Hacker News new | past | comments | ask | show | jobs | submit login

I hate studies like this. The researchers controlled for age, sex, body mass index, race, deprivation, family history of hypertension, education, smoking status, blood pressure, blood lipids, inflammation, blood glucose, kidney function and use of medications to lower cholesterol or blood glucose levels. But there are a million other potential causal factors that could be correlated with talking on the phone. For instance, stressors in your life can be a reason to talk on the phone, and stress is linked with hypertension. I think these types of studies can be fine if it's clear that this is just a correlation, but the researchers seem to be presenting this as if they've established a causal relationship.

> Professor Qin said: “Our findings suggest that talking on a mobile may not affect the risk of developing high blood pressure as long as weekly call time is kept below half an hour. More research is required to replicate the results, but until then it seems prudent to keep mobile phone calls to a minimum to preserve heart health.”

I really wish more work was done to establish a causal relationship before publishing something like this.




Anything that is less than double-blind is essentially a well documented hunch. In some situations this enough to merit action. This study is interesting enough to merit more science to build stronger evidence.

Merits of this study:

* It was large, twelve years and over 200,000 participants.

* The impact was dose responsive. People that reported talking longer on the phone had greater chances of hypertension.

* The impact is large enough to cause concern.

* In the published article, a mechanism of action was purposed (RF-EMF), and other studies related to this mechanism were cited.

* It fits the expected amount of 'controlling for' third variables you would expect, even if we cannot control for everything.

Weaknesses of this study:

* Mobile usage was self reported. Using empirical data such as cellphone bills or data from cellphone would have been much more accurate. There could be a bias where those with higher blood pressure perceive themselves as talking on mobile more frequently.

* There was no determination of the nature of the calls. Personally, my blood pressure feels very different when I talk to family for 30 minutes than if I'm talking to Comcast for 30 minutes.

* About 3/5 of the study participants dropped out. We'd expect lots of dropouts to occur with such a long study, but it could skew the data if dropouts had some statistical significance on measured values and weren't random. Dropouts are almost never random.

* The article admits those with mobile phone usage seem to be at a higher natural disposition for hypertension. More likely to smoke, many other high variables. It seems plausible that some other risk factors were associated with that cluster that did not get accounted for.

* No direct measurement of the purposed mechanism of action was performed.


> There was no determination of the nature of the calls. Personally, my blood pressure feels very different when I talk to family for 30 minutes than if I'm talking to Comcast for 30 minutes.

In my opinion that's the explanation. Even if not considering the extremes you mentioned, the more you're interrupted the less relaxed you are. The metric here is minutes talked but they're probably linear with the number of unexpected calls.


They proposed RF-EMF as the mechanism of action, but they didn't provide any evidence that RF-EMF was the cause of the hypertension.


Competing theory. External life events which are correlated with stress are also correlated with increased number of phone calls.

I’m gonna go out on a limb and say my theory is more plausible on its face.


Yeah, it seems highly unlikely. Having to talk a lot on the phone gives me stress. It's odd they don't even mention the most probably reason for their finding, and invent some unproven theory about emf affecting blood pressure. The only study I can find is this:

https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

It's a 7 person study, and it looks like they didn't correct for multiple comparisons. i.e. nothing to see here.


The evidence is that increased mobile use means increased RF-EMF dosage. It's not solid, but it's not bad for a survey.


It’s much worse than not solid, it’s terrible.

They’re beating the drum for an RF/EMF mechanism based on self reported usage.


Is the RF/EMF hypothesis disproven? I haven't looked too deep into it, but to me the biggest objection I have to it is "I don't want it to be true"


Sure, but, like the drop outs, all human behavior is correlated with a million other things. Myself and others have mentioned stress as a possible factor, but there are likely random things you wouldn't even think of like: people who use the phone more are more likely to talk to telemarketers, and talking to telemarketers is correlated with lower intelligence, and lower intelligence is correlated with eating more salt. I'm not suggesting that as a possible cause, just illustrating how complicated the true causal relationship could be. I think it's really a stretch to treat this study as establishing a causal relationship. There are also some open questions like why other types of phone usage don't seem to trigger this effect (the author mentions other studies purely on smartphone use have been inconsistent).

If the authors really believe RF-EMF can cause hypertension, then this study is a fine first step. But they really should find a way to test the mechanism of action before saying they've found a causal relationship.


I think this goes a long way towards making an argument that RF-EMF is a plausible risk factor for hypertension. I'd ideally like to see some follow up studies. Some rat model studies. Some humans with other increased RF-EMF exposure also having higher risks. There is a long distance between providing "any" evidence and "found a casual relationship", so I feel a bit like the goalpost is moving.

Even if RF-EMF isn't the cause, the effect is large enough that it begs the question "Why", which is where the other possible factors suggested come in.


It begs the question "Why", but I think there's already a few very plausible possibilities (more plausible than RF-EMF in my opinion).

In my mind, increased calls are linked with increased cases of emergencies (because much communication is done through text chat and equivalents), so calls are for when that's not sufficient or for when people are away from their normal groups for extended periods. Both of those tend to induce stress. Does stress increasing hypertension account for this?


> Anything that is less than double-blind is essentially a well documented hunch.

Not true! Economists have devoted person-millenia developing the theory of instrumental variables, which lets you extract causal relationships from historical non-experimental data.

That said, it's damn hard to find a good instrument. Many economists use it badly, and most doctors don't even know the theory.


I have my doubts about IVs in practice, but I agree they theoretically are an acceptable substitute.

The problem is we conduct science on systems we don't fully understand. Because we don't fully understand it, it's easy to miss important information we didn't know is important.

The double blind with random selection allows us to control for all important information, determine causation, and is the best known method for controlling for bias. We don't have to worry about missing important variables because the nature of random selection means those variables are randomly distributed.


When possible, experiments are better, yes. But that should not cause us to give up hope for causal information in areas where experiments are impossible.


Even subject level fixed effects would have made this a much better study. Doesn’t need to go all the way to IV to be a thoughtful design.


> People that reported talking longer on the phone had greater chances of hypertension.

The study depended on self-report? The solution is obviously to stop reporting that one talks on the phone.


>But there are a million other potential causal factors that could be correlated with talking on the phone.

This is basically the logic that cigarette companies used to deny that they were causing lung cancer. A longitudinal study is about as good as you can get for testing these sort of hypotheses.


True, but in the case of tobacco establishing a causal relationship was studied, e.g. https://pubmed.ncbi.nlm.nih.gov/13116124/ (1953).

Trying the same with RF-EMF would be an obvious next step. Just a personal opinion, but I don't think the results will hold in direct study.


No. That was the mechanism the cigarette companies used to deny they were causing heart disease. It turns out there’s really no such thing as a Type A personality. Those were just smokers.


> Years of use or employing a hands-free set-up had no influence on the likelihood of developing high blood pressure

So… radio waves cannot be the cause.


Yeah I imagine the average dosage of RF across populations has been increasing enough overtime along with rollout of new technologies that you’d be able to correlate some function of adoption of these technologies to population level effects if there was any real significance.


Hands free being what? Keeping using the speaker and setting the phone on a table? Using Bluetooth ear plugs with the phone in a breast pocket? So much we don't know...


Yikes. Yeah I'm of two minds about things like this. Usually it is meant to provide some interesting data for follow-up studies, so it has to be published.

But until the actual science is performed by making this recommendation and tracking the results, there's no advice or conclusion that can be drawn from stuff like this. We need more science literacy --in journalism--.


> For instance, stressors in your life can be a reason to talk on the phone, and stress is linked with hypertension.

Nevermind that, maybe the % correlation they found is simply roughly the proportion that find phone calls themselves stressful.


Yes and no. I understand that a causal relationship is the goal, but short of an experimental design, anything done is going to have similar holes punched through it. My view, as others have noted, is that this is a just one correlation. I'll feel that it's more causal once we see it pop up again and again in different studies, with different populations, controls, analyses, timescales, etc. It's ok to document correlations, but researchers need to temper their language when discussing correlation studies in general.


Pretty much this. Maybe the sample size were all wall street traders.

Regardless, it just sounds like someone publishing to not perish.


It's 2023. Now that we've got SMS, we don't call anyone on our phones anymore.


The frequency of shark attacks at the beach is correlates with the number of ice cream sales, therefore, ice cream must cause shark attacks! /s




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: