A couple of people commented on my last post that they
sometimes find it difficult to critically analyse a primary research paper or
necessarily trust it. Here is a short list of the different study types and
their various advantages and disadvantages.
Studies are normally divided into what we term descriptive
studies or analytic studies. So there isn’t an information overload, I will
just discuss descriptive studies on this post and will discuss analytic studies
on my next post.
Descriptive studies on a population level are termed
‘ecological studies’ and on an individual level are termed ‘case reports’,
‘case series’ or’ cross sectional studies’.
A good example of an ‘ecological study’ was that conducted
by Friedman and colleagues in 1967 which found that deaths from coronary heart
disease in 44 US states correlated with the number of cigarettes sold in those
states; i.e. the states selling the most cigarettes had the highest levels of
coronary heart disease. The advantages
of such a study are that they are helpful in finding correlations between risk
factors and disease outcomes. The disadvantages are that we do not get
information about individuals so it is impossible to tell if the people with
the risk factor (i.e. those buying the cigarettes) are the ones getting the disease (coronary
heart disease). This is termed the ‘ecological fallacy’. The other problem is
‘confounding’ – i.e. there may be other risk factors at play which aren’t being
considered.
Case reports and Case series are examples of ‘ecological
studies’ on an individual level. A good example of this is the case series
which emerged from the Centre for Disease Control and Prevention in 1981
reporting the occurrence of the rare lung infection PCP (Pneumocystic carinii
pneumonia) in a group of 5 otherwise healthy men in LA. This heralded the
beginning of awareness of AIDS in the USA. Relevant to MS are the case reports
that came out around 2005 showing that natalizumab (Tysabri) can result in JC
virus induced PML .
The important thing to take into account when looking at a
case series is to ensure that the patient’s being reported are representative
of all the patients with the condition so that conclusions can be generalised.
The advantages of such studies are that they are good for forming hypotheses
and describing clinical experience. It is also a good way to inform the medical
community of the first cases of what could be an emerging condition. The
disadvantages are that the finding may be pure chance or the patient selection
may be biased making generalisations difficult.
The final example of an ‘ecological study’ is the ‘cross
sectional study’ also termed a ‘prevalence study’. The object of this design is
to estimate how common a disease is within a population. In reality, one can’t
test the whole population so a sample of the population is taken. The
advantages of such studies are that they are good for common diseases (like
high blood pressure/high cholesterol) as you only need a small sample and the
study is normally of short duration. It’s not useful for rare diseases as they
are less likely to be accurately represented in a sample of the population. The
other point to take into account is that those who normally volunteer for such
studies tend to be well educated and health conscious so they may not be
representative of the population – we term this ‘confounding’.
I’ll post about analytical studies in the next couple of
days but you can see the links to the studies I mentioned below.
Cigarette
smoking and geographic variation in coronary heart disease mortality in the
United States
http://www.sciencedirect.com/science/article/pii/0021968167900896
PCP Case Series
http://www.cdc.gov/mmwr/preview/mmwrhtml/june_5.htm
Progressive Multifocal Leukoencephalopathy after Natalizumab Therapy for
Crohn's Disease