Brilliant presentation!

you did a great job !! Keep it up :)

Thanks man very helpful

Thank you so so so much ,it’s the best video of this topic in KZbin

Its amazing and helpful… thanks you❤️

Too difficult to understand you as you are talking more and irrelevant

Congrats Lucia, really proud of you, Go Lords 😅

Great content! Thank you very much!

Great review

Thank you for this video. Very educative. Im surprised I'm the first to comment in 2 years. More people need to see this

Love it!!!

Mam their is a test GRE is this iss benificll who wanna do futhr study???

kzbin.info/www/bejne/jqDGnIqBeKtrpKM (Car price prediction with multiple models .. Randomforest regression, decion tree regression , linear regression etc 🙂)

Thanks Prof. Felipe Fregni, happy new year, wish you many blessings

Congratulations dear Haneesha your achievement is your hard work. Keep rocking. God bless you

Muchas Felicidades Dr. Marco, Abrazos y Éxitos!.

Felicitaciones al Dr. E. BEDOYA y a todos en general!!!

Here is a question about group allocation: what if instead of using blocks in order to obtain equal groups, we just repeated simple randomization until we got equal groups for the first time? Would that produce biased results?

Very interesting presentation.

I was in trouble when doctor told me that I have been diagnosed with HIV... I though about my children and my wife, I know my children will face a serious problem when I'm gone, I lost hope and I wept all day, but one day I was searching the internet to see if there 's any way I I was in trouble when doctor told me that I have been diagnosed with HIV... I though about my children and my wife, I know my children will face a serious problem when I'm gone, I lost hope and I wept all day, but one day I was searching the internet to see if there 's any way I can be cured, so I found Dr Omini email and contact number, I called him and he guided me, I asked him for solutions and he started the remedies for my health, he told me amount for the herbal remedy, which was a change to me, all i needed was just to get cured, and stay healthy with my family, i Thank God how everything went, now everything is fine with me, I was cured by Dr Omini herbal medicine, I'm very thankful to Dr Omini and am very happy with my wife and children, Do not lose hope too soon, you can Email him @ [email protected] for easy and fast communication you can also call or add him on whats-app with this number 👉+2347016600744 stay bless Dr Omini is also a specialist in cureing 1. HIV/AIDS 2. HERPES 3. CANCER 4. ALS 5. HEPATITIS B 6. DIABETES 7. HUMAN PAPILOMA VIRUS DISEASE(HPV) 8. ALZHEIMER 9. LUPUS (Lupus Vulgaris or Lupus Erythematosus)''''''

Amazing experience, such a great peers and friends I made in PPCR 2020.

Thanks for such an insightful video, I had recently applied for the Bell Postdoctoral Fellowship. It would be great if you could share some informative videos about that Fellowship Programme. What are the criterion for a successful candidature

Excellent menssage.

I am doctor Regan I got effected with Hiv in the process of attending to my HIV patient I tried all I can to get cured but all to no avail, until i saw a post on KZbin about a herbalist man called Dr ukabuo who prepare herbal medication to cure all kind of diseases including HIV virus, at first i doubted if it was real but decided to give it a try, when i contact this herbalist via his email drukabuoherblhome@ gmail. com and he prepared a HIV herbal cure and sent it to me via fed-ex delivery company service, when i received this herbal cure, he gave me step by directions on how to apply it, when i applied it as instructed, i was totally cured of this deadly disease within 10 days of usage, I am now free from the deadly disease called HIV, all thanks to Dr ukabuo. Contact this great herbal spell caster. kindly contact him on +23481 552 241 53 or whatsapp him +234815 52241 53 Dr.ukabuo cure all kinds of sickness or diseases such as: 1. HERPES VIRUS 2. LASSA FEVER 3. GONORRHEA 4. HIV/AIDS 5. LOW SPERM COUNT 6. MENOPAUSE DISEASE 7. EPILEPSY 8. Hypertension 9. CANCER 10. Kidney problems 11. PREGNANCY PROBLEM 12. SHORT SIGHTEDNESS PROBLEM 14. Male menopause 15. Menopause - male 16. Menopause - peri 17. Menstruation problems 18. Mercury Poisoning 19. Migraine 20. Miscarriage1 21. Stroke etc drukabuoherbalhome.weebly.com

Congratulations Abd Amoniam Mohamed Ali Ahmed

Muito orgulhosa de você Róbson Norberto da Silva!

Parabéns minha irmã linda!! Adriana Barrichello

Parabéns Aline Matis

That’s was a great year! It was a pleasure to learn with you all 🙏🏻

John is amazing. This talk was so informative, thank you! 💯

amazing content friend

Thank you, Professor Felipe, for your message. We try to be ambassadors of your knowledge, taking this message to all those trying to be part of research and innovation in the education world.

Thank you Proffesor! It's Rinaldy BTW 😉

Thank you for the answers! They were really helpful.

Great lecture! Thank you so much! In case it helps, for those who want to do it straight in STATA, I replaced the missing data applying the model with the following commands: first I generated the new variable using the STATA command: "gen mom_depMissLinearRegression = mom_depMiss" // After that you can replace the missing values by applying the model (the same formula shown in the lecture, but in STATA): "replace mom_depMissLinearRegression = (9.075023 + (son_school_prop * 0.149823) + (house_org * -0.1325776) + (son_happy * -0.3355) + ( hub_sup * -0.4333)) if (mom_depMissLinearRegression >= .)" . Now you can just run the ttest with "ttest mom_depMissLinearRegression, by(eat_tapioca)" :)

Very thorough presentation. It was really helpful!

Hi, Camila! Why is truncated data important in the survival analysis? It is important to keep them in mind since there exist adequate statistical methods to deal with/impute values from truncated data. Best, John

Hi, Inara! There's no limit for the amount of censored data used in survival analysis. For example, you could follow the occurrence of an event, and decide to finish the study at a given point in time, by the time you decided to finish the study, only p25 of your sample experienced the event you were studying, and everyone else, at that point, would be right-censored. In the end, you were still able to use the information the provided, for the time you followed them. Best, John

Hi, Maria Jose! a) What if the constant proportional hazards over time assumption is not met, could we use survival analysis methods? Important to know that survival analysis methods, is quite a broad term, and includes both descriptive and inferential statistics. In short, yes, if said assumption is not met for a particular scenario, there are other methods you could use - for instance, this is an assumption to take into consideration when using a log-rank test, if not met, you could use Gehan-Breslow-Wilcoxon, and so on. b) Interpretation of HR for continuous variable: these are interpreted as unit increase/decrease; for instance, for every unit increase in x, this increases/decreases the risk of experiencing event by y times. c) Threshold/level for censored data in a trial: there's no limit for the amount of censored data used in survival analysis. For example, you could follow the occurrence of an event, and decide to finish the study at a given point in time, by the time you decided to finish the study, only p25 of your sample experienced the event you were studying, and everyone else, at that point, would be right-censored. In the end, you were still able to use the information the provided, for the time you followed them. Best, John

Hi, Natalia! There's definitely a relationship between survival analysis and sample size -> important to know that when calculating sample size it depends on the primary analysis (the outcome of interest), as in - if you're evaluating/comparing proportions you might think along the lines of a Fisher and/or Chi-square methods for sample size estimation/calculation. However, when considering a time-to-event variable, and obtaining a measure of risk (point estimate) is important or covariate adjustment is needed, then to adequately power your sample size you need to take into consideration that a Cox PH model is the potential best method and calculate sample size accordingly. Best, John

Hi, Paulo! Although we normally use Cox PH models when accounting for covariates, it is important to mention that it does provide valuable information in terms of the effect size (in terms of providing a point estimate such as HR). When you carry out a log-rank test you mostly obtain a p-value which tells you nothing about the effect size and its direction (increase/decrease in risk). Best, John

Hi, Reynie! When having competing risks, like the ones you have mentioned, you could apply a special type of survival analysis known as "competing-risk survival regression" which is a useful alternative in the setting of having one or more competing risks. I think this might be the best way to go (use of the aforementioned method, or any other that might allow accounting for the competing risks) - should you decide to drop patients just because they experience a competing risk, could potentially induce some sort of selection bias, jeopardizing the validity of your results (among other problems). Best, John

Hi, Arthur! Cox PH models are a type of regression analysis that deals with time to event variables. As you have mentioned, it does take into consideration the time it takes to develop a given event (as do other methods, both descriptive and inferential, such as Kaplan Meier curves and log-rank tests, respectively). Important to mention that you can account for covariates (i.e. confounders/event modifiers) when conducting a multivariate analysis; in addition, it allows you to get a point estimate and an idea of the effect size, should you decide to run it as a standalone univariate analysis (for which a log-rank test could also be used but won't provide a point estimate). Best, John

Dear Inara, thanks for your question. I agree with Rui's comment (and John's, via gmail). Indeed, there's no precise limit to the amount of censoring in a survival analysis of a study, since, no matter how many subjects actually suffered an event during the predefined follow-up period, the estimated probability of survival (in your time-to-event curve) will take into account the time elapsed until the censored data for each study subject. In fact, this is one of the main reasons for which censored data is not considered "missing data"!...

Excellent question, dear Inara. Survival analysis can be conducted if all the data are censored.

Dear Reynie, excellent question and opinion. I agree with you that competing risks sometimes happen in survival analysis. Please refer to the paper, Peter C. Austin, et al, Introduction to the Analysis of Survival Data in the Presence of Competing Risks, www.ncbi.nlm.nih.gov/pmc/articles/PMC4741409/#:~:text=Competing%20risks%20occur%20frequently%20in,causes%20is%20a%20competing%20risk. It will totally solve your question with a cardiovascular case study.

Important question, dear Paulo. We use Cox regression in order to assess whether explanatory factors are associated with survival. So, we can use any case if we want to assess whether explanatory factors are associated with survival. However, I suppose the covariates we want to do so can be regarded as really important ones. What do you think?

Great question, dear Natalia. There are complicated and advanced formulas for the calculation. For example, one of the survival analysis formula for sample size calculation requires to assume that survival time follows the exponential distribution, and then get the sample size. Also we can use the command, ‘power cox’ for cox sample size calculation, and ‘power logrank’ for logrank sample size calculation by STATA. Please refer to STATA HP, www.stata.com/features/overview/power-analysis/

Dear Maria Jose, great doubt. Regarding your question 1, if the assumption is not met, we cannot use survival analysis. So, to use survival analysis in that case, we can use an approach to divide the period in order to keep the assumption and conduct the test each. Another advanced method would be using restricted mean survival time. Regarding your question 2, if HR is less than 1, we can interpret the predictor is protective (for example, associated with improved survival) on the contrary the predictor is associated with increased risk, in other words decreased survival when HR is greater than 1. About your question 3, important question, first, we can conduct survival analysis when all the data is censored. So, I suppose we should take the characters of censored data into consideration into judge when we read papers. For example, there are a lot of loss to follow up as censored data, then it is difficult for us to regard the result reliable.

I enjoy her sense of humour 😄 This talk was so informative filled with practical advice. Super helpful! 👍🏼