Distortions: Bias and perspective. Aka: follow the money

In this business, the research business, there is a general understanding that money can affect bias. Most responsible scientific publications these days require researchers to declare conflicts of interest. This is important so readers can place the research they are reading into some kind of context, and identify if a bias might exist.

So for example, if a researcher is studying diabetes and finds that a certain drug helps to reduce problems, but that researcher has funding from the company that makes the drug, readers can take that conflict of interest into account. It does not mean that the research is invalid, but that one should pay closer attention to the methodology, maybe frame the conclusions with some doubt, or even try to reproduce the research themselves (if they are not funded by that company) to see if they get the same results. It is key to the pursuit of knowledge that we know the context in which the knowledge producers are located.

The alcohol research field is no different. If a researcher or a writer were funded by the booze industry, and if he or she found some kind of questionable benefits to alcohol consumption or provided evidence to question claims of the problems of alcohol consumption, his or her research could be suspect. It would not mean it is invalid, but, like the hypothetical diabetes researcher discussed above, it would be important for readers to be cautious about accepting the research without question.

Implicit within this need to declare conflict of interest is the understanding of the corrupting nature of money. We as a society assume that for-profit businesses (like drug companies or liquor companies) want to skew information in their favour so they can make money (or limit the losses). We further assume that researchers will be “influenced” by generous funding from these companies, and will therefore publish results that are biased towards their benefactors.

This is entirely reasonable, and let’s keep this in mind: Money can affect outcomes.

It is worth noting that in these cases, the researcher may not be receiving an income from the corporations involved; he or she may simply have received funding for his research, to pay for research assistants, software, materials, time, possibly release from teaching load to allow for more time to research. Nevertheless, even if the researcher is not enriched by the funding, their research results are questioned, and rightly so.

It certainly is something I have faced numerable times since I began writing with concern about the CCSA’s new guidance on so-called low-risk drinking. When I was interviewed on an AM radio station in Kitchener-Waterloo late last year the host asked me at the end of the interview if I had any funding or connection with the liquor industry. When the producer at CBC’s The Current was finishing her pre-interview with me, she asked the same thing. When my op ed in the Globe and Mail was published, comments included questions about my funding. (Well they were not so much questions as assumptions that I must have alcohol funding)

My response to all of these questions (except the commentator on the Globe’s website, to whom I did not reply) was: I have no funding to do this writing at all. The only connection I have with the liquor industry is as a customer. In other words, I give them money.

But normally I also follow this up with a caution and an observation: we assume that only industry funding can bias outcomes. This assumption is erroneous.

Just as industry funding might bias research results, if we accept the statement above (money can affect outcomes) then we have to accept that other types of funding can also affect outcomes.

It is important to understand a bit about the research ecosystem to get what I mean.

Think tanks (like the CCSA) and research centers (like the Canadian Institute for Substance Use Research at the University of Victoria) depend upon funding for their operations. You can’t do the sort of complex research without it. Researchers do not work for free. Administrators are not volunteers. Someone has to pay for the electricity and the heat; the travel and the research assistants; the computers and the software; the websites and journals…. You get it.

Normally this money comes from government sources like Health Canada and the Canadian Institutes of Health Research (CIHR), but it can come from other places. For example, private benefactors making charitable donations to universities can identify where they want their money to go and if they like what you’re doing they might send it to your institute.

This is all nice, you might say, but what has this to do with bias? Well, a few things.

First, in the university system we are assessed on our research, teaching, and service to the scholarly community. Annually my dean wants to know 1) what I have published 2) how much money I have brought in in the form of grants–and the most valued are big government grants–and 3) something about my teaching and admin work.

Deans do not normally ask about what kind of public outreach I have done, how my policy advocacy work has been going, and if I have dedicated my time to work for positive social change. Not only is it difficult to track, but it’s impossible to quantify.

Our Deans do ask about things like talks we have done, media contacts we have had, and publications we do in non-academic sources like newspapers, but they are not as important as academic publications and grant funding (this blog means nothing to my dean).

There is nothing wrong with this (although some of us would like our community work valued more than it might be). It’s the system in which we were trained and the system in which we have agreed to operate. It does mean some of my brilliant colleagues don’t bother to provide information to the media or do more accessible public outreach because it doesn’t matter to their research – which is too bad because their insights would be valued – but it is the system we have, and we accept as part of the gig.

Again, you might say, so what? Let’s summarize before the big reveal.

  1. Funding and publications drive careers
  2. Funding from government is crucial
  3. Benefactors will give money to your institute to help pay the bills.

If all thing were equal, this would not be a useful topic for this blog post. But in the alcohol studies field, all things are not equal.

This is because the priorities of funders shape the direction of research. And in the drug and alcohol field, the priorities of government funding are harm: dangers of addiction, negative health impacts of over consumption, violence or injury relating to consumption, and so on.

If you are a researcher hoping to do work on the positive social benefits of alcohol, you will go broke. Unless, that is, like some of my clever colleagues who don’t look at alcohol but look at some other aspects of social life like social connectedness, and can add questions about alcohol consumption into them.

So here is the bottom line: alcohol researchers, hoping to build their career and remain employed, are incentivized to look for harm. They are harm focused. Their job is to look at problems, and it should be no surprised when they find them.

On top of this we have the sources of data. Since people only end up in hospitals, morgues, and police records as a result of some bad thing happening, and if that bad thing is connected to drinking, then it is easy to trace the harmful outcomes of drinking. This is a form of confirmation bias: we are funded to look for harm, the data is skewed to track harm, and surprise, surprise, surprise: harm!

Well duh. Excessive drinking can cause bad things… but not only bad things.

The problem is, how do you quantify that? It’s hard, but not impossible.

So how do you research that? Well, it’s not impossible, but it’s expensive.

Crunching data is expensive enough. Drawing data from existing health surveys, hospital records, police records, death records, etc is time consuming and therefore costly (remember nobody works for free, although some students may do work placements that exploit them in the name of education). Hence you need big grants and good infrastructure to do it.

If you wanted to do something like track people’s drinking behaviour, link it to sense of well being, and then connect it to positive health outcomes… how do you do it? (This is entirely do-able, actually. But it is neither cheap nor fast).

So if you can’t find funding to do this, you will not be able to do it to a degree that provides rigorous and convincing data. You could do a small cohort study, but with small numbers comes small evidentiary strength (this is why data types refer to the total number of data points as representing the “power” of a study. The more data you have, the stronger the statistical relevance and thus the power of the evidence).

Drawing numbers from existing data that is collected as a matter of policy (again, hospital records, accident records, police records, vital statistics) provides lots of data points and high “power.” An expensive study that links drinking and social well being but has far fewer participants is often considered (by those health researchers) as of low quality due to the numbers involved and the methods used to gather information. (How do you measure “well being” for example, which is subjective, as opposed to, say, death rates, which are objective).

On top of this, scientists who look at things like the positive benefits of wine on the heart are similarly disadvantaged. I was discussing this with a colleague who studied resveratrol, the substance in wine that has been linked to lower levels of cardiovascular disease. She told me she stopped doing her research because there was no funding. As a result, you get researchers at the CCSA saying the evidence for the protective effects of red wine being weak or nonexistent, when really it is a yawning funding and research gap between people looking for benefits, and people looking for harms. (BTW my colleague could have gone looking for funding from the wine industry, but she knew that this would make her research look biased. Instead, she changed her research focus.)

So let’s wrap up: nobody works for free; the funding environment is skewed to favour research that looks for harms; people therefore seek funding to look for harm; and the vast bulk of the research on alcohol consumption is about harm.

Is it any wonder, then, that there is a large cohort of researchers arguing that all the best, “high quality” research shows harm? And they can be flown around the world to conferences at holiday destinations, put up in nice hotels, and paid substantial honorariums for their time by public health oriented government or non governmental agencies to talk about the harms, and they are pure as a newly christened baby, a triple distilled neutral spirit, or, if you must, the driven snow.

But, as a colleague of mine once noted, “share a lift with someone from the local brewery…, and your moral compass explodes.”

Next time you wonder about the funding and bias of researchers, ask if funding that is aimed at finding harm can lead to skewed results. It can. I’m not saying it does. I’m not accusing my colleagues of being corrupt. I’m just pointing out the fact that not having industry money does not exempt you from having a bias.

And while you’re at it, ask why someone who has no funding to write about this is seen as more biased and assumed to be corrupt, as opposed to those on whose careers such research depends.

(c) 2023 Dan Malleck

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s