The epidemiology of mental disorder

Question 1

What is epidemiology?

What does it involve?

Answer 1

“Epidemiology is the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems.”

• defining “disorders”
• counting and mapping “disorders” in place and time
• linking the occurrence of “disorder” to potential causative factors
• identifying appropriate (public health) interventions

Question 2

What are the two broad strands of epidemiology?

Answer 2

> Descriptive epidemiology

• prevalence
• incidence
• variations between populations
• comorbidities
• burden of disorder

> Analytical epidemiology
- causes

> Epidemiology also underlies the design of interventions studies

Question 3

What was John Snow’s (1854) “Broad Street Pump” use of epidemiological thinking?

Answer 3

John Snow and the Broad Street Pump
> 1854 - cholera in London
> Theory: cause = drinking water
> Mapped the incidence of cholera
> Clustering of cases around Broad Street Pump
> Cholera cases related to where victims drew their water
> Further data support hypothesis that sewage is a source of cholera

Question 4

What was Doll’s and Hill’s use of epidemiological thinking, studying the association between tobacco and lung cancer?

Answer 4

> They studied the relationship between tobacco and lung cancer among British male doctors between 1951 - 1961
Found a positive correlation between number of cigarettes and cancer deaths

-> This study led to a public health revolution: we are now aware of the risks associated with smoking… very few doctors now smoke

Question 5

What is psychiatric epidemiology?

Answer 5

Distribution of mental disorders in a population

• onset, course and outcomes of mental disorders
• measure burden on society
• identify causal mechanisms underlying the development of mental disorders, with aims of prevention and modifying outcomes

Question 6

What are caseness, incidence and prevalence?

Answer 6

> Caseness: define the presence of a mental disorder as defined by operational criteria
- complex and contested

> Incidence: transition to caseness over a defined time period
- usually a year

> Prevalence: proportion of people who are cases at the time
- e.g. one-year period prevalence ; lifetime prevalence

Question 7

What does the descriptive perspective in the epidemiology of mental disorder consist of?

Answer 7

Provide estimates of caseness, diagnosis and the burden of illness within a population

• identity
• sample
• structured interviews (data on symptoms and disability)

Question 8

In epidemiological studies , how are definitions and diagnostic criteria, two contested issues, used?

Answer 8

> For the purposes of these studies, decisions need to be made about how to define mental disorders
Diagnostic criteria are reliable, but have contested validity
Majority of epidemiological studies use the DSM structure

Question 9

What does the analytical perspective in the epidemiology of mental disorder consist of?

Answer 9

> Breaking population down in a structured fashion
identify variables (correlates and causes) of mental disorder
Apply methodology
- e.g. case-control studies (comparing identified cases with non-cases for the presence of a variable of interest)

Question 10

What are the three classic US epidemiological studies of mental disorder?
What were their limitations?

Answer 10

> 1980s: Epidemiological Catchment Area
1990s: National Comorbidity Study (NCS)
2000s: NCS Replication

-> They used diagnostic tools developed in relation to DSM-III, DSM-III-R, DSM-IV

> Significant limitations:

• they don’t tell anything about low-prevalence disorders like schizophrenia
• > we need alternative sources of information for such studies
• they have missed out mental disorders in the elderly: a very important source of societal burden in advanced countries

> HOWEVER, they are very important

Question 11

What are the headline findings from the National Comorbidity Study Replication (NCSR)?

Answer 11

1. Mental disorders are common
   - 26.2% for 12-month period prevalence
   - minimum 46.2% for lifetime prevalence
2. Early onset
   - 75% of disorders have begun by the age of 24
3. Serious or moderate severity
4. Comorbidity is common
   (presence of more than one mental disorder at the same time -> lack of validity of classification systems used?)
5. Access to treatment is limited
   - relative poor quality if and when it happens, often very delayed

Question 12

Despite very divergent and fairly crude assessment strategies, what does the available data of the National Comorbidity Study Replication (NCSR) consistently demonstrate?

Answer 12

1. An association of all mental disorders with a considerable disability burden, in terms of the number of work days lost
2. Generally low utilisation and treatment rates
   - only 26% of all cases had any consultation with professional health care services
   - > considerable degree of unmet need

Question 13

What does the WHO World Mental Health Survey (2007) suggests?
How is the data lacking plausibility?

Answer 13

Wide discrepancy of prevalence across countries, but these differences aren’t entirely plausible
- e.g. people in Spain seem twice as mentally healthy as people in France

> Suicide data suggests a variance in societal expression of mental health
- some countries have very high suicide rates

> Comparing the two sets of data (lifetime prevalence of mental disorders and suicide rates):
- in Japan: low prevalence of disorder, while a very high suicide rate… not entirely plausible

Question 14

In epidemiological studies on the prevalence of mental disorder, what does the information have to include to be comprehensive?
What is specific about people with schizophrenia for such studies?

Answer 14

> Population-based surveys
“administrative” data that takes into account what we know about treated population with severe mental illness/disorder
e.g. people with schizophrenia tend to not answer the door when canvassed in population-based studies

Question 15

What is the summary of Witchen’s and Jacobi’s (2005) paper on epidemiological EU data on a wide range of mental disorders?

Answer 15

> Multimethod approach

• 27 eligible studies with variable designs
• 150,000 subjects from 16 European countries

> Prevalence on the basis of meta-analytic techniques, re-analysis of selected data:

• 27% of the adult EU population affected by mental disorder in past 12 months (2005) = 82.7 million cases
• little evidence for considerable cultural or country variation

> Most frequent disorders:

• anxiety disorders
• depressive
• somatoform and substance dependance

> 12-month prevalence estimates of mental disorders

• psychosis: 1%
• bipolar disorder: 1%
• major depression: 7%

> The distribution of prevalence estimates is wide
-> authors used the medians

Question 16

What did Witchen’s and Jacobi’s (2005) paper on epidemiological EU data on a wide range of mental disorders say about disability, treatment and future research?

Answer 16

Despite very divergent and fairly crude assessment strategies:
> Considerable disability burden in terms of number of work days lost
> Low utilisation and treatment rates
- only 26% of all cases ( of the 27%) (all mental disorders)

> Considerable future research needs to determine the degree of met and unmet needs of services
-> taking into account severity, disability, and comorbidity

> Research most pronounced for:

• new EU members
• adolescent and older populations

Question 17

On what are the Eu and US data, from epidemiological studies, consistent on?

Answer 17

> There is a lot of mental disorder

> Only a small proportion of people with mental disorder are receiving treatment

Question 18

In analytical epidemiology, what is a relative risk (RR)?

Answer 18

The ratio of the probability of an event occurring in an exposed group to the probability of the event occurring in a comparison group.
e.g. the study of Doll and Hill with smokers vs. non-smokers (association of cigarettes and lung cancer)

RR < 1 means the risk of an event occurring is reduced
RR = 2 means the risk of an event occurring is twice likely to happen

Question 19

What data do we have on relative risk (RR) of psychosis, from Bebbington et al. (2011) and Jablensky (2003)?

Answer 19

> Pre-morbid cannabis use RR: 2.0
Urban birth RR: 2.4
Children of emigrants to UK RR: 7.0
History of psychosis in a first-degree relative RR: 9.3

> History of child sexual abuse and psychosis odds ratio: 10.1

Question 20

Historically, where is the focus for understanding the causal mechanisms for psychosis?
What about recently (Keyes and Susser, 2014)?

Answer 20

> Historically, focus on social factors (immigration, urbanicity, psychosocial stresses, life events) on the prevalence of mental disorders

> Recently, focus on mechanisms through which social forces result in the onset of mental disorder(s) (Keyes and Susser, 2014)
-> a more psychiatric epidemiology

Question 21

What is currently studied for the causal mechanisms for psychosis and common mental disorders?

Answer 21

> Biomarkers
> Gene-environment interactions
> Impact of social environment on neurobiology
> Genome-wide research
> Epigenetics