I am grateful to have the health insurance I have, and grateful for the payments they've made to resolve problems I've had. Nonetheless, I can't help but be astounded at the never-ending flow of expensive, incompetent, annoying and utterly useless interaction I have had with the company's computer systems. It's small potatoes in the overall scheme of things. It's also simple stuff. Why can't they (and others like them) get it right?

The answer is simple: the company's leaders, like most enterprise companies, want to be leaders in technology. Today, that means funding big, publicized  initiatives in AI and ML. Initiatives that will, of course, transform healthcare. Soon. Getting email right? Getting paper mail right? Trivial stuff. Wins no awards, gets no attention. It's unworthy of attention, like the way the rich owners of a grand house with a dozen servants wouldn't stoop to paying attention to the brand of cleaning products they used.

The email

An email from Anthem showed up in my inbox with the subject "Schedule your checkup now — at no extra cost." Naturally I open it. Right away there's a graphic, demonstrating that it wasn't just the software team on this job:

The message with the graphic repeats the message in the subject line, strengthening it — don't just schedule a checkup, schedule it early. Why should I do this? "It's a good way to stay on top of current health issues and take care of any new ones early, before they become more serious."

Sounds good! Except that the very next thing in this email urging me to "schedule [my] checkup early this year. There's no extra cost." is this:

My plan "usually" covers it?? WTF?? Right after telling me "There's no extra cost," as in There IS no extra cost??

Then comes "You may pay a copay, percentage of the cost, or deductible if you've already had your physical for the year or if the visit is to diagnose an issue and set a plan for treatment or more tests."

I'm supposed to schedule it "early." I last had an annual physical six months ago.  Is a physical I schedule now, in March, free or not? At the bottom of the email there is a nice big box that says in big type "Schedule your checkup today." It then says "To find a doctor or check what your plan covers, please use the Sydney Health mobile app or visit Anthem.com."

I've already done the Sydney trip, describing it here. Not going there again. I'll go to the main site. I'll spare you the details. They don't know who my primary care doctor is and don't let me tell them. They give me a big list of doctors I could visit, most of whom are pediatric — oh, yeah, good suggestions, Anthem! They must think I'm young for my age … or something.

Then I try to find out what my plan covers, as they suggest. Nothing about annual checkups being free of charge; it's all about co-pays. Maybe it's there somewhere, but I can't find it. As usual, the link Anthem provides is to the front door of the site, not a cool new twenty-year-old technology called "deep linking," which brings me right to the relevant place. Maybe next year. Or decade. Or century.

What could have happened

There's a concept that's been around in the industry for a couple decades called "personalization." It includes things like

• when you send an email, address it to the person, instead of making the email be like a brochure.
  
• reflect basic knowledge of the person, like whether they had an annual checkup last year — if they did, maybe they already think it's a good thing, and the message should be to be sure to do it again
  
  • They've got my history — they could praise me for getting checkups for the last X years, and reminding me to keep up the good work.
    
• Is the checkup "no cost" or not? Anthem has my account information, name, address and the rest. They have my plan. They know whether it's free or not. They just don't bother to check.
  • Taking my history into account, they could say that, just as last year's checkup was 100% free, this one will be too.
• As it happens, a week before getting the email I saw my primary care physician and then a specialist who submitted pre-auths for tests. Anthem has the visit claims and pre-auths. I'm doing exactly what they want me to do, as they said in the email, "take care of any new ones early, before they become more serious." Instead, what I hear from Anthem is 100% clueless — exhorting me to do something that the slightest bit of effort on their part would tell them I'm already doing! Blanketty-blank it!

This is customer interaction 1.01. It's also common sense. It's standard practice for companies whose tech and marketing teams have progressed past the year 2000 into the current century.

The Postcard in the mail

You might think it couldn't get worse. You'd be wrong.

After I got the email, a postcard showed up in the regular US mail. A full-color postcard from my friends at Anthem! Here's the front of it, showing a person who looks just like me having a virtual doctor visit.

Anthem cares about my health and really wants me to get that checkup — today! They care about it so much that they appear to have two whole departments, one for email and one for postal paper mail, each charged with getting me to get that checkup.

So what do they tell me on the back? Take a look:

Here's what the email said:

It's a good way to stay on top of current health issues and take care of any new ones early, before they become more serious.

Here's what the postcard said:

Having a checkup is one of the best ways to stay on top of current health issues and take care of any new ones early, before they become more serious.

Notice the similarities and the subtle differences — it's clear that each department wanted to assert its independence and word the exhortation in the way it thought best. The email modestly said "it's a good way," while the postcard went all the way, saying it's "one of the best ways." How much time in meetings was spent getting the wording exactly right, do you think?

Last but not least is the issue of cost. Like with the email, the postcard strongly asserted that the cost is completely covered. But then there's that little asterix, hinting that you might want to look at the tiny little print at the bottom of the page, where you find maybe it's not free after all. At least there was no mention of Sydney. I guess the paper mail department is jealous, and wants to avoid promoting the thing those snotty folks in IT keep yammering on about.

Anthem Leads the way

You might think from this that Anthem is incapable of going beyond the 1-2 punch of emails AND mass paper mailing. Incapable of doing basic software of the kind I was writing in high school, software that is little but common sense. I will let the evidence speak for itself.

Whatever Anthem may or may not be doing in terms of keeping up with paper mail and adding an electronic version, a little searching reveals that Anthem is spending huge amounts of time and money on "advanced digital" whatever, fashionable things like AI, ML and the rest of the lah-de-lah.

To discover Anthem's strategy, you have to find and sift through an array of websites that aren't the Anthem.com one you would think.

Here is part of what the Anthem CEO says in the most recent annual report: "The traditional insurance company we were has given way to the digitally-enabled platform for health we are becoming. This platform strategy is grounded in data and deploys predictive analytics, artificial intelligence, machine-learning and collaboration across the value chain to produce proactive, personalized solutions for our consumers, care providers, employers, and communities." I guess that means they're working on getting AI and ML to send me an email that's "personalized" sometime soon. Maybe.

Anthem has a Chief Digital Officer. Here's what he said in that same annual report: "At Anthem, we have built the industry’s largest platform, integrating our immense data assets, proprietary AI, and machine-learning algorithms." Is this just a lab project? No! "It’s through this platform that we are able to digitize knowledge and create a more agile and seamless experience for our consumers, customers …" I guess digitizing my name and slipping it into an "agile and seamless" email to me is right around the corner!

In May 2020 Anthem signed a major "digital transformation" deal with IBM. According to Anthem's CIO Tim Skeen, "We are seeing a dynamic change in the healthcare industry, requiring us to be more agile and responsive, utilizing advanced technology like Artificial Intelligence (AI) to drive better quality and outcomes for consumers." Sounds good! If IBM's Watson AI can beat the world champion Ken Jennings at Jeopardy, I guess it's just a matter of time until it figures out how to personalize emails.

A glowing article last year quoting the Anthem Chief Technology Officer described how Deloitte and AWS are helping Anthem deliver "measurable benefits" such as "capabilities that use AI/ML, cognitive, analytics, and data science" to implement their strategic vision, one of whose key tenets is " 'n=1' personalization through consumer-driven whole-health products and services." Is it possible that the strategic vision of "n=1" personalization will enable them to send me an email that's to me, instead of a brochure? We'll see.

At yet another website of Anthem's I discovered that they have a Product Management and Strategy Lead who talks about how Anthem is "using predictive models and machine learning to provide consumers with the unique information, programs, and services they need …" There's a VP of AI Technology who is "harnessing machine learning and AI …" There's a VP of Innovation who is "… implementing innovative solutions …"

What a wealth of important people and efforts, all bringing digital transformation to Anthem! With all this industry-leading technology, it's only a matter of time before I receive something from Anthem that isn't a postcard with the added bonus of a digital brochure, do you think?

Conclusion

See this post and the summary at the end for links to other amazing achievements of the Anthem software team — which extends from bad communications to it's-really-bad cyber-security involving massive losses of customer personal information.

It's clear that Anthem, like most companies of its kind, pays huge amounts of attention to the current "thing," whatever that is, making sure everyone knows they're leading the way. Meanwhile, they largely ignore trivial things that are "beneath" them, things like treating customers moderately well. It starts with avoiding paying attention to the foundation of everything, which is data. Then it's compounded by the perverse status hierarchy in software in general and data science in particular; the hierarchy is simple: the farther you are away from real human customers, the higher your status. I hope this will change, but I'm not betting on it. Meanwhile, I remain grateful for the payments they make for the health care services I receive.

I personally experienced the roll-out of Anthem BC-BS’s new patient billing initiative. As is well known, and as I’ve discussed in detail, patient billing, both from providers and payers, is a nightmare for everyone involved. Fixing it appears to be tough – it’s worse than being between a rock and a hard place; billing is between a rock, a hard place, the devil and the deep blue sea.

As I’ve said many times, Anthem is one of the best health insurance companies out there, full of hard-working, well-meaning people who just want to do the right thing. Their many efforts at making things better demonstrate this. But there’s clearly something about being a giant organization with lots of extensive, complex software that seems to make things go wrong. Small, entrepreneurial organizations can sometimes avoid the usual traps and turn out great, high-quality software quickly. But the big, lumbering bureaucracies just can’t seem to follow the clear patterns of success demonstrated by the winning the upstarts. Everything is against winning. They follow the regulations, the best advice of the best experts from industry and academia, and the results are consistent. Consistently bad.

Getting things badly wrong in software is most often a team effort. When it’s wrong, it’s rarely an ooops, how did that happen? I’ll fix it and then we’ll be OK! It’s more often pervasive, multi-dimensional badness. Anthem’s billing break-through is clearly the typical case.

In this post, I’ll describe the break-down, the fact that the new billing effort is broken. Again: this is not about Anthem specifically. It’s an example of a pervasive issue.

What led me to discover the new patient billing feature was an email I got from Anthem. That’s a story in itself, which I tell here.

While I was on the site, I had occasion to carefully examine the welcome page:

I noticed something new since last time I was there – the ability to pay provider bills. Wow! I’d never heard of any insurer offering this capability before. Of course I want to try out this amazing bill-paying break-through.

I’d better step back for a moment and explain the billing situation in healthcare, because when you just glance at what you’re about to see, you might think it’s no big deal.

In US healthcare insurance, there’s something called a “co-pay.” This innocent-sounding term is the cheerful face of a nightmare for all involved – patients, providers and insurers. At some point, I have no idea when, someone decided that insured patients should pay a part of their cost of their care, a bit like the well-known general insurance of a deductible, but fancier. The idea is that, for each service a patient gets, insurance would pay part of it, if the service is covered at all. Then, depending on a whole bunch of other things like whether a provider is “in-network,” the patient should pay something to the provider, an amount that the insurance company calculates using arcane formulas that are secret. Here is a detailed example.

So get this: you have insurance. You go to a covered provider to get a covered service. The insurance company pays some of the provider’s bill. The provider generates a bill for you, sometimes when you check out, but often much later, in the mail. The provider’s bill includes an amount you have to pay directly to the provider. That’s the co-pay.

This is the context that helps you understand Anthem’s billing break-through: instead of getting random bills in the mail from random providers and having to write checks and mail them various places, Anthem is letting you go to one central place and take care of them. Even electronically! A great convenience.

Health insurance companies are ideally positioned to provide this service! They already have the complex systems set up to transfer money to providers (doctors and hospitals) electronically, because they’re already paying them! Setting this kind of system up from scratch is hard, and maintaining it with all the changes to bank accounts, etc. takes a lot of work. Leveraging all that infrastructure to enable patients to pay their bills is a great idea.

Even better, the insurance company knows how much the patient is supposed to pay the provider! Even though the normal procedure is that the providers bills the patient, all the numbers ultimately come from the insurance company. So all the required data and systems are already in place! What a fabulous opportunity!

Naturally, I dove right in to trying it out. Here’s what I saw right away:

Now I’m getting really pumped – it’s exactly what’s needed!

I ask for the list and see a juicy target to pay:

Perfect, a typical co-pay. I click to see the details of the claim. Typical stuff, as expected:

Of course, it's a bill that no one but a clueless health insurance company would have the nerve to create. It doesn't tell me who provided the service, when the service was provided, where it was provided or what the service was. Other than that, it's great! Oh and by the way, they don't seem to know whether I actually owe this money or whether I've already paid it. In fact, I happen to know I have already paid it!

I know, complain, complain. But still, moving ahead…

I’ll spare you the details of how I had to provide payment information. I did it and authorized payment. Then I got this lovely screen:

Yes, that’s the whole screen. The technical term for this kind of screen is “face plant.” A “face slap” is when there’s an error, but the software knows it and can give you a nice error message with something soothing about how things aren’t as bad as they look. When you’re on a site labelled Anthem and then suddenly you’re no longer on their site and some other company’s logo appears, you know that the two bodies of software are in all-out war. You know the war has gone nuclear when the dreaded “500” error appears. That’s not an error message generated by software that is confused and knows it’s in trouble – that’s an internal website server error. This means things have gone so bonkers that the software that’s supposed to handle things has gone AWOL, and all the server can do is say, in technical terms, is OOPS.

So I re-grouped, went back to the start, picked another bill, put in my credit card information (again!), and authorized payment. Here’s what I got for my trouble:

Things are getting better – we’ve advanced from a full-on face plant to a relatively mild face slap error – no one else’s logo is in sight, no servers are complaining (at least visibly). It just plain doesn’t work.

Anthem presents bills that fail nearly all the criteria required of a reasonable bill. Their bill payment mechanism doesn't work. Even if the final step of paying the bill worked instead of face-planting or face-slapping, it would still be near-worthless! What good is a bill that doesn't actually show what you owe? Doesn't show when, where or by whom the service was provided, or even what the service was?

Here is the simple conclusion: Anthem claims to have a revolutionary method to enable patients to pay their providers, a great convenience. If it works. Which it doesn’t, to varying levels of embarrassment.

We all know that incentives work. That's why you always read about the "low prices" and the "sale" about to start, or the "limited-time offer." They're incentives to buy this or that. A server at a restaurant is incented to provide good service to get a good tip, and a salesperson is incented to sell by getting a commission. What a good idea it must be to apply this idea to healthcare, right?

Maybe the idea of incentives applies to healthcare. I take no position on that subject. But I do know that when the sprawling bureaucracy of a health insurance company tries to apply the idea, it turns into yet another costly bit of overhead that yields no benefits beyond allowing top executives who float blissfully above the facts and reality to claim that they're modern and innovative. Right.

Incentives in business

A business has strong incentives — to get incentives right! A giant commercial or government bureaucracy has NO incentive to get them right. The business is aware that it's spending money to get customers, money that could be spent on an endless number of good things, from advertising to improving the product/service, to improving customer service to get more repeat customers, and on and on.

Here's the key thing: incentives are old news in businesses that have customers and need to make a profit. There is a long history of giving incentives, measuring how effective they are, and adjusting accordingly.

For example, in retail there were specialists who carefully controlled each season's products and set the incentives based on the experience of measuring the results. Already in the 1990's software products began to emerge that would take line-item POS (point-of-sale terminal) data from the last few years, and predict what would be the best time to start a sale, on which products, in which exact stores, and how deep the cuts should be. After side-by-side testing in multiple retail chains, the math-driven sales proved to be more effective than the ones generated by the best, most experienced people. So the industry transitioned to the algorithmic approach — it's now malpractice if you don't use algorithmic sale incentives in retail.

Retail has gotten amazingly effective with incentives, following the classic path to excellence in AI/ML as I described in this series of posts.

Incentives by health insurers

Given this background, when the health insurance giants finally decided to apply incentives to their insured population, naturally they carefully studied incentives in other fields and adapted state-of-the-art techniques to their situation, right? WRONG!! In every case I've seen, they've done the dumbest things possible, not just starting from incentives 1.01, but screwing up so badly that any internal measurement system (which of course was NOT in place) should have resulted in the prompt cancellation of the project and demoting everyone concerned to a starting position opening mail. None of which happened, of course.

Incentive Case Study: Anthem

Anthem is an excellent health insurance company. It and its managers strive to be industry-best and provide great service to the employers and individual customers it serves. I am using Anthem for the case study here NOT because they're the worst — far from it! I'm featuring them simply because I'm a customer, and so get a ground-level view of how things work there.

Sadly, as a giant, heavily-regulated bureaucracy, Anthem lumbers along and, along with its peers, gets really important things wrong. I wrote about the mess they made a couple years ago, first by allowing themselves to be hacked, and second by responding to the hack somewhat, ahem, ineptly.

Some time ago Anthem decided that Wellness and Health Incentives were something they should dive into. They now appear to be thoroughly committed to it, as they say loudly and clearly on their website:

Some of these programs may be truly wonderful. I make no comment on them. But I doubt that the one that was pushed onto me was exceptional, so let's dive in.

I decided it had been too long since I'd had a general health check, so I signed up for one and got it. After a while, I got a packet of stuff in the mail. Here's the top page:

Anthem clearly had gotten the claim for my visit and auto-enrolled me in their incentive program to get me to do the stuff they say, including getting such check-ups regularly. Wow, the incentive-program-babies who designed this program thought, we'll pay him some money for getting a check-up and maybe he'll do it again next year, hoping to get another card in the mail — though of course we won't breathe a word about that.

Let's check out the rest of the package. Next page I get a wonderful, inspiring picture  of how exercise makes me healthy:

Next page, reality starts to hit. No more nice pictures and color. Just the facts, ma'am (the card itself was glued onto this page):

It's a gift card kind of thing! Except it's "pre-paid," and is accepted where Visa debit cards are accepted, but when you use it, you've got to lie and say it's "credit." Hmmm. And NOWHERE does it say how much money is on the card! I have to call or go online to find out. And of course the card doesn't say Anthem, it says SVM. Who are they?

Maybe I'll find out. Let's keep on. Next page:

Great. When I get a regular credit card, I call a number to make it live, and then away I go. For this one I have to read something and then check somehow what's on it. I wonder how much IS on it? I hope eventually they'll tell me. First I better read the rest.

Man, this print is getting mighty small! I wonder if I have to pass a test before I'll be allowed to use the card that may contain some secret amount of money that no one will tell me. Minus whatever fees and other stuff that they jam into it when they feel like.

Reading carefully, I find out that, even though it's a debit card, like for your bank, I can't just get the cash out of the card! What the &*&&$&*()$ is THAT about?? It's an incentive, darn it! A incentive that's a money incentive — and I can't get the cash and spend it?? What is this, if I go to Burger King and order something they don't approve of the card won't work?? Who knows??!!

Reading more, I can't use it at an automated gas pump. Or at most restaurants. And there's a PIN, which I have to call to set — in TINY print in the middle of the TINY PRINT page? I thought I was supposed to select CREDIT, and credit cards don't have PIN's. What's going on?

Maybe the next page will help:

Or maybe it won't. Or the next 3 after that, which are more of the same, and which I refuse to read.

Finally, last page, bigger print:

This appears to be the cheat sheet. This I can read. It appears I really do have to go online and enter a bunch of stuff, and remember all the conditions on using it, including the ones they don't repeat on this sheet. Maybe I can even find out how much money they're giving me!

If I were a regular person with some mild interest in whatever this incentive is, I would have dropped out by now. But I'm a fanatic and want to see how this story ends, so I'm going on to the next step. On-line we go! First step:

Get my card, copy the numbers in. Next step:

Put some more numbers in. Next step:

I appear to be in, but not really. I have to go back, enter the security code again and also enter the hard-to-read code they put in to stop robots. Wow — all I can think is, this incentive must be huge. Why else would they be making it so hard??

Finally, I get logged in:

I think I've said "wow" a few too many times, but I should have held off for now. Tucked away in the upper right corner is the size of the golden goose I've spent all this time and effort seeking: $50! Maybe. Sort of. Except not in cash. And minus whatever fees. And not at restaurants or gas stations unless you follow the rules. And maybe there's a PIN, check the amount before each time you use it because you might have been dinged a fee, and…

And before I can touch any of it anywhere, I've got lots more information to enter. I'm outa here!

Conclusion

This incentive card program is one of the more bone-headed, dysfunctional things I've encountered in a while. Lots of lawyers, bureaucrats, managers and even publicity/image people contributed to it, but did anyone with, you know, real knowledge of how things like this work ever get a shot at it? The number of steps to get rewarded and all the uncertainty and conditions are guaranteed to produce maximum drop-out. In the reward card business this is called "breakage," and unethical rewards people try to maximize breakage; Anthem should be up for "rookie of the year" in the breakage stat.

It's one thing to try to create good behavior by giving an incentive. It's another to dangle the promise of an incentive, trick people into going through a horrific maze that most won't make it through, to get the incentive (of unknown value!), with the primary result that your basic impression of the health insurance company as incompetent and wasteful and something you should ignore whenever possible is strengthened.

Is anyone out there listening?