Some Links

A few interesting and useful links for Health Econ.

1. The sharp slowdown in health care expenditures. How much is because income growth slowed and how much is due to structural changes in the healthcare sector?

2. Obamacare rules increase transparency on how much pharma and medical device companies kick back to doctors. It probably won't change consumer's behavior, but may change doctor's.

3. Sex ratios matter. The worst cities for single females. I bet they act more like men.

Airbag Helmet

Improving mortality statistics by reducing deadly accidents.

Narco

America's healthcare system is different than every other in the world, and so is our behavior. Take this fact:

America consumes 80% of the world opioid supply (99% of the world hydrocodone supply), but has about 5% of the world’s population. If you don’t think America has some kind of opioid problem, then move along because this rational, evidence-based, experience-laden way in which I’m going to discuss opioid use and misuse will not interest you.

The article has much more. Read it.

Skin Cancer Scammer App

From junkcharts:

The Wall Street Journal hyped a research article about mobile apps that supposedly "detect skin cancer". While the tone of the article is quite balanced, I cringed when the reporter wrote: "the best-performing app accurately identified cancerous moles 98.1% of the time."

A nicely misleading bit of marketing.

Those who have been reading this blog hopefully will immediately wonder... if the app produces few false negatives, would it produce lots of false positives? It's almost guaranteed because there is a trade-off between those two types of errors.

Another question you might have is: assuming the app tells me the mole is malign, what is the probability that I have skin cancer? Notice this is the reverse of sensitivity. Sensitivity is the probability that the app tells me the mole is malign assuming that I have skin cancer.

Sorry to pop the bubble. The so-called positive predictive value is between 33 and 42%. This means that of those people whom the app claims have skin cancer, less than half of them actually does.

The 98% number is pretty much useless. It's the 40% number that we need to be worrying about.

P-Hacking

A reminder of the cult of statistical significance. I fear my teaching sometimes leads students to honor the wrong idols.

Ted Talk Secrets

Here is an article on what makes Ted Talks so great.

3. How can anybody learn to tell stories that are as enthralling as a TED Talk?

The first thing to know is that there's no one formula or way to give a TED Talk. They're as unique as the speakers themselves. The real key is authenticity. TED Talks work when speakers share both their personality and their original ideas with the audience. You can't just lecture. You have to be passionate; you have to be willing to be vulnerable. This doesn't mean giving the audience jazz hands or shedding crocodile tears. It means sharing your curiosity and your excitement, your failures along with your successes. It’s also about telling stories that take the audience on a journey. And allowing your own contagious enthusiasm to inspire the audience in their own work. And helping the audience to see the world through your eyes. If you can get people to see the world differently, then that’s a talk that will be truly memorable.

All that said, a great TED Talk also takes PRACTICE. With very limited time (only 18 minutes or less), you have to know your material inside and out, so you can present in a way that feels polished but -- importantly -- not rehearsed. You have to practice with real audiences so you get the feel for the rhythm of a talk: When are they riveted? When do you lose them? It's only by test-driving a talk that you can learn what works and what doesn't.

In the end, TED Talks are about great, old-fashioned human storytelling.

Here are the top 20 most watch Ted Talks to date.

Graphs

Are the meant to enlighten or to mislead? What is the problem with this graph? Does it mislead, or inform?

Here are some more.

And here is a post on the historical changes in the goals of data visualization.