I fit the same model with no
weight normalization. The estimated parameters are equivalent to what would be
scaled from the weight-normalized model, but there is no covariance matrix (not
surprising). It becomes problematic when the bootstrap results are considered
* beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018). Again, this is not
surprising given that the covariate model is not centered.
I have attempted to make several compromises, including reporting the parameter
estimates in both median weight-normalized terms and normalized per kg. I have
also included scaled CL estimates for typical patients at several ages and body
weights. This hasn’t met the approval of the investigator, who is now insisting
that I report the model building procedure from the median weight model, but
report scaled parameters only on a per kg basis. This is wrong in my opinion
and is actually more confusing to someone who is trying to understand the model.
Can I get the group’s opinion on this? Am I being stubborn looking at the world
through a modeler’s point of view?
Thanks,
John Mondick PhD
Research Assistant Professor
Division of Clinical Pharmacology and Therapeutics
The Children's Hospital of Philadelphia
Tel (267) 426-2292
FAX (215) 590-7544
Email: [EMAIL PROTECTED]
Reporting Modeling Results
An option to report your results to appease non-modelers in your case might be
reporting your CL values in a table with different age and weight.
Alan
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of John Mondick
Sent: Wednesday, October 24, 2007 11:18 AM
To: [email protected]
Subject: [NMusers] Reporting Modeling Results
I would like to get some feedback from the group concerning the reporting of
modeling results. I have a Pop PK model developed from data arising from 124
pediatric patients, age 1 to 48 months. All of the structural parameters have
been scaled allometrically, with the median body weight used as the reference
value. After accounting for body size, a covariate model was incorporated to
describe maturational changes in CL for young children. The maturation of
clearance was modeled using an exponential model proposed in:
Andersen et al. Population clinical pharmacology of children: modelling
covariate effects. Eur J Pediatr. 2006
Two parameters are estimated as part of this model * the fractional change in
CL for a typical one month old patient (beta - estimated to be 0.76 (0.589,
0.96) for this analysis) and a maturational half-life (TCL - 3.82 (1.57, 6.95)
months). CI’s are from the bootstrap.
The problem that I am running into is how to report the modeling results. It
seems very natural to me to report the model results normalized to median body
weight (L/h/10.4 kg^0.75). One of the study investigators disagrees with me and
would like to report the results on a per kg basis (L/h/kg^0.75). This seems
to be counterintuitive to me, as I tend to think about what represents the
“typical patient.” It also makes no sense to me to represent the CL in a one
kg child. The argument is that reporting in this manner makes more sense to
clinicians and that there is no such thing as a typical child.
So in an attempt to appease the investigator, I fit the same model with no
weight normalization. The estimated parameters are equivalent to what would be
scaled from the weight-normalized model, but there is no covariance matrix (not
surprising). It becomes problematic when the bootstrap results are considered
* beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018). Again, this is not
surprising given that the covariate model is not centered.
I have attempted to make several compromises, including reporting the parameter
estimates in both median weight-normalized terms and normalized per kg. I have
also included scaled CL estimates for typical patients at several ages and body
weights. This hasn’t met the approval of the investigator, who is now insisting
that I report the model building procedure from the median weight model, but
report scaled parameters only on a per kg basis. This is wrong in my opinion
and is actually more confusing to someone who is trying to understand the model.
Can I get the group’s opinion on this? Am I being stubborn looking at the world
through a modeler’s point of view?
Thanks,
John Mondick PhD
Research Assistant Professor
Division of Clinical Pharmacology and Therapeutics
The Children's Hospital of Philadelphia
Tel (267) 426-2292
FAX (215) 590-7544
Email: [EMAIL PROTECTED]
Hi John,
I think you can safely separate statistics/mathematics and clinical use. I would fit the model in the shape and form suitable to get the best results (normalized to a typical patient in your case) and then report the results in the form most convenient for the clinical use. If this is per-kg values, then report it as they request. If you look in the literature, results are routinely reported as V/kg or CL/m^2, etc.
On a side not, I am actually surprised that you got different results with different scaling. For the allometric scaling with fixed power, two parameterizations:
CL=TCL*WT^0.75 and CL=TCL*(WT/10)^0.75
differ by the fixed factor
(1/10)^0.75 = 0.18
I am not sure how this can influence your model CI so strongly. I would check how you stratify the bootstrap data sets. Could it be that stratification on something else depend on parameterization?
If you would estimate the power:
CL=TCL*(WT/10)^THETA()
then parametrization would be more likely to affect CI, but for the fixed power I would look for other explanations of the differences. It would be easier to discuss the model if you would attach the PK block of the nonmem code.
Leonid
--------------------------------------
Leonid Gibiansky, Ph.D.
President, QuantPharm LLC
web: www.quantpharm.com
e-mail: LGibiansky at quantpharm.com
tel: (301) 767 5566
John Mondick wrote:
> I would like to get some feedback from the group concerning the reporting of
> modeling results. I have a Pop PK model developed from data arising from 124
> pediatric patients, age 1 to 48 months. All of the structural parameters have
> been scaled allometrically, with the median body weight used as the reference
> value. After accounting for body size, a covariate model was incorporated to
> describe maturational changes in CL for young children. The maturation of
> clearance was modeled using an exponential model proposed in:
>
> Andersen et al. Population clinical pharmacology of children: modelling
> covariate effects. Eur J Pediatr. 2006
>
> Two parameters are estimated as part of this model * the fractional change in
> CL for a typical one month old patient (beta - estimated to be 0.76 (0.589,
> 0.96) for this analysis) and a maturational half-life (TCL - 3.82 (1.57, 6.95)
> months). CI’s are from the bootstrap.
>
> The problem that I am running into is how to report the modeling results. It
> seems very natural to me to report the model results normalized to median body
> weight (L/h/10.4 kg^0.75). One of the study investigators disagrees with me and
> would like to report the results on a per kg basis (L/h/kg^0.75). This seems
> to be counterintuitive to me, as I tend to think about what represents the
> “typical patient.” It also makes no sense to me to represent the CL in a one
> kg child. The argument is that reporting in this manner makes more sense to
> clinicians and that there is no such thing as a typical child.
>
> So in an attempt to appease the investigator, I fit the same model with no weight normalization. The estimated parameters are equivalent to what would be scaled from the weight-normalized model, but there is no covariance matrix (not surprising). It becomes problematic when the bootstrap results are considered * beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018). Again, this is not surprising given that the covariate model is not centered.
>
> I have attempted to make several compromises, including reporting the parameter
> estimates in both median weight-normalized terms and normalized per kg. I have
> also included scaled CL estimates for typical patients at several ages and body
> weights. This hasn’t met the approval of the investigator, who is now insisting
> that I report the model building procedure from the median weight model, but
> report scaled parameters only on a per kg basis. This is wrong in my opinion
> and is actually more confusing to someone who is trying to understand the model.
>
> Can I get the group’s opinion on this? Am I being stubborn looking at the world
> through a modeler’s point of view?
>
> Thanks,
>
> John Mondick PhD
> Research Assistant Professor
> Division of Clinical Pharmacology and Therapeutics
> The Children's Hospital of Philadelphia
> Tel (267) 426-2292
> FAX (215) 590-7544
> Email: [EMAIL PROTECTED]
John,
Leonid correctly points out that it makes no difference in terms of parameter
estimation what weight is used for parameter 'normalization' because this is
just a linear scaling factor.
My colleague, Brian Anderson, and I have preferred to consider this
'normalising' weight as a standard weight. We do not try to make anything
'normal' but simply choose a scale that will provide us with a parameter that
represents a standard human. We use a weight of 70 kg as the standard (Holford
1996). This makes it easier to compare parameters estimated in different
populations, e.g. adults and children, because the parameters are scaled to the
same standard size. Recently we have pointed out that use of a size standard
with a suitable model for maturation allows the simple prediction of adult
clearances from data collected in children (Anderson et al 2007b). We note that
this cannot be done in the reverse direction i.e. adult data cannot be used to
predict the maturational changes in clearance which occur in very young humans.
Leonid mentions that results are "routinely reported as V/kg or CL/m^2, etc"
but this is simply tradition without any discernible scientific rationale -
especially the use of the square metre as the standardising factor. Drugs are
not eliminated to any important extent via the skin so there is no mechanistic
reason for this. The surface area method is a hangover of discredited theories
of allometric scaling. The per kg method of scaling clearance is also a problem
because it leads to the misguided viewpoint that clearance is larger in
children in adults (Anderson & Holford 1997).
Perhaps you can help your investigator colleague to read some of the published
literature in this area so that he/she can get a clearer understanding of the
size scaling issue.
The bottom line:
'We conclude with the proposal that, at least in terms of pharmacokinetics, the
widely quoted aphorism Children are not small adults should be changed to
Children are small adults babies are young children. ' Anderson & Holford
2007a
Nick
1. Holford NHG. A size standard for pharmacokinetics. Clinical Pharmacokinetics
1996;30:329-332
2. Anderson BJ, McKee AD and Holford NHG. Size, myths and the clinical
pharmacokinetics of analgesia in paediatric patients. Clin Pharmacokinet
1997;33:313-27
3. Anderson BJ, Holford NH. Mechanism-Based Concepts of Size and Maturity in
Pharmacokinetics. Annu Rev Pharmacol Toxicol 2007a Oct 3; [Epub ahead of print]
4. Anderson BJ, Allegaert K, Van den Anker JN, Cossey V and Holford NH.
Vancomycin pharmacokinetics in preterm neonates and the prediction of adult
clearance. Br J Clin Pharmacol 2007b;63:75-84
Leonid Gibiansky wrote:
>
> Hi John,
> I think you can safely separate statistics/mathematics and clinical use.
> I would fit the model in the shape and form suitable to get the best
> results (normalized to a typical patient in your case) and then report
> the results in the form most convenient for the clinical use. If this is
> per-kg values, then report it as they request. If you look in the
> literature, results are routinely reported as V/kg or CL/m^2, etc.
>
> On a side not, I am actually surprised that you got different results
> with different scaling. For the allometric scaling with fixed power, two
> parameterizations:
>
> CL=TCL*WT^0.75 and CL=TCL*(WT/10)^0.75
>
> differ by the fixed factor
> (1/10)^0.75 = 0.18
>
> I am not sure how this can influence your model CI so strongly. I would
> check how you stratify the bootstrap data sets. Could it be that
> stratification on something else depend on parameterization?
>
> If you would estimate the power:
>
> CL=TCL*(WT/10)^THETA()
>
> then parametrization would be more likely to affect CI, but for the
> fixed power I would look for other explanations of the differences. It
> would be easier to discuss the model if you would attach the PK block of
> the nonmem code.
>
> Leonid
>
> --------------------------------------
> Leonid Gibiansky, Ph.D.
> President, QuantPharm LLC
> web: www.quantpharm.com
> e-mail: LGibiansky at quantpharm.com
> tel: (301) 767 5566
>
> John Mondick wrote:
> > I would like to get some feedback from the group concerning the reporting
> > of modeling results. I have a Pop PK model developed from data arising from
> > 124 pediatric patients, age 1 to 48 months. All of the structural
> > parameters have been scaled allometrically, with the median body weight
> > used as the reference value. After accounting for body size, a covariate
> > model was incorporated to describe maturational changes in CL for young
> > children. The maturation of clearance was modeled using an exponential
> > model proposed in:
> >
> > Andersen et al. Population clinical pharmacology of children: modelling
> > covariate effects. Eur J Pediatr. 2006
> >
> > Two parameters are estimated as part of this model * the fractional change
> > in CL for a typical one month old patient (beta - estimated to be 0.76
> > (0.589, 0.96) for this analysis) and a maturational half-life (TCL - 3.82
> > (1.57, 6.95) months). CIs are from the bootstrap.
> >
> > The problem that I am running into is how to report the modeling results.
> > It seems very natural to me to report the model results normalized to
> > median body weight (L/h/10.4 kg^0.75). One of the study investigators
> > disagrees with me and would like to report the results on a per kg basis
> > (L/h/kg^0.75). This seems to be counterintuitive to me, as I tend to think
> > about what represents the typical patient. It also makes no sense to me
> > to represent the CL in a one kg child. The argument is that reporting in
> > this manner makes more sense to clinicians and that there is no such thing
> > as a typical child.
> >
> > So in an attempt to appease the investigator, I fit the same model with no
> > weight normalization. The estimated parameters are equivalent to what would
> > be scaled from the weight-normalized model, but there is no covariance
> > matrix (not surprising). It becomes problematic when the bootstrap results
> > are considered * beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018).
> > Again, this is not surprising given that the covariate model is not
> > centered.
> >
> > I have attempted to make several compromises, including reporting the
> > parameter estimates in both median weight-normalized terms and normalized
> > per kg. I have also included scaled CL estimates for typical patients at
> > several ages and body weights. This hasnt met the approval of the
> > investigator, who is now insisting that I report the model building
> > procedure from the median weight model, but report scaled parameters only
> > on a per kg basis. This is wrong in my opinion and is actually more
> > confusing to someone who is trying to understand the model.
> >
> > Can I get the groups opinion on this? Am I being stubborn looking at the
> > world through a modelers point of view?
> >
> > Thanks,
> >
> >
> >
> >
> > John Mondick PhD
> > Research Assistant Professor
> > Division of Clinical Pharmacology and Therapeutics
> > The Children's Hospital of Philadelphia
> > Tel (267) 426-2292
> > FAX (215) 590-7544
> > Email: [EMAIL PROTECTED]
> >
> >
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090
www.health.auckland.ac.nz/pharmacology/staff/nholford
Nick, I think, in the big picture, it is important to remember why we do all this. It isn't for our own entertainment, or to congratulate each other on how insightful we are. It is to provide useful information to providers. We are not (we hope) the real audience for our work. It probably isn't realistic to expect non-oncologist pediatricians to scale doses allometrically. It seems to be unrealistic for other physicians (except neurologists) to scale doses at all. Dose/kg may be the best we can hope for. I'm quite sure it is unrealistic to expect drug companies to request labels for non! -chemotherapy drugs based on allometric scaling (really bad marketing move) Doing so - although perhaps scientifically correct, would likely lead to even more dosing errors that we currently see, with undemonstrated clinical benefit. So, how to report depends on who your audience is. If it is a bunch of nerds who know what eignevalues are, allometric scaling is great, if it is people who have 12 minutes to examine, diagnose and treat a patients, maybe we can keep it simple. Doing otherwise puts use are risk for irrelevance. Mark Sale MD Next Level Solutions, LLC
www.NextLevelSolns.com
919-846-9185
> -------- Original Message -------- Subject: Re: [NMusers] Reporting Modeling Results From: Nick Holford <[EMAIL PROTECTED]> Date: Thu, October 25, 2007 12:20 am To: nmusers <
>
> [email protected]
>
> > Cc: Brian Anderson <[EMAIL PROTECTED]> John, Leonid correctly points out that it makes no difference in terms of parameter estimation what weight is used for parameter 'normalization' because this is just a linear scaling factor. My colleague, Brian Anderson, and I have preferred to consider this 'normalising' weight as a standard weight. We do not try to make anything 'normal' but simply choose a scale that will provide us with a parameter that represents a standard human. We use a weight of 70 kg as the standard (Holford 1996). This makes it easier to compare parameters estimated in different populations, e.g. adults and children, because the parameters are scaled to the same standard size. Recently we have pointed out that use of a size standard with a suitable model for maturation allows the simple prediction of adult clearances from data collected in children (Anderson et al 2007b). We note that this cannot be done in the reverse direction i.e. adult data cannot be used to predict the maturational changes in clearance which occur in very young humans. Leonid mentions that results are "routinely reported as V/kg or CL/m^2, etc" but this is simply tradition without any discernible scientific rationale - especially the use of the square metre as the standardising factor. Drugs are not eliminated to any important extent via the skin so there is no mechanistic reason for this. The surface area method is a hangover of discredited theories of allometric scaling. The per kg method of scaling clearance is also a problem because it leads to the misguided viewpoint that clearance is larger in children in adults (Anderson & Holford 1997). Perhaps you can help your investigator colleague to read some of the published literature in this area so that he/she can get a clearer understanding of the size scaling issue. The bottom line: 'We conclude with the proposal that, at least in terms of pharmacokinetics, the widely quoted aphorism “Children are not small adults” should be changed to “Children are small adults — babies are young children.” ' Anderson & Holford 2007a Nick 1. Holford NHG. A size standard for pharmacokinetics. Clinical Pharmacokinetics 1996;30:329-332 2. Anderson BJ, McKee AD and Holford NHG. Size, myths and the clinical pharmacokinetics of analgesia in paediatric patients. Clin Pharmacokinet 1997;33:313-27 3. Anderson BJ, Holford NH. Mechanism-Based Concepts of Size and Maturity in Pharmacokinetics. Annu Rev Pharmacol Toxicol 2007a Oct 3; [Epub ahead of print] 4. Anderson BJ, Allegaert K, Van den Anker JN, Cossey V and Holford NH. Vancomycin pharmacokinetics in preterm neonates and the prediction of adult clearance. Br J Clin Pharmacol 2007b;63:75-84 Leonid Gibiansky wrote: > > Hi John, > I think you can safely separate statistics/mathematics and clinical use. > I would fit the model in the shape and form suitable to get the best > results (normalized to a typical patient in your case) and then report > the results in the form most convenient for the clinical use. If this is > per-kg values, then report it as they request. If you look in the > literature, results are routinely reported as V/kg or CL/m^2, etc. > > On a side not, I am actually surprised that you got different results > with different scaling. For the allometric scaling with fixed power, two > parameterizations: > > CL=TCL*WT^0.75 and CL=TCL*(WT/10)^0.75 > > differ by the fixed factor > (1/10)^0.75 = 0.18 > > I am not sure how this can influence your model CI so strongly. I would > check how you stratify the bootstrap data sets. Could it be that > stratification on something else depend on parameterization? > > If you would estimate the power: > > CL=TCL*(WT/10)^THETA() > > then parametrization would be more likely to affect CI, but for the > fixed power I would look for other explanations of the differences. It > would be easier to discuss the model if you would attach the PK block of > the nonmem code. > > Leonid > > -------------------------------------- > Leonid Gibiansky, Ph.D. > President, QuantPharm LLC > web:
>
> www.quantpharm.com
>
> > e-mail: LGibiansky at quantpharm.com > tel: (301) 767 5566 > > John Mondick wrote: > > I would like to get some feedback from the group concerning the reporting of modeling results. I have a Pop PK model developed from data arising from 124 pediatric patients, age 1 to 48 months. All of the structural parameters have been scaled allometrically, with the median body weight used as the reference value. After accounting for body size, a covariate model was incorporated to describe maturational changes in CL for young children. The maturation of clearance was modeled using an exponential model proposed in: > > > > Andersen et al. Population clinical pharmacology of children: modelling covariate effects. Eur J Pediatr. 2006 > > > > Two parameters are estimated as part of this model * the fractional change in CL for a typical one month old patient (beta - estimated to be 0.76 (0.589, 0.96) for this analysis) and a maturational half-life (TCL - 3.82 (1.57, 6.95) months). CI’s are from the bootstrap. > > > > The problem that I am running into is how to report the modeling results. It seems very natural to me to report the model results normalized to median body weight (L/h/10.4 kg^0.75). One of the study investigators disagrees with me and would like to report the results on a per kg basis (L/h/kg^0.75). This seems to be counterintuitive to me, as I tend to think about what represents the “typical patient.” It also makes no sense to me to represent the CL in a one kg child. The argument is that reporting in this manner makes more sense to clinicians and that there is no such thing as a typical child. > > > > So in an attempt to appease the investigator, I fit the same model with no weight normalization. The estimated parameters are equivalent to what would be scaled from the weight-normalized model, but there is no covariance matrix (not surprising). It becomes problematic when the bootstrap results are considered * beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018). Again, this is not surprising given that the covariate model is not centered. > > > > I have attempted to make several compromises, including reporting the parameter estimates in both median weight-normalized terms and normalized per kg. I have also included scaled CL estimates for typical patients at several ages and body weights. This hasn’t met the approval of the investigator, who is now insisting that I report the model building procedure from the median weight model, but report scaled parameters only on a per kg basis. This is wrong in my opinion and is actually more confusing to someone who is trying to understand the model. > > > > Can I get the group’s opinion on this? Am I being stubborn looking at the world through a modeler’s point of view? > > > > Thanks, > > > > > > > > > > John Mondick PhD > > Research Assistant Professor > > Division of Clinical Pharmacology and Therapeutics > > The Children's Hospital of Philadelphia > > Tel (267) 426-2292 > > FAX (215) 590-7544 > > Email:
>
> mondick @email.chop.edu
>
> > > > > -- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
>
> n.holford @auckland.ac.nz
>
> tel:+64(9)373-7599x86730 fax:+64(9)373-7090
>
> www.health.auckland.ac.nz/pharmacology/staff/nholford
That's even better by including the recommended dosing regimens for different
age/weight if the model results are directly used to guide dose administration
for clinicians -after all, CL is just an intermediate parameter.
Alan
Quoted reply history
-----Original Message-----
From: Bruce Charles [mailto:[EMAIL PROTECTED]
Sent: Wednesday, October 24, 2007 7:49 PM
To: Xiao, Alan; John Mondick; [email protected]
Subject: RE: [NMusers] Reporting Modeling Results
We have used the conversion-table approach a number of times before in trying
to negotiate the translation from model to clinic. See Table 5 in the attached
paper as an example.
Good luck.
Bruce CHARLES, BPharm(Hons), PhD, GradDipBusAdmin, MPS
Reader
School of Pharmacy
The University of Queensland, 4072 Australia
[University Provider Number: 00025B]
TEL: +61 7 336 53194
FAX: +61 7 336 51688
http://www.uq.edu.au/pharmacy/brucecharles/charles.html
[EMAIL PROTECTED]
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Xiao, Alan
Sent: Thursday, October 25, 2007 1:28 AM
To: John Mondick; [email protected]
Subject: RE: [NMusers] Reporting Modeling Results
An option to report your results to appease non-modelers in your case might be
reporting your CL values in a table with different age and weight.
Alan
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of John Mondick
Sent: Wednesday, October 24, 2007 11:18 AM
To: [email protected]
Subject: [NMusers] Reporting Modeling Results
I would like to get some feedback from the group concerning the reporting of
modeling results. I have a Pop PK model developed from data arising from 124
pediatric patients, age 1 to 48 months. All of the structural parameters have
been scaled allometrically, with the median body weight used as the reference
value. After accounting for body size, a covariate model was incorporated to
describe maturational changes in CL for young children. The maturation of
clearance was modeled using an exponential model proposed in:
Andersen et al. Population clinical pharmacology of children: modelling
covariate effects. Eur J Pediatr. 2006
Two parameters are estimated as part of this model * the fractional change in
CL for a typical one month old patient (beta - estimated to be 0.76 (0.589,
0.96) for this analysis) and a maturational half-life (TCL - 3.82 (1.57, 6.95)
months). CI's are from the bootstrap.
The problem that I am running into is how to report the modeling results. It
seems very natural to me to report the model results normalized to median body
weight (L/h/10.4 kg^0.75). One of the study investigators disagrees with me and
would like to report the results on a per kg basis (L/h/kg^0.75). This seems
to be counterintuitive to me, as I tend to think about what represents the
"typical patient." It also makes no sense to me to represent the CL in a one
kg child. The argument is that reporting in this manner makes more sense to
clinicians and that there is no such thing as a typical child.
So in an attempt to appease the investigator, I fit the same model with no
weight normalization. The estimated parameters are equivalent to what would be
scaled from the weight-normalized model, but there is no covariance matrix (not
surprising). It becomes problematic when the bootstrap results are considered
* beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018). Again, this is not
surprising given that the covariate model is not centered.
I have attempted to make several compromises, including reporting the parameter
estimates in both median weight-normalized terms and normalized per kg. I have
also included scaled CL estimates for typical patients at several ages and body
weights. This hasn't met the approval of the investigator, who is now insisting
that I report the model building procedure from the median weight model, but
report scaled parameters only on a per kg basis. This is wrong in my opinion
and is actually more confusing to someone who is trying to understand the model.
Can I get the group's opinion on this? Am I being stubborn looking at the world
through a modeler's point of view?
Thanks,
John Mondick PhD
Research Assistant Professor
Division of Clinical Pharmacology and Therapeutics
The Children's Hospital of Philadelphia
Tel (267) 426-2292
FAX (215) 590-7544
Email: [EMAIL PROTECTED]
I have received a lot of great feedback in response to my original email, both
on and off the board. But I think my original point was missed by some (or I
wasn't very clear in making it). There is no question that I would like to make
the results of an analysis accessible to any caregiver administering drug to a
patient. I also would never suggest that dosing guidance be based on
calculating a patient's dose from an allometric expression, particularly when
there is an age effect confounding the matter.
What I am attempting to resolve is the issue of developing a model using a
particular methodology (normalized to median body weight), but reporting only
results normalized to 1 kg. As stated in my original email, I am not opposed to
transforming parameters to reflect the per kg value (i.e., CL in L/h/kg^0.75).
I have also supplied parameter values for various ages and weights.
As far as reporting parameters in a linear fashion, the modelers here at our
institution have had great success in convincing researchers that the
traditional way of reporting CL in L/m2 or any other linear fashion is not
appropriate. What I can't seem to overcome is the thinking that parameters
should be reported only as normalized to one kg, even when scaled allomtrically
and even when the modeling procedure hasn't reflected this. My frustration lies
in the fact that this is the only way that this investigator would like to see
the results. This isn't the end of the world for me, I have reported parameters
in this fashion previously when I have been able to describe the data with an
appropriately structured covariate model.
As far as using 70 kg as a standard weight, that has met even more opposition.
The curious thing about that is that in reality, we compare PK in children to
that of adults and attempt to explain differences routinely. So I will stop
beating this dead horse and move onto the other issue.
Leonid - here is the code from the PK block. The bootstrap was stratified to
include a representative number of subjects above and below 20 months
(approximately 5 CL maturation half-lives). When I originally saw the boostrap
results from the model with parameters normalized to one kg, I chalked it up to
numerical instability (due to estimating an age effect on CL in a one kg child
which doesn't exist). Maybe I wasn't thinking correctly. Interestingly,
structural model parameters are fine, it's the CL maturation parameters that
yield wide CI's.
$PK
TVCL = THETA(1)*(WT/10.4)**0.75
BETA = THETA(5)
TCL = THETA(6)
FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
TVCL2 = TVCL*FCL
CL = TVCL2*EXP(ETA(1))
TVV1 = THETA(2)*(WT/10.4)
V1 = TVV1*EXP(ETA(2))
TVV2 = THETA(3)*(WT/10.4)
V2 = TVV2*EXP(ETA(3))
TVQ = THETA(4)*(WT/10.4)**0.75
Q = TVQ
Thanks for all of your input - JM
>>> Nick Holford <[EMAIL PROTECTED]> 10/25/2007 2:40:49 PM >>>
Mark,
We have a different view of the big picture. Regulatory and marketing are minor
shadows in the big picture of using medicines to improve health.
My job as a university based researcher is not to fill forms for companies to
send to regulatory agencies to tick boxes and write regulatory labels.
I work with clinicians to develop practical insights into treating children.
Brian Anderson is a paediatric specialist who applies and teaches the concepts
of rational dosing to clinicians. We work together to help understand PKPD in
children and find ways to bring it to clinical use.
The idea is to define the science first and then to apply it eg. with tables of
doses at different ages and weights. These tables are simple to generate and
distribute for clinicians to use. We certainly dont expect people to apply the
complex models that have been used to describe and understand the biology at
the bedside.
I would hope that John Mondick can be supported by contemporary scientific
literature to educate his colleague about science. After that they can find a
way to apply the science in a suitable format for busy clinicians.
Nick
>
>
>
> Nick,
> I think, in the big picture, it is important to remember why
> we do all this. It isn't for our own
> entertainment, or to congratulate each other on how insightful
> we are. It is to provide useful
> information to providers. We are not (we hope) the real
> audience for our work. It probably isn't
> realistic to expect non-oncologist pediatricians to scale doses
> allometrically. It seems to be
> unrealistic for other physicians (except neurologists) to scale
> doses at all. Dose/kg may be the best
> we can hope for. I'm quite sure it is unrealistic to expect
> drug companies to request labels for non!
> -chemotherapy drugs based on allometric scaling (really bad
> marketing move) Doing so - although
> perhaps scientifically correct, would likely lead to even more
> dosing errors that we currently see,
> with undemonstrated clinical benefit.
> So, how to report depends on who your audience is. If it is
> a bunch of nerds who know what
> eignevalues are, allometric scaling is great, if it is people
> who have 12 minutes to examine,
> diagnose and treat a patients, maybe we can keep it simple.
> Doing otherwise puts use are risk for
> irrelevance.
>
>
>
> Mark Sale MD
> Next Level Solutions, LLC
> www.NextLevelSolns.com
> 919-846-9185
>
Quoted reply history
> -------- Original Message --------
> Subject: Re: [NMusers] Reporting Modeling Results
> From: Nick Holford <[EMAIL PROTECTED]>
> Date: Thu, October 25, 2007 12:20 am
> To: nmusers <[email protected]>
> Cc: Brian Anderson <[EMAIL PROTECTED]>
>
> John,
>
> Leonid correctly points out that it makes no difference
> in terms of parameter
> estimation what weight is used for parameter
> 'normalization' because this is just a
> linear scaling factor.
>
> My colleague, Brian Anderson, and I have preferred to
> consider this 'normalising'
> weight as a standard weight. We do not try to make
> anything 'normal' but simply
> choose a scale that will provide us with a parameter
> that represents a standard
> human. We use a weight of 70 kg as the standard (Holford
> 1996). This makes it
> easier to compare parameters estimated in different
> populations, e.g. adults and
> children, because the parameters are scaled to the same
> standard size. Recently
> we have pointed out that use of a size standard with a
> suitable model for
> maturation allows the simple prediction of adult
> clearances from data collected in
> children (Anderson et al 2007b). We note that this
> cannot be done in the reverse
> direction i.e. adult data cannot be used to predict the
> maturational changes in
> clearance which occur in very young humans.
>
> Leonid mentions that results are "routinely reported as
> V/kg or CL/m^2, etc" but
> this is simply tradition without any discernible
> scientific rationale - especially the
> use of the square metre as the standardising factor.
> Drugs are not eliminated to any
> important extent via the skin so there is no mechanistic
> reason for this. The surface
> area method is a hangover of discredited theories of
> allometric scaling. The per kg
> method of scaling clearance is also a problem because it
> leads to the misguided
> viewpoint that clearance is larger in children in adults
> (Anderson & Holford 1997).
>
> Perhaps you can help your investigator colleague to read
> some of the published
> literature in this area so that he/she can get a clearer
> understanding of the size
> scaling issue.
>
> The bottom line:
>
> 'We conclude with the proposal that, at least in terms
> of pharmacokinetics, the
> widely quoted aphorism "Children are not small adults"
> should be changed to
> "Children are small adults * babies are young children."
> ' Anderson & Holford
> 2007a
>
> Nick
>
> 1. Holford NHG. A size standard for pharmacokinetics.
> Clinical Pharmacokinetics
> 1996;30:329-332
> 2. Anderson BJ, McKee AD and Holford NHG. Size, myths
> and the clinical
> pharmacokinetics of analgesia in paediatric patients.
> Clin Pharmacokinet
> 1997;33:313-27
> 3. Anderson BJ, Holford NH. Mechanism-Based Concepts of
> Size and Maturity
> in Pharmacokinetics. Annu Rev Pharmacol Toxicol 2007a
> Oct 3; [Epub ahead of
> print]
> 4. Anderson BJ, Allegaert K, Van den Anker JN, Cossey V
> and Holford NH.
> Vancomycin pharmacokinetics in preterm neonates and the
> prediction of adult
> clearance. Br J Clin Pharmacol 2007b;63:75-84
>
>
> Leonid Gibiansky wrote:
> >
> > Hi John,
> > I think you can safely separate statistics/mathematics
> and clinical use.
> > I would fit the model in the shape and form suitable
> to get the best
> > results (normalized to a typical patient in your case)
> and then report
> > the results in the form most convenient for the
> clinical use. If this is
> > per-kg values, then report it as they request. If you
> look in the
> > literature, results are routinely reported as V/kg or
> CL/m^2, etc.
> >
> > On a side not, I am actually surprised that you got
> different results
> > with different scaling. For the allometric scaling
> with fixed power, two
> > parameterizations:
> >
> > CL=TCL*WT^0.75 and CL=TCL*(WT/10)^0.75
> >
> > differ by the fixed factor
> > (1/10)^0.75 = 0.18
> >
> > I am not sure how this can influence your model CI so
> strongly. I would
> > check how you stratify the bootstrap data sets. Could
> it be that
> > stratification on something else depend on
> parameterization?
> >
> > If you would estimate the power:
> >
> > CL=TCL*(WT/10)^THETA()
> >
> > then parametrization would be more likely to affect
> CI, but for the
> > fixed power I would look for other explanations of the
> differences. It
> > would be easier to discuss the model if you would
> attach the PK block of
> > the nonmem code.
> >
> > Leonid
> >
> > --------------------------------------
> > Leonid Gibiansky, Ph.D.
> > President, QuantPharm LLC
> > web: www.quantpharm.com
> > e-mail: LGibiansky at quantpharm.com
> > tel: (301) 767 5566
> >
> > John Mondick wrote:
> > > I would like to get some feedback from the group
> concerning the reporting of
> modeling results. I have a Pop PK model developed from
> data arising from 124
> pediatric patients, age 1 to 48 months. All of the
> structural parameters have been
> scaled allometrically, with the median body weight used
> as the reference value.
> After accounting for body size, a covariate model was
> incorporated to describe
> maturational changes in CL for young children. The
> maturation of clearance was
> modeled using an exponential model proposed in:
> > >
> > > Andersen et al. Population clinical pharmacology of
> children: modelling
> covariate effects. Eur J Pediatr. 2006
> > >
> > > Two parameters are estimated as part of this model *
> the fractional change in
> CL for a typical one month old patient (beta - estimated
> to be 0.76 (0.589, 0.96)
> for this analysis) and a maturational half-life (TCL -
> 3.82 (1.57, 6.95) months).
> CI's are from the bootstrap.
> > >
> > > The problem that I am running into is how to report
> the modeling results. It
> seems very natural to me to report the model results
> normalized to median body
> weight (L/h/10.4 kg^0.75). One of the study
> investigators disagrees with me and
> would like to report the results on a per kg basis
> (L/h/kg^0.75). This seems to be
> counterintuitive to me, as I tend to think about what
> represents the "typical
> patient." It also makes no sense to me to represent the
> CL in a one kg child. The
> argument is that reporting in this manner makes more
> sense to clinicians and that
> there is no such thing as a typical child.
> > >
> > > So in an attempt to appease the investigator, I fit
> the same model with no
> weight normalization. The estimated parameters are
> equivalent to what would be
> scaled from the weight-normalized model, but there is no
> covariance matrix (not
> surprising). It becomes problematic when the bootstrap
> results are considered *
> beta = 0.78 (0.005, 0.995), TCL = 3.90 (0.001, 6.018).
> Again, this is not
> surprising given that the covariate model is not
> centered.
> > >
> > > I have attempted to make several compromises,
> including reporting the
> parameter estimates in both median weight-normalized
> terms and normalized per
> kg. I have also included scaled CL estimates for typical
> patients at several ages
> and body weights. This hasn't met the approval of the
> investigator, who is now
> insisting that I report the model building procedure
> from the median weight model,
> but report scaled parameters only on a per kg basis.
> This is wrong in my opinion
> and is actually more confusing to someone who is trying
> to understand the model.
> > >
> > > Can I get the group's opinion on this? Am I being
> stubborn looking at the
> world through a modeler's point of view?
> > >
> > > Thanks,
> > >
> > >
> > >
> > >
> > > John Mondick PhD
> > > Research Assistant Professor
> > > Division of Clinical Pharmacology and Therapeutics
> > > The Children's Hospital of Philadelphia
> > > Tel (267) 426-2292
> > > FAX (215) 590-7544
> > > Email: [EMAIL PROTECTED]
> > >
> > >
>
> --
> Nick Holford, Dept Pharmacology & Clinical Pharmacology
> University of Auckland, 85 Park Rd, Private Bag 92019,
> Auckland, New Zealand
> [EMAIL PROTECTED] tel:+64(9)373-7599x86730
> fax:+64(9)373-7090
> www.health.auckland.ac.nz/pharmacology/staff/nholford
>
>
>
>
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090
www.health.auckland.ac.nz/pharmacology/staff/nholford
Hi John,
The code seems to be good, and the results, in my opinion, should not depend on scaling. One idea: if you started both sets of problems (with and without scaling) from the same initial conditions (while the solutions differ by factor 5 or so), nonmem could have difficulties finding the correct minimum when started far from optimum. If your initial conditions were coming from 10.4 kg-normalized solution, then it could explain wider CI for the non-normalized problem: some of those runs did not converged or converged to a local minima. If this is true, you may want to repeat the non-normalized set with the initial conditions closer to the solution (if you choose to use non-normalized problem as the final model).
Leonid
--------------------------------------
Leonid Gibiansky, Ph.D.
President, QuantPharm LLC
web: www.quantpharm.com
e-mail: LGibiansky at quantpharm.com
tel: (301) 767 5566
John Mondick wrote:
> $PK TVCL = THETA(1)*(WT/10.4)**0.75
>
> BETA = THETA(5)
> TCL = THETA(6)
> FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
> TVCL2 = TVCL*FCL
> CL = TVCL2*EXP(ETA(1))
>
> TVV1 = THETA(2)*(WT/10.4)
> V1 = TVV1*EXP(ETA(2))
>
> TVV2 = THETA(3)*(WT/10.4)
> V2 = TVV2*EXP(ETA(3))
>
> TVQ = THETA(4)*(WT/10.4)**0.75
>
> Q = TVQ
Hi Leonid,
The aspect of the results that John describes as being dependent on
scaling are the (bootstrap) confidence intervals, not the parameter
estimates. I actually agree with John's original statements about not
being surprised by the inflated standard errors when the model is not
centered. Removal of normalization induces covariance in the estimation
of affected parameters. This covariance naturally leeds to inflated
standard errors (w/o bootstrapping). And as we all know, high
correlation between parameters can result in matrix singularity and
consequently failing covariance steps, consistent with John's
statements.
Best regards,
Jeroen
J. Elassaiss-Schaap
Scientist PK/PD
Organon NV
PO Box 20, 5340 BH Oss, Netherlands
Phone: + 31 412 66 9320
Fax: + 31 412 66 2506
e-mail: [EMAIL PROTECTED]
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Leonid Gibiansky
Sent: Thursday, 25 October, 2007 23:30
To: John Mondick
Cc: [email protected]
Subject: Re: [NMusers] Reporting Modeling Results
Hi John,
The code seems to be good, and the results, in my opinion, should not
depend on scaling. One idea: if you started both sets of problems (with
and without scaling) from the same initial conditions (while the
solutions differ by factor 5 or so), nonmem could have difficulties
finding the correct minimum when started far from optimum. If your
initial conditions were coming from 10.4 kg-normalized solution, then it
could explain wider CI for the non-normalized problem: some of those
runs did not converged or converged to a local minima. If this is true,
you may want to repeat the non-normalized set with the initial
conditions closer to the solution (if you choose to use non-normalized
problem as the final model).
Leonid
--------------------------------------
Leonid Gibiansky, Ph.D.
President, QuantPharm LLC
web: www.quantpharm.com
e-mail: LGibiansky at quantpharm.com
tel: (301) 767 5566
John Mondick wrote:
>
>
> $PK
>
> TVCL = THETA(1)*(WT/10.4)**0.75
> BETA = THETA(5)
> TCL = THETA(6)
> FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
> TVCL2 = TVCL*FCL
> CL = TVCL2*EXP(ETA(1))
>
> TVV1 = THETA(2)*(WT/10.4)
> V1 = TVV1*EXP(ETA(2))
>
> TVV2 = THETA(3)*(WT/10.4)
> V2 = TVV2*EXP(ETA(3))
>
> TVQ = THETA(4)*(WT/10.4)**0.75
> Q = TVQ
>
This message, including the attachments, is confidential and may be privileged.
If you are not an intended recipient, please notify the sender then delete and
destroy the original message and all copies. You should not copy, forward
and/or disclose this message, in whole or in part, without permission of the
sender.
Mark, Nick
I think there are two issues here.
First,
> I would, of course, have to disagree that regulatory and
> marketing are minor shadows in the big picture - that is how
> new medicine come to improve health. I would suggest that
> better use of existing medicine is not much more than a minor
> shadow,
I think we may have to agree to disagree here. There is mounting evidence
in a variety of therapeutic areas that post-marketing optimization of
existing drug treatments by skilled clinical staff can improve patient
outcomes. Indeed, there is growing research that indicates that the
numbered needed to treat (NNT) may be in the order of 10-15 patients (i.e.
treat 10 patients with optimized care on existing drugs to save 1 additional
event). This value of NNT is about as good as any new drug is usually able
to claim (treating AFib with warfarin has an NNT of about 100).
So - I think the pre-marketing and regulatory aspects are as important as
the post-marketing clinical aspects. Clearly we need new drugs - but we
also need to use them better. And I don't think we can hope to learn
everything there is to know about a drug during the drug development
process.
Second,
> and while I think you're almost certainly right (as
> usual), I can think of only a couple of examples where this
> has been empirically shown that pk/pd informed dosing insight
> GENERATED AFTER APPROVAL is better.
There are also a growing number of examples where dosing that has arisen out
of PKPD studies, that were gained after marketing, has provided significant
patient benefits. We have seen this in patients who are obese (and often
not included in pre-marketing trials) and with a variety of disease
pathologies.
It is (for me) without question that industry, regulatory and academia all
play equally important roles in improving patient care.
Regards
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
Steve, As usual, beating things to death. I certainly agree that post-marketing optimization is very useful. However, I said post-marketing (and intended to say outside the regulatory process) pk/pd. Great examples abound for post-marketing research, much of it in academics (the CAST study, recent work on Avandia - with apologies to former collegues at GSK, so much work on Coumadin in a-fib, once daily aminoglycosides, ACE inhibitors in heart failure, the list is very long). WRT studies such as special population (obese, elderly, young, renal failure), I'm interested in how many of these were done by ! academics, in their relentless search for truth, and how many were done by industry for regulatory or marketing reasons, and how many demonstrated clinical benefits, as opposed to just a pk difference. Mark Mark Sale MD Next Level Solutions, LLC
www.NextLevelSolns.com
919-846-9185
> -------- Original Message -------- Subject: RE: [NMusers] Reporting Modeling Results From: "Stephen Duffull" <[EMAIL PROTECTED]> Date: Thu, October 25, 2007 6:52 pm To: "'Mark Sale - Next Level Solutions'" <[EMAIL PROTECTED]> Cc: "'nmusers'" <
>
> [email protected]
>
> > Mark, Nick I think there are two issues here. First, > I would, of course, have to disagree that regulatory and > marketing are minor shadows in the big picture - that is how > new medicine come to improve health. I would suggest that > better use of existing medicine is not much more than a minor > shadow, I think we may have to agree to disagree here. There is mounting evidence in a variety of therapeutic areas that post-marketing optimization of existing drug treatments by skilled clinical staff can improve patient outcomes. Indeed, there is growing research that indicates that the numbered needed to treat (NNT) may be in the order of 10-15 patients (i.e. treat 10 patients with optimized care on existing drugs to save 1 additional event). This value of NNT is about as good as any new drug is usually able to claim (treating AFib with warfarin has an NNT of about 100). So - I think the pre-marketing and regulatory aspects are as important as the post-marketing clinical aspects. Clearly we need new drugs - but we also need to use them better. And I don't think we can hope to learn everything there is to know about a drug during the drug development process. Second, > and while I think you're almost certainly right (as > usual), I can think of only a couple of examples where this > has been empirically shown that pk/pd informed dosing insight > GENERATED AFTER APPROVAL is better. There are also a growing number of examples where dosing that has arisen out of PKPD studies, that were gained after marketing, has provided significant patient benefits. We have seen this in patients who are obese (and often not included in pre-marketing trials) and with a variety of disease pathologies. It is (for me) without question that industry, regulatory and academia all play equally important roles in improving patient care. Regards Steve -- Professor Stephen Duffull Chair of Clinical Pharmacy School of Pharmacy University of Otago PO Box 913 Dunedin New Zealand E:
>
> stephen.duffull @otago.ac.nz
>
> P: +64 3 479 5044 F: +64 3 479 7034 Design software:
>
> www.winpopt.com
Hi Jeroen,
I cannot see any extra covariance introduced by scaling by a fixed numbers. Moreover, I even not sure that the term "centered" is applicable to this model. It came from the linear model where the model
Y=ax+b
was called centered if presented in the form
Y=a1*(x-meanX)+b1
(here and below a and b are parameters, THETAs, while x is random)
Indeed, it is more stable in the second, centered form. If you assume that a1 and b1 are normally distributed without correlation, then
a = a1 and
b = b1-a1*meanX
are correlated, and correlation is higher for large meanX.
In this regard, the model that consist of several equations of the type
Y=b*x^0.75
is equivalent to the model with
Y=b'*(x/10)^0.75,
no extra correlations added.
However, the model
Y=b*x^a
is more like the linear model because it can be presented in the log scale as
log(Y)=a*log(x) + log(b)
and then it is better be centered:
log(Y)= a1*(log(x)-mean(log(x))) + log(b1)
or equivalently
Y=b1*(x/x0)^a1
where x0 is geometric mean of x.
Thus, these two (examples only!) models need to be centered:
CL=THETA()+(WT-70)*THETA()
or
CL=THETA()*(WT/70)^THETA()
but this one can be used in any shape and form
CL=THETA()*(WT/10)^0.75 = (THETA()/10^0.75) WT^0.75 =~ THETA() WT^0.75
This is just a scheme, not exact derivation/proof, but the bottom line is that scaling by a constant should not influence the standard errors, there should be another reason (may be, just insufficient sample size of the bootstrap process? or bounds on the parameters?) for the discrepancy. May be we need to stratify it into more than two age groups, to guarantee sufficient coverage of the entire age range of interest in each and every bootstrap set.
Leonid
Elassaiss - Schaap, J. (Jeroen) wrote:
> Hi Leonid,
>
> The aspect of the results that John describes as being dependent on
> scaling are the (bootstrap) confidence intervals, not the parameter
> estimates. I actually agree with John's original statements about not
> being surprised by the inflated standard errors when the model is not
> centered. Removal of normalization induces covariance in the estimation
> of affected parameters. This covariance naturally leeds to inflated
> standard errors (w/o bootstrapping). And as we all know, high
> correlation between parameters can result in matrix singularity and
> consequently failing covariance steps, consistent with John's
> statements.
>
> Best regards,
> Jeroen
>
> J. Elassaiss-Schaap
> Scientist PK/PD
> Organon NV
> PO Box 20, 5340 BH Oss, Netherlands
> Phone: + 31 412 66 9320
> Fax: + 31 412 66 2506
>
> e-mail: [EMAIL PROTECTED]
>
Quoted reply history
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
> On Behalf Of Leonid Gibiansky
> Sent: Thursday, 25 October, 2007 23:30
> To: John Mondick
> Cc: [email protected]
> Subject: Re: [NMusers] Reporting Modeling Results
>
> Hi John,
>
> The code seems to be good, and the results, in my opinion, should not depend on scaling. One idea: if you started both sets of problems (with and without scaling) from the same initial conditions (while the solutions differ by factor 5 or so), nonmem could have difficulties finding the correct minimum when started far from optimum. If your initial conditions were coming from 10.4 kg-normalized solution, then it
>
> could explain wider CI for the non-normalized problem: some of those runs did not converged or converged to a local minima. If this is true, you may want to repeat the non-normalized set with the initial conditions closer to the solution (if you choose to use non-normalized problem as the final model).
>
> Leonid
>
> --------------------------------------
> Leonid Gibiansky, Ph.D.
> President, QuantPharm LLC
> web: www.quantpharm.com
> e-mail: LGibiansky at quantpharm.com
> tel: (301) 767 5566
>
> John Mondick wrote:
>
> > $PK TVCL = THETA(1)*(WT/10.4)**0.75
> >
> > BETA = THETA(5)
> > TCL = THETA(6)
> > FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
> > TVCL2 = TVCL*FCL
> > CL = TVCL2*EXP(ETA(1))
> >
> > TVV1 = THETA(2)*(WT/10.4)
> > V1 = TVV1*EXP(ETA(2))
> >
> > TVV2 = THETA(3)*(WT/10.4)
> > V2 = TVV2*EXP(ETA(3))
> >
> > TVQ = THETA(4)*(WT/10.4)**0.75
> >
> > Q = TVQ
>
> This message, including the attachments, is confidential and may be privileged.
> If you are not an intended recipient, please notify the sender then delete and
> destroy the original message and all copies. You should not copy, forward
> and/or disclose this message, in whole or in part, without permission of the
> sender.
Hi Leonid,
You are right that BW normalization with a fixed slope (exponent)
shouldn't influence the estimation accuracy of the theta of clearance
directly. But the covariate AGE of course is correlated with body
weight, and a body weight of 10 corresponds roughly to a point somewhat
above the AGE of 1 - close the center of the AGE effect. The whole age
effect is scaled to the TVCL.
The effect of an excentric (1 kg is below even a newborn) basal
clearance will be a different shape of the AGE relationship because of
the non-linear relationship between AGE and BW - the deviation from a
linear relation will be larger for younger children/babies. Perhaps the
exponential form allows this shape change without large differences
between parameters (a bit more shallow, a bit less steep) but has the
effect of inducing parameter correlation and thus uncertainty.
John, would perhaps be so kind to inspect the correlation between the
bootstrapped parameters? That should provide some insight on the cause
of the inflated confidence intervals.
Best regards,
Jeroen
Quoted reply history
-----Original Message-----
From: Leonid Gibiansky [mailto:[EMAIL PROTECTED]
Sent: Friday, 26 October, 2007 5:27
To: Elassaiss - Schaap, J. (Jeroen)
Cc: John Mondick; [email protected]
Subject: Re: [NMusers] Reporting Modeling Results
Hi Jeroen,
I cannot see any extra covariance introduced by scaling by a fixed
numbers. Moreover, I even not sure that the term "centered" is
applicable to this model. It came from the linear model where the model
Y=ax+b
was called centered if presented in the form
Y=a1*(x-meanX)+b1
(here and below a and b are parameters, THETAs, while x is random)
Indeed, it is more stable in the second, centered form. If you assume
that a1 and b1 are normally distributed without correlation, then
a = a1 and
b = b1-a1*meanX
are correlated, and correlation is higher for large meanX.
In this regard, the model that consist of several equations of the type
Y=b*x^0.75
is equivalent to the model with
Y=b'*(x/10)^0.75,
no extra correlations added.
However, the model
Y=b*x^a
is more like the linear model because it can be presented in the log
scale as
log(Y)=a*log(x) + log(b)
and then it is better be centered:
log(Y)= a1*(log(x)-mean(log(x))) + log(b1)
or equivalently
Y=b1*(x/x0)^a1
where x0 is geometric mean of x.
Thus, these two (examples only!) models need to be centered:
CL=THETA()+(WT-70)*THETA()
or
CL=THETA()*(WT/70)^THETA()
but this one can be used in any shape and form
CL=THETA()*(WT/10)^0.75 = (THETA()/10^0.75) WT^0.75 =~ THETA() WT^0.75
This is just a scheme, not exact derivation/proof, but the bottom line
is that scaling by a constant should not influence the standard errors,
there should be another reason (may be, just insufficient sample size of
the bootstrap process? or bounds on the parameters?) for the
discrepancy. May be we need to stratify it into more than two age
groups, to guarantee sufficient coverage of the entire age range of
interest in each and every bootstrap set.
Leonid
Elassaiss - Schaap, J. (Jeroen) wrote:
> Hi Leonid,
>
> The aspect of the results that John describes as being dependent on
> scaling are the (bootstrap) confidence intervals, not the parameter
> estimates. I actually agree with John's original statements about not
> being surprised by the inflated standard errors when the model is not
> centered. Removal of normalization induces covariance in the
estimation
> of affected parameters. This covariance naturally leeds to inflated
> standard errors (w/o bootstrapping). And as we all know, high
> correlation between parameters can result in matrix singularity and
> consequently failing covariance steps, consistent with John's
> statements.
>
> Best regards,
> Jeroen
>
> J. Elassaiss-Schaap
> Scientist PK/PD
> Organon NV
> PO Box 20, 5340 BH Oss, Netherlands
> Phone: + 31 412 66 9320
> Fax: + 31 412 66 2506
> e-mail: [EMAIL PROTECTED]
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
> On Behalf Of Leonid Gibiansky
> Sent: Thursday, 25 October, 2007 23:30
> To: John Mondick
> Cc: [email protected]
> Subject: Re: [NMusers] Reporting Modeling Results
>
> Hi John,
> The code seems to be good, and the results, in my opinion, should not
> depend on scaling. One idea: if you started both sets of problems
(with
> and without scaling) from the same initial conditions (while the
> solutions differ by factor 5 or so), nonmem could have difficulties
> finding the correct minimum when started far from optimum. If your
> initial conditions were coming from 10.4 kg-normalized solution, then
it
>
> could explain wider CI for the non-normalized problem: some of those
> runs did not converged or converged to a local minima. If this is
true,
> you may want to repeat the non-normalized set with the initial
> conditions closer to the solution (if you choose to use non-normalized
> problem as the final model).
> Leonid
>
> --------------------------------------
> Leonid Gibiansky, Ph.D.
> President, QuantPharm LLC
> web: www.quantpharm.com
> e-mail: LGibiansky at quantpharm.com
> tel: (301) 767 5566
>
>
>
>
> John Mondick wrote:
>>
>> $PK
>>
>> TVCL = THETA(1)*(WT/10.4)**0.75
>> BETA = THETA(5)
>> TCL = THETA(6)
>> FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
>> TVCL2 = TVCL*FCL
>> CL = TVCL2*EXP(ETA(1))
>>
>> TVV1 = THETA(2)*(WT/10.4)
>> V1 = TVV1*EXP(ETA(2))
>>
>> TVV2 = THETA(3)*(WT/10.4)
>> V2 = TVV2*EXP(ETA(3))
>>
>> TVQ = THETA(4)*(WT/10.4)**0.75
>> Q = TVQ
>>
>
>
> This message, including the attachments, is confidential and may be
privileged. If you are not an intended recipient, please notify the
sender then delete and destroy the original message and all copies. You
should not copy, forward and/or disclose this message, in whole or in
part, without permission of the sender.
>
>
This message, including the attachments, is confidential and may be privileged.
If you are not an intended recipient, please notify the sender then delete and
destroy the original message and all copies. You should not copy, forward
and/or disclose this message, in whole or in part, without permission of the
sender.
Hi Steve,
I think it would be a good idea for you to cite the randomized studies
that back up your points in order to move this discussion forward.
When citing a number need to treat, we also need to consider the
actually cost of doing so. It's not that the staff in hospitals around
the world are not skilled, it's just that they are already very busy.
I'm trying to imagine a world where every hospital has a specialist
department for individually optimizing drug dosing for a wide variety of
drugs. What staff and equipment would it require? Are these staff
available? Would we require new procedures in the hospital to implement
all of the different drug protocols? What are the actual numerical
costs?
I am not asking these question rhetorically, I just think that this is a
very complex economic (rather than scientific) question and I would
value additional insight.
Best regards, James
James G Wright PhD
Scientist
Wright Dose Ltd
Tel: 44 (0) 772 5636914
www.wright-dose.com
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Stephen Duffull
Sent: 25 October 2007 23:53
To: 'Mark Sale - Next Level Solutions'
Cc: 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
Mark, Nick
I think there are two issues here.
First,
> I would, of course, have to disagree that regulatory and
> marketing are minor shadows in the big picture - that is how
> new medicine come to improve health. I would suggest that
> better use of existing medicine is not much more than a minor
> shadow,
I think we may have to agree to disagree here. There is mounting
evidence in a variety of therapeutic areas that post-marketing
optimization of existing drug treatments by skilled clinical staff can
improve patient outcomes. Indeed, there is growing research that
indicates that the numbered needed to treat (NNT) may be in the order of
10-15 patients (i.e. treat 10 patients with optimized care on existing
drugs to save 1 additional event). This value of NNT is about as good
as any new drug is usually able to claim (treating AFib with warfarin
has an NNT of about 100).
So - I think the pre-marketing and regulatory aspects are as important
as the post-marketing clinical aspects. Clearly we need new drugs - but
we also need to use them better. And I don't think we can hope to learn
everything there is to know about a drug during the drug development
process.
Second,
> and while I think you're almost certainly right (as
> usual), I can think of only a couple of examples where this
> has been empirically shown that pk/pd informed dosing insight
> GENERATED AFTER APPROVAL is better.
There are also a growing number of examples where dosing that has arisen
out of PKPD studies, that were gained after marketing, has provided
significant patient benefits. We have seen this in patients who are
obese (and often not included in pre-marketing trials) and with a
variety of disease pathologies.
It is (for me) without question that industry, regulatory and academia
all play equally important roles in improving patient care.
Regards
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
I guess I missed one point with a quick glance. The SE estimate for covariate
coefficient may change but the SE for the CL should not change significantly,
if the data is informative and representative and the model is stable and
number of bootstrapping is large enough and....
Alan
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Xiao, Alan
Sent: Friday, October 26, 2007 8:45 AM
To: Leonid Gibiansky; Elassaiss - Schaap, J. (Jeroen)
Cc: John Mondick; [email protected]
Subject: RE: [NMusers] Reporting Modeling Results
I would imagine that when scaling is different, the absolute magnitude of the
estimate of the coefficient for a covariate would be different and therefore
its standard error (SE) would be different accordingly, although the SEM%
(SE/Mean*100) might not be.
Alan
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Leonid Gibiansky
Sent: Thursday, October 25, 2007 11:27 PM
To: Elassaiss - Schaap, J. (Jeroen)
Cc: John Mondick; [email protected]
Subject: Re: [NMusers] Reporting Modeling Results
Hi Jeroen,
I cannot see any extra covariance introduced by scaling by a fixed
numbers. Moreover, I even not sure that the term "centered" is
applicable to this model. It came from the linear model where the model
Y=ax+b
was called centered if presented in the form
Y=a1*(x-meanX)+b1
(here and below a and b are parameters, THETAs, while x is random)
Indeed, it is more stable in the second, centered form. If you assume
that a1 and b1 are normally distributed without correlation, then
a = a1 and
b = b1-a1*meanX
are correlated, and correlation is higher for large meanX.
In this regard, the model that consist of several equations of the type
Y=b*x^0.75
is equivalent to the model with
Y=b'*(x/10)^0.75,
no extra correlations added.
However, the model
Y=b*x^a
is more like the linear model because it can be presented in the log
scale as
log(Y)=a*log(x) + log(b)
and then it is better be centered:
log(Y)= a1*(log(x)-mean(log(x))) + log(b1)
or equivalently
Y=b1*(x/x0)^a1
where x0 is geometric mean of x.
Thus, these two (examples only!) models need to be centered:
CL=THETA()+(WT-70)*THETA()
or
CL=THETA()*(WT/70)^THETA()
but this one can be used in any shape and form
CL=THETA()*(WT/10)^0.75 = (THETA()/10^0.75) WT^0.75 =~ THETA() WT^0.75
This is just a scheme, not exact derivation/proof, but the bottom line
is that scaling by a constant should not influence the standard errors,
there should be another reason (may be, just insufficient sample size of
the bootstrap process? or bounds on the parameters?) for the
discrepancy. May be we need to stratify it into more than two age
groups, to guarantee sufficient coverage of the entire age range of
interest in each and every bootstrap set.
Leonid
Elassaiss - Schaap, J. (Jeroen) wrote:
> Hi Leonid,
>
> The aspect of the results that John describes as being dependent on
> scaling are the (bootstrap) confidence intervals, not the parameter
> estimates. I actually agree with John's original statements about not
> being surprised by the inflated standard errors when the model is not
> centered. Removal of normalization induces covariance in the estimation
> of affected parameters. This covariance naturally leeds to inflated
> standard errors (w/o bootstrapping). And as we all know, high
> correlation between parameters can result in matrix singularity and
> consequently failing covariance steps, consistent with John's
> statements.
>
> Best regards,
> Jeroen
>
> J. Elassaiss-Schaap
> Scientist PK/PD
> Organon NV
> PO Box 20, 5340 BH Oss, Netherlands
> Phone: + 31 412 66 9320
> Fax: + 31 412 66 2506
> e-mail: [EMAIL PROTECTED]
>
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
> On Behalf Of Leonid Gibiansky
> Sent: Thursday, 25 October, 2007 23:30
> To: John Mondick
> Cc: [email protected]
> Subject: Re: [NMusers] Reporting Modeling Results
>
> Hi John,
> The code seems to be good, and the results, in my opinion, should not
> depend on scaling. One idea: if you started both sets of problems (with
> and without scaling) from the same initial conditions (while the
> solutions differ by factor 5 or so), nonmem could have difficulties
> finding the correct minimum when started far from optimum. If your
> initial conditions were coming from 10.4 kg-normalized solution, then it
>
> could explain wider CI for the non-normalized problem: some of those
> runs did not converged or converged to a local minima. If this is true,
> you may want to repeat the non-normalized set with the initial
> conditions closer to the solution (if you choose to use non-normalized
> problem as the final model).
> Leonid
>
> --------------------------------------
> Leonid Gibiansky, Ph.D.
> President, QuantPharm LLC
> web: www.quantpharm.com
> e-mail: LGibiansky at quantpharm.com
> tel: (301) 767 5566
>
>
>
>
> John Mondick wrote:
>>
>> $PK
>>
>> TVCL = THETA(1)*(WT/10.4)**0.75
>> BETA = THETA(5)
>> TCL = THETA(6)
>> FCL = 1-BETA*EXP(-(AGE-1)*0.693/TCL)
>> TVCL2 = TVCL*FCL
>> CL = TVCL2*EXP(ETA(1))
>>
>> TVV1 = THETA(2)*(WT/10.4)
>> V1 = TVV1*EXP(ETA(2))
>>
>> TVV2 = THETA(3)*(WT/10.4)
>> V2 = TVV2*EXP(ETA(3))
>>
>> TVQ = THETA(4)*(WT/10.4)**0.75
>> Q = TVQ
>>
>
>
> This message, including the attachments, is confidential and may be
> privileged. If you are not an intended recipient, please notify the sender
> then delete and destroy the original message and all copies. You should not
> copy, forward and/or disclose this message, in whole or in part, without
> permission of the sender.
>
>
James, Mark
I don't have time to go through all RCTs in this manner.
But I will provide a single recent example that I was involved with. For
Mark, the special population was obesity and renal impairment. In both
situations a continuous dosing scale was used rather than the recommended
"cut-off" value idea (i.e. change dose if CLCR < 30ml/min) ...
All studies were performed in academia, post-marketing, without industry
support
* Background study (model for LBM): Janmahasatian et al. Clin PK
2005;44:1051-1065
* PKPD studies: Green et al BJCP; 2002:54:96-103 & Green et al BJCP
2004;59:281-290
* Clinical study: (an RCT individualising dose to LBM) Barras, CPT 2007
[advance online publication 10 October 2007]
NNT: 4-8 (to reduce adverse outcomes)
Requirements, a dose calculator or nomogram to calculate LBM (see background
study), Cockcroft and Gault formula to calculate CLCR (based on LBM).
Otherwise limited impact on already busy clinical pharmacy services.
There are other examples (but I'm marking exams right now) but I'm sure
others can comment here too.
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
Quoted reply history
> -----Original Message-----
> From: James G Wright [mailto:[EMAIL PROTECTED]
> Sent: Friday, 26 October 2007 10:22 p.m.
> To: 'Stephen Duffull'; 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
> Hi Steve,
>
> I think it would be a good idea for you to cite the
> randomized studies that back up your points in order to move
> this discussion forward.
>
> When citing a number need to treat, we also need to consider
> the actually cost of doing so. It's not that the staff in
> hospitals around the world are not skilled, it's just that
> they are already very busy.
> I'm trying to imagine a world where every hospital has a
> specialist department for individually optimizing drug dosing
> for a wide variety of drugs. What staff and equipment would
> it require? Are these staff available? Would we require new
> procedures in the hospital to implement
> all of the different drug protocols? What are the actual numerical
> costs?
>
> I am not asking these question rhetorically, I just think
> that this is a very complex economic (rather than scientific)
> question and I would value additional insight.
>
> Best regards, James
>
> James G Wright PhD
> Scientist
> Wright Dose Ltd
> Tel: 44 (0) 772 5636914
> www.wright-dose.com
>
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]
> On Behalf Of Stephen Duffull
> Sent: 25 October 2007 23:53
> To: 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
>
> Mark, Nick
>
> I think there are two issues here.
>
> First,
>
> > I would, of course, have to disagree that regulatory
> and marketing
> > are minor shadows in the big picture - that is how new
> medicine come
> > to improve health. I would suggest that better use of existing
> > medicine is not much more than a minor shadow,
>
> I think we may have to agree to disagree here. There is
> mounting evidence in a variety of therapeutic areas that
> post-marketing optimization of existing drug treatments by
> skilled clinical staff can improve patient outcomes. Indeed,
> there is growing research that indicates that the numbered
> needed to treat (NNT) may be in the order of
> 10-15 patients (i.e. treat 10 patients with optimized care on
> existing drugs to save 1 additional event). This value of
> NNT is about as good as any new drug is usually able to claim
> (treating AFib with warfarin has an NNT of about 100).
>
> So - I think the pre-marketing and regulatory aspects are as
> important as the post-marketing clinical aspects. Clearly we
> need new drugs - but we also need to use them better. And I
> don't think we can hope to learn everything there is to know
> about a drug during the drug development process.
>
> Second,
>
> > and while I think you're almost certainly right (as usual), I can
> > think of only a couple of examples where this has been empirically
> > shown that pk/pd informed dosing insight GENERATED AFTER
> APPROVAL is
> > better.
>
> There are also a growing number of examples where dosing that
> has arisen out of PKPD studies, that were gained after
> marketing, has provided significant patient benefits. We
> have seen this in patients who are obese (and often not
> included in pre-marketing trials) and with a variety of
> disease pathologies.
>
> It is (for me) without question that industry, regulatory and
> academia all play equally important roles in improving patient care.
>
> Regards
>
> Steve
> --
> Professor Stephen Duffull
> Chair of Clinical Pharmacy
> School of Pharmacy
> University of Otago
> PO Box 913 Dunedin
> New Zealand
> E: [EMAIL PROTECTED]
> P: +64 3 479 5044
> F: +64 3 479 7034
>
> Design software: www.winpopt.com
>
>
Mark,
> I don't mean - at all - to suggest the pk work done in academics
> (post-marketing, without
> industry or regulatory support), but I had to take exception with the
> assertion:
> >Regulatory and marketing are minor shadows in the big picture of using
> medicines to improve
> >health.
> IMHO, there are a lot of very talented people in pharm companies and
> regulatory agencies,
> working very hard to figure out the best way to use drugs. I think that
> their contributions are not
> minor shadows in figuring out how to use medicine. They are the front line
> in doing this, not just
> box ticking exercises.
I agree with you that there are talented PKPD people in drug companies and even
at some regulatory agencies. But it is hard to discern any evidence of benefit
on outcome that can be attributed to regulatory agencies and marketing
activities. That is why I said regulatory and marketing are minor shadows in
the big picture. If you can point to such evidence in the published literature
I am willing to be educated.
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090
www.health.auckland.ac.nz/pharmacology/staff/nholford
Mark / James,
Just to clarify. The RCT performed on the back of the PKPD modeling papers
Steve highlighted showed decreased adverse events (bleeding) without loss in
effectiveness. So, I think we did show improved outcomes for an example
where there was a huge clinical need to do the work due to a less than
adequate drug label for the obese and renally impaired. Even with extremely
talented people in the industry and the regulatory authorities, dosing for
special populations with this drug was inadequately addressed to the extent
that many physicians were refusing to use the drug.
I also think that it is worth noting that many 'Clinical Pharmacists' are
employed specifically to dose individualize, and are often funded by medics
to do the job. I speak for the UK as that is my main experience, but I am
sure these services are offered worldwide. Specifically, warfarin, HIV,
lipid, heart failure clinics are commonly led by pharmacists to dose
optimize. Resources are available in hospitals in the UK as such
interventions have been shown to reduce readmission rates. In a typical UK
hospital about half the pharmacists are now funded from outside of the
'Pharmacy' budget.
Cheers
Bruce
----------------------------------------------------------------------------
--------------
Bruce Green, School of Pharmacy, University of Queensland, Australia
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Mark Sale - Next Level Solutions
Sent: Saturday, October 27, 2007 7:22 AM
Cc: 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
Steve,
Thanks for the example, there are lots of examples like that, I now
realize, all you have to do is go to the annual meeting of ASCPT. and while
I certainly believe that good pk targeting is useful, I assume you didn't
actually show an improved outcome. But, there are examples of pk/pd clin
pharm studies that do show outcome, like Lou Cantellinas work with
terfenadine (and grapefruit juice/emycin).
I don't mean - at all - to suggest the pk work done in academics
(post-marketing, without industry or regulatory support), but I had to take
exception with the assertion:
>Regulatory and market! ing are minor shadows in the big picture of using
medicines to improve >health.
IMHO, there are a lot of very talented people in pharm companies and
regulatory agencies, working very hard to figure out the best way to use
drugs. I think that their contributions are not minor shadows in figuring
out how to use medicine. They are the front line in doing this, not just box
ticking exercises.
Mark Sale MD
Next Level Solutions, LLC
www.NextLevelSolns.com
919-846-9185
Quoted reply history
-------- Original Message --------
Subject: RE: [NMusers] Reporting Modeling Results
From: "Stephen Duffull" <[EMAIL PROTECTED]>
Date: Fri, October 26, 2007 2:56 pm
To: "'James G Wright'" <[EMAIL PROTECTED]>, "'Mark Sale - Next
Level Solutions'" <[EMAIL PROTECTED]>
Cc: "'nmusers'" <[email protected]>
James, Mark
I don't have time to go through all RCTs in this manner.
But I will provide a single recent example that I was involved with. For
Mark, the special population was obesity and renal impairment. In both
situations a continuous dosing scale was used rather than the recommended
"cut-off" value idea (i.e. change dose if CLCR < 30ml/min) ...
All studies were performed in academia, post-marketing, without industry
support
* Background study (model for LBM): Janmahasatian et al. Clin PK
2005;44:1051-1065
* PKPD studies: Green et al BJCP; 2002:54:96-103 & Green et al BJCP
2004;59:281-290
* Clinical study: (an RCT individualising dose to LBM) Barras, CPT 2007
[advance online publication 10 October 2007]
NNT: 4-8 (to reduce adverse outcomes)
Requirements, a dose calculator or nomogram to calculate LBM (see background
study), Cockcroft and Gault formula to calculate CLCR (based on LBM).
Otherwise limited impact on already busy clinical pharmacy services.
There are other examples (but I'm marking exams right now) but I'm sure
others can comment here too.
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
http://email.secureserver.net/pcompose.php#Compose
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
> -----Original Message-----
> From: James G Wright [mailto:[EMAIL PROTECTED]
http://email.secureserver.net/pcompose.php#Compose ]
> Sent: Friday, 26 October 2007 10:22 p.m.
> To: 'Stephen Duffull'; 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
> Hi Steve,
>
> I think it would be a good idea for you to cite the
> randomized studies that back up your points in order to move
> this discussion forward.
>
> When citing a number need to treat, we also need to consider
> the actually cost of doing so. It's not that the staff in
> hospitals around the world are not skilled, it's just that
> they are already very busy.
> I'm trying to imagine a world where every hospital has a
> specialist department for individually optimizing drug dosing
> for a wide variety of drugs. What staff and equipment would
> it require? Are these staff available? Would we require new
> procedures in the hospital to implement
> all of the different drug protocols? What are the actual numerical
> costs?
>
> I am not asking these question rhetorically, I just think
> that this is a very complex economic (rather than scientific)
> question and I would value additional insight.
>
> Best regards, James
>
> James G Wright PhD
> Scientist
> Wright Dose Ltd
> Tel: 44 (0) 772 5636914
> www.wright-dose.com
>
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
http://email.secureserver.net/pcompose.php#Compose
> [mailto:[EMAIL PROTECTED]
http://email.secureserver.net/pcompose.php#Compose ]
> On Behalf Of Stephen Duffull
> Sent: 25 October 2007 23:53
> To: 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
>
> Mark, Nick
>
> I think there are two issues here.
>
> First,
>
> > I would, of course, have to disagree that regulatory
> and marketing
> > are minor shadows in the big picture - that is how new
> medicine come
> > to improve health. I would suggest that better use of existing
> > medicine is not much more than a minor shadow,
>
> I think we may have to agree to disagree here. There is
> mounting evidence in a variety of therapeutic areas that
> post-marketing optimization of existing drug treatments by
> skilled clinical staff can improve patient outcomes. Indeed,
> there is growing research that indicates that the numbered
> needed to treat (NNT) may be in the order of
> 10-15 patients (i.e. treat 10 patients with optimized care on
> existing drugs to save 1 additional event). This value of
> NNT is about as good as any new drug is usually able to claim
> (treating AFib with warfarin has an NNT of about 100).
>
> So - I think the pre-marketing and regulatory aspects are as
> important as the post-marketing clinical aspects. Clearly we
> need new drugs - but we also need to use them better. And I
> don't think we can hope to learn everything there is to know
> about a drug during the drug development process.
>
> Second,
>
> > and while I think you're almost certainly right (as usual), I can
> > think of only a couple of examples where this has been empirically
> > shown that pk/pd informed dosing insight GENERATED AFTER
> APPROVAL is
> > better.
>
> There are also a growing number of examples where dosing that
> has arisen out of PKPD studies, that were gained after
> marketing, has provided significant patient benefits. We
> have seen this in patients who are obese (and often not
> included in pre-marketing trials) and with a variety of
> disease pathologies.
>
> It is (for me) without question that industry, regulatory and
> academia all play equally important roles in improving patient care.
>
> Regards
>
> Steve
> --
> Professor Stephen Duffull
> Chair of Clinical Pharmacy
> School of Pharmacy
> University of Otago
> PO Box 913 Dunedin
> New Zealand
> E: [EMAIL PROTECTED]
http://email.secureserver.net/pcompose.php#Compose
> P: +64 3 479 5044
> F: +64 3 479 7034
>
> Design software: www.winpopt.com
>
>
<<image001.png>>
All
To follow on from Bruce's comments:
"So, I think we did show improved outcomes for an example where there was a
huge clinical need to do the work due to a less than adequate drug label for
the obese and renally impaired. "
It seems we have come almost full circle back to an original component of
the thread. Which included discussion about how to display dosing
recommendations to prescribers. It seems, that oversimplifying the dosing
regimen to meet perceived limitations of the clinical staff may not always
be in the patient's best interests.
So perhaps (to misquote Einstein) - "the dosing regimen should be as simple
as possible but not simpler..."
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
Nick, Tacrine, My understanding (correct me if I'm wrong), was that the traditional analysis was unconvincing wrt efficacy and the model you developed ultimately helped justify approval. Presumably, some patients had an improved outcome from the use of the drug. My experience with end of phase IIa meetings is that they are quite helpful in dose selection (often supporting the position that clin pharm took within the company). WRT the lamictal example, while the analysis was not accepted as sufficient support for label change, the dosing algorithm that was developed was used in a trial ! that led to label change. There are lots of example within companies, or analyses that led to rational dose decisions, that didn't end up published. This is what you've been preaching for 20 years or so - are you refusing to acknowledge your success? Mark Mark Sale MD Next Level Solutions, LLC
www.NextLevelSolns.com
919-846-9185
> -------- Original Message -------- Subject: Re: [NMusers] Reporting Modeling Results From: Nick Holford <[EMAIL PROTECTED]> Date: Fri, October 26, 2007 7:26 pm To: "'nmusers'" <
>
> [email protected]
>
> > Mark, I can't see how these examples show any influence of regulatory agencies or marketing on outcome. Can you explain in each case why you think this is so? Nick I wrote: > > I agree with you that there are talented PKPD people in drug companies and even > at some regulatory agencies. But it is hard to discern any evidence of benefit on > outcome that can be attributed to regulatory agencies and marketing activities. > That is why I said regulatory and marketing are minor shadows in the big picture. > If you can point to such evidence in the published literature I am willing to be > educated. You replied: > Nick, > I might start with a piece of work that got a drug approved for a disease that, at the time was without any treatment, you may be familiar with this. >
>
> http://www.pnas.org/cgi/reprint/89/23/11466.pdf
>
> > > and move on the the current end of phase IIa process, which is mostly focused around using > modeling to select dosing regimens, and finish with a purely regulatory and marketing piece of > work that I'm familiar with: >
>
> http://www.aesnet.org/Visitors/AnnualMeeting/A
>
> ! bstracts/dsp_Abstract.cfm?id=2578 > (this got published as a real article, but I can't find a link to the paper). -- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
>
> n.holford @auckland.ac.nz
>
> tel:+64(9)373-7599x86730 fax:+64(9)373-7090
>
> www.health.auckland.ac.nz/pharmacology/staff/nholford
Mark,
Your examples illustrate PKPD to aid drug development and the rational use of
medicines. I don't disagree with this.
But you keep missing my point that the actions of regulators and drug marketers
do not discover how to use medicines. Many groups in industry, academia and
clinical medicine contribute to this knowledge by designing, performing,
analysing and reporting clinical trials -- it is not the role of regulators and
marketers to do these things. This is why I said "Regulatory and marketing are
minor shadows in the big picture of using medicines to improve health."
Nick
> Nick,
>
> Tacrine,
> My understanding (correct me if I'm wrong), was that the
> traditional analysis was unconvincing
> wrt efficacy and the model you developed ultimately helped
> justify approval. Presumably, some
> patients had an improved outcome from the use of the drug.
> My experience with end of phase IIa meetings is that they are
> quite helpful in dose selection (often
> supporting the position that clin pharm took within the
> company).
> WRT the lamictal example, while the analysis was not accepted
> as sufficient support for label
> change, the dosing algorithm that was developed was used in a
> trial that led to label change.
>
> There are lots of example within companies, or analyses that
> led to rational dose decisions, that
> didn't end up published. This is what you've been preaching
> for 20 years or so - are you refusing
> to acknowledge your success?
>
> Mark
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090
www.health.auckland.ac.nz/pharmacology/staff/nholford
Hi Steve and Bruce,
As you are holding up this enoxaparin work as a generalizable example, I
think your bold claims merit at least a token challenge. When I review
academic studies, the first thing I check for is patients who are
excluded from the data analysis.
4 of 60 patients in the new enoxaparin individualization regime arm were
excluded post-study, whereas there were zero exclusions in the 62
patients tried-and-tested per-label individualization arm. A single
event in one of the excluded subjects would render the p-value for the
primary endpoint non-significant. The p-value against all 4 exclusions
occuring in the novel arm by chance is less than .05, so this begs the
question...
Did excluding these 4 subjects alter the results of the trial?
Best regards, James
James G Wright PhD
Scientist
Wright Dose Ltd
Tel: 44 (0) 772 5636914
www.wright-dose.com
Quoted reply history
-----Original Message-----
From: Stephen Duffull [mailto:[EMAIL PROTECTED]
Sent: 26 October 2007 19:57
To: 'James G Wright'; 'Mark Sale - Next Level Solutions'
Cc: 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
James, Mark
I don't have time to go through all RCTs in this manner.
But I will provide a single recent example that I was involved with.
For Mark, the special population was obesity and renal impairment. In
both situations a continuous dosing scale was used rather than the
recommended "cut-off" value idea (i.e. change dose if CLCR < 30ml/min)
...
All studies were performed in academia, post-marketing, without industry
support
* Background study (model for LBM): Janmahasatian et al. Clin PK
2005;44:1051-1065
* PKPD studies: Green et al BJCP; 2002:54:96-103 & Green et al BJCP
2004;59:281-290
* Clinical study: (an RCT individualising dose to LBM) Barras, CPT 2007
[advance online publication 10 October 2007]
NNT: 4-8 (to reduce adverse outcomes)
Requirements, a dose calculator or nomogram to calculate LBM (see
background study), Cockcroft and Gault formula to calculate CLCR (based
on LBM). Otherwise limited impact on already busy clinical pharmacy
services.
There are other examples (but I'm marking exams right now) but I'm sure
others can comment here too.
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
> -----Original Message-----
> From: James G Wright [mailto:[EMAIL PROTECTED]
> Sent: Friday, 26 October 2007 10:22 p.m.
> To: 'Stephen Duffull'; 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
> Hi Steve,
>
> I think it would be a good idea for you to cite the randomized studies
> that back up your points in order to move this discussion forward.
>
> When citing a number need to treat, we also need to consider the
> actually cost of doing so. It's not that the staff in hospitals
> around the world are not skilled, it's just that they are already very
> busy. I'm trying to imagine a world where every hospital has a
> specialist department for individually optimizing drug dosing for a
> wide variety of drugs. What staff and equipment would it require? Are
> these staff available? Would we require new procedures in the
> hospital to implement
> all of the different drug protocols? What are the actual numerical
> costs?
>
> I am not asking these question rhetorically, I just think that this is
> a very complex economic (rather than scientific) question and I would
> value additional insight.
>
> Best regards, James
>
> James G Wright PhD
> Scientist
> Wright Dose Ltd
> Tel: 44 (0) 772 5636914
> www.wright-dose.com
>
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]
> On Behalf Of Stephen Duffull
> Sent: 25 October 2007 23:53
> To: 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
>
> Mark, Nick
>
> I think there are two issues here.
>
> First,
>
> > I would, of course, have to disagree that regulatory
> and marketing
> > are minor shadows in the big picture - that is how new
> medicine come
> > to improve health. I would suggest that better use of existing
> > medicine is not much more than a minor shadow,
>
> I think we may have to agree to disagree here. There is mounting
> evidence in a variety of therapeutic areas that post-marketing
> optimization of existing drug treatments by skilled clinical staff can
> improve patient outcomes. Indeed, there is growing research that
> indicates that the numbered needed to treat (NNT) may be in the order
> of 10-15 patients (i.e. treat 10 patients with optimized care on
> existing drugs to save 1 additional event). This value of NNT is
> about as good as any new drug is usually able to claim (treating AFib
> with warfarin has an NNT of about 100).
>
> So - I think the pre-marketing and regulatory aspects are as important
> as the post-marketing clinical aspects. Clearly we need new drugs -
> but we also need to use them better. And I don't think we can hope to
> learn everything there is to know about a drug during the drug
> development process.
>
> Second,
>
> > and while I think you're almost certainly right (as usual), I can
> > think of only a couple of examples where this has been empirically
> > shown that pk/pd informed dosing insight GENERATED AFTER
> APPROVAL is
> > better.
>
> There are also a growing number of examples where dosing that has
> arisen out of PKPD studies, that were gained after marketing, has
> provided significant patient benefits. We have seen this in patients
> who are obese (and often not included in pre-marketing trials) and
> with a variety of disease pathologies.
>
> It is (for me) without question that industry, regulatory and academia
> all play equally important roles in improving patient care.
>
> Regards
>
> Steve
> --
> Professor Stephen Duffull
> Chair of Clinical Pharmacy
> School of Pharmacy
> University of Otago
> PO Box 913 Dunedin
> New Zealand
> E: [EMAIL PROTECTED]
> P: +64 3 479 5044
> F: +64 3 479 7034
>
> Design software: www.winpopt.com
>
>
James
I had hoped we would move on :-)
> As you are holding up this enoxaparin work as a generalizable
> example,
It is just a single example - there are many others.
> I think your bold claims merit at least a token
> challenge. When I review academic studies, the first thing I
> check for is patients who are excluded from the data analysis.
I am absolutely astounded that you consider academic studies in a more
critical light than industry driven studies. Shouldn't all studies be taken
on their merit - or are you suggesting that academic studies are naturally
flawed in some way?
> Did excluding these 4 subjects alter the results of the trial?
Simply the patients were removed before analysis. They did not meet
protocol requirements (e.g. some received unfractionated heparin).
It is my best belief that if we continued to recruit patients that we would
see the same signal as the trial found. The trial was adequately powered so
we are not expecting bias inherent in underpowered studies (compare to the
underpowered APPROVe study wrt CVS observations).
In addition, the purpose of this part of the thread, for me, was to show
that a) studies do arise in academic settings that improve patient care [to
respond to Mark's comments] and b) that simplifying the dosing on the drug
label to make the drug "easier to use" doesn't necessarily result in better
patient outcomes.
Regards
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
Hi James and Steve
I too am astounded that James thinks academic studies should be considered in a
more critical light than industry studies.
Which reminded me of this little snippet from an abstracting journal earlier
this week!
"Funding source of clinical studies affects authors' interpretation of results
The funding source of clinical studies affects authors' interpretation of
results, an analysis of 504 studies examining the effects of inhaled
corticosteroids suggests (Archives of Internal Medicine 2007:167:2047).
Manufacturer-funded trials are less likely to identify adverse effects of
inhaled cortico-steroids than studies without pharmaceutical company funding, a
finding partly explained by differences in trial design.
In addition, authors of manufacturer-funded trials are less likely to conclude
that statistically significant adverse effects are clinically important than
authors of non-pharmaceutical-funded studies, researchers found"
Cheers
Carl
Carl Kirkpatrick BPharm(Hons) PhD MPS
School of Pharmacy
University of Queensland
St Lucia 4073
Brisbane, Australia
Phone: +617 33653227
Fax: +617 33651688
URL: http://www.uq.edu.au/pharmacy/ckirkpatrick/kirkpatrick.htm
"Doing things the way we always do them isn't the answer." Richard Soley
University Provider Number: 00025B
Hi James,
The 4 subjects were excluded immediately after consent, not at the data
analysis level. Steve has provided further comments so I will move on.
I would actually like to know if you believe drugs are dosed in practice
according to the label? I ask because the results from our trial are
actually an under-representation of the benefits of individualized dosing,
which you might not be aware of. You said:
"in the 62 patients 'tried-and-tested per-label' individualization arm"
In the real world, drug doses prescribed by practitioners will use the label
as a guideline. Copying the label verbatim is an administrative task that
does not require degrees and health care training. Individualized care uses
knowledge about the patient and the drug to develop treatment strategies
that should yield improved outcomes. In this trial, physicians continued to
exercise this knowledge in the 'tried-and-tested per-label arm' and did not
copy the label verbatim. I presume this was due to fear of overdose and was
very happy with the result. After all, why would you want to follow the
label when you know it is inadequate and would harm patients?
Cheers
Bruce
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of James G Wright
Sent: Monday, October 29, 2007 8:44 PM
To: 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
Hi Steve and Bruce,
As you are holding up this enoxaparin work as a generalizable example, I
think your bold claims merit at least a token challenge. When I review
academic studies, the first thing I check for is patients who are
excluded from the data analysis.
4 of 60 patients in the new enoxaparin individualization regime arm were
excluded post-study, whereas there were zero exclusions in the 62
patients tried-and-tested per-label individualization arm. A single
event in one of the excluded subjects would render the p-value for the
primary endpoint non-significant. The p-value against all 4 exclusions
occuring in the novel arm by chance is less than .05, so this begs the
question...
Did excluding these 4 subjects alter the results of the trial?
Best regards, James
James G Wright PhD
Scientist
Wright Dose Ltd
Tel: 44 (0) 772 5636914
www.wright-dose.com
-----Original Message-----
From: Stephen Duffull [mailto:[EMAIL PROTECTED]
Sent: 26 October 2007 19:57
To: 'James G Wright'; 'Mark Sale - Next Level Solutions'
Cc: 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
James, Mark
I don't have time to go through all RCTs in this manner.
But I will provide a single recent example that I was involved with.
For Mark, the special population was obesity and renal impairment. In
both situations a continuous dosing scale was used rather than the
recommended "cut-off" value idea (i.e. change dose if CLCR < 30ml/min)
...
All studies were performed in academia, post-marketing, without industry
support
* Background study (model for LBM): Janmahasatian et al. Clin PK
2005;44:1051-1065
* PKPD studies: Green et al BJCP; 2002:54:96-103 & Green et al BJCP
2004;59:281-290
* Clinical study: (an RCT individualising dose to LBM) Barras, CPT 2007
[advance online publication 10 October 2007]
NNT: 4-8 (to reduce adverse outcomes)
Requirements, a dose calculator or nomogram to calculate LBM (see
background study), Cockcroft and Gault formula to calculate CLCR (based
on LBM). Otherwise limited impact on already busy clinical pharmacy
services.
There are other examples (but I'm marking exams right now) but I'm sure
others can comment here too.
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com
> -----Original Message-----
> From: James G Wright [mailto:[EMAIL PROTECTED]
> Sent: Friday, 26 October 2007 10:22 p.m.
> To: 'Stephen Duffull'; 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
> Hi Steve,
>
> I think it would be a good idea for you to cite the randomized studies
> that back up your points in order to move this discussion forward.
>
> When citing a number need to treat, we also need to consider the
> actually cost of doing so. It's not that the staff in hospitals
> around the world are not skilled, it's just that they are already very
> busy. I'm trying to imagine a world where every hospital has a
> specialist department for individually optimizing drug dosing for a
> wide variety of drugs. What staff and equipment would it require? Are
> these staff available? Would we require new procedures in the
> hospital to implement
> all of the different drug protocols? What are the actual numerical
> costs?
>
> I am not asking these question rhetorically, I just think that this is
> a very complex economic (rather than scientific) question and I would
> value additional insight.
>
> Best regards, James
>
> James G Wright PhD
> Scientist
> Wright Dose Ltd
> Tel: 44 (0) 772 5636914
> www.wright-dose.com
>
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]
> On Behalf Of Stephen Duffull
> Sent: 25 October 2007 23:53
> To: 'Mark Sale - Next Level Solutions'
> Cc: 'nmusers'
> Subject: RE: [NMusers] Reporting Modeling Results
>
>
> Mark, Nick
>
> I think there are two issues here.
>
> First,
>
> > I would, of course, have to disagree that regulatory
> and marketing
> > are minor shadows in the big picture - that is how new
> medicine come
> > to improve health. I would suggest that better use of existing
> > medicine is not much more than a minor shadow,
>
> I think we may have to agree to disagree here. There is mounting
> evidence in a variety of therapeutic areas that post-marketing
> optimization of existing drug treatments by skilled clinical staff can
> improve patient outcomes. Indeed, there is growing research that
> indicates that the numbered needed to treat (NNT) may be in the order
> of 10-15 patients (i.e. treat 10 patients with optimized care on
> existing drugs to save 1 additional event). This value of NNT is
> about as good as any new drug is usually able to claim (treating AFib
> with warfarin has an NNT of about 100).
>
> So - I think the pre-marketing and regulatory aspects are as important
> as the post-marketing clinical aspects. Clearly we need new drugs -
> but we also need to use them better. And I don't think we can hope to
> learn everything there is to know about a drug during the drug
> development process.
>
> Second,
>
> > and while I think you're almost certainly right (as usual), I can
> > think of only a couple of examples where this has been empirically
> > shown that pk/pd informed dosing insight GENERATED AFTER
> APPROVAL is
> > better.
>
> There are also a growing number of examples where dosing that has
> arisen out of PKPD studies, that were gained after marketing, has
> provided significant patient benefits. We have seen this in patients
> who are obese (and often not included in pre-marketing trials) and
> with a variety of disease pathologies.
>
> It is (for me) without question that industry, regulatory and academia
> all play equally important roles in improving patient care.
>
> Regards
>
> Steve
> --
> Professor Stephen Duffull
> Chair of Clinical Pharmacy
> School of Pharmacy
> University of Otago
> PO Box 913 Dunedin
> New Zealand
> E: [EMAIL PROTECTED]
> P: +64 3 479 5044
> F: +64 3 479 7034
>
> Design software: www.winpopt.com
>
>
Steve,
Retrospectively excluding treated subjects from an RCT is a basic
methodological error (no matter who does it). Doing so casts very
fundamental doubts on the reported results. However, if these 4
exclusions didn't in fact bias the results of your study, then your
conclusions may still be correct.
Did any adverse events occur in the 4 excluded subjects in the novel
treatment arm? This is all I wish to know.
Best regards, James
James G Wright PhD
Scientist
Wright Dose Ltd
Tel: 44 (0) 772 5636914
www.wright-dose.com
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Stephen Duffull
Sent: 29 October 2007 20:21
To: 'James G Wright'; 'nmusers'
Subject: RE: [NMusers] Reporting Modeling Results
James
I had hoped we would move on :-)
> As you are holding up this enoxaparin work as a generalizable example,
It is just a single example - there are many others.
> I think your bold claims merit at least a token
> challenge. When I review academic studies, the first thing I check for
> is patients who are excluded from the data analysis.
I am absolutely astounded that you consider academic studies in a more
critical light than industry driven studies. Shouldn't all studies be
taken on their merit - or are you suggesting that academic studies are
naturally flawed in some way?
> Did excluding these 4 subjects alter the results of the trial?
Simply the patients were removed before analysis. They did not meet
protocol requirements (e.g. some received unfractionated heparin).
It is my best belief that if we continued to recruit patients that we
would see the same signal as the trial found. The trial was adequately
powered so we are not expecting bias inherent in underpowered studies
(compare to the underpowered APPROVe study wrt CVS observations).
In addition, the purpose of this part of the thread, for me, was to show
that a) studies do arise in academic settings that improve patient care
[to respond to Mark's comments] and b) that simplifying the dosing on
the drug label to make the drug "easier to use" doesn't necessarily
result in better patient outcomes.
Regards
Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: [EMAIL PROTECTED]
P: +64 3 479 5044
F: +64 3 479 7034
Design software: www.winpopt.com