Curs 3 CB Metodologia Cercetarii 2008

7/30/2019 Curs 3 CB Metodologia Cercetarii 2008

1/96

Corin Badiu, 2008

Metodologia cercetrii tiinifice

Colectarea datelor

Strategii de msurare


2/96

Obiective Stabilirea tipurilor de variabile

Surse IT de dateIdentifica si localizeaz seturile corecte de date

In instituie In afara instituiei

Strategii de msurare

Analiza, interpretarea si raportarea rezultatelor Selecteaz programul adecvat: Excel, SPSS Folosete programele pentru analize statistice simplesi prezentarea grafica a rezultatelor

Interpreteaz rezultatele


3/96

Predictor * Rezultat

Variabile deconfuzie*

Modificatori ai efectului*

Tipuri de variabile

*Considerate general ca expunere la factori de risc


4/96

Tipuri de studii clinice

Studii fr variabile Studii de caz, serii de cazuri, editoriale,

opinii / comentarii, rapoarte review Studii cu o singura variabilaStudii descriptive

Studii cu 2 variabile ExperimenteStudii observaionale Meta-analize si review-uri sistematice


5/96

Ierarhia tipurilor de studii clinice

Studii clinice

Descriptive Analitice

Experimentale Observationale

Cohorta Caz-control Cross-sectional


6/96

Variabila predictor(independenta)

Variabila rezultat(dependenta)

Variabile


7/96

Evidence Based Medicine

Metode de tratament susinute dedovezi clinice si de cercetare.

Necesita integrarea celor mai bunedovezi din cercetare pentru diagnostic sitratament cu experiena clinica.

Ia in considerare ce este optim pentrufiecare pacient precum i preferineleacestuia.


8/96

Realitatea

Cercetarea asupra eficientei de tratamentface subiectul unui numr mic de articole.

Evidence based medicine este considerat unconcept ce folosete baze de date inclusivstudii sistematice de caz pentru a ghidaintervenii terapeutice.

Dovezile trebuie evaluate intr-un contextterapeutic efectiv, Ce tip de interveniecapt sens pentru mine ca practician?


9/96

ntrebri clinice

Care este cea mai bun alegere terapeutic pentrupacientul meu?

Este acest program suficient de bine documentatteoretic? Funcioneaz acest program terapeutic? Ct timp dureaz acest program?

Ce urmeaz dup aceea?


10/96

Utilitatea clinicii!

ClinicieniiSunt n liniaI a asistentei medicale

Au expertiza clinica necesara i cunosc bine pacienii Sunt formai pentru gndire tiinifica Sunt obinuii cu colectarea datelor tiu sa identifice rezultate


11/96

Msurtori de baz

O Linie de baz este o msurtoare aratei de rspuns in absena tratamentului

Linii de baz Stabilete necesitatea tratamentuluiPermite mbuntirea documentaiei Ne permite s mbuntim studiul


12/96

Date de baz

Se creeaz un set de exemplare ale intelor studiului

Pregtim o fi de nregistrare Folosim msurtori conform criteriilor de referin


13/96

Strategii de Colectare a Datelor

ntotdeauna alegei mai mult dect o msurtoare

Verific acurateea datelor de baz

Selecteaz designul cercetrii clinice


14/96

Design de Cercetare Clinic

ABA designTest, tratament, test

ABAB designTest, tratament, test, tratament

Design serial n timpStabilete o linie de baz

ncepe tratamentulMsoar rezultatele tratamentului

Design cu linii de baz multiple Obine un numr de linii de baz diferite Fiecare linie de baz trebuie s fie independent de altele Trateaz doar o variabil odat


15/96

Cerine

Fiabilitate Valide Sensibilitate Universale Fr erori

Instrumente de colectare a datelor


16/96

fiabilitate? Instrumentul va putea msura cu aceeai acuratee:

n situaii diferite de testare? Testeaz - retesteaz fiabilitatea

Experimentatori diferii? Fiablitate Inter - operator



17/96

Este valid?

Instrumentul careva fi folosit poatemsura datele pe

care le dorim?



18/96

Este Sensibil? Instrumentul ar trebui sa fie la fel de sensibil indiferent

de prezenta sau nu a variabilei msurate i de cantitateaacesteia (linearitate).

Trebuie sa se caracterizeze valorile: Fals-pozitivi:

Subiectii sanatosi, prezentati ca bolnavi.

Fals-negativi:Subiecii bolnavi, prezentai ca sanatosi.



19/96


20/96

Ar trebui sa nu existe diferene intre valoarea reala i valoarea pecare un investigator o obine prin msurare altele dect diferentedate de variabilitatea de eantionare.

Este fara erori?



21/96

Metode de colectare

1. Colectare Randomizata (Simpla)

2. Colectare Sistematica

3. Colectare Stratificata

4. Colectare pe grupe / segmente

5. Colectare pe alte criterii

6. Alte tipuri de colectare


22/96

Variabile Calitative i Cantitative

Exemple de variabile calitative sunt ocupatia, sexul, status marital, etc.

Variabilele care permit cuantificarea i care pot fi msuratesunt considerate variabile cantitative. Exemple de variabilecantitative sunt greutatea, naltimea, vrsta.

Variabilele cantitative pot fi clasificate ca discrete sau continue


23/96

Tipuri de variabile

1. Variabile de tip categorii (e.g., Sex, Status Marital,religie, naionalitate, profesie )

2. Variabile de tip continuu (e.g., vrsta, venit,greutate, inaltime)

3. Variabile discrete (e.g., Numar de copii)

4. Variabile Binare sau Dihotomice (e.g., ntrebri lacare se da raspuns prin Da sau Nu)


24/96


25/96

Variabile din protocol

TIPURI de VARIABLEindependentedependente

intermediareDe confuzie


26/96

Variabila Independent

O caracteristica pentru care se face omsurtoare i/sau observaie, careinflueneaz un eveniment sau rezultat(variabila dependenta).

NOTAVariabila independenta nu este influenatde eveniment sau rezultat, dar o poate determina sau contribui la variaia sa.


27/96

O variabil a crei valoare este dependent deefectul altor variabile (variabile independente)in raportul de cauzalitate studiat.

Sinonime: variabila rezultat sau rspuns .

NOT Un eveniment sau rezultat a crui variatie

ncercam sa o explicam prin influenta variabilelor independente.

Variabila Dependent


28/96

Variabile Intermediare

O variabila care apare intr- o relaie de cauzalitate de lao variabila independenta la o variabila dependenta.

Sinonime: intervenie, mediere

NOTE produce variaii in variabila dependenta i variaz

datorita variabilei independente.O asemenea variabila este asociata att cu variabila dependenta cat i cu variabilaindependenta.


29/96

Un factor (care este el insusi determinant alrezultatului), care perturba efectul aparent al uneivariabile de studiu asupra rezultatului.

NOTAUn asemenea factor poate fi distribuit inegal

intre cei expui i neexpui i astfel influeneaz magnitudinea aparenta i chiar direcia efectului.

Variabile de confuzie


30/96

Organizarea Datelor

1. Tabel de frecventa2. Histograma de frecventa

3. Histograma de frecventa relativa4. Histograma5. Poligon de frecventa6. Poligon de frecventa relativa

7. Grafic cu coloane8. Grafic placinta 9. Tabel Trunchi cu ramuri 10.Box plot


31/96

Tabel de frecventa

Suppose we are interested in studying the number of children in the families living in a community. The

following data has been collected based on a randomsample of n = 30 families from the community.

2, 2, 5, 3, 0, 1, 3, 2, 3, 4, 1, 3, 4, 5, 7, 3, 2, 4, 1, 0, 5, 8, 6,

5, 4 , 2, 4, 4, 7, 6Organize this data in a Frequency Table!


32/96

X=No. of Children Count(Freq.)

Relative Freq.

0 2 2/30=0.067

1 3 3/30=0.1002 5 5/30=0.1673 5 5/30=0.1674 6 6/30=0.200

5 4 4/30=0.1336 2 2/30=0.0677 2 2/30=0.0678 1 1/30=0.033

Tabel de frecventa


33/96

0

1

2

3

4

5

6

0 1 2 3 4 5 6 7 8

Freq.

Grafic de frecventa


34/96

Now suppose we need to construct a similar frequency table for theage of patients with Heart related problems in a clinic.

The following data has been collected based on a random sample of n = 30 patients who went to the emergency room of the clinic forHeart related problems.

The measurements are: 42, 38, 51, 53, 40, 68, 62, 36, 32, 45, 51, 67,53, 59, 47, 63, 52, 64, 61, 43, 56, 58, 66, 54, 56, 52, 40, 55, 72, 69.

Tabel de frecventa


35/96

Age Groups Frequency RelativeFrequency

32 -36.99 2 2/30=0.06737- 41.99 3 3/30=0.10042-46.99 4 4/30=0.13447-51.99 3 3/30=0.100

52-56.99 8 8/30=0.26757-61.99 3 3/30=0.10062-66.99 4 4/30=0.13467-72 3 3/30=0.100

Total n=30 1.00

Tabel de frecventa


36/96

Msurarea tendinei centrale

Unde este centrul distributiei?

1. Media2. Mediana

3. Modulul


37/96

Empiric

For a Normal distribution approximately,

a) 68% of the measurements fall within one standarddeviation around the mean

b) 95% of the measurements fall within two standard

deviations around the mean

c) 99.7% of the measurements fall within threestandard deviations around the mean


38/96

Concepte necesare

Scale de msura Distribuie, tendina centrala, variabilitate,

probabilitate Prevalenta i incidenta bolilor Rezultate ale bolilor (mortalitate, morbiditate) Asocieri (corelaie sau covariana)

Impact asupra sanatatii (diferene de risc irapoarte) Sensibilitate, specificitate, valoare predictiva


39/96

Scale

Nominal qualitative classification of equal value: gender, race, color, city

Ordinal - qualitative classificationwhich can be rank ordered:socioeconomic status of families

Interval - Numerical or quantitativedata: can be rank ordered and sizescompared : temperature

Ratio - interval data with absolutezero value: time or space


40/96

Distribuie, Tendina centrala Mean


41/96

Variabilitate, Probabilitate

Mean Median Mode Standard deviation Statistical Significance p < .01


42/96

Interval de ncredere


43/96

Semnificaia statistica - erori

Ho True Ho False

Reject Ho Type I error Correctdecision

Do Not RejectHo

Correctdecision Type II error

Null Hypothesis = Ho


44/96

Manual statistic Online

The Statistics Homepage http://www.statsoftinc.com/textbook/stathome.html
http://www.statsoftinc.com/textbook/stathome.htmlhttp://www.statsoftinc.com/textbook/stathome.html


45/96

Prevalena i Incidena

Prevalenceprobability of disease in entire population at

any point in time2% of the population has diabetes

Incidence

probability that patient without diseasedevelops disease during interval0.2% or 2 per 1000 new cases per year


46/96

Sensibilitate, Specificitate

sensitivity =a / (a+c)

specificity =d / (b+d)

Patientswithdisease

Patientswithoutdisease

Test ispositive a bTest isnegative

c d


47/96

Valoare Predictiva

Positive predictivevalue = a / ( a+b)

Negative predictivevalue = d / (c+d)

Post-test probability ofdisease given positivetest = a / (a+b)

Post-test probability ofdisease given negativetest = c / (c+d)

Patientswithdisease

Patientswithoutdisease

Test ispositive a bTest isnegative

c d


48/96

Resurse

An Introduction http://www.poems.msu.edu/InfoMastery/default.htm

Diagnosis Sensitivity and specificity Predictive values

Likelihood ratios InfoRetriever Calculators: Epidemiology, Diagnostic test
http://www.poems.msu.edu/InfoMastery/default.htmhttp://www.poems.msu.edu/InfoMastery/default.htm


49/96

Eroarea n trialuri cliniceAreas in which bias can occur

Systematic error in . . . Allocation Response Assessment


50/96

Allocation or Susceptibility Bias

Can occur when patient assignments to a

trial group are influenced by an investigatorsknowledge of the treatment to be received.

Can result intreatment groupsthat have differentprognoses.

Eroarea n trialuri clinice


51/96

Allocation or Susceptibility Bias

Treatment groups must have similar prognoses, which is achieved by:

Randomization of patients

Prospective evaluation of patientsWell-defined inclusion and exclusion criteria



52/96

Occurs when patientsare assigned totreatments by meansof a mechanism thatprevents both thepatients and theinvestigator fromknowing which

treatment is beingassigned.

Randomizarea n trialuri clinice


53/96

Beneficiile Randomizarii

Prevents the systematic introduction of bias.

Minimizes the possibility of allocation bias.

Balances prognostic factors for treatmentgroups.

Improves the validity of statistical tests usedto compare treatments.


54/96

Response & Assessment/Recording Bias

Can occur when a patient reports a treatmentresponse or when an investigator assesses thatresponse either person can be influenced byknowing the treatment.

A patient or an investigator may have apreconceived idea of which treatment is better.The patient may also want to please theinvestigator.



55/96

Blinding

To minimize Response & Assessment/Recording Bias

Single Blind (patient blinded): protects againstresponse bias.

Double Blind (patient and investigator blinded):protects against assessment/recording bias as wellas response bias.



56/96

Transfer bias

Occurs when patients are lost to follow-up.Must be minimized.

Performance bias

Can occur with a single surgeon or withmultiple surgeons.



57/96

Confuzia - factori

Relationship between coffee andpancreatic cancer, BUT

Smoking is a known risk factor for pancreatic cancer

Smoking is associated with coffeedrinking but it is not a result of coffeedrinking.


58/96

Detectarea confuziei

If an association is observed betweencoffee drinking and pancreatic cancer

Coffee actually causes pancreatic cancer,or The coffee drinking and pancreatic cancer association is the result of confounding bycigarette smoking.


59/96

Evaluarea confuziei

If you know something is a possibleconfounder, in the data analysis use

Stratification, or Adjustment

Fear the unknown!


60/96


61/96

Randomizarea

Random allocationknown chance receiving a treatment

cannot predict the treatment to be given Eliminate selection bias Similar treatment groups


62/96

Un factor Var. Ind.

Randomization tries to ensure that ONEfactor is different between two or more

groups. Observe the Consequences Attribute Causality


63/96


64/96

Simple Blocked Randomization Stratified Randomization

Tipuri de randomizare


65/96

Randomizare simpla

Randomize each patient to a treatmentwith a known probability

Corresponds to flipping a coin Could have imbalance in # / group or

trends in group assignment

Could have different distributions of atrait like gender in the two arms


66/96

Randomizare n bloc

Insure the # of patients assigned toeach treatment is not far out of balance

Variable block size An additional layer of blindness

Different distributions of a trait likegender in the two arms possible


67/96

Randomizare stratificata

A priori certain factors likely important(e.g. Age, Gender)

Randomize so different levels of the factor are balanced between treatment groups Cannot evaluate the stratification variable


68/96

For each subgroup or strata perform aseparate block randomization

Common strataClinical center, Age, Gender

Stratification MUST be taken intoaccount in the data analysis!

Randomizare stratificata


69/96

Rezultat

Introductory Statistical Definitions What is Randomization? Randomized Study Design Experimental vs. Observational Non-Randomized Study Design Stat Software, Books, Articles


70/96

Tipuri de studii randomizate

Parallel Group Sequential Trials Group Sequential trials Cross-over Factorial Designs


71/96

Grup paralel

Randomize patients to one of ktreatments

ResponseMeasure at end of studyDelta or % change from baseline

Repeated measuresFunction of multiple measures


72/96

Studiul Ideal - Gold Standard

Double blind Randomized Parallel groups

2 studii


73/96

2 studii

Study 1 A U.S. study (2000) compared 469 patients with brain cancer to 422 patients who did not have brain cancer. The patientscell phone use was measured using a questionnaire. Thetwo groups use of cell phones was similar.

Study 2 An Australian study (1997) conducted a study with 200transgenic mice. One hundred were exposed for two 30minute periods a day to the same kind of microwaves with

roughly the same power as the kind transmitted from a cellphone. The other 100 mice were not exposed. After 18months, the brain tumor rate for the exposed mice was twiceas high as that for the unexposed mice.


74/96


75/96

?

Why do the results of different medicalstudies sometimes disagree?

Could the second study be performedon human beings?


76/96

Evaluarea faptelor

Suppose a friend recently diagnosedwith brain cancer was a frequent cell

phone user. Is this strong evidence thatfrequent cell phone use increases thelikelihood of getting brain cancer?

Informal observations of this type are called _____________ _____________.You should rely on reputable researchstudies, not anecdotes.


77/96

Cum putem obine informaii

Observational StudyThe researcher observes values of the responseand explanatory variables for the sampled subjectswithout imposing any treatmentsExample:

ExperimentThe researcher assigns experimental conditions(also called treatments) to subjects (also calledexperimental units) and then observes outcomeson the response variable.Treatments correspond to values of theexplanatory variableExample:


78/96

Tipuri de studii observationale

RetrospectiveObservational studies that look back in time This is sometimes done to find risk factors for

certain diseases Cross-Sectional

Observational studies that take a cross

section of the population at the current time ProspectiveObservational studies in which subjects arefollowed into the future

Studii Experimentale /


79/96

Studii Experimentale /Observaionale

In an observational study, there can alwaysbe confusion variables affecting the results.

This means that observational studies can _________ show causation. It is easier to adjust for confusion variables in

an experiment. In general, we can study the effect of an

explanatory variable on a response variablemore accurately with an experiment than withan observational study.


80/96

Dezavantaje ale Experimentului

They can be ____________ to perform on thesubjects in which you are interested.

It can be difficult to monitor subjects to ensure thatthey are doing what they are told.

They can take many years, even decades, tocomplete.

Results of experiments that use animals do not ______________ to humans.

They are unnecessary the question of interest doesnot involve trying to assess _____________.


81/96

Design al colectrii datelor pentru

studii observaionale Simple Random Sampling (SRS) A simple random sample of n subjects from a population isone in which each possible sample of that size has the

_______ chance of being selected.


82/96

Stratified Sampling A stratified random sample divides the population intoseparate groups, called ________, and then selects an SRSfrom each stratum.

Colectarea datelor n studiiobservaionale

l d l d


83/96

Cluster Sampling A cluster random sample can be used if the target populationnaturally divides into groups, each of which is representative of theentire target population. In this method, a SRS of groups (or strata)is taken. Every member of the selected groups is put into the

sample.

Colectarea datelor n studii

observaionale


84/96

Systematic Sampling A systematic sample selects every k th

person from the sample frame. Theresearcher randomly selects a number between 1 and k in order to know whichperson to select first, then selects every k th

person after this.

Colectarea datelor n studii observaionale


85/96

Avantaje ale colectrii

Simple Random Sampling (SRS)It is the easiest most widespread form of sampling.Each subject has an _______ chance to be in thesample.The sample enables us to determine how likely it isthat descriptive statistics (like the sample mean)fall close to corresponding values for which wewould like to make inference (like the populationmean).


86/96

Stratified SamplingIt ensures that there are enough _________ ineach group that you want to compare.

Cluster SamplingIt does not require a sampling frame of subjects.

It is less ___________ to implement.

Avantaje ale colectrii


87/96

Eroarea n colectarea datelor

A sampling method is _________ if The sample tends to favor some parts of thepopulation over others.In other words, the results from the sample are notrepresentative of the population.

Obviously, __________ samples are our

goal.


88/96

Tipuri de erori

UndercoverageOccurs when a sampling frame leaves out some groups inthe population

Nonresponse biasOccurs when some sampled subjects cannot be reached,refuse to participate or fail to answer some questions

Response biasOccurs when the subject gives an incorrect response or when the question wording or the way the interviewer asksthe questions is confusing or misleading


89/96

Elemente pentru un studiu corect Control group

Gives us something to compare againstEnables us to control the __________ _______ The placebo effect occurs when patients seem to improve

regardless of the treatment they receive. RandomizationEliminates ______ that can result when researchers assigntreatments to the subjectsBalances the group on variables that you know affect theresponseBalances the group on _________ variables that may beunknown to you


90/96

BlindingIncreases reliability of the results

_________-blind: subjects do not know thetreatment assignment _________-blind: neither the subjects nor those

in contact with the subjects know the treatment

assignment

Elemente pentru un studiu corect


91/96

Exemplu A pharmaceutical company has developed a new drug for

treating high blood pressure. To determine the effectiveness of the drug, the company conducted an experiment in whichsubjects with a history of high blood pressure were treated with

the new drug. A later experiment randomly divided subjects with a history of

high blood pressure into two groups. Group A was treated withthe new drug as before. Group B received the most popular drug on the market at that time. The subjects were unaware of which treatment they received. 60% of the patients in Group Aimproved, while 63% of the patients in Group B improved.

The __________ experiment is better because


92/96

Exemplu To investigate whether antidepressants help smokers to quit smoking,

one study used 429 men and women who were 18 or older and hadsmoked 15 cigarettes or more per day in the previous year. They wereall highly motivated to quit and in good health. They were assigned toone of two groups: one group took an antidepressant called Zyban,while the other group did not take anything. At the end of a year, thestudy observed whether each subject had successfully abstained fromsmoking.


93/96

Studii comparative randomizate

Randomization ensures that the groups of subjectsare similar in all respects before the treatments areapplied.

Using a control group for comparison ensures thatexternal influences operate equally on both groups. If the groups are large enough, natural differences in

subjects will average out.

This means that there be little difference in the resultsfor the groups unless the treatments themselvesactually cause the difference.


94/96

Studii i erori Types of studies:

Observational studies and experimentsExperiments control for lurking variables

Sampling designs:SRS, stratified random samples and cluster samplesSRS is the preferred method

Potential sources of bias:UndercoverageResponse biasNonresponse biasConvenience samplingVoluntary response sampling

Elements of good experiments:Control group, randomization and blinding


95/96

Precauii If a group is underrepresented in the sample, we cannot

make inference about it.

We must be careful when interpreting the results of

observational studies. For comparison of several treatments to be valid, you must

apply all treatments to similar groups of experimental units.

Interesting questions are usually pretty tough to answer.This is due in part to the fact that no single experiment or observational study can determine causation.


96/96

Stop and Think!!! Write the study!

Describe & classify thevariables. Instruments for measure?

Bias?

Prepare to analyze data!

Date post:	04-Apr-2018
Category:	Documents
Upload:	anda-madalina-zaharia
View:	222 times
Download:	1 times

Curs 3 CB Metodologia Cercetarii 2008

Documents