Not all scientific studies are created equal - David H. Schwartz

Nisu sve naučne studije iste - Dejvid H. Švarc (David H. Schwartz)

Prevodilac: Anja Saric Lektor: Mile Živković

Studies have shown that

Studije su pokazale da je

taking vitamins is good for your health

uzimanje vitamina dobro za vaše zdravlje

and bad for your health.

i loše za vaše zdravlje.

That newly discovered herb can improve your memory

Novootkrivena lekovita biljka može da poboljša vaše pamćenje

or destroy your liver.

ili da uništi vašu jetru.

Headlines proclaim a promising new cancer treatment

Naslovi obljavljuju obećavajući novi lek za rak

and never mention it again.

i posle ga više ne pominju.

On a daily basis,

Svakodnevno nas

we are bombarded with attention-grabbing news,

bombarduju vestima koje privlače pažnju

backed up by scientific studies,

i koja su podržana naučnim studijama,

but what are these studies?

ali šta su te studije?

How are they performed?

Kako su one izvedene?

And how do we know whether they're reliable?

I kako da znamo da li su pouzdane?

When it comes to dietary or medical information,

Kada je reč o dijetetskim ili medicinskim informacijama,

the first thing to remember

prva stvar koju treba zapamtiti

is that while studies on animals or individual cells

je da iako studije nad životinjama ili pojedinačnim ćelijama

can point the way towards further research,

mogu dati smernice ka budućem istraživanju,

the only way to know how something will affect humans

jedini način da znamo da će nešto imati efekta na ljudima

is through a study involving human subjects.

je kroz studiju koja obuhvata ljudske subjekte.

And when it comes to human studies,

A kada je reč o ljudskim studijama,

the scientific gold standard is

zlatno pravilo naučnika je

the randomized clinical trial, or RCT.

nasumična klinička studija, ili NKS.

The key to RCTs is that the subjects are randomly assigned

Ključ NKS-a je u tome da su subjekti nasumično određeni

to their study groups.

u svoje ispitivačke grupe.

They are often blinded to make them more rigorous.

Često su zaslepljeni, kako bi bili rigorozniji.

This process attempts to ensure

Ovim procesom pokušavamo da osiguramo

that the only difference between the groups

da je jedina razlika između grupa

is the one the researchers are attempting to study.

ona koju naučnici pokušavaju da izuče.

For example,

Na primer, kod testiranja

when testing a new headache medication,

novog leka protiv glavobolje,

a large pool of people with headaches

velika grupa ljudi sa glavoboljom

would be randomly divided into two groups,

je nasumično podeljena u dve grupe,

one receiving the medication

od kojih jedna prima lek,

and another receiving a placebo.

a druga placebo.

Ako je nasumična podela dobro izvedena,

With proper randomization,

the only significant overall difference

jedina značajna razlika

between the two groups

između dve grupe

will be whether or not they received the medication,

biće u tome da li uzimaju lek ili ne,

rather than other differences that could affect results.

umesto drugih razlika koje bi mogle da utiču na rezultat.

Randomized clinical trials are incredible tools,

Nasumične kliničke studije su neverovatni instrumenti,

and, in fact, the US Food and Drug Administration

i zapravo, Administracija hrane i lekova SAD-a

often requires at least two to be conducted

često zahteva bar dve ovakve studije

before a new drug can be marketed.

pre nego što dozvoli da se novi lek izbaci na tržište.

But the problem is that an RCT is not possible

Problem je što NKS nije moguće izvesti

in many cases,

u mnogim slučajevima,

either because it's not practical

ili zbog praktičnih razloga,

or would require too many volunteers.

ili zato što bi zahtevala previše volontera.

In such cases,

U takvim slučajevima,

scientists use an epidemiological study,

naučnici izvode epidemiološke studije,

which simply observes people going about their usual behavior,

koje prosto posmatraju ljude kroz njihovo uobičajeno ponašanje,

rather than randomly assigning active participants

bez nasumičnog deljenja aktivnih učesnika

to control invariable groups.

u kontrolne nepromenljive grupe.

Let's say we wanted to study

Recimo da želimo da istražimo

whether an herbal ingredient on the market

da li lekovita biljka na tržištu

causes nausea.

izaziva mučninu.

Rather than deliberately giving people something

Umesto da dajemo ljudima nešto

that might make them nauseated,

od čega bi im pripalo muka,

we would find those who already take the ingredient

pronašli bismo one koji već uzimaju sastojak

in their everyday lives.

u svom svakodnevnom životu.

This group is called the cohort.

Ova grupa naziva se kohorta.

We would also need a comparison group

Takođe bi nam trebala grupa za poređenje

of people who do not have exposure to the ingredient.

sačinjena od ljudi koji nisu izloženi sastojku.

And we would then compare statistics.

I onda bismo poredili statističke rezultate.

If the rate of nausea is higher in the herbal cohort,

Ako je nivo mučnine viši u kohort grupi,

it suggests an association

to sugeriše povezanost

between the herbal supplement and nausea.

između lekovite biljke i mučnine.

Epidemiological studies are great tools

Epidemiološke studije su odlični instrumenti

to study the health effects of almost anything,

za izučavanje zdravstvenih efekata skoro bilo čega,

without directly interfering in people's lives

bez direktnog mešanja u ljudske živote,

or assigning them to potentially dangerous exposures.

ili njihovog izlaganja potencijalno opasnim stvarima.

So, why can't we rely on these studies

Zašto ne možemo da se oslonimo na ove studije

to establish causal relationships

prilikom uspostavljanja veze

between substances and their effects on health?

između supstanci i njihovih efekata na zdravlje?

The problem is

Problem je u tome

that even the best conducted epidemiological studies

što čak i najbolje epidemiološke studije

have inherent flaws.

imaju inherentne mane.

Precisely because the test subjects

Baš iz razloga što subjekti

are not randomly assigned to their groups.

nisu nasumično podeljeni u grupe.

For example, if the cohort in our herbal study

Na primer, ako se kohort u našoj studiji sa lekovitim biljem

consisted of people who took the supplement

sastoji od ljudi koji uzimaju suplement

for health reasons,

zbog zdravstvenih razloga,

they may have already had higher rates of nausea

oni možda već imaju višu stopu mučnine

than the other people in the sample.

od drugih ljudi u uzorku.

Or the cohort group could've been composed of

Ili bi kohort grupa mogla biti sastavljena

people who shop at health food stores

od ljudi koji kupuju u prodavnicama zdrave hrane

and have different diets

i imaju različitie režime ishrane

or better access to healthcare.

ili bolji pristup zdravstvu.

Ovi faktori koji mogu uticati na rezultat,

These factors that can affect results,

in addition to the factor being studied,

pored faktora koji se izučava,

are known as confounding variables.

zovu se zbunjujuće varijable.

These two major pitfalls,

Ove dve najveće zamke, u kombinaciji

combined with more general dangers,

sa opštim opasnostima,

such as conflicts of interest or selective use of data,

kao što su konflikti interesa ili selektivno korišćenje podataka,

can make the findings of any particular epidemiological study suspect,

mogu da naprave sumnjivim rezultate bilo koje epidemiološke studije,

and a good study must go out of its way

i dobra studija mora uložiti napor da dokaže

to prove that its authors have taken steps

da su njeni autori preduzeli sve korake

to eliminate these types of errors.

kako bi eliminisali ovakve greške.

But even when this has been done,

Ali čak i kada je to obavljeno,

the very nature of epidemiological studies,

sama priroda epidemioloških studija,

which examine differences between preexisting groups,

koje istražuju razlike između već postojećih grupa,

rather than deliberately inducing changes within the same individuals,

a ne izaziva namerne promene kod istih pojedinaca,

means that a single study

znači da jedna studija

can only demonstrate a correlation

može samo da prikaže korelaciju

between a substance and a health outcome,

između supstance i zdravstvenog ishoda,

rather than a true cause and effect relationship.

a ne i pravi uzrok i posledice te veze.

At the end of the day,

Na kraju krajeva,

epidemiological studies have served as excellent guides to public health,

epidemiološke studije služe kao odlični vodiči kroz javno zdravstvo,

alerting us to critical health hazards,

opominjući nas

such as smoking, asbestos, lead, and many more.

na zdravstvene opasnosti kao što su pušenje, azbest, olovo i mnoge druge.

But these were demonstrated through

One su prikazane kroz

multiple, well-conducted epidemiological studies,

višestruke, dobro obavljene epidemiološke studije,

all pointing in the same direction.

koje sve pokazuju u istom pravcu.

So, the next time you see a headline

Tako da kad sledeći put vidite naslov

about a new miracle cure

o novom čudotvornom leku

or the terrible danger posed by an everyday substance,

ili strašnoj opasnosti od svakodnevne supstance,

try to learn more about the original study

pokušajte da saznate više o prvobitnoj studiji

and the limitations inherent in any epidemiological study or clinical trial

i inherentnim organičenjima bilo koje epidemiološke ili kliničke studije

before jumping to conclusions.

pre nego što donesete prerane zaključke.