1
WEEK 4: Multidisciplinary mental health care
Lecture 7 (video lecture): Multidisciplinary mental health care research
Unidimensional ontology
• Multidisciplinary health care: to understand mental health care from a MDO point of view we have to look at
the organization of mental health care, but also take a closer look at how we conduct psychiatric research.
At the moment most of the research is unidimensional, which leads to a unidimensional approach in clinical
practice.
• Ontology = theory of being → our worldview, how the world is organized, what is truth
• Epistemology = theory of knowledge → how we, as human beings, can understand the world
• In Medical Model in Psychiatry
o Ontological premise (assumption) is one universal (predominantly physical) world that behaves
regularly
o Epistemic premise is that we can explain mental disorders in a systematic way (disease model)
▪ Research: observe symptoms systematically, develop treatment protocols infer causal
impact of treatment protocol on presence disorder
▪ Treatment: protocolled treatment (well-defined intervention with a lot of detailed steps,
regardless of which therapists gives the treatment), dose-response, stepped care =
“Symptom-reduction model” (Van Os et al., 2019) → the core of evidence based treatment.
Treatments are distributed if there is empirical support for their efficacy in reducing
symptoms for the majority of people with these symptoms. This guides the way we do
research.
o This approach of treatment as a protocol gives the assumption that a treatment as such, as a tool,
has working effects. Medical model expects a dose response relationships with medications, and
this is also expected in psychiatry and psychological treatments. Different types of severity of the
same disorder receive different amounts of treatment sessions (stepped-care model). If the amount
of sessions was not enough, sessions can be added. This goes hand in hand with the ‘lijnen’
system in the Netherlands (eerstelijns (BGGZ), tweedelijns (SGGZ), derdelijns (klinische opname)
Evidence in psychotherapy research
• Evidence-based treatment: methodological procedure as warrant
for evidence → see pyramid
• Everything below the RCT is done, but it’s not taken as seriously
• Randomized Controlled Trials on top of ‘evidence hierarchies’
(e.g., Cartwright, 2007)
• The golden standard
o RCT: Systematic comparison of group level symptom
development pre-post treatment. Ideally a waiting list
group versus a treatment group, but this is not always
seen as ethical. Instead of waiting list group treatment as
usual or another form of treatment. Comparing symptom
reduction in both groups
o Methodological requirements:
▪ Comparable populations → homogeneous samples (otherwise you do not know if
differences are due to the intervention or something else. To prevent systematic
differences people are randomly assigned to a group)
▪ Well-defined outcome variable (dependent variable) → eligibility criteria: presence of target
symptomatology, no comorbidity (because this could confound with the outcome variable)
▪ Systematic and well-defined intervention (independent variable) → protocolled treatment
▪ Keep all possibly interfering factors as even over samples as possible → random allocation
to interventions
▪ Keep expectancy effects limited → placebo control; but in reality often comparative trials
with TAU or another EBT due to ethical downsides
Methodology shapes finding
• Methodological designs and the associated with research influence what we can find when conducting
research
• Assumption of malleability (to change): “The longer the therapy, the more variability, the less one can
draw causal conclusions. As we argue, the preference for brief treatments is a natural consequence of
, 2
efforts to standardize treatments to bring them under experimental control” (Westen et al., 2004, p. 633) →
this makes it more difficult to state causal conclusions, so research designs and treatment protocols
become shorter. We do not have proof that short intervention are better than long interventions, and this is
difficult to research because of the methodological requirements of research
• Paradox of pure samples: ‘Interventions to adress such [polysymptomatic/comorbid conditions] will, as
simple result of methodological preconditions, never be identified if investigators routinely start with less
complex cases and focus studies in the community on interventions previously validated in RCTs’ (Westen
et al., 2004, p. 637) → our methodology does not allow us to research this. This is problematic because it’s
something we encounter during clinical practice and it’s not studied
• The paradox of manualization: we protocol treatment so we can clearly define our independent variable
(the intervention). It’s necessary to specify your treatment, so you know that what you do is effective and
draw causal conclusions about this. Manualization is a very important precondition of this RCT/causal
intervention model. A consequence is that we only find evidence for manualized treatments. Non-
manualized treatments are not researchable within our research paradigm. This led to the assumption that
manualized treatments are better than non-manualized treatments, but we can’t conclude this because the
comparison is not made. Studies that did research this, found no differences or differences in favor of non-
manualized care because it’s more person-centered.
Ontology & Epistemology: Erklaren
• The discourse of Erklären – Explanation:
o Ontology: there is a real, universal, regular (lawlike), outside world…
o Epistemology: … that we can explain by deducing universal patterns of phenomena
o Goal: explain phenomena causally
o Method: experimental, deductive, top-down, measurement & calculation
• Measurement is a very systematic and rational way to compare phenomena.
The measurable human being
• Measurement as prerequisite for gold standard. It’s central to our research design
• Systematic assessment of group level symptom development pre-post treatment
o Systematic assessment = use of standardized symptom-focused measures, mostly self-reports
o Symptom development = standardized measurement of target symptoms pre- and post-treatment
o Pre-to-post-treatment = before the start and directly after treatment (ideally: follow-up)
o Group level: aggregated data over sample of people
• “The data” will tell the story of treatment efficacy
Measurement in psychology: self-report questionnaires
• We quantify content of statements or questions
• We put this data together and start calculating
Efficacy number as outcome of aggregation & analyses
• →
• This set-up is based on group-level symptom reduction
model. We look into the mean effect for a whole group
of people. We don’t know the scope on which this
mean is based, which can be clinically problematic. We
do not now for who it works, for who it can be aversive,
what we can do when its aversive. The group level model is good for a general evidence based treatment,
but its clinically not very helpful.
• We also limit our findings to pre- and posttreatment level. But this is basically snapshot research. It can
make quite a difference at what moment you measure this. Alternatives for snapshot (unidimensional
research:
o Beyond singular/linear outcome variables:
▪ Domain based → e.g., Research Domain Criteria (dimensional, multi-level developmental.
The basic assumption is that several domains work together when talking about mental
health. Critique: in practice the emphasis is on the two lower domains which are more
physical aspects). It enforces multidisciplinary research:
• Interdisciplinary research → not just a psychiatric framework, but a collaboration
between social work, sociologists, anthropologists, neurologists, policy makers..
• “No practice of inquiry is ruled out a priori; multiple goals of inquiry are rendered
plausible and multiple methodological pathways may claim a situated legitimacy.”
, 3
• Advantage of RDoC: Multidimensional content (=ontology) asks for a pluralist
(different perspectives coming together) research approach & mixed methodology
(=epistemology)
• The topic dictates the research method instead of the other way around (with
RCTs being the golden standard)
▪ Complex Dynamic Systems → e.g., network theory in diagnostics. In the DSM-V we speak
of isolated disorders and comorbidity between them. However, is this the reality?
Symptoms and disorders seem to interact with each other, which means they are
interrelated.
o Beyond gold standard research design
▪ Efficiency versus efficacy designs (less strict inclusion, more process focus, multiple
variables, downside: less causal, upside: way more close to actual clinical practice)
▪ Process research → e.g., Routine Outcome Measurement (Boswell et al., 2013)
• In pre-post research designs with ROMs we assume linearity, which is almost
never not the case (scores go down and up). This is clinically very relevant,
because what is happening with the person? This leads to that we should not just
measure before and after treatment, but also in between.
▪ Clinically significant change → Reliable Change Index (Jacobson & Truax, 1991) →
differentiate between the groups below
• Recovery
• Improvement
• No change
• Deterioration
▪ Individual differences research → e.g., within-group Individual differences
▪ Personalized measurement (Elliott et al., 2016)
o Beyond measurement: ‘numbers are words too’ (Truijens, 2019)
▪ Mixed methodology approach
▪ Qualitative stance (Hesse-Biber, 2010)
Voices in data
• How participants can speak in self-report measures → pre-structured, quantified & aggregated
o What is participant allowed to say?
o Do participants understand instructions and quantification as we intended?
• Who is ‘the average patient’?
o The statistical mean is not a person
o A priori → can we learn something not-anticipated?
• Voice = noise?
o Noise = clinical process! Clinical practice is messy
o ‘Numbers are words too’ (Truijens, 2019; Stiles, 2006)
Ontology & Episemology: Verstehen
• The discourse of Verstehen – Understanding:
o Ontology: people make meaning of phenomena in their lived experience…
o Epistemology: … that want to understand by listening, thematically or discursively analysing, and
synthesizing
