What you want is Cox regression (proportional hazard) models to perform survival analysis. It is years since I have used SPSS and would encourage consideration for the use of R (Python is a bit more complicated). There is a commercial go-between which is easier to use than R but more computationally efficient than SPSS, but I cannot remember its name at all.
I have no idea whether the information is correct, but it looks simple enough to point and click.
Background
A proportional hazard model is one of numerous survival analysis which model unit risk, e.g. mortality, over time considering the risk e.g. mortality of the baseline population (covariates). Generally a hazard model will comprise a hazard function of the risk of events per unit time against the baseline covariates, such as age, smoking that sort of thing. Thus you wouldn't want to consider the mortality risk of a 25 year old to comparable to that of a 90 year old patient for a given factor, because there's loads of confounders.
At a minimum SPSS could be slow and this is a regression analysis, but you will find out soon enough.
It appears there's a forward prediction being sought here. A simple regression may be able to capture this, but the gist of the question appears to be a precise temporal forward prediction. Precise forward prediction are complicated and whilst there are definitely R packages that do this, at a precision level of forecast e.g. 2 vs 5 year this definitely ain't SPSS, a long way from its capabilities. The OP will need to learn R to do this study, it isn't trivial, or approximate from a hazard style regression. For this style of calculation significant epidemiological and clinical information is required. For example, these are often used in cancer studies, but precisely what the background epi/clinical is will be important as to whether a given R library is relevant.
Summary The answer to the question is SPSS can't do this and learning R is the way forward, because the OP will need to use an established library and this only place I'm aware that such libraries exist. The only alternative is to approximate using classical survival analysis.
What you want is Cox regression (proportional hazard) models to perform survival analysis. It is years since I have used SPSS and would encourage consideration for the use of R (Python is a bit more complicated). There is a commercial go-between which is easier to use than R but more computationally efficient than SPSS, but I cannot remember its name at all.
I have no idea whether the information is correct, but it looks simple enough to point and click.
BackgroundA proportional hazard model is one of numerous survival analysis which model unit risk, e.g. mortality, over time considering the risk e.g. mortality of the baseline population (covariates). Generally a hazard model will comprise a hazard function of the risk of events per unit time against the baseline covariates, such as age, smoking that sort of thing. Thus you wouldn't want to consider the mortality risk of a 25 year old to comparable to that of a 90 year old patient for a given factor, because there's loads of confounders.
At a minimum SPSS could be slow and this is a regression analysis, but you will find out soon enough.
It appears there's a forward prediction being sought here. A simple regression may be able to capture this, but the gist of the question appears to be a precise temporal forward prediction. Precise forward prediction are complicated and whilst there are definitely R packages that do this, at a precision level of forecast e.g. 2 vs 5 year this definitely ain't SPSS, a long way from its capabilities. The OP will need to learn R to do this study, it isn't trivial, or approximate from a hazard style regression. For this style of calculation significant epidemiological and clinical information is required. For example, these are often used in cancer studies, but precisely what the background epi/clinical is will be important as to whether a given R library is relevant.
Summary The answer to the question is SPSS can't do this and learning R is the way forward, because the OP will need to use an established library and this only place I'm aware that such libraries exist. The only alternative is to approximate using classical survival analysis.
Hi there! thank you for your response. You are absolutely right, it's as easy as that to perform survival analysis. However, this results in a comparison in overall survival, whereas I'm looking specifically for a 2-years survival analysis but I just can not seem to find out how to do this.More
What you want is Cox regression (proportional hazard) models to perform survival analysis. It is years since I have used SPSS and would encourage consideration for the use of R (Python is a bit more complicated). There is a commercial go-between which is easier to use than R but more computationally efficient than SPSS, but I cannot remember its name at all.
There is a weblink for using Cox regression in SPSS: https://www.scalestatistics.com/cox-regression.html
I have no idea whether the information is correct, but it looks simple enough to point and click.
Background A proportional hazard model is one of numerous survival analysis which model unit risk, e.g. mortality, over time considering the risk e.g. mortality of the baseline population (covariates). Generally a hazard model will comprise a hazard function of the risk of events per unit time against the baseline covariates, such as age, smoking that sort of thing. Thus you wouldn't want to consider the mortality risk of a 25 year old to comparable to that of a 90 year old patient for a given factor, because there's loads of confounders.
At a minimum SPSS could be slow and this is a regression analysis, but you will find out soon enough.
It appears there's a forward prediction being sought here. A simple regression may be able to capture this, but the gist of the question appears to be a precise temporal forward prediction. Precise forward prediction are complicated and whilst there are definitely R packages that do this, at a precision level of forecast e.g. 2 vs 5 year this definitely ain't SPSS, a long way from its capabilities. The OP will need to learn R to do this study, it isn't trivial, or approximate from a hazard style regression. For this style of calculation significant epidemiological and clinical information is required. For example, these are often used in cancer studies, but precisely what the background epi/clinical is will be important as to whether a given R library is relevant.
Summary The answer to the question is SPSS can't do this and learning R is the way forward, because the OP will need to use an established library and this only place I'm aware that such libraries exist. The only alternative is to approximate using classical survival analysis.
What you want is Cox regression (proportional hazard) models to perform survival analysis. It is years since I have used SPSS and would encourage consideration for the use of R (Python is a bit more complicated). There is a commercial go-between which is easier to use than R but more computationally efficient than SPSS, but I cannot remember its name at all.
There is a weblink for using Cox regression in SPSS: https://www.scalestatistics.com/cox-regression.html
I have no idea whether the information is correct, but it looks simple enough to point and click.
BackgroundA proportional hazard model is one of numerous survival analysis which model unit risk, e.g. mortality, over time considering the risk e.g. mortality of the baseline population (covariates). Generally a hazard model will comprise a hazard function of the risk of events per unit time against the baseline covariates, such as age, smoking that sort of thing. Thus you wouldn't want to consider the mortality risk of a 25 year old to comparable to that of a 90 year old patient for a given factor, because there's loads of confounders.
At a minimum SPSS could be slow and this is a regression analysis, but you will find out soon enough.
It appears there's a forward prediction being sought here. A simple regression may be able to capture this, but the gist of the question appears to be a precise temporal forward prediction. Precise forward prediction are complicated and whilst there are definitely R packages that do this, at a precision level of forecast e.g. 2 vs 5 year this definitely ain't SPSS, a long way from its capabilities. The OP will need to learn R to do this study, it isn't trivial, or approximate from a hazard style regression. For this style of calculation significant epidemiological and clinical information is required. For example, these are often used in cancer studies, but precisely what the background epi/clinical is will be important as to whether a given R library is relevant.
Summary The answer to the question is SPSS can't do this and learning R is the way forward, because the OP will need to use an established library and this only place I'm aware that such libraries exist. The only alternative is to approximate using classical survival analysis.
More
VOTE
VOTE