Are Doctors Racial Profiling?

[Robert Toy]

"It's absolutely proven through studies that a black man and a white man going to the hospital with the same complaint will be treated differently," Dr. Neil Calman, a family physician and president of the Institute for Family Health in New York, said. Calman is also Reid's regular physician.

For example, a 2005 study found African-American cardiac patients were less likely than whites to receive a lifesaving procedure called revascularization, where doctors restore the flow of oxygen to the heart. The study authors at RTI International, a research institute, noted that all of the patients had Medicare, which covers the cost of revascularization.

In a study conducted in 2007, Harvard researchers showed doctors a vignette about a 50-year-old man with chest pain who arrived at the emergency room, where an EKG showed he'd had a heart attack. Sometimes the researchers paired the medical history with a photo of black man and other times with a photo of a white man.

The doctors were significantly more likely to recommend lifesaving drugs when they thought the patient was white than when they thought the patient was black.

http://www.cnn.com/2009/HEALTH/07/23/doctors.attitude.race.weight/index.html

I have no idea why the different treatment, any ideas?

[Zoombie]

Well, there is a genetic difference between different ethnicities, and some drugs work better with some races than with others.

...and there also might be an element of racism, which still survives to this day, which always bugs me.

[icerose]

This was a study on taking meds and controlling cholesterol. It's interesting to see that white women are on the same level as black men and black women are the worst in the study group. And this was a study to see who best controlled their cholesterol on identical treatments.

http://www.medscape.com/viewarticle/569614_3

[Robert Toy]

This was a study on taking meds and controlling cholesterol. It's interesting to see that white women are on the same level as black men and black women are the worst in the study group. And this was a study to see who best controlled their cholesterol on identical treatments.

http://www.medscape.com/viewarticle/569614_3
link requires a login

[icerose]

Weird, when I got it off google, I had no problem getting in. That freakin' sucks.

[backslashbaby]

Doctors can be strange with what all they bring to the table. I used to go to the doc in sweats, cos I was sick and felt awful. Don't do that, at least at the docs I saw. They don't spend any time with you/listen to you the same way as when you're dressed for work. I was floored by that when it happened.

Getting a doc to spend time, listen, and take you seriously is actually one of the major hurdles in getting good care. Obviously, a doc that doesn't listen or spend more than 2 seconds with you can't give the proper level of care.

Also, if he/she thinks you exaggerate, just complain because you are obese - or black - and other lovely prejudices, the stats will bear that out on a large enough scale, imho.

[Kitty Pryde]

In a college class we learned about a big study conducted at a doctor's conference (this would have been 2004-ish). A bunch of doctors were shown one of four identical videos starring moderately attractive older people. There was a black man, black woman, white man, and white woman. They each read the exact same speech in the exact same manner, dressed the same way, describing their symptoms and history. Differences in treatment decisions were significant: the white man was most likely to be treated, then the white woman, then the black man, then the black woman. (I believe the doctors themselves were American, both men and women of varying ethnicities.)

It's hidden or unconscious bias. It's not a top secret concept or anything. http://www.tolerance.org/hidden_bias/tutorials/02.html It's basically a case of unintentionally discriminating, inadvertent but still harmful. These hidden biases can be uncovered and mitigated to a certain extent (and I would hope people like doctors, policemen, social workers etc. all do work to eliminate their own hidden biases), but it can be hard for a person to get rid of all of them.

[Roger J Carlson]

http://www.cnn.com/2009/HEALTH/07/23/doctors.attitude.race.weight/index.html

I have no idea why the different treatment, any ideas?
I hate articles like this. They make broad generalizations and always have some quote like: "If I had been a white man, I'm sure..." In their haste to make a good story, they always imply that doctors just don't care.

However, there is evidence that when such racial differences are pointed out to doctors, they do respond. In this article (http://www.journals.elsevierhealth.com/periodicals/jac/article/PIIS0735109797000892/abstract), it shows how rate of by-passes, angio-plasty, and cardiac catheterizations increased between 1980 and 1993 after such racial discrepancies were pointed out. Doctors and hospitals as an overall group do care.

Here's a better article (http://jama.ama-assn.org/cgi/content/full/297/22/2489) about racial differences with "revasuclarization" (by-pass and angioplasty) by JAMA.

In it's abstract, it comes to this conclusion: Conclusions Black patients admitted to hospitals with and without coronary revascularization services are less likely to receive coronary revascularization. The higher long-term mortality of black patients may reflect the lower use of revascularization or other aspects of AMI care.

However, if you read to the end it makes some important caveats in the Comments
The study has several limitations. First, unmeasured aspects of medical decision making that may affect the process of care, including patient and family preferences, cannot be ascertained from administrative data. Although previous studies suggest that the proportion of patients refusing myocardial reperfusion is small and does not vary by race, patient preferences may play a more prominent role in more delayed revascularization treatments. Similarly, although a previous single-center study found no racial difference in physician recommendations for coronary revascularization, data on physician decision making in AMI treatment are scarce.

Second, the study was unable to account for important clinical indications for transfer and revascularization.... We were also unable to capture the use of thrombolytics..., which has been shown to differ for black and white patients,19 and were unable to examine the potential impact of economic factors, such as the availability of supplemental health insurance.

Third, the development of risk-adjustment models based on administrative data has inherent limitations, and the reliability of individual diagnosis codes may vary across study hospitals; such data may not capture important long-term prognostic factors, such as body mass index, smoking, and follow-up medication treatment. Last, the analysis was restricted to elderly Medicare fee-for-service beneficiaries and thus may not be generalizable to younger patients or to patients enrolled in Medicare managed care.

But the bit about Administrative Data is very important. Since my job is to collect and evaluate administrative data from multiple hospitals, I know a little bit about it.

Administrative data is the data the hospital uses to bill the patient, insurance, Medicare, or what have you. It consists of demographic data (race being one), procedure codes, diagnosis codes, and charge codes.

Procedure and diagnosis codes are fairly standard, but they must be interpreted by people (called coders) who input the data from the charts into the billing systems. So there is an inherent translation problem. One coder might code something differently than another coder or one hospital may have a different standard for coding something than another.

Charge codes are charges that are billed to a patient. If a patient is billed for a cardiac cath, they probably got one. Charge codes are even worse because each hospital has it's own charge code system and it can be difficult to reconcile these systems to each other.

But the case is rarely that a single code (either a procedure, diagnosis, or charge code) represents a particular diagnosis or procedure performed. Most of the time, we look for one of a number of codes with exceptions based on other codes and possible charges.

Demographics are another problem because their is no standard on what "race" actually is. Some hospitals have White, Black, Asian, and Other. While others may have White, Black, Hispanic (white), Hispanic (non-white), Native American, and so forth. Comparing the stastics against such different definitions is problematic.

Another problem with race is, who decides what race the patient is? In some hospitals, the patient is asked to self-identify. In others, the race is just assigned by the person registering them. Some hospitals have transitioned from the second system to the first, so their statistics can vary even within the same hospital over time just based on how they determined race.

Those are some of the "inherent limitations" of administrative data the article mentions. A lot about the reasons for a diagnosis or recommendation for a procedure are completely missing. As the article also points out: Despite these limitations, the current study provides evidence that racial differences in the use of revascularization after AMI ... These differences could be due to unmeasured clinical or socioeconomic factors, patient preferences, and unmeasured aspects of medical decision making...

[Fran]

Would the level of medical insurance a patient has make a difference as to what tests are ordered and what procedures are carried out? Are black people more or less likely to have health insurance than white people? Is that even an issue?

As always your labyrinthine insurance rules fascinate me. I apologise in advance if I'm being really thick about this :)

[Kitty Pryde]

In response to the previous two posts, insurance isn't really a factor, nor is the labyrinthine goings-on of hospital billing and coding.

These biases show up in studies done on doctors' assessments of videos or photos of pretend patients. Numerous studies done not based on real, messy patients show that there is a trend for doctors to prescribe treatment more to certain groups of people than to others.

IRL, yes, insurance affects who gets what treatment. Which is in most cases detrimental. Rate of being uninsured is 20% for blacks, 12% for whites (and 33% for Hispanics!) ( http://covertheuninsured.org/content/raceethnicity ). BUT the research shows that insurance is not the only issue.

(Also, if a doctor looks at a hypothetical black patient and thinks, 'He's black, so he's probably uninsured. I won't order Expensive Lifesaving Test,' then that is a case of hidden bias. If a doctor looks at a hypothetical uninsured patient and thinks, 'He's uninsured, so I won't order Expensive Lifesaving Test,' that's a separate problem!)

[Roger J Carlson]

In response to the previous two posts, insurance isn't really a factor, nor is the labyrinthine goings-on of hospital billing and coding.In the study I cited, it did, as it would with any study that looked at Administrative data, which is most of them.

[Kitty Pryde]

In the study I cited, it did, as it would with any study that looked at Administrative data, which is most of them.

Again, IN SOME STUDIES, hospital administrative stuff has clouded the data. But IN MANY OTHER STUDIES, hospital administration WAS NOT INVOLVED IN ANY WAY. I can pull up a list of them if needed. My point being that a shortcoming in one study is not a shortcoming in them all. Many studies without this shortcoming show the same results.

[Roger J Carlson]

Again, IN SOME STUDIES, hospital administrative stuff has clouded the data. But IN MANY OTHER STUDIES, hospital administration WAS NOT INVOLVED IN ANY WAY. I can pull up a list of them if needed. My point being that a shortcoming in one study is not a shortcoming in them all. Many studies without this shortcoming show the same results.The majority of these studies will be done on administrative data because it offers the greatest statistical relevence. The study I cited was of 1.2 million patients over 4600 hospitals over a 5 year period. That's not just "one study" to be dismissed. It's a major study published by JAMA.

Studies which measure doctor's reactions are by their very nature limited in scope. They can be indicative and corroborative, but not conclusive.

I'm not trying to say differences don't exist. The study I cited even concludes they do. However, it acknowledges that difference can exist for a variety of reasons, not all of them having to do with racial bias on the part of doctors.

Robert Toy asked for ideas on why the differences, and I gave him some.

[MacAllister]

Gently, folks.

[Robert Toy]

The majority of these studies will be done on administrative data because it offers the greatest statistical relevence. The study I cited was of 1.2 million patients over 4600 hospitals over a 5 year period. That's not just "one study" to be dismissed. It's a major study published by JAMA.

Studies which measure doctor's reactions are by their very nature limited in scope. They can be indicative and corroborative, but not conclusive.

I'm not trying to say differences don't exist. The study I cited even concludes they do. However, it acknowledges that difference can exist for a variety of reasons, not all of them having to do with racial bias on the part of doctors.

Robert Toy asked for ideas on why the differences, and I gave him some.
And I thank you

[backslashbaby]

It is fascinating that studies that have been as controlled as possible to see if race is really the unusual factor show that it is. Of course those studies don't speak to real-life examples, as their point is to control for as much as possible.

There's something there, and I'd be thrilled to hear of more of the profession taking proactive steps to rid themselves of any bias. It's pretty crucial to good health, imho.

[GeorgeK]

I have no idea why the different treatment, any ideas?

I didn't follow the link because I've never seen any of these studies have any practical application because they all seem to exist more as an inflammatory vehicle than anything else. If it's the same thing I saw a couple days ago on the news on CNN it was a very one sided piece.

They confuse confounding variables as physician bias. Are there physicians who are bigots? Yes, but as a proportion of the population my experience says that physicians are less likely to be bigotted than the population as a whole. I can't comment on the cardiology stuff but in Urology race often has a big impact on care because of the diseases. Metabolically, there are different considerations for a black man age 59 with a kidney stone as opposed to a white man, or the same two men with an otherwise identically elevated PSA. Races are subsets of the population arbitrarily linked to skin color. Each subset of the population will vary in some ways to another subset; have different diseases, strengths, weaknesses. I can only assume that there is the same thing in subsets of Cardiology patients.

Also be wary of studies that use physicians and videotapes or actor patients. The physicians know they are being scored and will not behave in a study the same way they do in practice. At meetings I've seen it done where they have audience participation wireless gizmos where you can punch in A-E and it is tabulated in seconds and the responses show up on the video screen. They have 5 questions that they use for their study and 20 or so that are "controls". The control questions are geared to alter your response slowly over a few patients while the moderator chimes in snippets that do not appear on the actual screen. It's the old line of statistics don't lie but statisticians do. CNN is just the opposite slant of Fox. Neither can be trusted 100% of the time. Also don't trust studies funded by a group that has a vested interest in furthering its own agenda. If group XYZ has a study to claim that carpenters should use hammers with blue heads and XYZ is the one who sells blue headed hammers then the study is at best suspect.

[TerzaRima]

The anecdote that starts out that article is pretty asinine. The young doctor recommended amputation, and she was wrong, and her consultant was right...so she's a racist? Physicians who make medical errors with nonwhites are now supposed to be bigots? How much does that chip weigh?

Yes, but as a proportion of the population my experience says that physicians are less likely to be bigoted than the population as a whole.

Mine too.

[Kitty Pryde]

The point is, (most of) these doctors aren't evil bigots. They are consciously not racist, they consciously don't discriminate against their patients for any reason.

Seriously go read this link. Unconsciously a lot of people do discriminate based on sex/race/age/etc, no matter how much they think they don't. When hidden bias is uncovered, one can work on eliminating it. But pretending it doesn't exist won't help. http://www.tolerance.org/hidden_bias/tutorials/02.html

[TerzaRima]

The other thing I wondered about was the demographics of the physicians in the study--race, sex, age. You tend to have hidden bias against those who are most different from you. But that doesn't make physicians any different from, say, dog groomers.