SPSS tutorials for beginners part 5 - Kaplan Meier, Cox regression & calculating follow-up time

In this SPSS tutorial you will learn how to calculate your own follow-up time. Also, I will teach you how to use Kaplan Meier analyses & the Log Rank test, and Cox Regression analyses. We will both be looking at univariate and multivariate analyses.
About me:
I am a registered clinical epidemiologist and working as a fellow in medical oncology in the Netherlands. I have published over 50 manuscripts and have received several large research grants. I am particularly interested in research in geriatric oncology and am a an active member of the International Society for Geriatric Oncology.
You can find more information about my work on my linkedin page:
www.linkedin.com/in/nienke-de-glas
And here is my full bibliography:
pubmed.ncbi.nlm.nih.gov/?term...
Disclaimer:
Views and opinions are my own. Examples from clinical research will always include either my own work, or previously published research. I will include references in the description box.

[@alexanderchrysos6366]

Very educational video, with clear instructions. Thank u Nienke!

[@melissapompilius9299]

Thank you for this and your other excellent videos. They are so helpful for learning more about the statistics discussed and also how to get the most from SPSS!

[@lakshitharanasinghe3896]

Great video. This included all the info that I needed. Thank you.

[@user-eu8mh3tq4v]

Thank you so much for this video. Your directions are totally clear and helped me a lot! I did my thesis analysis based on your video.

[@asmikul1]

Thank you so much for all the information regarding analsyis all the lectures are really helpful.

[@johnsonmshiu4837]

Thanks for the great video. You have saved my life

[@douglasanguloherrera8366]

Excellent class. Very detailed. Greetings from Venezuela.

[@Muuip]

Great concise presentation! Much appreciated!👍

[@nienkedeglas_mdphd]

Thanks a lot!

[@rupasingh5619]

I was having hard time to run the Kaplan Meier...Thank you very much for the well detailed explanation!

[@engelbertsyperillo]

This is a very helpful video! Thank you!!!

[@shafirehman4109]

May all be blessings with you! This is awesome and the best presentation of SPSS. Thanks

[@Duhaiby]

Agreed

[@muhammedhadedy4570]

Great video. I already started following your KZbin channel. Please keep up the great work and continue this beautiful series. Thank you so much.

[@nienkedeglas_mdphd]

Thanks a lot for your positive feedback!

[@maurisioviemmo118]

perfect thank you Dr. you are my angel

[@zr8282]

Thank you for a wonderful informative video.

[@audrajakobi7257]

Thank you so much! This was exactly what I needed!!!

[@nienkedeglas_mdphd]

you are welcome!

[@TheyfwLena]

Super clear and helpful, thank you very much!

[@mrcx6142]

Brilliant video! I'm just sitting in front of SPSS trying to do the statistics for my thesis properly and this helped a ton.

[@nienkedeglas_mdphd]

so happy to hear that, thanks for the feedback :)!

[@jonrrobinson]

Very informative, clear and concise, thank you very much 👍

[@nienkedeglas_mdphd]

you are welcome :)!

[@rafaelescate3618]

Thank you very much Nienke

[@nelly4147]

It is very helpful. ❤ Thank you a lot!

[@maryanka5465]

Thank you very much! Your videos really helped me!!

[@nienkedeglas_mdphd]

You are very welcome!

[@adunkz]

thank you for the information, it is very helpful 🤝

[@anettesandstrom64]

Thank you for a very clear explanation. The best I have found so far:) I would also be interested in knowing more about how to test proportional hazards. Looking forward to other videos from you.

[@nienkedeglas_mdphd]

thank you for your suggestion! I may do a follow-up video in the future on this topic.

[@AnilGuptadr]

thank you, easy to follow tutorial.

[@RawaMuhsin]

Thank you for the clear and simple explanation. I am working on a paper right now and this helped me understand a few more concepts. Will surely check out your other videos. Keep up the good work!

[@nienkedeglas_mdphd]

Thanks a lot for your feedback, great to hear that is is helping :)!

[@wiranugraha1114]

Great video, Thank you so much

[@hajiaman1900]

Thank you! It was a nice video.

[@kcbekenov735]

Thank you , it is very helpful ☺️

[@badermouna8202]

thank you, very helpful

[@tommygogh4243]

Thank you for your insight full video's Nienke! I really like your video because you also show your way of doing the analysis and things to look out for. Also the fact you show how to do it on SPSS. Maybe you could also tell us more, like what a hazard ratio means and what assumptions you're making when applying cox regression (cox proportional hazard assumption). Also, what to do with censored cases and how cox regression deals with that.

[@nienkedeglas_mdphd]

thanks a lot for your suggestions! I will definitely take those into account in followup videos, this was so far a very basic introduction but you are completely right that these factors deserve further explanation.

[@sohaildhanji6398]

This was amazing. Thank you

[@nienkedeglas_mdphd]

You are very welcome :)!

[@skavenos]

Thank you so much!

[@ghazal3340]

Best video!!!

[@dmitryvoznuk8344]

Thanks for the great video! It would be amazing if you can explain how to check assumptions behind Kaplan-Meier and Cox regression

[@nienkedeglas_mdphd]

Thanks for the suggestion, might do so in an upcoming video. This video was more meant to learn how to perform the analyses from a practical perspective but I will definitely look into it!

[@juncarlotorres6822]

thankyou for this,

[@SoulnMatter]

Excellent

[@YaSa__]

very useful video. It would be perfect if could you talk about plot interpretation as well.

[@khyatibhatia2552]

very useful video

[@abderrahmaneboumedienederk5293]

Wonderful explanation, can you share the data set used in the demonstration

[@chineduiwu522]

Great video, thank you so much. Please could you explain how to check the PH assumptions using SPSS? also can Schoenfeld’s test of residuals be done with SPSS?

[@naungcho9096]

Very helpful

[@aysenurkeske5463]

Thank you so much for the extremely helpful video! Do you take patients who are lost to follow up into account in Kaplan Meier analysis? Is there a way to define LTF as a category in SPSS different than "non-dead"?

[@noiyar338]

Thank you for your detailed explanations. Could you tell me how to get number at risk table in SPSS?

[@saik9980]

Great video and presentation...Thank you!!!! , when do you use 1 minus survival option

[@niceboowy]

nice video. thank you so much. Which data output by SPSS should I use to draw the ROC curve in Cox regression?

[@gobind5212]

Hi, great video. Can you explain how to do a time point survival analysis? Specifically, let’s say I am interested in determining overall survival of my cohort at 3 years. How would I do this?

[@sinank2308]

Hello, thanks a lot!! I have a question for cox regression. I have only one variable (non-categorical), 'change from previous step' p value is 0.02 but p value in the equation is 0.42. It is a bit confusing since ı have only one variable. Which one should I pick? Thank you for your time.

[@user-rd4ge7uy5d]

Hi, thank you so much for your explanations - they are very helpful! After conducting my analysis, I am wondering how to calculate the power of the log-rank test (and thus, the type II error), especially when comparing more than two groups. Do you know and could help us out? Best to you! Karla

[@turqueer6704]

Thank you for this very helpful video! Is it possible to use non-categorical , continuous variables as covariants in my analysis? How can I perform this?

[@ioannispatras6324]

Fantastic video! How would one convert days into months?

[@merveteker3330]

can we do the graphics in spss of hazard ratios of dfs or os as comparison two different type of treatments?

[@gedionasnake1946]

Could you please give me some advise on how to calculate incidence rate of loss to follow-up

[@fabrizioturrini5479]

Hi thank you very much for your brilliant video(s). They are very helpful to me. I have some questions for you regarding Cox analyses with SPSS. 1) How many covariates can you put into analyses in order not to lose power? 2) How can you choose the right covariates? Do you perform a univariate analyses before, to select the only significant ones? 3) When you run cox analyses with multiple covariates which method should I select? Stepwise? Backward? Wald?... 4) Once you get the result can you use Exp (B) value to build a score for prediction?

[@nienkedeglas_mdphd]

thank you for your questions! I tried to give some answers below: 1. this depends on your sample size and number of events. As a rule of thumb, you need 10 events per covariate that you are adding to the model. 2. this depends on your research question and possible confounders. Please also see my video " adjusting for confounding". 3. just leave it at the default setting 4. this is a whole different topic, I will record a video on how to build prediction models but you can take a look at my playlist on prediction already to get a general idea.

[@fabrizioturrini5479]

@@nienkedeglas_mdphd Thank you sooo much! You are on the top of the list. Very very nice and helpful

[@seckinbilgic]

I don't know which one to choose as an indicator in Cox analysis. Moreover, in the calculation where there are 10 variables, the data changes with each iteration. P values ​​are changing. How can I solve this?

[@mahmutbilalsert9845]

Thanks a lot for a fantastic and a very nice educational video. I have 137 patient spss file and many variables I jest wonder if you could help me to analyse PFS and OS. Thanks.

[@ghzltnsr20]

do we calculate median follow-up from death minus date of diagnosis ... or date of enrolment in study??

[@jeromebuhay5152]

thank you very much! i learned a lot, may i request a sample data set which i will use for practice.

[@nienkedeglas_mdphd]

Thank you for the compliment! Unfortunately I am not able to share my data as they are real patient data. There are likely to be some files that you can find by googling on the internet though.

[@nesalasica]

Thank you for very detailed explanation. If we have a variable, lets say gender (female/male) and do the univariate cox regression, what would be the reference than?

[@nienkedeglas_mdphd]

You can choose this yourself, depending on how you would like to present your result. For example: let's say that the hazard ratio for women versus men would be 2.0, this would have been the opposite if you would change the reference category to men (meaning a hazard ratio of 0.5). For gender, it probably does not matter much, but for example for age categories it usually makes most sense to take the lowest (meaning youngest) category as a reference.

[@TheRainbow7227]

Hi, thanks for the great Video, its very helpful, but I have a problem with calculating my variable, it shows an Error about the Targetvariable name, and the calculation can not be made, thanks again

[@rongkagornchuntamongkol4187]

hey thanks so much! i have a problem with my data please. Wondering if you could give me advice please? my survival straight lines down to zero at the end date of my data - as if it says every one died on the same day (who survived).

[@nienkedeglas_mdphd]

Sorry for replying so late! Could there be an error in your data? It sometimes helps if you look at your life tables, there you can see exactly what goes wrong and at which timepoint.

[@laxmanbisht2638]

Nice video

[@nienkedeglas_mdphd]

Thanks!

[@asa7832]

Hi, how to calculate 1-year, 5-year and 10-year survival rate?

[@lisannestienstra8729]

Slay queen 🔥

[@luizrobertopaezdib1965]

Thank's for the content. Can i adress de hazard ratio in terms of %? If yes, would be proper to say "the hazard increases in 1,267%"? or "in 126,7%"? Thank's again!

[@nienkedeglas_mdphd]

this is not very common to do so, it is better to present the hazard ratio as it comes (so for example 1.26). If you want to translate this into percentages, you should think about the difference between relative risks and absolute risks, this is often forgotten.

[@RawaMuhsin]

Can you please explain (or direct me to another resource where it explains) how to choose covariates? I am working on a paper, and depending on how many and what covariates I choose, I get different p values for the individual components.

[@nienkedeglas_mdphd]

Thank you for your question! I would recommend to watch my video's on confounding factors. Essentially, you should come up with a good assumption on which covariates play a role in the association that you want to study, and not look at p-values. I explained this in more detail in my previous videos. Good luck! You can find these videos here: - kzbin.info/www/bejne/jnmvc2mrmK2Sh5o - kzbin.info/www/bejne/g17YoXpjacR4r9E

[@RawaMuhsin]

@@nienkedeglas_mdphd Thanks a lot for answering. I will check them out.

[@niceboowy]

great video. Thank you!! Which data output by SPSS should I use to draw the ROC curve in Cox regression?

[@nienkedeglas_mdphd]

Hi! For Cox it is a bit more complicated to make ROC curves. If you censor your follow-up at a certain time and define your outcome as mortality within time period (yes/no), you can make a ROC curve in SPSS (see my video on this topic) but if you want to incorporate the follow-up time in the ROC-curve, you will need more advanced software such as R.

[@niceboowy]

@@nienkedeglas_mdphd Thank you for your response. I will take your advice and analyze it.

[@alinuh5564]

Many thanks Good example and clear explanation I found number of steps quite intimidating Does spss ver 27 makes it any easier Will check

[@nienkedeglas_mdphd]

I understand that but you will see that once you are working with your own data, it is not as complicated as it may seem. Good luck!!

[@juliasjoberg7173]

Hello! What should I write as "the end of follow up time" for the patients that are dead? Is it the same year they died? Thank you.

[@nienkedeglas_mdphd]

Follow-up time is defined as the time between start of measuring follow-up (for example at the moment of diagnosis of a disease), until either the event of interest (in your case indead death), or loss to follow-up.

[@humbertorodriguez3336]

Hi Nienke, how can I obtain a Survival Table when doing COX regression on SPSS ? I want to get 5 year OS and 1 year survival from the table. Could you please orient me

[@esan120au]

09:05 15:55

[@halolhaha1]

GLAS helder

[@SoulnMatter]

Fantastic! However your speed is too fast😅