The Law Of Unintended Consequences

So yesterday’s post introduced the law of unintended consequences. Hopefully, this wasn’t the first time you’d heard of this idea, but just in case it was, let’s define what is meant: The law of unintended consequences is that every action has consequences. Imagine this as ripples in a pond when a rock is dropped in: rock creates the first splash, but that splash then ripples outward and the ripples can have impacts themselves, separate from the initial impact of the rock into the water. It’s worth noting that unintended consequences may not be unwanted consequences; sometimes, the ripple effect turns out to be exactly what you want. However, if you experimental design goes wrong, you can probably guarantee that the cause was one of these ripple effects gone wrong.

Let’s start by looking at a simple example: you’re sitting in a boat on a lake. You’ve got a full glass of an icy cold beverage. You pluck out one of the ice cubes and decide to toss it into the lake. It makes a nice splash, and you see it ripple outwards. You decide to try this again, but with a bigger chunk of ice. You set your very full glass of icy cold beverage down and reach into the cooler, fishing out an enormous chunk of unbroken ice. You toss this overboard with a decent amount of force. You get the same satisfying splash, but this time, the ripples are substantial enough that they rock your boat. This is enough to spill your icy cold beverage. As you sit back down, your pants get wet in the spilled icy beverage. Jumping up in alarm, you rock the boat further, and as this is actually just a little boat, it’s enough to overturn the boat. You’ve now fallen into the water, along with your cooler, your icy beverage, and everything else you had on the boat. Fortunately, you were in relatively shallow water, so you’re able to get to your feet and keep your head above water. You can right your boat and then begin to find all the things you spilled, putting them back into the boat, but you really don’t know if you can get back into the boat yourself without overturning it again.

Tossing the ice and seeing the ripples are the intended consequences. Everything else is unintended consequences, from the spilled drink, to the wet pants, the upended boat, the spilled contents. In an experiment, this long list of extra effects from ripples is an indication of the things that can go wrong unless you think things through in advance.

So, what might this look like in your lab? Let’s imagine you’re looking for the minimum bactericidal concentration of an antibiotic against a strain of E. coli that you suspect may have developed resistance to multiple antibiotics. You want to be certain that the drug of choice will be effective against the bacteria, so doing the MBC test correctly is important.

I haven’t discussed the how-to on doing an MBC before; I’ll be doing an entry on that later. For the moment, we’ll assume you know how: you culture your sample to ensure you have enough to do your test, do an isolation plate to make sure you test only the bacteria of interest, reculture that bacteria, do a count in the microscope to determine the concentration, then inoculate a series of broths made with the antibiotic of interest. Incubate for 24 hours, take the tubes without growth, and inoculate plates without any antibiotics. Whichever plate produces no growth from a broth with the lowest possible dose of the antibiotic determines the minimum bactericidal concentration. I go over these details (there are more) to give you a quick overview of the places where you can have unintended consequences from your decisions.

What if you decided to use tap water when you made up your culture broth? It’s possible that autoclaving the broth before its cultured would negate any unintended consequences of using water that wasn’t sterile, but some bacteria thrive on saline environments while others are inhibited by too much salinity. The use of deionized water would control for any variation in salinity that even autoclaving could not remove. That would help limit the impact on your bacteria’s growth either through enhancing the growth or inhibiting it.

But you used tap water, so the water has a little too much chlorine in it, compared to what would be in DI water. You don’t think this will be a big deal: gram negative bacteria are classically grown on MacConkey’s Agar, which has bile salts and NaCl to inhibit gram positive growth, so increased salinity shouldn’t be an issue. That is, unless the chlorine content is above even what the bacteria is able to tolerate compared to the MacConkey’s agar, leading to a reduced growth during the initial growth and culture, before you even attempt the MBC. That means before you’ve started testing the antibiotic, you’ve created a hostile environment that kills your bacteria – and invalidated your results.

You thought of that, which is why you used DI water instead, and made sure to sterilize everything in the autoclave before you started your test. In fact, everything went perfectly. You got the results you needed: this specific strain of E. coli is resistant to cephalosporins, but not to the quinolones. You pass this information on. What you don’t know is that the patient can’t safely take quinolones – the best drug to treat this patient’s infection isn’t safe for the patient. The unintended consequence here is that the doctor must decide to either treat with a less effective drug, or risk treatment with a drug the patient won’t react well to.

This is the challenge of rising antibiotic resistance, the challenge faced by researchers in the study yesterday. Even when science is done right, what works in the lab may not work in the clinic. This is the law of unintended consequences. The best way to try to prevent them is to always do more research. There’s no guarantee that research will find everything you need, but not doing the book work (or journal work, or internet work) will leave you hurting more often than it won’t.

I mentioned that sometimes the unintended consequences can be beneficial: aspirin is a classical example of this. When acetylsalicylic acid was first derived as an alternative to the salicylic acid from white willow bark (which caused digestive issues when used, another example of unintended consequences) it was used to treat pain. It was later found to serve as an anticoagulant as well, and has since gained widespread acceptance in treatment of heart attacks or strokes. This was a positive unintended consequence.

This isn’t a new idea: in 1936, Robert K. Merton listed possible causes of unintended consequences. See if you can identify which causes may be at play in our above scenarios.

1. Ignorance, or the inability to anticipate every possible outcome.
2. Errors of analysis or resulting from following habits that worked in past situations but do not necessarily apply to the current one.
3. Immediate interests overriding long term interests
4. Basic values that may prohibit certain actions over others (even if the resulting long term consequences could be unfavorable).
5. Self-defeating prophecies, or the drive to solve problems before they occur (possibly preventing such problems).

By being aware of the possible causes, you may be able to prevent mistakes yourself in the future. Research helps prevent ignorance related errors, along with errors of analysis. Being certain you understand the risks, benefits, and your own moral and ethical compass will help limit the consequences from immediate interests or basic value conflicts. Finally, remember that all scientists are human, and learn from your mistakes. Like the overturned boat, you gather yourself, pick up the pieces, and move forward.