Monday, November 16, 2020

Diversity, Inclusion and Equality (In T1D Research)

I think it is important that we all support the Black Lives Matter movement in whatever we do.  This posting harnesses my knowledge of type-1 diabetes research to discuss how we can overcome racism in type-1 diabetes research.
The first thing to understand is that there is racism in type-1 research and treatment (and the institution of medicine as a whole).  In this post, I'm going to focus on two areas:
  1. Under representation in clinical trials:  If a minority group is under represented in clinical trials, then they will be under helped by the medical progress that comes out of that research.  Clinical trials are the raw material from which medical progress emerges.  Fewer Black patients in clinical trials now results in worse outcomes in Black patients later, and this damage is done no matter why a minority group is under represented in research.

  2. Fewer doctor visits and less aggressive treatment by doctors:  The most important factor in good results from T1D treatment, is seeing your health care team often, and having a good therapeutic relationship with them.  If a patient doesn't visit, or doesn't trust their endocrinology team, then their long term health is going to suffer.  Again, it doesn't matter why someone sees an endo less often or why they have less trust.  If it happens, then it hurts. 

The Patients In Clinical Trials

I've reported on 100s of clinical trials over the last 12+ years.  They fall into two groups: the majority, which don't report the ethnic or racial makeup of the people who are enrolled, and a minority which do report, and show that Black and Hispanic patients are underrepresented.  Minority groups are enrolled as a much smaller percentage than their population percentage around the recruiting locations. 

The whole point of clinical trials is to test treatments on the same people who will eventually end up taking them.  The scientific phrase is "representative sample".   Therefore, if a study seriously under enrolls any group, that study is not medically effective in testing the treatment.  The question of why enrollment doesn't match population doesn't matter in terms of effectiveness of the study.  The study is failing its primary goal, regardless of "why". 

These are not new ideas.  In the end notes section, I discuss the 1979 Belmont Report, which laid the foundation (in the United States) of Justice as a basic requirement for human experiments.  It specifically required diversity of participation, so that everyone would benefit from medical research.  It is wrong to think that diversity requirements in clinical trials are a recent response or a modern sensitivity.  They are long standing requirements.

When I point this out, there are several defensive reactions that should be discussed.

"That doesn't really matter because human bodies react to T1D the same way, so we don't need to include Black Americans or Hispanic Americans to learn how to treat/cure them.  Diversity not really needed in clinical trials."

This is wrong for at least two reasons.  First, it is arrogant to say that T1D effects all racial and ethnic groups in the same way, to the point where we don't need to include the minority groups in studies.  Remember, this is not just saying that, so far as we know, all ethnic groups react to T1D the exact same way.  It is also saying that all future research will continue to show this, to the point where we don't even need to do the research in a way which would detect differences!  In addition to being arrogant, it is a profoundly unscientific attitude.

Second, a new drug is not just a physical treatment with possible side effects.  It is something that people will decide to use in a social context.  For example, a CGM device, which only comes in white, might be more comfortable for White people to wear.  The exact same device, but in black might have higher adoption rates by Black patients.  Testing that includes minorities might find that, but testing that doesn't include minorities will not.   Drugs have these kind of social issues as well.  A White kid may not care about taking an insulin needle to school.  However, a Black kid might need to consider the real danger of police over reaction to this exact same action (even if completely legal, and medically necessary).

"We don't discriminate when recruiting.  It's just that fewer minority patients take part in clinical trials."

This argument assumes that the only kind of discrimination is personal discrimination.  It assumes if the recruiting process is not explicitly racist, then there is no racism present at all. That is wrong because it ignores both historical racism and institutional racism.

But there is another problem here. This statement is about why there are fewer minorities in the study, and the "why" question doesn't matter in terms of effectiveness of the study.  What does matter is that large groups within the population are severely under represented.   As an example, if Hispanic patients are not enrolled in a clinical trial, the results will not be as useful to them as to the groups that were enrolled.  It is not just the results of that one study, but all future research based on that study will be less useful to the excluded group.  The question of why Hispanics (or any other minority) were under represented is helpful to fixing the problem, but it doesn't matter when measuring the size and importance of the problem. 

"We're in a White neighborhood, and there just aren't that many minorities near us."

First of all, as I've said above, it doesn't matter.  If you plan to publish a study that shows drug X has effect Y, and 90% of your participants are White, then you are really showing that drug X has effect Y in White people.  If that convention in writing titles were applied universally in scientific journal articles, the racism would be obvious.  The scientific process requires that the people you enroll represent the people you plan to treat.  There is no little footnote saying "if you work in an overwhelming White neighborhood then it's OK to test a drug on a non-representitive population". 

Also, this thinking assumes that location, as a cause of ethnic exclusion from studies, is an unsolvable act of nature.  It's not.  Researchers in a heavily White area can recruit at another site more convenient to minorities.  They can hire a shuttle, pay for transportation, hire minority recruiters, or advertise in specialty social media where minorities have a strong presence.  In short, they could spend extra effort to get a representative sampling of the whole American population.  In the past, spending less money to come out with results that were only applicable to Whites was an acceptable thing to do.  It shouldn't have been then, and it certainly isn't now.

Furthermore, the location of hospitals and clinics is often the result of systemic enduring racism, in many ways.  Rich philanthropists would build hospitals in their own neighborhoods.  Clinics and doctors would prefer to build in more wealthy areas with more people like themselves.  Even if made decades ago, these clinic location decisions skew research done today.   Of course, the existence of ethnic neighborhoods was often shaped by racism in housing, law, banking, policing, and society as a whole.

Managing Type 1 Diabetes

I don't think it is controversial to say that people with T1D have better outcomes when they see their endocrinology team more often, and have a good therapeutic relationship with that team.  A good relationship meaning that each side understands and trusts the other, etc.

Recent studies (both in ADA 2020 and previous ADA conferences) show very clearly that Black Americans with T1D visit their endo teams less often than other Americans, that their doctors suggest medical interventions less often, and that (generally) they trust their medical teams less. 

The results of these differences are less effective treatments, earlier and worse complications, and higher death rates.  Two recent studies showing worse results can be seen here, but there are dozens more:
The first included over 10,000 people and found racial differences even after adjusting for poverty.
The second, of 200+ people found racial differences even after adjusting for insurance coverage.

As before, there are several defensive reactions that should be discussed.
"That's just economics.  Black people are poorer than White people (on average) so they end up going to doctors less often for everything, not just T1D."
There are several ways to respond to this, but the first one that pops into my mind is: so what?  Sure relative wealth contributes to the issue, but the issue is caused by many different things, and the situation would be improved by improving any one of the causes.  The fact that poverty is part of the problem does not mean we should ignore the racist part of poverty.   We can, and should, try to fix all the problems, and if we can't or won't fix one (such as poverty), maybe that means we should try even harder to fix the others (like racism).
Second, this argument implies that poverty is not itself caused by racism.  Obviously, that is not true.  If there is one thing we have learned from all the recent videos of racism and of police brutality, it is that discrimination is alive and well and having a big impact in the day-to-day lives of people all over America.   We are now seeing videos where everyday people viciously discriminate against minorities by refusing them service, stopping them from walking down the street or delivering packages, calling the cops on them, assuming they are criminals, and many other forms of discrimination, including murder.  And if that is bad now, think what is was like a few decades ago!  It should be obvious to everyone that this kind of day-to-day interference in the lives of minorities is a major cause of poverty.  I don't think anyone should say "the cause is poverty not racism".  They should say "the cause is racism manifesting as poverty", or "the cause is racism and poverty".
Lastly, I want to repeat that many studies (including the two linked above) show bias even when poverty or insurance are taken into account.
"It's not the fault of the medical team, if Black patients do not see them as much as White patients."

One of the findings in this year's ADA conference was that medical professionals are less likely to recommend interventions to Black patients than to White patients.  Obviously, this has bad effects in type-1 management because it means that Black patients will (on average) have fewer tools to manage their type-1 diabetes.  But it also sends a message that seeing a doctor is less important.  After all, one of the reasons we see a doctor is to get recommendations for new equipment, so if your doctor tends not to give you those, then there is less reason to visit.  And for people who think "doctors don't do that any more" remember that habits are built in the past, so if doctors provided lower service to minorities 20 or 30 years ago, those same minorities are now deciding (as middle aged adults) to see their doctors less often or not at all.   And maybe bring their children in less often, because of their very real bad experiences from their own childhood.  Racist actions in the past cause racist results in the present via history, habit, stereotypes, "community knowledge" and in many other ways. 

Changing How I Report On Clinical Trials

I am convinced that the first step towards reducing racism in clinical trials is to measure and report on racism in clinical trials.  If a clinical trial is not including minorities, then it is contributing to racism in medicine, and if I report on the "results" of that trial without including the racial makeup of the study, then I am contributing to the racist result.
So, I'm going to change the way I report on clinical trials.  Starting now, for all Phase-II and later studies that I report on, I'm going to include information on the diversity of the patients enrolled in the study, and I'm going to specifically call out studies that don't report on the ethnic or racial make up of their patients at all.  I'm hoping that if science reporters in general do this, we will encourage researchers to report on the composition of their clinical trials and then actually improve that composition until it is representative of the general population where they do their research.

I don't think this is a big contribution, and I wish I could do more, but I can do this, so I will do it.  We all need to do what we can.  I'm also looking for more that I can do, so if you have suggestions in this regard, please do send them to me.

Measuring Progress To Decrease Racism

What should the racial or ethnic composition of the research population be?  I did not want to get sidetracked by this discussion above, so I'm putting it down here.  I focus mostly on studies in the US, and the participation of Black and Hispanic patients is so low that discussing what it should be often is used as a "red herring" to sidetrack the discussion.

But I think there are two good answers to this question.  One is to say that the makeup of all the studies done, taken together, should be the same as the makeup of the country as a whole.  The second is to say that the makeup of each study should be the same as the makeup of the area where the study is recruiting.  Area defined broadly.  I'm not talking about within 5 miles of the research site, but maybe 20 miles, or the entire metro area where the research site is located.   I think that both of these pieces of data should be reported on.
Every major funder of research (the FDA, NIH, JDRF, etc.) should report on the overall makeup of the patients in the clinical research they fund in the US.  This should be compared to the nation as a whole.  More regional funders (such as DRI, the Sansum Institute, etc.) should publish numbers and work towards matching the populations in their regions.
Finally, each trial should publish the makeup of their patients, and these should be compared to the recruitment area(s).  This is something that the JDRF, FDA, NIH, etc. could change immediately.  All they need to do is make a condition of funding, that the researchers publish the racial composition of their participants.  When that data is published, it will provide it's own pressure to fix the problems which cause it.
End Notes
Historical Note: In the United States, the idea that the whole population must be represented in clinical research has been well established (on paper) since the Belmont Report of the late 1970s.  This federal report is part of the chain of events which led to the Federal Policy for the Protection of Human Subjects: Nuremberg Code (1947), Helsinki Declaration (1964), Belmont Report (1979), and finally the Common Rule (1991).  It listed the three fundamental ethical principles for using any human subjects for research as:
  • Respect for persons: protecting the autonomy of all people and treating them with courtesy and respect and allowing for informed consent. Researchers must be truthful and conduct no deception;
  • Beneficence: the philosophy of "Do no harm" while maximizing benefits for the research project and minimizing risks to the research subjects; and
  • Justice: ensuring reasonable, non-exploitative, and well-considered procedures are administered fairly — the fair distribution of costs and benefits to potential research participants — and equally.

The principal of "Justice" required both that minority groups must not be targeted for potentially dangerous clinical trials, nor could they be ignored by potentially beneficial ones.   Since all clinical trials are potentially dangerous and potentially beneficial, minority groups should not be over or under represented. 

More reading: and

View Point: This blog posting is very US-centric.  I'm a White American, and a software engineer working in Silicon Valley, and these set my social context.  My views on who is a minority and who is the majority doesn't cover the whole world, and the examples I use are American.  If you're reading this in most of the rest of the world, they may not make sense to you, and I'm sorry for that.  The basic messages I'm trying to get across are universal, even if the details do change in different regions. 

Personal Note: Racism's huge impact and wide effect was "brought home" for me when I was buying my house.  It had a covenant prohibiting "any non-Caucasian person" from owning it.   This covenant has been unenforceable since 1953, but is still part of the deed of the house, and there is no legal way to remove it.   These covenants used to be relatively common in California, especially for single family homes.  There are plenty of people alive today who's housing choices were limited by these racist covenants.  The banking, policing, and social impact lasted long past 1953.  More reading:

I will say that reading these racist housing covenants in a news article is horrible, but reading them in a document you are about to sign as part of buying a house, is even worse.  Even though they are completely unenforceable, it made me feel so dirty, so unclean, and to think of the lives they ruined.

Note on language: English is undergoing a change in how we refer to ethnic and racial groups.  Both terminology and capitalization are in a state of flux.  For this blog posting I have capitalized ethnic and racial descriptors, and used them as adjectives rather than nouns.  For example writing "White people" and not "Whites" or "whites".

Thanks to reviewers: As you might expect, this blog went through more review than is usual, and it has benefited from everyone who read it and gave me feedback.  I want to thank all the reviewers.

Joshua Levy
publicjoshualevy at gmail dot com
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF or JDCA news, views, policies or opinions. My daughter has type-1 diabetes and participates in clinical trials, which might be discussed here. My blog contains a more complete non-conflict of interest statement. Thanks to everyone who helps with the blog.