Mahalul, unfortunately the terms "predictor" and "risk factor" are used interchangeably in epidemiology. But i think the term "predictor" makes more sense in prediction modeling, while  the term "risk factor" may be more relevant for questions of etiology. 

Mahalul, unfortunately the terms "predictor" and "risk factor" are used interchangeably in epidemiology. But i think the term "predictor" makes more sense in prediction modeling, while  the term "risk factor" may be more relevant for questions of etiology. 

And to add to Christophers' answer. If studying etiology you will probably want to adjust for all known confounders. However, if studying prediction, you probably want to be more conservative  in adjusting for confounding variables or you might even want to run your prediction models unadjusted

Technically there is none. Regarding the statistics the more proper terms would be independent variable (aka risk factor or predictor). It's context related, if you call an independent variable a predictor or a risk factor (see Christophers answer).

Although they are often used synonymously as Christopher mentioned, there is a slight difference in epidemiological studies. Predictor variables are those that are significantly associated with a particular outcome whether they are beneficial or risk fact

Risk factor is an event, circumstance or characteristic that is present in a subject, that is common in sufferers of a particular disease. A predictor is a circumstance, characteristic or event that occurs while an action is taking place, that favors one particular outcome (positive or negative).For example when I do an experimental work about lets say examining a sample or a drug type sometimes I predict using one experimental method rather than the other (but I don't really really know which one is the best to use.. that was in the first year), so here you can see my results at the end of the work are not really perfect (that sometimes due to me limited experience). But when I know that there is a high risk of damaging the sample or the drug I am testing when using that experiment I will not be using it.

https://www.google.com.tw/url?sa=t&rct=j&q=&esrc=s&source=web&cd=8&cad=rja&uact=8&ved=0ahUKEwjzysqU6pjQAhWGopQKHSdTAM4QFghaMAc&url=http%3A%2F%2Fwww.ncbi.nlm.nih.gov%2Fpubmed%2F18854808&usg=AFQjCNFYrVe8EH5kSVvLBYVKw52Sz3AJ8Q

https://www.google.com.tw/url?sa=t&rct=j&q=&esrc=s&source=web&cd=12&cad=rja&uact=8&ved=0ahUKEwiymNOP65jQAhWCf7wKHVImAo04ChAWCCAwAQ&url=http%3A%2F%2Fjac.oxfordjournals.org%2Fcontent%2F50%2F3%2F375.long&usg=AFQjCNE1Z-rHRz5w-G7-ZJ9fyCmWq-h84w

I think the two terms are interchangeable for statistical purposes, as Martin Eichler noted, but solely under this respect; their practical significance  is deeply different.

A good predictor of an adverse effect (the value of an exposure to a defined agent or mix of agents, an exposure circumstance such as "working in the shoe industry" , a descriptor of a possibly relevant personal habit like the number of cigarette smoked each day or the assumption of fruit and vegetables , the value of a single parameter in blood or urine or a set of such parameters, and so on) shows a defined (positive or negative) correlation with the occurrence of this effect, in the better cases with a recognizable dose vs response relationship: but not necessarily a good predictor is the real risk (or protective) factor.

A good predictor could be just a proxy for a causative, poorly known exposure (e.g. urinary 1-OH-pyrene for overall exposure to PAH's),  a substitute of a causative, poorly known exposure  (e.g., looking for lung cancer risk, the presence and extension of pleural plaques standing for overall asbestos exposure)  or simply a condition that, in the studied population, is not causally associated to the really causative exposure (e.g. high social status with uterus' neck cancer risk). 

I hope this be useful to you, tell if you need more.             

Agree with others above, it is often used interchangeably and a predictor is often used in predictive models in prognosis studies for example. Risk factors in the end are the same  in practice, things that increase the probability of a particular condition, i.e have a potential causative role. The predictor may not be causative but is often associated with a particular condition. 

Agreed with all. Predictor is used with regards to an outcome of a disease while risk factor is talked about in the etiology of the disease..

The term predictor is often mis-used and misleading and I would suggest to avoid it when you can - except for real prediction studies e.g. risk equation models.

Even the term risk factor could be misunderstood as it suggests causality and I would use a more neutral term such as "associated variable" or "independent variable" as suggested above.

best

Christophe

As per my opinion the term risk factor is used for etiology of diseases and predictor is used for outcome of a diseases

Agree with all; to add in simple terms ; risk is preconceived especially in case control or even in cohort studies and what we do is reaccertain the association using RR; predictors terminology becomes more relevant in a descriptive or longitudinal study wherein several factors are assessed, confounding is adjusted and maybe population specific predictors identified. However i agree that somewhere the causality assessment makes these terms to be used interchangeably and perhalps depending on the study design and objectives.

 Great answers above. Another concept to consider- "risk" has a negative connotation. Predictors can be for good or bad outcomes/events.

In current life studies, ongoing right now in the US political sphere...poor election predictors have led to an extremely high probability of negative risk factors - not just domestically, but unfortunately, globally. I don't think we'll need to run any stats on these though...its pretty universally accepted.

The predictive factors allow to anticipate the occurrence of an event, while the risk factors allow to establish a condition, characteristic or exposure that increases the probability of suffering an event.

 A risk factor can be postulated a cause of a disease, a prognostic factor or a predictor factor. A prognostic factor is a risk factor identified on the occasion of a prognostic study. Predictor factor influences the course of  a disease when this factor is associated with the administration of  a treatment.

Literally spoken the answer of Margarita is a good description, technically spoken I would´t see much difference between the two terms.

Dear sir,

i am attached the below link to download the books. 

H Hu, K Kurotani, N Sasaki… - BMC public …, 2016 - bmcpublichealth.biomedcentral.com

A Eranti, AL Aro, T Kenttä, A Holkeri… - Scandinavian …, 2016 - Taylor & Francis

HY Choi, DJ Joo, MK Song, MS Kim, HC Park, YS Kim… - Medicine, 2016 - journals.lww.com

Good Luck

A risk factor is an exposure, an attribute, an element or a condition which is associated with change (mostly increase) in the risk of disease occurrence or mortality. A risk factor is either modifiable that is, changeable  with intervention like smoking, body weight,  lifestyle etc,  or non-modifiable like aging, sex and others. Risk factor may be used as predictor of  future changes in health. Predictors may be risk factors but may be not. These are factors which can be utilized to predict(anticipate) the probability of occurrence or the probability of existence  of disease or features of disease or prognosis of disease. Not all predictors are risk factors. 

Great contributions from researchers and experts; kudos!

Nice to hear from you all. All variables are risk factors for different outcomes. When you run a bivariate analysis on risk factors, you may pick some predictor variables which needs to be analyzed further under multivariable analysis to remove the confounders and some modifiers. Risk factors which survive whole this process becomes the real predictors of an outcome.

Great thanks to all for the valuable contribution and brain storming; the terms are now clear to all.

Regards

Hi Mahalul,

From an epidemiological perspective, their's no major difference between the two terms. However, the specific context may influence the term used.

Regards,