FW: Evidence-based and "Mix and Match" Programs

["Don Cohon" <cohond@xxxxxxxxxxxxx>]

Forwarding comments on this thread by the Director of Programs and Strategy
at my agency.

Don Cohon, Ph.D., Director
Institute for the Study of Community-Based Services
Edgewood Center for Children and Families
1801 Vicente Street
San Francisco, CA 94116

(415) 383-2273


-----Original Message-----
From: Ken Epstein [
Sent: Tuesday, December 30, 2008 11:58 PM
To: Don Cohon--UCSF
Subject: FW: Evidence-based and "Mix and Match" Programs

Yes this is a good discussion but it needs to have an interface with
Program people.  I fear that the conversation does not address the
realities of implementing programs in the fiscal and overlitigious
environment we are in and that it does not mention the costs associated
with buying the components or the manualized practice, having a research
group do the research, having supervisors and mentors, maintaining staff
and building sustainable funding for programs.  Our funders more and
more want us to provide evidence based programs and less and less give
us the dollars and time to build these processes.  I also agree
wholeheartedly with the notion that more is not good. However I disagree
that we need to forego addressing poverty or racism or poor education as
a source.  If we continue to support environments that increasingly
expose our youth and adults to more and more violence, then we create
more and more candidates for programs with less and less resources.
Simply put violent crime is down in the US except with African American
Teens where it has increased dramatically.  

 

I fear that the evidence revolution will ultimately pick and choose good
candidates for success with limited interventions that are tested but
cannot be widely replicated or sustained and are not effective with the
most neediest and most challenging families.  


-----Original Message-----
From: bounce-3426746-6833904@list.cornell.edu
[mailto:bounce-3426746-6833904@list.cornell.edu] On Behalf Of Ben
Saunders
Sent: Monday, December 29, 2008 10:20 AM
To: Child Maltreatment Researchers
Subject: Re: Evidence-based and "Mix and Match" Programs

 

Sandra,

Absolutely agree with your comments (other than the "titan" 

characterization).  Without question, neighborhood and other contextual 

factors have an enormous impact not only on the functioning of people 

living in them, but also on the range of options available to helping 

professionals trying to do their job.  Our challenge is, how can we best
do good child welfare work in very bad contexts with families with 

multiple problems.  Do we have to cure poverty, crime, unemployment, 

poor education, housing shortages, and discrimination in order to do 

quality child welfare work?  I hope not.  But do these social ills make 

it tough work? Absolutely.  I think this is the focus part Mark is 

speaking about and I am agreeing with.  (Mark, I am sure you will feel 

free to correct me.)  It may be that in trying to mitigate all problems 

with multiple, specific interventions/programs we actually decrease the 

potential for the positive impact of any of them.  Giving direction for 

the most effective mix of interventions with which families is one of 

the jobs of research.  The good news is that we seem to be learning more
about this.  It may be that by actually doing one or two things that 

effectively attack a limited set of problems, we will have a more 

positive impact not only on the targeted problems but also on many other
problems compared to trying to do a lot of things all at once. The 

burden of multiple interventions imposed on families (and workers) may 

reduce their individual effectiveness, perhaps because they are never 

properly delivered.  There also may be negative interactions between 

interventions that inherently reduce their impact even when families 

participate in everything assigned.  These are all hypotheses to be 

tested.  However, the evidence does seem to be mounting that there is 

not simply a linear, additive effect to interventions so that more is 

better.  A primary challenge for researchers right now probably is not 

to develop and test yet more interventions and programs.  Rather, we 

might want to examine the "reach" or generalization effects to 

nontargeted problems of some of our more effective existing 

interventions and the interactions positive or negative, between them.  

Mark's work with PCIT is a great example of that sort of work.  Without 

this sort of research to base their work on, I think you are right, 

frontline workers will continue to see awful situations and think trying
to give the family everything is the way to go.  Without reliable 

information to the contrary, why would they not?  As you say, it feels 

right to do so and the stakes often are very high.

 

I also agree that any system that does not provide effective training, 

ongoing supervision, and ready access to useful information and practice
tools for its front-line workforce is unlikely to produce good work, and
the child welfare and mental health systems are no different.  Our 

challenge as researchers is to figure out better ways of doing that 

effectively and efficiently.  The emerging dissemination and 

implementation research literature can be of help here, as well as the 

substantial dissemination work going on around the country.  I am 

optimistic.  In our statewide EBT dissemination and implementation 

project (Project BEST) we find that the practice wisdom of frontline 

caseworkers is not contrary to the existing research, as my training 

example in the previous post illustrated.  Caseworkers have a pretty 

good idea which programs work and which do not.  They lament that 

ineffective programs are often part of the "standard" treatment plan 

institutionalized on the agency computer system and that effective, 

evidence-based programs often are not available in their community.  

They know treatment plans with multiple interventions often fail even 

with highly motivated families because they are unrealistic for the 

context.  After relatively brief training, they are excited about the 

notion of evidence-based treatment planning, and for the most part have 

a good feel for titrating interventions and capitalizing on the 

strengths of families in treatment plans.  With ongoing consultation, 

practice tools, and administrative support, they are able to put it in 

practice.  Our state is rethinking its traditional approach to new 

worker training as a result.  However, more research is needed to better
understand optimal approaches to training, implementation and 

sustainability.  How best to get what is known incorporated into the day
to day work of the frontline worker is yet another challenge for 

researchers.  And this is a tough one.

 

Ben

 

Sandra Azar wrote:

> Mark and Ben,

> 

> I am hesitant to get into the dialogue of two such titans in the
field, but

> am struck as I always am about the missing elements in our treatment

> discussions. Both of you make incredibly important points and I too
believe

> deeply in science guiding our planning of treatment. But this
discussion

> leaves out two important elements in child protection work -- that I
believe

> we often we as researchers ignore -- WHO is actually doing the work
and the

> CONTEXT in which it is done.  Coupled with the risk of a child death,
both

> may explain the desire to add and add and add.

> 

> Who - Most parent home visitors and caseworkers are not trained to do
the

> kind of work we want them to do.  Indeed, I have seen home visitors
who are

> bank tellers or human resource workers or ones who did housing work.
None

> with prior training more than a workshop here and there. So we can
develop

> whatever intervention we like --we are still expecting someone to
carry it

> out. In many contract agency settings, supervision is minimal and

> backgrounds of staff do not match the requirements of our protocols.

> 

> Where - In urban settings when I have visited homes to conduct
research

> interviews or in the past to do home interventions, I often have gone
to

> neighborhoods where a shooting has taken place just a week ago
(setting off

> a sense of vigilance even though I am seasoned in going to such

> neighborhoods). I might have to step over a drunken neighbor who is
passed

> out on the stoop or the half drunk one, who insists on having a
conversation

> with me before I can get past them to get into the home. I then enter
homes

> where at times I am treated to a shouting argument between a man and a
woman

> that at moment feels like it will erupt into a physical fight.  Other
men

> enter the home during the visit and stare rather belligerently at me
or join

> in taking the boyfriend's or husband's side.  The woman may have no
food and

> the child appears sick. She is crying.  This is the context where the
work

> is being done. Cognitive science tells us that cognition is disrupted
under

> such conditions.

> 

> Put a frightened young worker and this context together and add high
risk

> for a child and the worker is overwhelmed and wants all the help she
can get

> and probably is shaking too much to carry out anything in any
standardized

> way. It FEELS like more is better -- because the situation has so many

> elements. She may habituate to this over time and with good
supervision grow

> into a decent practitioner who can carry out focused approaches. More
likely

> if she is good, she will leave the work for a job in a setting where
she can

> feel safe and carry out her work more effectively.

> 

> Yes -- in a the context of a workshop and under ideal conditions with
the

> right coach, caseworkers can come up with a good treatment plan that
is

> focused and may be enough to produce results -- but that is with a
coach

> like one of you and the safety of sitting in a room with nothing to
lose and

> no threats hanging over them.

> 

> Treatment development work in child maltreatment is unlike other
treatment

> development areas where most interventions are carried out by
professionals

> who have much education and in the context of an office setting. I
want the

> field to develop science based protocols, but until we deal with
staffing

> issues and focus our attention on methods for coping for staff -- the
"add

> and add and add" approach will continue to feel safer for most.

> 

> Sandra

> 

> ------------

> Sandra T. Azar, Ph.D.

> Professor

> Psychology Department, Moore 541

> Pennsylvania State University

> University Park, PA l6801

> 814-863-6019 (office)

> sta10@psu.edu

> 

> -----Original Message-----

> From: bounce-3424753-6833833@list.cornell.edu

> [mailto:bounce-3424753-6833833@list.cornell.edu] On Behalf Of Saunders
PhD,

> Benjamin E

> Sent: Friday, December 26, 2008 11:44 PM

> To: Child Maltreatment Researchers

> Subject: RE: Evidence-based and "Mix and Match" Programs

> 

> Mark,

> Thank you for the excellent summary of several very important issues.
Two

> points are particularly critical for future research.  IMHO, the
current

> excitement in some quarters over "components" approaches to treatment
vs.

> manualized "protocols" is, from a research perspective, a red herring.
As

> you note, both must have some sort of decision rules about what to do
next

> at certain points in treatment.  If not, they just become virtually
random

> in nature.  So, the results of those decisions will need to be tested

> empirically whether it means following a "protocol" or decision rules
about

> using components.  Frankly, when one scratches the surface, the two

> approaches sound suspiciously similar.

> 

> Some have suggested that components approaches are more efficient
because

> they use only the "active ingredient" components of protocols at key
points

> in the treatment and skip the unnecessary stuff.  Unfortunately there
is

> precious little dismantling research discerning exactly what those
active

> ingredients are, and whether or not they only get active when the
other

> "unnecessary" components have been used as well (what one might call

> conditional component efficacy).  It may turn out that the components
that

> have been picked to be used actually do have the most impact even when
not

> used in concert with other techniques.  Or not.  This hypothesis
remains to

> be tested for most approaches.

> 

> Others have suggested that "components" approaches are more palpable
to

> clinicians because then they can use their clinical judgment when to
do what

> rather than following the strict rules of a protocol.  However, as you
point

> out, many of the components approaches then proceed to teach elaborate
rules

> for when and how to use the particular components chosen and end up
being

> more complicated than protocols (but without the outcome research to
support

> their efficacy).  While the whole components vs. protocols debate is
an

> interesting pastime for some of us, from an empirical testing
standpoint, it

> may be a debate without a difference.  The empirical question still
is,

> "When therapists do this, do clients get better compared to when
therapists

> do that?"  Call it what you will, components, protocols, or whatever,
you

> still have to define the "this" (aka independent variable) in
sufficiently

> replicable manner.

> 

> You second point about "focused" vs. "comprehensive" treatment
planning also

> is absolutely critical.  At a recent training for about 50 CPS
workers, I

> gave them all a case we had seen recently in our clinic and asked them
to

> break into groups and come up with a treatment plan.  the case was a
typical

> train wreck, multiproblem, abusive family.  We then wrote on a flip
chart

> all of the interventions, treatments, programs and meetings they
thought the

> family should receive and go to.  It took 4 large flip chart sheets to
write

> them all down.  I then asked the workers two questions.  First, did
they

> think any family in the world, much less this family, could get their
child

> to even half of the appointments they were recommending.  Second, did
anyone

> in the room believe this was an effective treatment plan that would

> accomplish the goals we had set for this family.  No one in the room

> believed any family (even their own) could accomplish half of the
treatment

> plan, and not one person thought it was an effective treatment plan.
Yet

> they wrote it.  They agreed that they have been trained and
acculturated to

> simply add and add and add and add to treatment plans to the point of
being

> ridiculous.  The good news is that at the end of the day, when
challenged to

> come up with a feasible treatment plan composed of evidence supported

> interventions and programs, they were able to do it and the plan was
about

> 1/3 of a flip chart page.

> 

> The notions that more is better and that doing an untested something
is

> always better than doing nothing permeates the system and needs to be

> challenged.

> 

> Again, thanks for the elegant thoughts.

> 

> Ben

> 

> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> Benjamin E. Saunders, Ph.D.

> National Crime Victims Research and Treatment Center

> Medical University of South Carolina              843-792-2945  Phone

> Charleston, SC  29425                                 843-792-7146
Fax

> 

> Visit our web sites:   www.musc.edu/ncvc       www.musc.edu/tfcbt

>                                www.musc.edu/ctg
www.musc.edu/saprevention

> ________________________________________

> From: bounce-3422423-6832002@list.cornell.edu

> [bounce-3422423-6832002@list.cornell.edu] On Behalf Of Chaffin, Mark
J.

> (HSC) [Mark-Chaffin@ouhsc.edu]

> Sent: Wednesday, December 24, 2008 8:31 PM

> To: Child Maltreatment Researchers

> Subject: RE: Evidence-based and "Mix and Match" Programs

> 

> Tom,

> 

> Yaiiii....where to start.  There is considerable interest in the idea
of

> extracting common elements from across evidence based models, then
applying

> these depending on assessed case characteristics and some systematic

> algorithm.  Probably the most detailed system of how this process
might be

> undertaken has been described by Chorpita and colleagues.  Note,
however,

> that the processes for identifying both the elements themselves and
the

> matching algorithm as described by Chorpita are NOT just a matter of

> logic-model 'mix and match' eyeballing based on the clinician's gut or

> personal preferences.  It is a quite structured and quantitative
process.

> It is still a protocol (and a complicated one at that), and not at all
the

> same thing as free-styling, fly-by-the-seat-of-the-pants services
often

> advocated by the anti-EBT crowd.  If anything, it is a system probably

> requiring more expertise and training than most simple EBT's.  How
well

> these algorithm-driven, elements-based systems work in practice is a
subject

> of ongoing study.  The jury is not in.  Some EBT's such as MST, have
done

> essentially this same thing for years and its worked out well.  MST,
for

> example, is not a single protocol, but is assessment driven and
tailored,

> yet focuses the intervention on what we know matters, using basic
elements

> that are known to work  In this sense, there are fairly complex EBT's,
such

> as MST, which target broad populations, and more specific EBT's, such
as

> TF-CBT for example, which are specific for particular types of
well-defined

> problems (i.e. PTSD).

> 

> The elements-based idea does have a clear appeal--EBT's are probably
not

> monolithic entities incapable of being subdivided.  Most share common

> elements within a given domain (e.g. most evidence-based parenting
programs

> share many common elements such as use of labeled praise, application
of

> structured time-out protocols, etc.).  However, I would be very
skeptical of

> efforts to use this rationale to neuter EBT's, or to simply say "oh,
we're

> already doing all that" or "this is the same thing" when it really
isn't the

> case.  One need look no further than the Blueprints implementations to
find

> evidence of this, where well-intended shooting from the hip led to
blending

> in ad hoc crap with EBT's and spoiled the results.  The point is that
the

> elements-based approach is not an excuse to allow anything and
everything to

> come into the intervention or to mix-and-match without some fairly
tight

> limits, such as those described by Chorpita and colleagues.

> 

> The area of most concern in your question lies in the idea of whether
more

> is better.  There is excellent evidence at this point that this is not
only

> false but that more can become harmful.  For example, where parenting

> interventions are concerned, it appears that adding additional
services to a

> parenting program actually poisons the benefits (Kaminsky, et al.

> meta-analysis).  CPS and courts are notorious for this
misconception--often

> prescribing so many services that whatever benefits any of the
services

> might have offered may be quickly lost in the confusion.  The analogy
to

> "polypharmacy" in psychiatry is not a bad one.  Focus (i.e. directly

> behaviorally targeting the top priority that needs to change) rather
than

> comprehensiveness (trying to fix everything), is the new watchword in
many

> service systems.   Exactly how much is too much, and what demands the

> highest priority is an unanswered empirical question.  But the
emerging

> science in this area does, IMO, suggest a couple of general
principles.

> First, tailoring might work...IF the elements are selected carefully
based

> on scientific evidence and are clear essential common components
across

> EBT's.  Second, there are clearly both practical and therapeutic
limits to

> how many things can be done well at once, and this point is reached
rapidly

> and past that point quickly begins to ruin the overall service
benefit.  So,

> "focus" not "comprehensive" should be the watchword.  We need to
emphasize

> this watchword because practitioners have been so imbued with the idea
that

> "comprehensive" is necessary that it takes considerable effort to
disabuse

> them of this unfortunate misconception.  This is another reason why
any

> novel algorithm-driven, elements-based protocols need to be
structured.  And

> why it needs to be rigorously evaluated.  Keep in mind that
implementations

> of novel blended or elements-based programs cannot properly be called

> evidence-based just because the sources for the elements were

> evidence-based. For that matter, it cannot even be presumed to be
effective,

> altough we might predict that it would be.

> 

> Mark

> 

> 

> ________________________________________

> From: Tom Hanna [tph3@cornell.edu]

> Sent: Thursday, December 18, 2008 8:25 AM

> Subject: Evidence-based and "Mix and Match" Programs

> 

> On another list, there is active discussion underway on starting an

> "ancillary" parenting education program to an existing "core" home

> visitation program.  The conversation quickly turned to the topic of

> "evidence-based", and then to funders and their requirements.

> 

> The picture quickly got cloudy for me:

> 

> 1.  Some folks who already have an ancillary parenting education

> program reported that "blending" aspects of two evidence-based

> programs allowed them to tailor the trainings to the specific needs

> of their "home visited" parents.  Others quickly pointed out that

> this is "wrong" and should not be done -- neither evidence based

> program is being followed precisely, and therefore both are

> "contaminated."  Funders frown.

> 

> 2. No one has said what additive effect, if any, is expected from

> providing a parenting program on top of a home visitation program.

> The underlying assumption is that families will be better off with

> two distinct interventions instead of one. (In fact, many centers in

> this home visiting network have many ancillary programs that serve

> some if not many of their home visited families.)

> 

> 3. I know that lots of funders are demanding that agencies use

> "evidence-based" programs.  But I now learn that funders are pushing

> implementation of a "matrix" of "evidence-based" programs.  The

> underlying assumption is that "if one evidence-based program is good

> for families, then many are better."

> 

> My question: Is there any research that helps multi-service agencies

> make their way through this minefield when working with a cohort of

> families?

> -- Any study of the "deterioration of effects" of the blending of two

> evidence based models for the same intervention?

> -- Any classical studies of "additive effects" of multiple targeted

> interventions?

> -- Any evidence that a "matrix" of evidence based programs has a

> stronger effect than a "pure" one-program approach?

> -- Any analysis that shows that evidence based programs in different

> interventions (home visitation vs parenting ed vs therapy groups) are

> (or are not) internally consistent?  (ie, my doctor gave me one

> instruction about diet, my nutritionist gave me a contradictory

> instruction, and my home visitor's instruction differed from the

> other two.)

> 

> TIA

> Tom

> 

> 

> 

> 

> --

> --

> Tom Hanna, Director

> Child Abuse Prevention Network

> www.child-abuse.com

> tom@child-abuse.com

> tph3@cornell.edu

> off 607.275.9360

> cel 607.227.4524

> fax: 415.962.0510

> --

> 

> 

> 

> 

>   

 

-- 

Benjamin E. Saunders, Ph.D.

National Crime Victims Research and Treatment Center

Medical University of South Carolina

165 Cannon Street, MSC852         843-792-2945 Phone

Charleston, SC  29425             843-792-7146 Fax

 

Visit our web site:         www.musc.edu/ncvc

Take our web-based courses: www.musc.edu/tfcbt   www.musc.edu/ctg

 

 

 

 

No virus found in this incoming message.

Checked by AVG - http://www.avg.com 

Version: 8.0.176 / Virus Database: 270.10.1/1868 - Release Date:
12/29/2008 10:48 AM


No virus found in this incoming message.
Checked by AVG - http://www.avg.com 
Version: 8.0.176 / Virus Database: 270.10.2/1873 - Release Date: 1/3/2009
2:14 PM