Saturday, 13 September 2014

Classification in Psychiatry presentation

Classification in Psychiatry - Presentation

Classification In Psychiatry

Categorisation II
Describe what you have just seen
Probably used words like “goal” “football”   maybe even “Scotland”  or for the more advanced “good triumphing over evil”
We experience raw streams of data that our brain subdivides into categories and concepts to help us make sense of this 
We think abstractly using the same method of concepts/ categories
This happens at basic animal levels of consciousness – “do we mate with it/ eat it/avoid or go round it/run away from it”?

Categorisation III
Doctors often see patients in time critical situations e.g. brief appointments
Other stressors could be e.g. seeing patient at 3am, being aware of multiple patients to see after this one, caseloads of 100s
Often have to make decisions about patient such as admit/not admit, what treatment to start or change etc
Need a robust system to help them make good decisions under alll kinds of pressure

A,B, C, D Model of Patient attributes
A - What patient shares with all other patients
B - What patient shares with some other patients
C – What is unique to this patient
{D- What patient shares with everybody else}
If concentrate therapeutic efforts solely on A, offer same treatment to everyone
If concentrate therapeutic efforts solely on C, can’t extrapolate from this patient how to treat other patients
If concentrate purely on B lose sight of the person
Must make therapeutic efforts on A,B,C and D

Why Do We Need Classification?
Allows communication between professionals
Allows groups to be defined for research to gain information on
Range of likely outcomes (prognosis)
Response to treatment
Important correlates such as bio psycho social features
Causes of the classified disorder (and explanation)
Statistics for administration, management and budgeting
Eligibility for care both from services but also family/society
Facilitate appropriate payments for care
Reduces potential for abuse e.g. of psychiatric diagnosis

The Status of a Diagnosis
Diagnosis is an abstract concept used to summarise complex information about a patient’s health state
It doesn’t exist as a “thing”
It is ALWAYS provisional
People may be given a diagnosis but they are NOT the diagnosis
It can have powerful effects on people’s lives. Not all intentional or harmless

Essentialism & Carving Turkeys
“Essentialism” suggests that for example illnesses can be separated into “natural kinds”
These separate entities are distinguished by points of rarity and the categories have validity
(Pneumonia is different from myocardial infarction, and can be further subdivided into a variety of aetiological causes)
So diagnosis is supposed “to carve nature at the joints”
BUT nature is not a giant multi-limbed turkey
Mental disorder particularly hard to separate from each other

Promiscuous Reality and Elephants

What is Promiscuous Reality?

Multiple viewpoints can describe a “piece” of the whole truth
Putting all these viewpoints together can increase your knowledge of the whole though maybe not even the totality of whole truth
Different viewpoints can have different purposes
Multiple “models” of mental of disorder – biological, social, cognitive, psychodynamic and so forth
These can all be useful in understanding and guiding help for patients
They do not necessarily exclude each other
Pragmatic realists recognise different classifications can be used depending on their Utility (usefulnesss)

Hepatic Lobule: multiple views
Can look at anatomical view : hexagonal structure
Can look at biliary flow point of view : centred on bile duct drainage area

All views valid and can choose which view suits your purpose.

What is Validity? I
Construct validity. Construct validity refers to the extent to which a particular measure relates to other measures consistent with theoretically derived hypotheses
Content validity refers to the extent to which a measure represents all facets of a given social construct (non-statistical expert opinion)
Convergent validity refers to the degree to which a measure is correlated with other measures that it is theoretically predicted to correlate with e.g. intelligence test correlate with e.g. exam results

What is Validity? II
Discriminant validity tests whether concepts or measurements that are supposed to be unrelated are, in fact, unrelated i.e. separates “natural kinds”
Predictive validity measures how well the construct can predict future events ie. Prognosis  as well as treatment response
Criterion validity. Criterion validity is measuring something that is external to the measuring instrument itself, called the criterion, that is representative of the construct. In medicine we often use biomarkers  e.g. BP, blood tests, X –rays etc

What is Validity? III
Concurrent validity refers to the degree to which the construct description/ measures correlates with other measures of the same construct that are measured at the same time
Face validity is a non-statistical  estimate of whether a test appears to measure a certain criterion; it does not guarantee that the test actually measures phenomena in that domain.
Procedural validity. Procedural validity refers to the adequacy of a new diagnostic procedure in replacing or simulating some existing procedure e.g. use of standardised interviews to detect mental disorder over non-standardised. Not validity of construct itself.
Validity of Psychiatric Disorders (Robbins & Guze 1970)
1) Clinical Description (including symptom profiles, demographic characteristics, and typical precipitants),
2) Laboratory studies (including psychological tests, radiology and post mortem findings) extended to other external validators e.g. molecular biology, cognitive neuroscience
3) Delimitation from other disorders (by means of exclusion criteria)
4) Follow-up studies (including evidence of diagnostic stability), 
5) Family studies showing familial aggregation, specificity for same or related disorders

Validity of Psychiatric Disorders (Kendler 1980)
Antecedent validators 
familial aggregation
premorbid personality
precipitating factors 
Concurrent validators 
psychological tests (?external criterion?)
 Predictive validators 
diagnostic consistency over time
rates of relapse and recovery
response to treatment

Lumpers v Splitters
Lumpers emphasise lack of distinction between separate categories and high degree of comorbidity
Splitters emphasise differences between members of bigger classes because of differences in treatment response / prognosis

Problems of classifying Mental Disorders
Brain is most complex organ in body by far
Functioning of brain poorly understood : what we do know is still difficult to understand how it translates into “outputs” such as thoughts, feelings, memories 
(Physiology and pathophysiology poorly understood)
“Outputs” of brain extraordinarily complex and affected by external factors such as culture and very hard to objectively measure
Given complexity of outputs, the role of external factors such as culture, lack of knowledge of functioning of brain, difficulty objectively measuring outputs an aetiological classification is very difficult 

“Without Reliability there can be NO Validity (OR UTILITY)”
If a diagnosis can’t be reliably agreed upon it has low utility
Inter-Rater Reliability in terms of diagnosis can be defined as the degree to which different observers can agree on the presence of a diagnosis in the same patient
Often measured using the Cohen’s kappa concordance statistic : the degree of agreement between 2 raters, greater than chance
Also can be “weighted” for multiple categories, taking into account how close categories are together for agreement (e.g. yellow/orange more disagreement than red/violet)

Measuring Reliability
Kappa Agreement
< 0 Less than chance agreement
0.01–0.20 Slight agreement
0.21– 0.40 Fair agreement
0.41–0.60 Moderate agreement
0.61–0.80 Substantial agreement
0.81–0.99 Almost perfect agreement
(DSM 5 used much lower thresholds to justify diagnostic categories as having good reliability)

Inter-Rater reliability of ICD10 Diagnosis
For schizophrenia is substantial 0.79 – 81 but less so for subtypes 0.40- 0.74
For Personality disorder subtypes only slight to fair 0.07-0.43

Improving Inter-Rater Reliability of Diagnostic Agreement
Standardising how and where patients are assessed
Standardising what questions are asked
Standardising training of raters
Standardising how answers and observations are interpreted e.g. definitions of symptoms
Standardising diagnostic definitions
Standardising the rules how diagnosis are made e.g. in terms of co-morbid symptoms, co-morbid diagnosis, co-morbid aetiological factors e.g. substances
NB Test-retest reliability 

Many medical disorders exist on a spectrum which includes normality, the extremes of which clearly represent illness but with a "contentious" zone at more central values. An "expert committee" is usually formed which decides the best compromise for deciding a threshold. This threshold is often decided on the basis of associated poor outcomes such as death and other disease events.  Sometimes there is a threshold at only one end of the spectrum but often at both ends of a spectrum too

Examples include blood pressure, blood glucose, bronchial reactivity, depression/anxiety symptoms 

Within medical disorders often there can be mixed spectrum of clinical features and people can be put into separate diagnostic categories depending on which clinical features predominate . (Examples include mixes of bronchitis and emphysema, or psychosis and mania).

Utility of Colours
Number of colours varies with culture e.g. Newton said 6, some cultures recognise only 2, others up to 7.
Colours are cultural constructs separating spectrum into categories
However people from all cultures are equally able to recognise difference in hues/shades 
Recognising differences in colours has usefulness : in natural and artificial world. 
PANTONE is a system of allocating codes to different shades so that manufacturers, printers, artists etc can communicate with each other and be consistent when using colours

Utility without Validity : SCZ and Mania

Can reliably separate: ICD 10 kappa for schizophrenia 0.79-0.81 and for bipolar, manic disorders to be 0.69-0.77
Has predictive value for outcome: AESOP study if presented with Mania had 3-4 times better chance of symptom & functional recovery at 3 and 10 years than presenting with Schizophrenia
Has predictive value for treatment: Lithium effective in bipolar mania (and bipolar, depression) but not schizophrenia. Antipsychotics can be effective for  both bipolar mania and schizophrenia but not bipolar, depression (quetiapine low dose may be effective for bipolar dep)
Bipolar on average less cognitive impairment than SCZ
More likely to have same type of disorder inherited (not purely tho)

Diagnosis 0.0
These are problems that may be the focus of clinical attention or may be referred to clinicians for help
Not usually “illnesses” per se with a “disease” cause but often mixed groups and very culturally based
Often need a diagnostic “code” to enable someone to access help and for the clinician to be remunerated (or allowed to help)
Maybe contentious as some do not feel should be “medical”
Examples include relationship problems, Oppositional Defiance Disorder,  anger management issues, ?paraphillias, some plastic surgery problems

Diagnosis 1.0
Descriptions of clinical states that are problematic and widely thought to need medical attention
Often describe  syndromes i.e. collections of clinical features occur together , often little validity to the distinctions
Not aetiological i.e. not based on cause of the disorder so members of this group often have different causes (known or unknown)
As not aetiological not able to “explain” and treatment responses /outcomes highly variable, lack of lab tests; cultural limitations
Many psychiatric disorders e.g. depression, schizophrenia, ADHD, also “functional” medical disorders and e.g. “fever”

Diagnosis 2.0
Fully aetiological basis known and external validating criteria such as lab tests to confirm diagnosis and monitor progress
More able to sort in to truly valid categories
As aetiological basis known, better able to predict outcomes and treatment responses, develop new treatments
Explanation of disorder known : greater patient information and public health interventions possible
Many medical conditions but also e.g. alzheimer’s disease, wernicke-korsakoff, steroid induced depression

Stigma and other drawbacks 
By concentrating on the diagnosis can lose whole picture especially if “overforce” into ill matching category
Stigma: people often get associated with their diagnosis e.g. epileptic, diabetic
This can lead to self-stigma, stigma from other people, and from healthcare professionals based on stereotypes of that illness
This stigma can be even more distressing than the disorder itself
Reification “over estimate” the degree of “predictive knowledge” and “validity” of the diagnostic construct e.g. SCZ is an illness not a disease (as pathology unknown in most cases)

Comparison with General Medical Diagnosis  Szasz fallacy
Disease can only mean something people "have," while behavior is what people "do". Diseases are "malfunctions of the human body, of the heart, the liver, the kidney, the brain" while "no behavior or misbehavior is a disease or can be a disease. That's not what diseases are.”
Says mental illnesses are metaphors
Ignores the conceptual basis of general medicine
Assumes all general medical problems have identified pathology
Ignores value judgements made even in “objective” lab tests
Assumes perfect knowledge of pathology and physiology

Diagnosis in Medicine – Problems
60% of MDDUS negligence claims v GPs  are due to failure to diagnose
Diagnostic errors are the most common source of the medical malpractice payments, the most costly and the most dangerous when it came to patients’ health
A systematic review of studies of the autopsy calculated that in about 25% of autopsies a major diagnostic error will be revealed. 
This rate has decreased over time and the study projects that in a contemporary US institution, 8.4% to 24.4% of autopsies will detect major diagnostic errors.

Causes of Diagnostic Errors 1
Anchoring bias (COMMONEST) – locking on to a diagnosis too soon and failing to adjust to new information.
Availability bias – thinking that a similar recent presentation you have encountered is also happening in the present situation.
Confirmation bias – looking for evidence to support a pre-conceived opinion or notion, rather than looking for information to prove oneself wrong.
Diagnosis momentum – accepting a previous diagnosis without applying sufficient scepticism.

Causes of Diagnostic Errors 2
Overconfidence bias – Over-reliance on one’s own ability, intuition and judgement.
Premature closure – similar to confirmation bias, but more like jumping to a conclusion.
Search-satisfying bias – A “eureka” moment that stops all further thought on the matter
I would also add 
treating the test result not the patient
“over-forcing” patient into a diagnostic category they don’t fit well into

Comparison with Medical Diagnosis Reliability
“…most indexes of agreement between pathologists ranged from poor (needle biopsy 0.21) to moderate (necropsy/surgery 0.57).” 
160 cases reviewed  3 pairs of neurologists “with a special interest in stroke.”8 The kappa was only “fair to good” in most categories of ischemic stroke, with a high of 0.70 for oral contraceptive-related stroke and a low of 0.28 for lacunar infarcts; the average kappa for all categories was 0.53 
Knee osteoarthritis kappa = 0.1 renal artery stenosis kappa =0.43
that interrater reliability is no higher in many nonpsychiatric medical specialties than in psychiatry 

Comparison with Medical Diagnosis Validity
Many medical problems have unknown pathology and would fail to meet “validity” tests
Headaches, migraines, CFS/ME, Fibromyalgia Irritable Bowel Syndrome
Up to 40% of referrals to different medical OPs have these “functional disorders”
May medical problems have “cultural” overlays in presentation e.g. myocardial infarction

  • Even when pathology known and "validity gap", often general medical treatments have similar effectiveness to psychiatric treatments

Comparison with Medical Diagnosis  Spectrum
Many accepted medical disorders lie on a spectrum with “normality”
Examples include hypertension, non insulin dependent diabetes, asthma
Many medical disorders don’t have clear cut boundaries with similar disorders
Example: bronchitis/emphysema, unstable angina/myocardial infarction

History of Classification – Early Days
Initially mental disorders thought to be somatic in origin
Gradually introduction of ideas of “mental” disorders with varying degree of somatic input 
Every country, every hospital sometimes even every psychiatrist had a separate diagnostic system
Research papers often had to include lengthy descriptions of each patient so that readers could compare to own patients e.g. Aubrey Lewis paper on depression 40-50 pages had 20 pages on describing each patient
Psychodynamic classifications important in US from early-mid 20th C

US/UK Diagnostic Project
Noted high rates of schizophrenia in US especially New York compared to UK
Group of London psychiatrists visited NY in late 65-70 and using standardised psychiatric interviews and a standardised way of making psychiatric diagnosis
Found high rates of schizophrenia in New York due to their use of a “psychodynamic” diagnostic criteria, many of these cases would be diagnosed with mood disorders or personality disorders in the UK
Highlighted the variability of diagnostic practices reduced its’ usefulness for research or communication or statistics or clinical practice

Nomenclature & Classification
Nomenclature refers to the naming of objects in  a classification
These objects are given a description , often with a definition of what is unique to that object
The combination of nomenclature of objects with their descriptions is called a glossary 
The classification ideally should be mutually exclusive and jointly exhaustive : every example (e.g. of clinical cases) described and examples can only be put in 1 place
A Taxon is a group of objects with obvious and important similarities 
Taxonomic hierarchy often used e.g. classification of animals

ISPS (1967) and DOSMED (1992)
Multiple international centres of research in broad spread of developed and developing countries
Used standardised semi-structured interviews and diagnostic procedures
“Broad” definition of psychosis  incidence varied widely by country 
“Narrow” definition of psychosis roughly similar incidence across globe- adding to cross-cultural validity of concept of “Schneiderian” SCZ
Outcomes generally better in developed countries ?less EE ?different aetiology?

Diagnostic Hierarchy
Choose diagnosis that is highest in hierarchy that can explain all symptoms 
  1. Organic disorders
  2. Mood Psychosis
  3. Schizophrenia
  4. Depression
  5. Anxiety Disorders
  6. Personality issues
Prototypic Matching/ Narrative v Operational Criteria

Prototypic matching/  narrative classifications describe “typical cases” (cf blue bloaters/pink puffers). Professionals match patient in front of them to closest description in classification 
Clinicians prefer this type as quick and convenient

Operational criteria list a set of rules to make a diagnosis in a particular patient (cf making a diagnosis of myocardial infarction) such as mandatory features, exclusion criteria etc. 
Researchers prefer this as more get more standardised homogenous groups

Multi-axial Classification
With most diagnosis often non-aetiological, often ignoring social circumstances, high rates of comorbidity
Can list important factors in different “axes”
Eg. DSM III & IV had
  1. Clinical syndromes/disorders
  2. Personality disorders, mental retardation
  3. Medical conditions
  4. Psychosocial and Environmental Stressors
  5. Global Assessment of Functioning

The International Classification of Diseases
The standard diagnostic tool for epidemiology, health management and clinical purposes.
Every country has to send health statistics using ICD codes to UN
In its 10th edition, with 11th coming out soon
6th edition started providing psychiatric codes but with no description (similar to other health problems) 
From 8th edition onwards a “glossary” produced with descriptions of each psychiatric code to improve standardisation of practice
10th has a main version of a  glossary with prototypic descriptions, a DCR with operational criteria and a simplified primary care version

Diagnostic Statistical Manual
Introduced to produce codes for variety of admin uses and descriptions of these codes for US 
Gives ICD codes for disorders and increasing similarity to ICD
DSM editions I,II used psychodynamic concepts and aetiology
DSM III revolutionary in 1980 : operational criteria, multi-axial classification, dropped Freudian aetiology, aimed for reliability
Latest edition DSM V controversial: felt to be too low thresholds to include too many “contentious” cases (Yanks worry though that too restrictive would deny people access to help)
Most commonly used system in research

Alternative Diagnostic Systems
France uses a psychodynamic influenced classification for child psychiatry
ICD as has to be used internationally incorporates many of different traditions of psychiatric diagnostic practices in their classification

Diagnosis as part of Formulation
Formulation is summary of all relevant clinical information including diagnosis
Should guide treatment and management of patient
Should tell you “why does this patient present with this type of problem at this particular time (and still does)?”
Can be divided into predisposing, precipitating and maintaining factors
The types of factors can be divided into biological, psychological and social

Alternatives to Diagnosis: Why?
Lack of validity
High rates of comorbidity
High levels of assumptions contained within diagnosis
Dislike of “medical” emphasis/language/system and perceived “biological orientation” “Disease based model”
Diagnostic system may not suit their practice: 
e.g. primary care where the cause of the stress (usually psychosocial) & most often mixed anxiety and depressive symptoms
Psychotherapists and psychologists with “office based practice”
Ignores social context & locates pathology in individual

Alternatives to Diagnosis : Symptoms
Can just use symptoms : no assumptions about relationships or associations so no concerns re : validity
Individual symptoms can have some predictive validity
“most scientific” as just concentrate on what is present and can measure these
Some treatments such as eg CBT focus on individual symptoms
Very time and cognitive resource intensive
Often “diagnosis by proxy” as have to take into account aetiology eg substance-related anxiety treated differently than free floating
Often medical treatments effect multiple symptoms at same time

Alternatives to Diagnosis : Dimensions
Dimensions are  Groups symptoms/clinical features  into categories of related symptoms 
Often describes patients better than simple diagnosis (1 bit v multiple bits) in terms of clinical picture, prognosis and treatment responses
Also requires additional aetiological information 
Often prone to diagnosis e.g. describing patients with high levels of hallucinations/delusions as “paranoid schizophrenia” in head
Complexity can be hard to hold and manipulate in head whilst making decisions as well as collecting statistics, research (due to large no. of groups so often use correlation)
Multiple dimensions become like diagnostic comorbidity 

Alternatives to Diagnosis : Purely Psychosocial formulations
DCP trying to develop “paradigm shift” 
Intent is to provide a narrative formulation with less emphasis on biological factors
Trying to identify psychological processes as basis of classification
More useful therefore for psychotherapy and “office based practice”
Unknown reliability and validity
Unlikely to be useful to “medical psychiatry”
Issues of communication errors if used exclusively by psychologists when talking to psychiatrists

Further Reading
ICD 10 (mandatory until ICD 11 comes out)
DSM V (know about)
“A Companion to the Classification of Mental Disorders” Cooper and Sartorius
Read chapter on Classification in your main textbook e.g. Shorter Oxford Textbook on Psychiatry 
“Distinguishing between the validity and utility of psychiatric diagnoses” Kendell, Robert;Jablensky, Assen The American Journal of Psychiatry; Jan 2003; 160, 1
“Models for Mental Disorder”  Tyrer and Steinberg

  • Philosophical Issues in Psychiatry Volume 2: Nosology

No comments:

Post a Comment

Note: only a member of this blog may post a comment.