case control
HOME DENTAL DIRECTORY FAQ DENTAL TERMINOLOGY STUDENTS CORNER GREETING CARDS FREE CONSULTATION CONTACT US


CASE CONTROL & COHORT STUDIES
In day to-day lilife the common tendency is to figure out why an incident has occcured, once it happened.
Here one starts with a group of individuals who suffered from a particular episode of ill health.
Inquaries are taken up to find out what caused the episode.
This has to be done conjunction with the study of another group of individuals who have not fallen prey to the condition.
This group is called "controls". This a retrospective study design.
It is economical but not foolproof.
Case control studies are primarily used To
  • assess risks and
  • to study causes of disease in general
Aetiological research is traditionally performed in three settings :--
  • clinical
  • community, and
  • occupational settings.
CASE CONTROL STUDIES
clinical studies often orginate in a case report curious enough to be noticed and later growing in number when more and more of the syndrome are noticed.
The resultant case series when subjected to a controlled analysis metamorphoses into a case control study
In a community setup however certain ecological associations would stand out and this leads to cross-sectional studies from within case and controls
occupational settings.:-- An increased proportional mortality rate for a specific condition would segregate an occupation from others and would again result in a controlled observational study in a occupational study.
Conduct Of Case Control Studies
  1. Identification of cases
  2. Selection of Controls
  3. Matchings
  4. Exposure rates
  5. Analysis
1. IDENTIFICATION OF CASES
Defining the case with respect to a realistic criteria laid down before the start of the investigation is the first task in case-control study.
This is especially true because this study design is suited to study of rare and chronic conditions, especially when resources are scarce. E.g.
  • identification of children with cleft palates could be based entirely upon a simple physical examination. Other cases may need to be defined more accurately in terms of a microscopic diagnosis.
However regardless of the method chosen, it is recommended that, freshly diagnosed cases be included in the study for reasons of bias.
Location of cases can either be from
  • medical records, registries;
  • from wards
  • clinics; or from
  • community and
  • workplaces etc.
2. SELECTION OF CONTROLS
Greatest callenge in a case-control study is often the identification of an appropriate control group.
Selection of controls is based on the assurance that information on study factors can be obtained from the control group in a manner similar to that by which it was obtained from the cases ( which may include investigations).
The controls are generally selected at an equal ratio to that of cases and should have matching characteristics with respect to general confounding variables such as age, sex and socioeconomic status unless if these variables themselves are not the exposure characteristics.
Controls can be derived from
  • medical facilities
  • community or
  • neighborhood of cases
3. MATCHING
sampling of subjects to certain levels of relevant co-factors. should be restricted
Partial restriction, known as matching, involves selecting a control group which is parallelt to the cases in co-factor distribution.
frequency matching is done usually such that control group is similar to cases in overall distribution of matching factors like age, sex etc.
On the other hand we may resort to pair-wise matching where for each case selected, one ( or more ) of control with similar attributes is selected.
4. EXPOSURE RATES
The next step in case control study is to estimate the exposure rate which might be
  • self reported
  • obtained from hospital records or
  • detected using special investigations.
The association of an exposure and a disease in a case control study is calulated by deriving the odds ratio ( or )
ASSOCIATION BETWEEN CHEWING TOBACCO & DEVELOPMENT OF ORAL CANCER
Risk factor ( chewing tobacco ) cases controls
Present 80 ( a ) 40 ( b )
Absent 10 ( c ) 50 ( d )
This is the ratio of odds of exposure among the cases ( 80 : 90 ) to the odds in favour of exposure among the controls ( 40 : 90 )
This indicares cases were 10 times more likely than the controls to have chewed tobacco in the past.
5. ANALYSIS
The traditional measure of an exposure - disease association is the relative risk.
Since case-control study is not one of a prospective design it is impossible to generate the true incidence ratio between exposed vs non-exposed group.
However, an indirect method of estimating relative risk was suggested by Cornfield (1951) called the "odds ratio"
Attributable risk proportion ( ARP )
Attributable risk proportion is the proportion of total disease in exposed persons which may be attributed to the exposure.
This can be calculated from OR as
when the Or is > 1, the ARP has a range of possible values from zero to unity.
when OR = 1 then none of the disease risk in exposed persons will be attributed to the exposure.
BIASES IN CASE CONTROL STUDIES
  1. Biases in selection of cases
  2. Biases in investigating controls
  3. confounding bias
  4. Problems due to over matching
  5. Bias in analysis
ADVANTAGES OF CASE CONTROL STUDIES
  1. Efficient sampling of a rare disease
  2. Rapid evaluation of chronic diseases
  3. Economy of expense and personnel
  4. may serve either explanatory purposes
LIMITATIONS OF CASE CONTROL STUDIES
  1. not practical for rare exposures
  2. subject sampling prone
  3. historical information often cannot be validated
  4. Relevant cofactors may be difficult to control
  5. temporal sequence of exposure and disease may be obscured