New Data. New Opportunities.

So, I'm working on a new analysis, and my plan is to describe what I'm doing, why I'm doing it, and what I'm learning along the way, as it happens.

Well, I guess it's not entirely "new", depending on how you count it, I've been working on this for years already.

It's an extension of prior work I've talked about a lot on this blog, the extraordinary finding that gay men are more likely to be in "excellent" health than heterosexual men.

I've looked at this issue in many datasets, and published a paper on it in one particular dataset, the Behavioral Risk Factor Surveillance Survey, where I looked at the self-reported general health of men in same-sex and mixed-sex couples, based on their marital status and household composition.

In that study, I found that about half the general health advantage of men in same-sex couples was explained by the facts that men in same-sex couples were younger on average, better educated, and wealthier. When I looked at some of the specific health characteristics, men in same-sex couples were less likely to be overweight or obese, but more likely to smoke. Taking all the health factors I could into account didn't explain why man in same-sex couples were healthier than men in mixed-sex couples.

But what do I want to do now, in this new analysis? Well, it may yet shift and wiggle from this idea, but I'd like to do a couple of things. First, there's a lot more data to work with now. Instead of having to look at marital status and household composition, we can now look at sexual orientation directly, because a number of states have started to ask people not just who you live with and your marital status, but also whether you think of yourself as heterosexual, gay, lesbian, bisexual, or something else.

Second, because we now have both sexual orientation and marital status, and same-sex couples can get married in a lot more places than when I did the earlier analyses, I can look at single gay, lesbian and bisexual people, bisexual people in same-sex and mised-sex relationships, and a variety of household composition structures.

Third, I'd like to try to fold three different surveys into a single analysis, the BRFSS (which is the largest of the surveys, but the sexual orientation questions are asked in only select states), the National Health Interview Surveys (which are smaller, but comprehensively asked all across the country), and the National Health and Nutrition Examination Surveys (which are smaller still, but have a great deal of depth to them).

There are other ideas I'd like to look at too, like the degree to which heterosexuals would be misidentified as being sexual minorities because they live in same-sex households (like my heterosexual roommate). And what geographic and demographic factors predict that misclassification.

Also on the issue of misclassification, the BRFSS asks interviewers to guess the sex of the person they're on the phone with, and these data give us an estimate of how often they guess wrong.

Another set of issues I could look at with these data are things like basic demographics - Who do gay men and lesbians live with? How many are in relationships? Believe it or not, these basic issues have barely been touched in the academic literature, because data like this has not been available before.

Trends in Motor Vehicle Accidents

Every once in a while, I like to see what's going on with motor vehicle accidents. It turns out there's a lot going on. This data is from the Fatal Accident Reporting System. I haven't done anything special with it, just graphed the rather bland spreadsheet there on the home page.

 The obvious thing that jumps out at me is that after decades of increases in motor vehicle deaths (the trend goes back to the very introduction of the automobile), we seem to have hit a turning point in 2005, and there were huge drops in motor vehicle fatalities in 2008 and 2009 especially.
The other thing that jumps out at me is the increase in the number of motorcyclists killed on the roads (the purple bars), and perhaps a decline in pedestrian deaths (green bars), and certainly a decline in passenger deaths (bright red bars).
The timing of the precipitous drop in 2008 and 2009 certainly suggests a connection to the recession - fewer vehicles on the roads = fewer deaths. That decline in vehicles would presumably come from three sources: fewer trucks delivering goods, fewer commuters, and fewer errands and pleasure trips. But why would there be more motorcyclist deaths? Perhaps the aging of the baby boom generation? And I haven't got a clue about why there would be fewer pedestrian deaths. It would be interesting to see whether the decline in pedestrian deaths is also linked to the 2008-2009 drops - and could that be attributed to fewer commuters? Or fewer errand and pleasure trips? The drop in passenger deaths seems to be pretty strongly linked to the recession - so is that about less car-pooling among the remaining commuters?
At any rate, graphing the number of deaths is a bit misleading, because the population keeps growing.

So, when you divide the number of deaths by the population (and multiply by 1,000,000), the peak year isn't 2005, but 1995. Actually, if you trace these trends back, the peak year on a per population basis is some time back in the 1920's, when cars were just mowing people down left and right, with very little effort to make the vehicles, the roads, or the drivers safer. What you see in the long term trends is a long slow decline in motor vehicle death rates, followed by a rapid decline in the 1970's, linked to that decade's recession, and also the high price of gas (much much higher than today once you take inflation into account), speed limit restrictions, the imposition of seat belts, investments in improved road infrastructure (guardrails etc.), and a radical shift in how we viewed drinking and driving. The slower decline continued in the 1980's through the late 2000's, especially due to air bags, improved vehicle construction, lighter vehicles that do less damage to others, and continuing trends in driver, vehicle, and road infrastructure safety. But that drop in 2008/2009 is still really dramatic, and I have to wonder if it can all be attributed to the recession.

Presumably, if the decline is due to the recession, it should be directly related to how many vehicles are on the roads. So, if you divide the deaths by 'vehicle miles traveled' instead, it should smooth out the trend...

And that seems to be the case. The long trend towards lower deaths per mile traveled dominates, but there is still an extra bump in 2008/2009, suggesting that the recession not only reduced the number of miles traveled, but also made the miles traveled safer, especially for passengers and pedestrians.

So, as we climb out of the recession, I'd expect to see the number (and rates) of motor vehicle deaths increase a litttle bit, maybe as high as 120 per million residents per year or 12 per billion miles traveled, and then continue the long slow decline.

So, here's another example of major improvements in health being made. Not as sexy a story as the latest fad in diet, but it's good to be reminded once in a while about what's going right.

Research Directions

    Hey there blogfriends, I'm super excited because I'm going to have a first-author paper coming out in a few days - about the racial distribution of trees and pavement across the US - and exploring a few reasons that may explain it, like segregation (yes) and poverty (no). It looks like there's going to be some press on it, so keep an eye out.
    And my next first-author paper is getting really close to submission - so it's probably six months to a year from publication. That one's about the influence of living in more segregated cities on the probability of experiencing racial discrimination. That one's pretty interesting - lots of studies within one particular city or another have found that experiences of racial discrimination tend to be less common among Blacks who live in predominantly Black neighborhoods, and more common among Blacks who live in predominantly White neighborhoods. As far as I can tell, ours is the first to look at the degree to which the overall segregated character of the city (and her suburbs) affects reporting of racial discrimination experiences. We're seeing pretty dramatic results in that more segregation results in more experiences of racial discrimination, for Blacks, Hispanics, Whites and Asians.

    But what I'm stymied with at the moment is where to go after my most recent first-author paper - showing that gay men are more likely to be in excellent health than straight men... I'd love to get another paper on TBLG health out there, relatively soon, but it's challenging, because I have to do the work on my own dime and my own time. So here's some ideas, and I'd love to hear your thoughts on what would be most helpful (helpful in any sense - informing policy, improving science, satisfying curiosity - whatever greases your gears).

ONE: Improving Identification of Same-Sex Couples in Large Probability Datasets
    I know. Boring title. But here's why this has been floating my boat lately. When I was working on gay men in excellent health, I looked at the biggest dataset I could lay my hands on, the BRFSS. There were a fair number of same-sex married couples, even before same-sex marriage was legal anywhere in the US, which struck me as odd. Another thing that was odd is that their demographics (how old they were, how many kids they have, whether they served in the military, etc.) were a lot like heterosexually married people. I figured that what was most likely happening was that a small number of heterosexually-married people were accidentally mis-coded - and ended up being counted as same-sex couples. So, I threw them out of the analysis.
    BRFSS is especially vulnerable to this kind of error, but the problem is ubiquitous in any of the large probability samples that get used for research on same-sex couples - and rarely acknowledged.
    So what this project would be about is systematically going through the major datasets and trying to estimate how many of the same-sex couples identified are really same-sex couples, and how many are mis-coded heterosexually-coupled people.
    The main reason that it's important to do this project is that there are a lot of publications out there claiming that same-sex married couples are "just like" heterosexually-married couples. That may be a comforting message, and there's probably something to it, but a likely explanation that is almost never discussed is that a lot of those same-sex married couples are in fact heterosexuals. If we want an accurate picture, we need actual same-sex couples.

TWO: BLG health in relation to voting on marriage restrictions
    OK, so my thesis (never was able to get it published) was about the occurrence of suicide in relation to heteronormativity - the more heteronormative an area is, the higher the suicide rate there - especially for young men. I measured heteronormativity in three ways: the legal status of employment discrimination; how people voted on restricting marriage; how many same-sex couples the Census counted in an area.
     Given that nobody seems to care about employment discrimination any more these days, I figure that I should focus on the voting thing. The way I see it, how people in an area vote on restricting marriage to "one man and one woman" is a pretty good heteronormativity thermometer. There are some complications in that the wording is different from State to State, and the change in public attitudes is so rapid that a 60% endorsement rate today probably corresponds to an 80% endorsement rate in 2004. But assuming I can figure out a way to handle that, the other part is finding a dataset that has good BLG health measures in it.
    For my thesis, I used the overall suicide rate, and I didn't particularly care whether the people who died of self-inflicted injuries were "gay" or not. In fact, I suspect that the highest suicide risk associated with being gay or bisexual is before one declares openly to anyone else, and even before having sex, so it would be kind of silly to try to figure out who's who after they're dead. But I think that's one of the reasons I had trouble getting anyone interested in publishing it - it seems like people want to know how BLG people are affected by homophobia. Well, I'm interested in how heterosexuals are affected also. I very much doubt that it's a zero-sum game where heterosexuals gain some advantage while BLG people pay the price. I suspect it's much more likely that heterosexuals, too, are harmed by heteronormativity. And since there are a lot more of them, it should be even easier to pin that down. But I digress.
    So, I need a dataset that A) is a probability (random) sample of the US, B) has a large sample size (ideally in the 10's of millions, but I'll have to settle for less), C) identifies who is gay, lesbian, bisexual, and heterosexual, D) has a high degree of spatial resolution so I can figure out what the local homophobia "temperature" is, and E) has decent temporal resolution so I can figure out when people were sampled relative to important dates, and F) has decent measures of health in it.
    There are some datasets that come close to fitting the bill, but it's a challenge.

THREE: Transgender health from large population datasets
    There's only one publication out there about transgender health based on a probability sample - from the Massachusetts BRFSS. But there's the potential to do so much more. There are seven States that have asked about transgender identity on BRFSS. I'd love to collect the data from all seven, compare the basic demographics of transgender-identified people across the different question wordings & hypothesize about which questions work best. And then get into the health outcomes, much like the Massachusetts study did, but with much more data. I suspect that all of the question wordings are going to have a significant problem much like the same-sex married people identified in large population datasets - that is, even a very small number of errors in the coding of cisgender people is going to be a major headache. There's really only one way to handle that that I can think of - call them back to verify it - but I really can't see that happening anytime soon.

FOUR: The Real Blood Donors of Gaytown, USA
    There are just so many things wrong with banning gay blood donors. It made sense in 1985 (and frankly, it would have made even more sense earlier). But it doesn't make sense now, and everyone knows it. Including lots of gay men who donate blood anyway, and increasing numbers of young straight people who won't donate because they don't feel right about the discrimination. I'd love to be part of qualitative research on gay men who give blood. Why do they do it? How does it make them feel? What 'rules' about donating have they made for themselves to decide when they should and should not donate?
    There's a lot of interesting policy angles to wrangle through on this issue, but I think getting to know these guys would be really interesting - and informative in coming up with better deferral guidelines.

FIVE: Wage Gap and Death
    Strangely enough, there are only a handful of studies out there measuring how sexism affects health at a population level. Most of them use some sort of complicated mash of different ideas into an "index", and I hate indices - you never know what's really going on in there. So I took a simpler approach, just looking at the wage gap between men and women. It varies a lot - there are some parts of the country where women make almost as much as men, and some parts where men make about twice as much as women. What I expected to see was that women's mortality would be higher in areas where men make more. But I saw something completely different: where men make more relative to women, they live longer, but women's mortality is unrelated to the wage gap. I basically put this project on ice because I can't figure out a narrative that makes sense. But I could go back to it if y'all have fresh ideas.

So let me know, what do you think I should work on? And if you're feeling especially generous, for only $62,000, you get to decide.

Research Worth Reading: the Regnerus Study

OK, I'm probably not going to make a lot of friends with this posting, but I think most people read this blog because they want to hear my contrarian viewpoints, not because they agree with me.

By now pretty much everyone's heard of the "Regnerus Study" or the "Family Structure Study". Praised by the religious right & a scorpion in the boot of the gay movement, the study leaves precious few without a strong opinion. If you have no idea what it's about, a good summary of the study and the controversy surrounding it was written by William Saletan at Slate.

I spend a lot of time listening to broadcasts from the religious right: Bryan Fischer at American Family Radio, Liberty Counsel, Family Research Council, National Organization for Marriage, etc. You may drink coffee to get up & going - I listen to these folks.
I think it's important to understand where they are coming from, to understand what arguments they use, what they assume to be true, what they believe about people like me, etc. Often people seem to think these folks are crazy, stupid, or both. I don't think that they are, for the most part. The major spokespeople are far from crazy or stupid. However, they are strong partisans, and have interpreted the Regnerus study with a very partisan bias. They have claimed that it proves that children do best when raised by their biologic mother and father, and that children raised by gay or lesbian parents do worse in most areas than children of single parents. It proves no such thing, but I think it is a valuable addition to the discussion.

As many others have pointed out before me, the study does not have a sample of children raised by gay or lesbian parents upon which to make these claims. They asked a bunch of adults some questions about their parents, and classified anyone who claimed to know that their parent had had a same-sex experience as having been raised by gay or lesbian parents. The study had less than a handful of respondents who had been raised by same-sex parents from infancy.

Many people who I agree with on the substance of family studies have said that the Regnerus study should be pulled, that it is fraudulent and academically dishonest. I don't think it is. I think he clearly and accurately described what he did, and although I encourage people to vehemently disagree with his interpretations and conclusions, that the methodology of the study is not inherently flawed, and was not dishonestly presented in publication. As a result, I don't think it should be pulled.

The controversy around this piece has got me thinking in a lot of different directions, so I hope you'll forgive the scattered nature of the next few paragraphs.

Reaction A: Lots of crap gets published. I'd say over 98% of the studies I read have major methodologic weaknesses, and/or come to conclusions not supported by the underlying data they report. And don't get me started on plagiarism - that problem is out of control, and can even be found in esteemed academic publications. When I started trying to write a series on 'research worth reading' about gay health, it was a real struggle to find anything worth encouraging others to read. I went through hundreds of abstracts, read dozens of papers, and came down to a small handful of papers I thought were 'worth reading'.
Which is a far cry from saying that there isn't a lot to learn from all the crud that gets published.
But it does make me reticent to say that the Regnerus study, with all its flaws, is out of bounds when compared to the vast majority of academic publications. Is it 'worth reading' from the perspective that it skillfully addresses the underlying research question with precisely targeted methodology and conclusions that are well founded in the work itself? No on all counts. But, it is worth reading because it presents a very different perspective than most of the family structure studies out there currently, and it provides a methodologic contrast to them that makes it worth thinking about how to build from the methodologic weakness of the entire field something that would be more reliable.
So, if the Regnerus study frosts your buns, as it should, get off your duff and do a better study. The gauntlet has been thrown down & there's no way to force them to pick it up again and say 'my bad'.

Reaction B: Religious right commentators have claimed that there is a strong liberal bias in this field, and that any study like Regnerus's that challenges the pro-LGBT bias is unlikely to get a fair chance at publication. I'm afraid that they may be right on the first of these, although I doubt the latter.
The larger field of marriage and family structure studies has been very heteronormative with respect to lesbian and gay families, to the point that even when there is a same-sex household included in these studies it usually gets classified as a mixed-sex household because the researchers don't even consider the possibility that there might be same-sex households. But among the small number of studies that do acknowledge same-sex parents, this small subfield has been conducted and interpreted largely by partisans on our side of the debate.
I don't know how many anti-gay studies have been precluded from publication, but I doubt it is very many, if any. It is more likely that these studies just haven't been done. A couple possible reasons: 1) our adversaries often claim that it is obvious common sense that lesbian or gay parents are harmful, so there is no reason to confirm common sense (I'm not agreeing with that, just trying to explain why I think only one anti-gay study has been done so far). 2) Lots of people on the right say that they are tired of talking about homosexuality - by which they mean they wish we would just go away and not ever be part of their lives - ouch! But that sentiment, that they are tired of talking about us, carries through to why they would be unlikely to do a scientific study of family structure, valid or otherwise. Why would you invest time and effort into such a study if you were tired of thinking about it and just wished it would go away? 3) Putting the time and effort into such a study thus requires a significant investment in a heteronormative worldview, an obsession that is unusual in society in general, and academia in particular. Gay and lesbian researchers have an obvious interest in this sort of work, but it takes a heterosexual with a real bone to pick to become similarly invested.

Reaction C: I've been perplexed by the widely-held beliefs among the religious right that they are being persecuted by homosexual activists, and that our gains in society have come at their expense. I know that there's no conspiracy to reign in the religious right because I've seen first-hand how LGBT folks organize. We are way too fractious to pull something like that off intentionally. By the same token, I'm deeply suspicious of claims that "the church" or "the Mormons" are acting in concert as often as we think they are.
In the 90's I tried to do a lot of activism around victimization, and I really think that is a self-defeating way to go. It makes you more paranoid and can become self-fulfilling. So my word to both sides - leave the persecution stuff off the table - it doesn't help anyone.

Reaction D: I wonder how a study on family structure could be done in a methodologically convincing way. It's not easy. Regnerus tried (and failed) to get something close to a random sampling of the general population. That's a tough approach to use because children of lesbian and gay parents are still pretty uncommon, and that's the main reason his method failed. The approach mainly used by our side is to find families headed by same-sex parents and try to find a comparable comparison group of mixed-sex-headed families. That's a tough approach because it is very hard to be sure that the comparison group really is comparable. I think the best approach that might be feasible in the short-term would be to piggy-back on some other very large random sample of Americans and do a follow-up survey with all the same-sex-headed households and a matched sample of mixed-sex-headed households. The Current Population Survey would be, I think, an ideal vehicle for such a call-back survey. They interview about 50,000 Americans every month, so there might just be enough same-sex-headed households contacted through that survey to make it feasible. The Behavioral Risk Factor Surveillance System might work too, but it would be a huge logistic challenge to get permission from each state to call people back. The American Community Survey could work too, but because that is done by the Census, we would first need to get Congress to admit that same-sex marriages do in fact exist, and are worth studying.

Reaction E: Why is it important to compare the children of same-sex to mixed-sex households? I'll admit that it is interesting from an academic perspective, but I think most of the interest is generated by the desire to use evidence in policy debates. But should it matter?
A lot of the debate so far has centered on whether the children of same-sex couples are more likely to "turn" lesbian or gay themselves. Most of the studies on 'our' side have claimed that the answer to that question was no -- because our opponents were so fiercely complaining about gay contagion. But I think it's safe to say that the evidence is that kids of lesbian and gay parents are in fact more likely to realize that they are gay, lesbian, and especially bisexual. In 2009, I heard a great talk by Clifford Rosky which really pushed the audience to ask, "So what?". So what if gay, lesbian, and bisexual kids are more comfortable, more self-realized, after growing up in our households? Isn't that a good thing? (The Regnerus study counts being openly GLB as a 'negative' outcome, by the way!)
And that leads me to wonder what possible relevance the Regnerus study, or the studies on our side, should have in regards to public policy. Of course it would be easy and convenient if the children of same-sex parents were equal in all regards to the parents of mixed-sex parents. But would it really matter if that weren't the case? Shouldn't we expect that the children of same-sex parents would be worse off in some ways and better off in others? And even if the impossible were true: that children of same-sex parents were, on average, worse off in every possible measure, should that preclude every same-sex couple from having children or being able to marry? I'm struggling to see the relevance. The children of wealthy parents are much more likely to be diagnosed with autism - should we sterilize the rich?

No easy way I can see to wrap this all up. Thanks for listening, and feel free to chime in!

Health Disparities: Getting Worse or Getting Better?

Yes. Both. Simultaneously. With the exact same data.

For my general rip on health disparities research, see the previous (lower) post.

So the other day, I heard someone say that in almost all cases, (racial) health disparities are getting worse, not better. Which, of course fits perfectly with the scare tactics that much of public health relies on these days. I think when we public health types hear that, we think it means we need more resources to address health disparities.

But in the back of your head, doesn't it also say: all the work that's been done on reducing health disparities has been an abject failure, nothing we do works, and it's disheartening to contemplate moving forward? Wouldn't it be an even more powerful motivator to garner resources to be able to say "Hey, look at this, things are getting better". If you were a funder, wouldn't you rather build on success?

But the numbers are the numbers, right?

Actually... depending on how you look at the numbers, whether things are getting better or worse is all a matter of perspective.

Arguably, the most important health indicator of them all is mortality.

So, I went to one of my favorite websites, wonder.cdc.gov, and looked up trends in mortality by race and ethnicity over the past few years, and then threw it into a graph.

When I look at that graph, there are a couple things that jump out - first that there are rather enormous disparities in mortality by race in the US. Second that Hispanics and Asian/Pacific Islanders have much lower mortality rates than Whites. (How often do you hear about that?) And third, death rates for all racial/ethnic groups are declining fairly rapidly.

It isn't immediately obvious from this graph whether the racial disparities are getting wider or narrower, the differences between these lines looks pretty similar over this 9 year period.

So, let's look at the disparity using the official method.

You pick the healthiest group, then divide the other racial/ethnic groups by that group to get the relative disparity. In this case, the healthiest group is the Asians and Pacific Islanders, so you get this graph:

It's pretty clear that relative to Asians and Pacific Islanders, the disparity is essentially unchanged for Hispanics, and has increased for Whites and Blacks.

But I find this way of looking at things strange.

It obscures the fact that Whites are the dominant group in the US, so I did the same analysis, but using Whites as the reference group.

When you hear the term "racial disparity", what is the reference group that springs to mind?

At any rate, this picture is a bit different. The Black:White relative disparity in mortality is actually declining slowly over time, while the relative health advantage of Hispanics and also Asians and Pacific Islanders is increasing over time.

But the interpretation of whether

health disparities are getting better or worse over time depends on something besides which racial group represents your baseline - it also matters whether you divide (the standard method) or subtract.

When you plot the difference between Asian and Pacific Islanders rather than the relative disparity, you actually get a different interpretation:

that the difference between Asians and Whites has not been changing much over time, but the disparity between Blacks and Asians has been improving, the exact opposite conclusion one would reach from looking at the relative disparity.

And when you use Whites as the reference group for the difference, these results are pretty similar:

The Black:White disparity is clearly diminishing over time, the health advantage of Hispanics has been increasing slowly, and there is no apparent change in the health advantage of Asians.

I don't want to bore you with the math that explains all this, but I guarantee I haven't done anything funny. It's just that when you subtract, most health disparities appear to be on the decline, but when you divide (which is the standard method), most of the health disparities appear to be increasing.

So both are true simultaneously. Which raises the question as to why dividing has become the standard? I have two explanations - one rooted in history, the other rooted in computerized estimation methods. Neither rooted in any logic the least bit related to health disparities.

The first reason has to do with the history of trying to figure out if smoking caused lung cancer. In post-war England, a couple doctors (Sir Austin Bradford Hill and Sir Richard Doll) did a series of studies to investigate whether smoking could be causally linked to lung cancer. Not surprising in retrospect, they saw that smoking had a big influence on the development of lung cancer. Whether you divide or subtract, lung cancer was clearly more common in smokers than non-smokers. But there was a pesky finding in relation to heart disease. When you subtract, it looks like smoking causes about as many heart attacks as it does lung cancer.

To their thinking, it made sense that smoking would affect the lungs, but why the heart? Fortunately, when you divide, rather than subtract, it looks like smoking has a huge influence on the risk of getting lung cancer, but is associated with only about a 20% increase in heart attacks.

From that, they concluded that dividing must be the right way to think about it.

But subtracting is just as valid mathematically. All these numbers say is that smoking causes about the same amount of heart attacks as it causes lung cancer, but because lung cancer is so rare in non-smokers, that makes a huge relative difference, while heart attacks are pretty common among non-smokers, so a small relative difference still adds up to a lot of heart attacks.

The other reason that public health types divide rather than subtract is that computers like to divide. Well, that's not really true, they don't care. But the complicated statistical methods that we ask computers to do almost all rely on dividing as the one and only way to compare two numbers. The statistical methods that rely on subtracting as much harder to implement, and are often impossible. So what I like to do is throw the complicated statistical methods out the window and calculate the differences using more labor-intensive methods.

All of which is by way of saying that next time you hear that health disparities are getting worse, take it with a grain of salt. It probably depends on how you do the numbers.

My Rip on Health Disparities Research

A few days ago, I gave my final lecture in my epi class, partly an overview of social epidemiology, and partly a rip on health disparities research.

The short version:

Identifying and describing health disparities:

- is a great way to get grant money.

- seems like a great way to raise awareness-

--- in order to mobilize greater resources to address the problem &

--- in order to mobilize members of the 'target population' to take preventive action.

- can be demoralizing to the 'target population'.

- can induce feelings of helplessness in the 'target population'.

- can make ill health seem inevitable & expected in the 'target population',

--- potentially reducing the urgency for action.

- describes the result, not the causes-

--- if you believe genetics is the cause, the existence of health disparities supports that notion

--- if you believe lifestyle is the cause, the existence of health disparities supports that notion

--- if you believe racism is the cause, the existence of health disparities supports that notion.

- don't identify potential public health actions to address health disparities

Furthermore, the almost exclusive focus on adverse health disparities:

- is demoralizing to the 'target population'.

- enhances the sense of social distance between the 'target' and 'majority' populations.

- obscures the true picture of a mixture of adverse health disparities, health similarities and advantageous health disparities.

Ignoring advantageous health disparities

- blinds us to the possibility of understanding the mechanisms for these advantages

--- which could lead to more effective public health measures built upon those mechanisms to address the adverse health disparities

--- or could lead to building on those mechanisms to improve the health of populations other than the 'target population', including the 'majority' population.

After I Left AIDS - Part III (more thesis)

I didn't want to study suicide.

Mainly because suicide is a bummer of a topic. It reminded me of unpleasant memories from adolescence. And whenever I talk about it, the first thing everyone does is get quiet - then they get concerned about my well-being. Which is nice and all, and I appreciate it, but after working on this stuff for a few years, I would forget the level of emotional charge the topic has, and get really excited about some finer point of data analysis, and come off sounding callous when really all I wanted to share was this exciting little piece of the puzzle.

On the other hand, epidemiologic studies of suicide go way back (to Durkheim in 1897, and before him Morselli in 1881), and unlike most health conditions associated with sexual orientation, suicide has been measured in a consistent way across the whole population for an extended period of time. So, in a sense I was stuck with it as the only health outcome that had both geographic and temporal scope, which is what I needed to look at normative heterosexuality.

So anyway, as I mentioned before, I wanted to look at how heteronormativity (a shared set of assumptions about sex, gender, and who ought to be having sex with whom) affected suicide rates.
At first, I wanted to find a data set where I could could compare gay men, lesbians, and bisexuals to heterosexuals. But the death certificates don't have that kind of information. And as I got to thinking about it, even if they did, how reliable could it be?
And that got me to thinking, maybe the sexual orientation of these people is really beside the point. Perhaps the stresses associated with dealing with assumptions of heterosexuality are greatest among people who don't identify as "gay" anyway.

So, the first study I did was to look at gay rights laws as a measure of heteronormativity, the idea being that in order to enact a gay rights law, politicians have to believe that public opinion is such that they'd be better off protecting sexual minorities from discrimination than not. The first gay rights laws were enacted in 1973, in San Diego and Austin, I believe. In 1981, Wisconsin was the first state to pass a gay rights law, and by 2003, most of the country's population lived in a jurisdiction with a gay rights law. (the gray map there has a nifty time-lapse).
I looked at three levels of gay rights protections, in order to get something like a dose-response curve - the red areas had no protections whatsoever, the green areas were protections for public sector workers only, and the blue areas had protections for both public sector and private sector workers.

And the results here are pretty compelling - at least for White males, particularly adolescents, young men, and the elderly.
Each color in this graph represents a different age group. So, among White males aged 15-19, suicide rates were 179 per million in areas with no gay rights protections, 155 in areas with protections limited to the public sector, and 131 in areas with protections for all workplaces. The only group without a step-wise dose-response was White men aged 45-64.

Among White women, the first thing to notice is that suicide is less frequent, and also doesn't increase among elderly white women, unlike men. The decline in suicide rates with increasing levels of gay rights protections is also not so pronounced, but there are declines in each of the age groups under 45.

Suicide is less common among Black men than White men in the US, but is still pretty high. And unlike White men, the peak incidence of suicide is in younger age groups. But what is strikingly different is that the highest suicide incidence among Black males is in areas with the highest levels of gay rights protections, which suggests to that public opinion among Black populations about homosexuality may not be strictly related to public opinion among White populations from the same area, and presumably the enactment of gay rights protections is, in most jurisdictions, reflective mostly of White public opinion. I'd love to do an analysis based on what might be a better measure of heteronormative assumptions in Black communities. Any ideas?

Among Black females, the incidence of suicide is lower than the other populations above, and like White females, declines among older women.
The differences between areas with and without gay rights protections are not large, but in general, suicide rates among Black women tend to be slightly higher in areas with gay rights protections. So these results also raise questions about whether gay rights laws are a good measure of heteronormativity for all populations. Or alternately, if the social forces leading to suicide are perhaps not identical among White and Black populations - perhaps heteronormative assumptions cause more distress in White populations, particularly among White males, while economic issues and racial discrimination play a larger role in Black populations.

Another consideration is that perhaps the stresses induced by heteronormativity are largely related to the performance of masculinity, which is why men turn violent against themselves under these pressures. Perhaps men under heteronormative pressures also direct violence outwards towards the women closest to them, and thus homicide, rather than suicide, might be a more strongly related outcome among women. That's foreshadowing to an analysis I'm thinking about doing next...

The patterns I noted are virtually unchanged after adjusting for a wide variety of potential confounders, namely population density, region of the country, unemployment rate, poverty rate, and measures of social isolation (proportion living alone, proportion who moved in the last five years).
Also, when I looked only at those areas that changed status (went from no protections to having gay rights protections), the same trends held up, so in order to explain these results, some other factor would have to be changing at the same times in the same places, which seems like too much of a coincidence to be possible.

The trends above are very similar when I looked at how people vote on the restriction of marriage to "one man and one woman" as a measure of heteronormativity, but as I mentioned before, the strong trend towards people being less likely to endorse a restrictive definition of marriage makes this measure a bit more complicated, so I'm trying to figure out how best to represent it.

After I Left AIDS - Part II (Thesis)

So, after I left AIDS, I got thinking about how homophobia, as a societal norm, affects health. Not just queers' health, but how it also affects the health of the whole population.

In my last post, I talked a bit about my journey through thinking about health disparities, and how nobody seemed to be measuring the causes of these disparities. That leads directly to my doctoral thesis, which was about how to measure normative heterosexuality, and from there, estimating the impact of it on suicide. Not just on "gay" suicide, but suicide in the whole population, and also in various sub-populations defined by sex, age, and race/ethnicity.

So, following the lead of thinking about residential segregation by race/ethnicity, and income inequities, I began thinking about how to measure normative heterosexuality, the presumed cause of the health disparites that epidemiologists had begun to document with greater and greater precision.
How do you measure the degree to which a group of people (a large group of people) share a rigid set of beliefs about sex, gender, who ought to be having sex with whom, and how? My first thought was that the frequency of hate crimes directed against gay men would be a good measure. If this set of rigid beliefs dominated a social setting, then the informal "enforcement" of those beliefs would be enacted through the commission of bias-motivated crimes, presumably mostly by young men with "something to prove".

When I pulled the data down off the FBI's Uniform Crime Reporting (UCR) System, I quickly realized something was amiss. San Francisco had by far the highest number of anti-gay hate crimes in the country, and several Southern and Mountian states reported not a single one.

I've put more recent statistics by state in a table, based on numbers from 2004 to 2008, the five most recently reported. Basically the same trend holds - bias-motivated crime tends to be higher in places we think of as gay-friendly, and extremely low in the deep South. Then there are also strange jurisdictional oddities - Pennsylvania for example appears to have an extraordinarily low rate of bias-motivated violent crime.

The way I've come to understand this data is that it represents not the phenomenon of crime occurring, but rather on two phenomena: 1) how comfortable victims feel about reporting a bias-motivation to law enforcement, and 2) local law enforcement customs and legal constraints about recording and validating these reports. If it was just the first of these, then one could use the reporting of hate crime as a measure of homophobia at a societal level, that is the more hate crime reported in an area is evidence of how little homophobia there is there, as perverse as that sounds. But alas that second factor, particularly the bit about jurisdictional quirks in how different local law enforcement agencies deal with the reports that are made to them, really throws the whole thing off.

So, I couldn't use hate crime statistics. But maybe I could use the presence or absence of a law for reporting hate crime statistics that specifically included sexual orientation. Or, how about the presence or absence of a law prohibiting discrimination on the basis of sexual orientation?

So, the next thing I looked at was which states had gay rights laws, and when they were enacted. Various of the states have enacted gay rights laws over the years, the first being Wisconsin in 1981, a few more in the late 1980's, and a lot during the 1990's. Recently, state-by-state gains have slowed considerably, as gay activists have pressed for a national law (ENDA), or been distracted by the marriage thingy.
The point for my purposes is that the enactment of state-wide gay rights laws has been a pretty hotly-contested issue, debated for years within each state's legislature, rather than by a small cadre of legalistic judges, or the flash of public opinion of a referendum. As a result, the enactment of a gay rights law represents something of a local watershed, the point in time at which the balance of adverse consequences for elected officials switches from a net negative to a net positive.
So, looking at the enactment of gay rights laws seemed to hold promise, at least from a theoretical perspective, as a good measure of the broad social environment of a State in regards to the level of normative heterosexuality.

Another potential measure of normative heterosexuality to be considered is public opinion polling. The gay rights law thing seems a bit crude - a yes-or-no variable to measure something which I claimed varied by degree from one place to another, and one time to another within those places. Public opinion polling, on the other hand, offered the promise of a finely-tuned measure of normative heterosexuality. There are some relevant questions that have been asked the same way for decades. For instance, Paul Brewer has examined the time trends in how Americans feel about the "wrong"-ness of same-sex sex, which increased during the AIDS years, followed by a precipitous drop recently, the majority of Americans now saying it is not "always wrong" (small consolation that!).
So, public opinion polling looks like it might be a better "thermometer" to gauge how people feel about homosexuality. And there is longitudinal data to work with, so I could look at changes over time.
On the other hand, public opinion polls, by design, ask the smallest number of people possible in order to get accurate results. Thus a "large" national poll might have only 500 respondents. The GSS from which the data above is generated is a good bit larger than that, but still it is only a few thousand in any given year. A few thousand sounds like a lot of people, but what I needed to do was compare across places, not just time. So a few thousand breaks down into a few dozen in some states, and in others, fewer than ten. It would be a stretch to characterize the whole State of Connecticut based on how 15 randomly chosen people answered a question (for the record, I'm pulling that number out of thin air, but that's about what it comes down to).

So I was stuck with public opinion polling, too. Good temporal trends, but lousy in terms of geographic specificity.

A related idea was to look at how people voted on anti-gay referenda, such as the Briggs Initiative in California in 1978, Measure 8 in Oregon in 1988, and Colorado's Amendment 2 in 1992. These explicitly anti-gay referenda had the advantage of high geographic specificity, presumably accurate down to the precinct level, but represented a snap-shot in time. Also, they represented a small number of states, and the questions addressed in each one were quite different.

While I was working on my thesis, though, another opportunity to think about voter referenda came up. The issue of same-sex marriage cropped up. Although same-sex marriage has been contested in U.S. courts since 1970, it had never gotten much notice one way or the other - the Christian right didn't feel threatened by it, and most gay acitivists thought marriage was a non-starter politically, or at any rate a horrid reminder of heterosexuality run amok that should not be emulated.
But in 1998, Hawaii and Alaska voters chimed in on same-sex marriage, a few more did in the 2000 and 2002 elections, and then the 2004 election was swamped with voter initiatives to restrict marriage, in part a cynical manipulation by Republic Party operatives in order to keep their guy at the helm.

These referenda share the problem that opinion polling data have, in that they are a snap-shot in time (except for a few states which have had multiple referenda on this issue), but there were major advantages. For one thing, the question being asked was nearly identical in every state, some slight variation on whether legal recognition of marriage should be restricted to "one man and one woman". As an aside, no state has yet offered to restrict marriage to "one woman and one man" - something to consider when thinking about marriage as a forum for liberty and equity. And, the geographic scope was huge, with most states chiming in on the issue one way or another. The map I made here shows how different areas voted, from strongly in favor of restricting marriage (red) to being against restricting marriage (dark green).
On the whole, this map comports more or less with what one would expect, there's more red in the rural areas, more green in urban centers and on the Pacific coast, and there seems to be a trend towards more green in the Northeast. But there are some unexpected spots, too, such as South Dakota, which was substantially less in favor of restricting marriage than its neighbors Nebraska and North Dakota, And Arizona, which was the first state to reject restricting marriage in 2006 (alas, they went to the dark side in 2008).
So, there are some tricky issues to deal with in using this data. I haven't quite figured out how to make it comparable across time periods.

The final method I've thought of for measuring normative heterosexuality is using counts of same sex couples. The number of same sex couples was counted (albeit inadvertently) by the U.S. Census in 1990. For the 2000 Census, they did a better job of it, and the upcoming 2010 Census is expected to do better yet.
In any event, the number of people who identify themselves as married same-sex partners and un-married same-sex partners in the Census is probably mostly a factor of three forces: 1) How comfortable people in same-sex couples feel identifying themselves as such on the Census forms; 2) The degree of selective in-migration and out-migration of people in same-sex couples (or destined to join one), and 3) The degree of confusion by people in mixed-sex couples who inadvertently identify themselves as same sex partners.
The first two of these factors (net migration and comfort identifing as a same-sex couple) are related to what I want to measure - how accepting an area is of homosexuality. The third factor is a pain in the butt, not in a good way. I've discussed that issue at length before.

So, counting same-sex couples has two huge advantages: it uses the same methodology for the entire United States, and you can get comparable data down the the neighborhood level (census tracts). On the other hand, the data itself has some big caveats - it doesn't identify young people, single people, or couples living in separate residences, and it is essentially useless when considering older people (for reason 3 above). And although there will soon be three time points to compare, the methodology has changed in each Census, and it remains to be seen if the 2010 Census data will be comparable to the 2000 Census data (probably not, but for the reason that the methods are becoming more accurate).

So, in the end, I decided to pursue three measures of normative heterosexuality further:
1) The enactment of gay rights laws,
2) How people voted on referenda to restrict marriage to one man and one woman, and
3) The proportion of same-sex couples identified in the Census.

More to come...

After I Left AIDS - Part I

About a month ago, I wrote about Why I Left AIDS, but didn't get around to what I'd moved into.

While I was working in gerontology, and started taking classes again in public health, I was trying to figure out what I wanted to do research on. I knew it wasn't HIV/AIDS, and most of the other health outcomes related to gay men (suicidality, depression, substance abuse) were kind of downers. The depression bit hit close to home, and the substance abuse felt completely foreign to me, so I didn't really know where to go.
At the time, in gerontology, I was working on a variety of measures of regional variation in social conditions to try to explain health disparities. We had noticed a big difference in the occurrence of pressure ulcers (bed sores) by racial identity. While it was interesting for me to crunch large datasets, and to work with colleagues to figure out a narrative that might explain the health disparity, documenting the disparity and theorizing about why it occurred seemed unsatisfying. I wanted to measure the cause, not just the effect.
I had also been a teaching assistant for the epidemiology class at Brown for many years at that point, and we always chose an article about the link between residential segregation along racial lines and some health outcome or another, usually birth weight or premature delivery. The idea was that racial segregation, the separation of people in space, reflected social segregation, or the history and current strength of racial hierarchical ideology.
So it was a natural connection to say, hey let's look at whether the health disparity in bed sores is larger in cities characterized by high levels of racial segregation than it is in cities where people are more evenly distributed.
Unfortunately, we never got around to writing that paper (I don't think we even got to the analysis stage before I moved on), but the point is, I spent long hours figuring out how to measure, in a quantitative sense, the racial segregation of where people live, and also the levels of economic disparity (gap between the rich and poor), and how these measures vary across the U.S.
Roughly at the same time, I began to think that I wasn't so interested in documenting the extent of health disparity there was between gay men and straight men, lesbians and straight women (and there was almost no literature on bisexuals, and even less on gender identity), I was interested in measuring what causes the health disparities that do exist.
At first, I tried to think about measuring homophobia in high schools. In my own life, high school was definitely the most homophobic environment I had survived, after all.
I figured that one way to do it was to ask a wide range of students who had graduated and gone on to college to rate their high school environment with regards to homophobia. Having gone to college, they would have at least one other environment to compare to, some perspective. And by asking them about the school environment, rather than their personal experience, then queer kids and straight kids would both have valuable and relevant insights and perceptions on the issue.
I developed a 20-some odd page questionnaire and tested it on maybe 40 or so Brown undergrads, queer, straight, and in-between. At that point, it was an exercise for a survey design class I was taking, so I wasn't particularly interested in scientifically important questions like inter-rater reliability, I had a much more mundane purpose - did recent high school graduates feel like these questions made sense? Were they salient? Were they getting at what I intended them to get at? and was I missing anything?
It was a great experience (for me, anyway). And the questions did make sense (most of them), they were on target (most of them), and there were a few things I had missed. I was convinced that it was worth taking it to the next stage.
I began thinking about how to use it. It was one thing if one could describe the social environment, it was another to use it to predict health or health behaviors. In conversation with a student (Marc), we had an idea - which was to measure the homophobia at a series of high schools where more than one student had died of self-inflicted injuries to high schools where no student had died of self-inflicted injuries in many years, and to measure the extent to which the school social environment was infused with homophobia in both sets of schools.
And here's an important point - whether the students were queer, straight, or in-between was not relevant to our plan. It wasn't going to be a study about who had killed themselves, but about what sort of environment drives people to the point of ending their lives.

So maybe you're seeing a thread here already - the vast majority of research literature on queer health is about documenting the bad things that queers (and usually gay men specifically) are at higher risk for. But I wanted to take a different tack - I wasn't so concerned with what the specific health outcomes were, but the cause of them, and specifically, the cause in the sense of the social environment.
And this opened up a new possibility - examining the influence of the perfusion of homophobia in social environments not just on queer people, but on the whole population, on straight people too.
My involvement with ActUp/RI was highly influential in getting me to think about homophobia as a health hazard, but in that context, I thought about it as the reason the government was letting gay men die without saying a word, literally. Or when words were spoken, they would be words of condemnation, threats of quarantine, of judicial prosecution for having an infections disease, of punishment for exposing the "general population" to a scourge that we deserved but they did not.
Instead, I was now thinking about homophobia as a threat to the whole population.

More to come...

Breast Cancer Screening Controversy

I'm going to be teaching two sections of epidemiology this Spring, one for grad students, one for undergrads.
The grad student version I'm pretty confident about, but I want to change a few things, especially the cumulative paper that I ask the students to write throughout the semester.
The other thing I'm thinking about is pulling in the breast cancer screening controversy, which seems to have long legs, re-appearing in the news on a regular basis. I had been thinking about H1N1, but to be perfectly honest, it hasn't been able to attract my attention (not the way the 1918 war-fueled epidemic did anyway).
For the undergrads, I'm trying out a new textbook (new to me, anyway), which has more pictures. I haven't been able to find a good textbook for undergrad epi, and the worst are the ones that say that that is their target audience.

Anyway, back to breast cancer screening. I think it's a great issue to tussle with. It has a lot of emotionally laden content in addition to "the science". The science itself is complex and fascinating, and really engages all forms of epidemiologic study designs, from case-control studies to massive experimental trials, and concerns epidemiologists have about sources of error and misleading results.

Also, one of the pioneering epidemiologic researchers was Janet Lane-Claypon, who did a case-control study comparing 500 women with breast cancer to 500 women without breast cancer, and confirmed most of the risk factors that we now know have a large influence on the development of breast cancer, in 1926. I like having a historical focus in my class, and it bugs me that that means reading exclusively male writers in a class that's predominantly made up of women.
I'd also like to include more of the large corpus of early writing from Spanish language authors, but I'm not familiar enough with it, and the few pieces I have seen translated just wouldn't fit well into my curriculum. (Perhaps it's time to expand my curriculum, then!)

But back to screening. I myself didn't think much about breast cancer screening, until my mom got a positive mammogram. It pretty well freaked her, and me, out. Weeks of anxious anticipation were not erased after minor surgery removed what turned out to be perfectly benign calcified lumps. But still, what if it had been cancer, wouldn't it have been good to know earlier rather than later?
The more I've thought and read about it, the more I've come around to a different point of view - it probably wouldn't have been better to know about it earlier. I know that sounds harsh to anyone with breast cancer, and easy for me, given that it wasn't breast cancer. But I don't say it glibly. The unnecessary anxiety, the unnecessary (if minor) surgery, these are not benign side effects. They may be mild inconveniences compared to mastectomy, chemo and/or radiation. But really how many unnecessary side effects are we generating with screening mammograms compared to how many treatable breast cancers that get detected (and wouldn't be equally treatable after they grew a bit and became diagnosed by other means)? How many breast cancers are detected and treated with highly toxic and invasive methods that, left alone, would never have caused a problem? Those are complicated questions that are technically challenging to answer.
Then, there's also an issue of where we, as a society, spend money. I don't think that costs should be a determinant of what health care people get. In a previous post, I lampooned the idea of doing a cost-benefit analysis of vaccination against HPV. The more effective a vaccine campaign is, the less cost-effective it would be, so it's just silly to do a cost-benefit analysis in the first place.
But at the same time, one wonders if all the attention paid to promoting mammograms as the one thing you can do to prevent breast cancer has crowded out other means of preventing breast cancer. Methods that may be less sexy, and less under an individual's control. Why does preventing breast cancer have to be something each woman does for herself? What about pesticides and environmental pollutants that probably have a very small influence on any one woman's risk of getting breast cancer, but by increasing all women's risks somewhat, have a large societal impact? What about the disparities in the levels of these pollutants that often mimic disparities in class and race in this country? What about addressing the structural poverty and disenfranchisement that keeps women from having symptomatic breast cancers dealt with early on when it's more treatable? These methods at least give men something to do!