Home > Flashcards > Print Preview
The flashcards below were created by user
Alba123
on FreezingBlue Flashcards. What would you like to do?

Definition of Biostatistics
 Science that studies biological phenomenons from a quantitative
 point of view (quantitative analysis
 of biological phenomenons)

 Inlcudes:
 n GATHERING OF DATA
 n DATA PROCESSING
 n ANALYSIS AND INTERPRETATION
 n VALIDITY AND GRADE OF TRUST TRIALS

Definition
of “variable”
 n ANY CHARACTERISTIC, STUDIED ON A POPULATION THAT
 CAN CHANGE AND WHOSE CHANGE CAN BE MEASURED.
 n Examples: age

Classification
of “variables”
 Qualitative
 (by order or name)
 Quantitative
 (continuous or non continous)

Definition of qualitative variable
examples
(!)
 characteristic of a population
 that due to its peculiarity , can not be measured with numbers (it can not be
 quantified by numbers).
 Examples: sex, race,
 religion, civil status, level of
 academic studies

Qualitative
variables: classification according to number of possible values it can
have.
Examples.
 n DICOTHOMIC (choosing between two): the
 studied variable can only have one value without hierarchy
 between different values (ex.: man/woman; son/daughter).One value excludes
 the other.
 n POLICOTHOMIC OR MULTICOTHOMIC (choosing between more than two): the studied variable can have multiple
 values
 either with or without hierarchy among
 them (socioeconomical level (low, medium, high ), blood
 groups ( A/B/AB/0), colour of the hair
 (black, brown, blonde)

Qualitative
variables classification depending on the
existence of an order (
ordinal,
nominal). Examples.
 Ordinal variable: variable where its values show an order,
 sequence or natural progression.
 Ex.:order:
 days of the week, months of the
 year
 natural
 progression or sequence:grade of undernourishment (light, moderate,
 severe), answer to a treatment (good/ bad ), socioeconomical level (low,
 medium, high), alcohol consumption (light, moderate, severe), dental hygiene (good/intermediate/bad)

 Nominal variable: variable where its values
 do not show hierarchy ( order, sequence or natural progression):
 Ex.: names
 of people, races, blood groups, civil status, types
 of teeth( incisor, canine, premolar, molar)

Definition of quantitative variables
examples
 characteristic
 of a population that can be measured with numbers.
 n Ex.:number of children, number of teeth

Quantitative variables classification
 continuous,
 non continuous

Definition
of quantitative continuous variable.
Examples.
 n It can have a whole (entire) or fractioned value (inside a numeric scale) (they refer to items that show a
 continuum)
 n Ex.: age, depth of a gum pocket (4mm, 5,5 mm, normal:3mm), mouth opening (3,5 cm, 2,7 cm )

Definition
of quantitative non continuous
variable. Examples
 n It can
 only have a whole (entire) value (inside a numeric scale) (they refer to items that can not
 be divided)
 n Ex.: number of teeth (20/32)

Classification
of variables according to cause
 n Independent: variable
 causing the effect (Streptoccoccus
 mutans causes dental decay)

 n Dependent: variable due to
 an independent variable (dental decay is caused by Streptoccoccus mutans)
 (dental pain is due to
 infection)(gum
 sore due to unadjusted dental prosthesis)

Definition
of “confusion variable”
 Definition:
 variables that have an influence on the effect, but they are not the main
 cause.
 Ex.:
 stress can increase dental pain due to pulpitis. (stress is a confusion
 variable) deficient dental hygiene can increase
 periodontitis
 (but the main cause are bacteria of the mouth), deficient dental hygiene is a
 confusion
 variable.

 kann
 auftreten, muss aber nicht, wenns auftritt verstärkt es den depending
 variable

Example for
confusion, depending, indepedendent variable
 confusion
 variable: stress
 depending:
 dental pain
 independent:
 pulpitis

 independent:
 bacteria of the mouth
 dependent:
 perodontitis
 confusion:
 deficiet dental hygiene

Frequency
measures
 =
 parameter, we measure how often a disease occur in a population
 incidence,
 prevalence, > morbidity,
 mortality,
 frequency,
 percentage, proportion, ratio, relative risk, rate

Definition
of incidence and incidence rate
 n Incidence: number of new
 cases of a disease in a certain period of time (ex.: 100 new cases of
 influenza per week)
 n Incidence Rate: number of
 new cases of a disease in a certain period of time/total population X 100 (
 ex.: Incidence Rate of influenza in the population of this city is 8%)

Definition
of prevalence and prevalence rate
 n Prevalence: total number of cases (new and previous) in a determined period
 of time.
 n Prevalence Rate: total number of cases ( new and
 previous) in a determined period of time/ total population X 100
 n Ex.: prevalence of influenza infection is 5%( it
 means that 5% of total population is affected)

Definition
of “absolute frecuency”
 n They quantify the importance of a disease in a
 community ( total number of people affected).
 n Ex.: absolute frequency of Influenza in the
 population of that city is 850 cases.
 n disadvantage: Absolute data do not provide
 information about the probability to develop the disease (relative risk).

Definition
of “relative frequency”
 they
 provide a more complete information about the event we are studying and about
 the risk of the event or disease to
 happen.(
 because they compare absolute data with other data ).

Types of relative frequencies
 proportions,percentages,
 ratios and rates

define proportion
 n Relation between 2 events of the same type, in a
 different geographical area
 n Ex.: number of car accidents in Madrid/ number of
 car accidents in Spain;
 n Number of influenza cases in Valencia/ number of
 influenza cases in Spain.

define percentage
 n Fraction of total population affected by a disease
 or event, in relation to total population x 100
 (ex.: 35% of adults over 50 years old have
 periodontitis)

define ratio
 n Relation between two events of the same type, one
 of them exposed to the risk factor, and the other one non exposed to the risk factor.
 n Each event can be expressed as an absolute or
 relative frecuency.
 n Ex.: number of children with dental decay taking
 sugar/number of children with dental decay not taking sugar.
 n The most common ratio in epidemiological studies
 is RELATIVE RISK (RR).
n Relative risk is a ratio
 n It intends to show the relation between risk
 factors and appearance of diseases.
 n Ex: relative risk to develop cancer of the
 mouth associated with unadjusted dental
 prosthesis.
 n Ex.: relative risk to develop periodontitis
 associated with deficient oral hygiene.

define rate
 n Rate: number of events in relation to total
 population X constant (1001000)
 n Ex.: mortality rate: number of people died in one
 year/total population X 1000
 n Ex.: birth rate: number of children born alive
 during one year/total population X 1000
 n Ex.: children mortality rate: number of children
 under one year old died during one year/total of born alive in one year X 1000.
 n Rates give us a more complete information about
 the event we are studying, than absolute numbers.They allow us to compare data from different
 populations. Ex.: we can compare mortality rate in Africa (13,2 per 1000 ) and
 in Europe (11,8 per 1000)

types of rate
  Raw Rate: referred to total population.
 Ex.: mortality rate in Spain is 8/1000. e.g. birth rate, mortaliy rate
  Specific Rate: referred to a certain part
 of total population. Ex.: mortality rate in age group 4050 years old in that
 country is 4/1000.
 e.g.
 mortality by age group, morbidity by cause of the disease

define relative risk
relative risk = morbidity rate in the group of exposed people / morbidity rate in the goup of non exposed people > a/a+b / c/c+d
a/a+b: people with the disease in relation to total of people exposed to the risk factor
c/c+d: people with the disease in relation to total of people non exposed to the risk factor
RR= 1 indicates that the probability to develop the disease is the same in both groups (exposed and non exposed)
RR> 1 indicates that exposed group has higher probability to develop disease than non exposed

Association
between variables
 real
 association, random errors (bias), misleading errors (sesgos), confusion
 variables

Epidemiological
studies
 Show if an
 association exist or not between differen variables

Precision
 Absence of
 error due to random (the larger the sample, the more precision; criteria =
 logical)

Definition
of “null hypothesis”
 (Ho): it
 does not exist association or relation
 between 2 studied variables.

Definition
of “alternative hypothesis”
 (Ha): it exists association or relation between two
 variables

Definition
of “hypothesis test”
 = test of
 statistical significance = comparison between the 2 hypothesis

Definiton of
"statistical significance"
 appears at
 the end of scientific articles when the author is trying to demonstrate that
 his results have a “quality"
 Big
 statistical significance > result is trustable/true

Hypothesis
test: how do they work?
 We look at the size (magnitude) of the
 difference ( in the result) between the 2 groups we are studying.

Synonyms of
statistical significance
  we reject Null Hypothesis (=no association
 between 2 variables)
  we accept Alternative Hypothesis
  enough evidence to doubt about Null
 Hypothesis
  result observed is not compatible with Null
 Hypothesis
  it is unprobable to obtain a result like
 the one observed if Null Hypothesis would be true
  It is unprobable ( unlikely) that the
 result observed would be due to random
  “P” is minor 0,05 ( p<0.05 )

Definition
of “p” value
 probability
 to accept “Alternative hypothesis” as being true, when the true hypothesis
 could be “null hypothesis”
 To
 accept/reject a hypothesis has a riskt what we quantify as p value
 The smaller
 the p value, the greater the statistical significance/the more secure is the
 alternative hypothesis

Interpretation
of “p”<0.05
 we have a security of 95% of alternative
 hypothesis being true

Interpretation of “p” value<0.01
 we have a
 security of 99% of alternative hypothesis being true

Interpretation of a small “p” value
 the smaller
 the “p” value is, the greater is the statistical significance of the result of
 the study

Interpretation of “p” value>0.05
 random can
 not be excluded as the cause of the association between 2 variables, that is,
 we can not reject Null hypothesis
 that says
 that both variables are not associated

What does
statistical significance depend on?
 magnitude
 of the difference (the larger the
 difference, the higher the statistical significance), size of the sample (the
 larger the size, the higher the statistical significance)

What is the
relation between “magnitude of the difference” and “statistical
significance”
 the larger
 is the difference in the result observed at both groups we are studying, the
 easier is to demonstrate statistical
 significance

What is the relation between “size of the
sample” and “statistical significance”?
 the larger
 the size of the sample, the easier is to detect differences in the result
 obtained in both groups we are studying, that is, the easier is to detect
 statistical significance in the result of our study

Types of
error in association between variables
 False
 positive, False negative

Definition
of error Type I ( false positive)
 This means
 that we say that “it exists a relation between 2 variables”, but that is
 false
 We make
 this error when we reject Null Hypothesis and we are wrong
 we accept
 Alternative Hypothesis but we are wrong.
 Example: We
 make a study and our results show us that there is a relation between
 sweets and
 dental decay found at children. But at this particular case, we are wrong

Definition
of error Type II ( false negative)
 This means
 that we say that “it does not exist a relation between 2 variables”, but we
 are wrong
 We make
 this error when we accept Null Hypothesis as real, but it is false.
 we accept
 Null Hypothesis but we are wrong
 Example: We
 make a study and our results show us that there is not a relation
 between
 bad oral
 hygiene and periodontitis found on adults. But in this particular case, we
 are
 wrong

Definition
of random error. Examples
 To happen
 by chance
 Ex.: one
 person could have a cancer by random: he is not exposed to risk factors, he
 is not genetically predisposed, he has not family history related to that
 disease…but cancer just happens.

biological
variability. example
 In the
 sample that we have chosen out of the general population, there exist a
 certain biological characteristic
 that could influence the result of the study.
 Ex.: a
 sample of people receiving an antihypertensive treatment have responded very
 well because they had a previous excellent condition of their vascular system
 (arteries and veins), and not only because the medicine we have given to them
 has such a fantastic effect to reduce hypertension.

misleading
error. example
 Misleading
 error (sesgo): in the study we are carrying out we introduce a condition in the observers or in the method of study
 we are using , that will alter the result of the study, thus , giving us a
 wrong information (false result)
 Ex.: If we
 want to diagnose dental decay in children, through the method of detecting a
 “white spot” on the enamel of teeth, and we are using an observer who has
 visual deficit , we are introducing a misleading error in the study.

confusion
variable. example
 Variables that can affect the result of the
 study, but they are not the main cause of the result observed.
 Ex.:poor
 oral hygiene can be a confusion variable when we study periodontitis, where
 the main cause of the disease are the bacteria of the mouth ( and not the
 lack of dental hygiene ).

