UCSF MSTP vs Gates Scholarship

[hello_world_doctor]

Hi Everyone!

I'm incredibly fortunate to have received acceptances from both UCSF and the Gates-Cambridge Scholarship for a PhD in Computer Science at Cambridge University (waitlisted UCSF MSTP). If I went to UCSF I would send a strong LOI and likely convert to MSTP later if not accepted initially.

Logistics wise, UCSF only allows one year deferrals so I may have to reapply and retake the MCAT in 3 years after my PhD if I pursue the Gates. I am trying to get UCSF to allow a three year deferral, or the Gates to allow me to switch degree programs (1 year MPhil before going to UCSF) or even having a 2 year Gates deferral and doing a 2-4-2 degree but with the PhD in the UK, but these options seem unlikely. I also received funding from the NIH OxCam scholars program, which would supplement my UCSF MD with 50k a year (basically half of what MSTP funding normally is).

I'm trying to get things to work out between the organizations, but in the good chance that it doesn't I have to choose between these two great programs.

My main goal is to become a physician scientist working on converting bleeding edge computer science breakthroughs into healthcare and biology. Of course, this may change over time as new opportunities arise.

Personally, I think it would be really nice to just go to UCSF and stay near my friends, and work with Atul Butte. But I am afraid it will be closing a really great opportunity for me to grow and become a better physician-scientist. Any advice is greatly appreciated, thank you for reading!

Gates Scholarship/NIH OxCam Scholars

Pros

• I can do the PhD first, then decide whether I want to do the MD afterwards still (more flexibility)
• Dual mentorship between both NIH and Cambridge
• Opportunities to intern at Tech companies due to flexible nature of program
• Travel and explore new countries
• A new perspective on health-tech outside the US, exposure to a different health system

Cons

• Not as rigorous training as in UCSF
• Haven't really found mentors that are as good here as in UCSF
• Not much tech buzz compared to SF
• Would need to reapply to medical school AND retake the MCAT .
• If I go to medical school, overall money is only half of what MSTP programs cover during the MD years (basically the same price as converting to MSTP your 2nd year).

UCSF MSTP

Pros

• I went to Stanford undergrad, so a lot of my friends and potential collaborators are working in the area, would be challenging to maintain relationships while abroad.
• Can work with Atul Butte who pretty much the best in the world for what my research interests are
• Tech buzz so lots of money going into my research field
• Opportunities to build relationships in the tech industry/academics in the area which will be valuable for future career options

Cons

• Slightly longer degree (probably ~1 year)
• PhD is in bioinformatics, which isn't as unique/impressive as something in computer science
• Gates name might open lots of doors/?? Hard to really "gauge" or quantify how much this is though.

Advice from mentors:

• Both my mentors push me toward the Gates, namely because they say its so unique and the name will help open lots of doors.

---

Members don't see this ad.

 

[Maebea]

UCSF for MSTP. Reasons? Better research fit, better connections to tech industry, more efficient path to clinician-investigator career, better career support, and greater ability to develop a network of colleagues in the country where your career will unfold. To be honest, no one much cares what you doctorate is in (computer science vs. bioinformatics), they care about what you produced. Sure, the Gates is prestigious, but it has no advantage in your career over the MD-PhD from UCSF. Do you have any other MD-PhD acceptances?

Regarding the OxCam program, I am not a big fan. PM me if you want to discuss.

 

[dl2dp2]

Congrats. Pretty great options to choose from overall. CS at Cambridge is quite strong, though much more on the theory side. UCSF probably has a stronger bioinformatics department overall. From a career perspective, it may be better to defer 1 year and get a MPhil at Cambridge, build some connections/relationships, take some courses, etc. then go back to UCSF. In particular, I suspect if you defer for a year and kiss up during that year you have a good shot to be admitted straight into the MSTP during the second year. Secondarily, you can always collaborate with people you get to know at Cambridge once you are done. On your resume you'd have a Cambridge MPhil degree and a UCSF MD/PhD degree. This would be a win-win IMO.

Reapply is NOT recommended. That's a no, just get rid of that possibility now. Unless you are really not that interested in MSTP. It's possible that in 2-3 years the pool is much more competitive and you won't be able to get into UCSF. At all.

I don't really understand the NIH OxCam program. Is that program where you go to UCSF but partially funded? Their website says you'd need to be on NIH campus part time. NIH is typically thought to be not great to train as a grad student.


From a career development perspective, UCSF's PhD programs tend to go on the long side. At that level, it's more about what you publish later on and not where you went. But this is a separate issue.

 

[751866]

UCSF for MSTP. Reasons? Better research fit, better connections to tech industry, more efficient path to clinician-investigator career, better career support, and greater ability to develop a network of colleagues in the country where your career will unfold. To be honest, no one much cares what you doctorate is in (computer science vs. bioinformatics), they care about what you produced. Sure, the Gates is prestigious, but it has no advantage in your career over the MD-PhD from UCSF. Do you have any other MD-PhD acceptances?

Regarding the OxCam program, I am not a big fan. PM me if you want to discuss.

No one cares what the PhD is in except that the degree program can make your productivity higher quality. "Bioinformatics" really should be restricted to working in EMRs and data extraction making databases (not excel sheets which people misunderstand to be "databases") with some elements of computer programming and information technology. Computer science will be more rigorous and will likely give the OP better tools to have high quality productivity in the sense that he will have a more intimate understanding how to integrate the technologies, how to create them, and how to more adequately use statistical methodology as most computer science degrees require decent mathematics background and they cover a lot of areas that overlap with statistics.

Bioinformatics is a flashy name without a whole lot of substance in many cases and they get jammed into statistician roles the same as DrPH or Epi without a requisite background.

I don't know much on the specifics of UCSF bioinformatics, however, so without my own knowledge of their program I can only speak in generalities of "bioinformatics" programs. It could be a very strong program and then I agree that the combined MD-PhD from UCSF will look great, but looking strong is different than being strong, so OP may want to investigate further based on detailed career plans.

If possible, the plan by @sluox sounds pretty good because you'd have a more rigorous background from Cambridge that would allow you to make yourself stronger where the UCSF bioinformatics is weaker (likely quantitatively, but again, I have no specific information on UCSF bioinformatics so I could be mistaken).

 

[hello_world_doctor]

It seems that I will have to choose between the two programs, both are strict on the requirements. Any advice would be greatly appreciated

 

[751866]

It seems that I will have to choose between the two programs, both are strict on the requirements. Any advice would be greatly appreciated

I reread a bit of the first post, but can you summarize the current option.

From my reading:
You're admitted to Cambridge PhD in CS.
You're wait listed at UCSF for MSTP; you are not accepted to the MSTP or the MD program at UCSF.

Is this accurate, or have I misunderstood?

 

[hello_world_doctor]

I reread a bit of the first post, but can you summarize the current option.

From my reading:
You're admitted to Cambridge PhD in CS.
You're wait listed at UCSF for MSTP; you are not accepted to the MSTP or the MD program at UCSF.

Is this accurate, or have I misunderstood?

I have been accepted to UCSF's MD only. So am deciding between these two options!

 

[751866]

I have been accepted to UCSF's MD only. So am deciding between these two options!

Man, as much as I think the PhD in CS would be fantastic and actually a rare asset in medicine, I just don't think it's good to turn down an MD acceptance from UCSF (especially if they won't allow a deferral). I think bioinformatics < CS or Stats especially from good programs, but if your goal is to be a physician, I think taking the MD acceptance at UCSF unless you have another good MSTP at another great program.

Only you can make the call, but you need to look at this as if your branch point is medicine vs CS, because you can't count on getting the PhD later in something real like CS and if you go for the PhD in CS, you can't count on another med school acceptance (at least not from a school like UCSF). I will say, however, if you did have to come back to one of them, you're less likely to be able to come back to medicine than CS, in my opinion. I was at a conceptually similar decision before medical school, and you will have to give up much more to get back on track for a med school acceptance and then you'll give up a lot to go through med school. If you wanted to leave medicine later, you're financially more set (assuming you paid off debt and have been saving) and many will look so favorably on the MD from UCSF. And you still may be able to squeeze into the MD PhD at UCSF if you take the MD seat, or you can take CS courses over MS1 summer at the least.

Just my opinion. I think if you're on the fence with medicine, this requires some deep introspection before finalizing. But if you know you love one more than the other, don't put off what you love.

(Cambridge is a fantastic school, too, by the way. I'm just not sure CS will close their doors to you like medicine might if you reapply after having a prior acceptance.)

 

[CheGuevaraMD]

I gotta say, I know that this is a MD forum, but the advice in this thread is pretty bad. Medicine is definitely on its way down in terms of money/prestige, and computer science is still rocketing upward. You'd be insane to pursue a MD (even at a place like UCSF) instead of a PhD in CS. Plus, a UK PhD is only 3 years, so you'd be on the tenure track or at a tech company by your late 20s, while with the MD/PhD route you'd be slogging until your mid 30s and probably drop out of research. Go to Cambridge!

 

[CheGuevaraMD]

Here are software engineer salaries at major companies: https://www.levels.fyi

Note that they are on par with MD salaries from the start, involve accruing no debt, and only go up.

 

[751866]

I gotta say, I know that this is a MD forum, but the advice in this thread is pretty bad. Medicine is definitely on its way down in terms of money/prestige, and computer science is still rocketing upward. You'd be insane to pursue a MD (even at a place like UCSF) instead of a PhD in CS. Plus, a UK PhD is only 3 years, so you'd be on the tenure track or at a tech company by your late 20s, while with the MD/PhD route you'd be slogging until your mid 30s and probably drop out of research. Go to Cambridge!

Your advice is really no better than anything said in the thread and assumes the OP is only in it for money/prestige. For anyone who got into medicine for money or prestige, their priorities were whack and they may now be unhappy. OP didn't ask about outlook of medicine or lifetime earnings potential.

I think only the OP can prioritize, but if OP is on the fence about medicine, OP has to decide what to do with it (although I think PhD in CS is a better choice especially in that case).

I know people who went into medicine being on the fence, and they only ended up continuing because there's no good way out of medicine until you're an attending, for most people anyway.

 

[EvoMed]

Dream outcome seems like if you could convince Gates to let you differ for 2 years while you finish first two years of med school, then convince USCF to let you do your PhD @ Cambridge. I don't know how easy convincing the Gates people will be, but I bet the second part is relatively doable (sanctioned versions of that type of schedule already exist: MD/PhD Partnership Program | National Institutes of Health Oxford-Cambridge Scholars Program)

 

[tortuga87]

Just don't pay for your medical school. That is the most important part. It's very hard to do research after residency if you have 300K in debt.

The biggest advantage to becoming a licensed physician with a computer science PhD is access to large datasets. This is the main bottleneck in healthcare data analysis. If data was more widely available, then there would already be many less jobs in healthcare because a lot of it can already be automated. Other than that, you can make a weak argument about drawing inspiration from patient contact, but that is moreso a reductionist biologist perspective.

If you had to pick one, I would actually encourage you to go to the UK. Money wise, the pay for computer science is much higher than the biomedical fields that traditionally enroll MD/PhD students both by intra-year and lifetime salary. Moreover, you can make more than a full time research physician (assistant professor) as an entry level research scientist in industry straight out of a prestigious PhD about a decade faster. Just make sure to publish in places like NeurIPs, ICML, ICLR, etc; avoid journals even though that is what the medical schools want. If you do go down the MD/PhD route, by the time you are an assistant professor on the research track (which is a huge challenge in and of itself), you will be close to 40 years old and hence already be at least a senior researcher or PI in industry making close to or the equivalent of a full time clinical physician. You also can focus more on actual research rather than writing multiple grants per year to get one funded. For this reason, many computer science MD/PhDs dont do residency and academic PIs struggle to recruit computer science post docs.

Time-wise, you can do about a decade more of a research as a computer science PhD because you can bounce between jobs doing research full time as opposed to clinical work.

The cultures of medicine and computer science are also polar opposites. Medicine is focused on tradition and hierarchy while computer science is whoever invents the best method is top dog. E.g., the equivalent of someone like Goodfellow does not exist in medicine because he would have been forced to work under a mentor for ~7 more years after his PhD, and his mentor would have gotten credit for generative adversarial nets. Then he would have to write a K award that would still require him to be "mentored." You can see where this is going...

 

[751866]

Just don't pay for your medical school. That is the most important part. It's very hard to do research after residency if you have 300K in debt.

The biggest advantage to becoming a licensed physician with a computer science PhD is access to large datasets. This is the main bottleneck in healthcare data analysis...

Access, and a far greater ability to do something meaningful with the datasets as most physicians lack a background with quantitative tilt and therefore, nearly all can't conduct a reasonable statistical analysis or interpret results meaningfully.

I really want to tell OP just skip the MD, but it depends why and how badly the OP wants the MD.

 

[Gastrapathy]

You want to be a computer scientist. Do that

 

[deleted496821]

I gotta say, I know that this is a MD forum, but the advice in this thread is pretty bad. Medicine is definitely on its way down in terms of money/prestige, and computer science is still rocketing upward. You'd be insane to pursue a MD (even at a place like UCSF) instead of a PhD in CS. Plus, a UK PhD is only 3 years, so you'd be on the tenure track or at a tech company by your late 20s, while with the MD/PhD route you'd be slogging until your mid 30s and probably drop out of research. Go to Cambridge!

I have to support what Che is saying here. To hello_world_doctor: I would strongly suggest simply doing a PhD in CS with an in-folded Masters in Biomedical Engineering if you so desire, as many Engineering schools will allow you to get a free Masters during your PhD. Cambridge is a great program, but so are Stanford and Berkely, and they are in your backyard where you are already connected and already have a PI picked-out in Butte.

I am a Physician (MD) and completed a Masters in Data Science after medical school, residency, and a clinical fellowship. I have left clinical medicine outside of independent telehealth consulting and am currently obtaining a PhD in BME (Neural Engineering) with an in-folded Masters in CS.

What you will learn if you go to medical school and then do a PhD after or a post-doc research fellowship is that clinical work/obligations are more of a distraction from your research than an amplification of knowledge that informs your research or provides connections that further your research. This will only continue to get worse. You will also incur significant financial debt (unless you are financed by family for your MD).

This is a forum for clinicians, and I do not think most clinicians have any idea how AI is going to significantly affect clinical medicine in the next 20-30 years. It will not be the pretty "doctors using AI to better interact with patients" scenario that Physician futurists are promoting (most of whom do not have any real technical AI knowledge). It would be best for you to be in front of the shift that is coming, which the PhD in CS will allow.

 

[enarque]

Echo the previous comments about doing the PhD in CS. I am a MD/PhD and my husband is a software engineer. There is tremendous opportunities right now for people in software to impact medicine. As a MD/PhD student, you will waste so much time learning/dealing with many things that are just tedious and contribute little if nothing at all to the type of career you aspire to achieve! I would say MD/PhD is still very good if you are more interested in therapeutics but if you look at the people who are publishing top papers in medical applications of machine learning for example, the MD is very unnecessary. Clinical training will simply set you back years.

And yes the compensation in engineering is higher than academic medicine, given you are an excellent software engineer and work for a top company