Member Level Rating, Challenge #1: Quoting New Group Business

Last week, I drilled a little into the workings of the ACA-mandated member-level rating methodology to set the stage for discussing the rather large downstream impacts this change will likely have on an insurance carrier's internal operations. Since I wrote that post, HHS published the final rule on Health Insurance Market Rules, and there are not substantive changes in the rating methodology put forth in the original proposal rule. So, it looks like this new rating method is going to stand. And, it's going to create quite a number of big challenges for carriers to implement.

We'll dig into each one step by step, and today we'll look at . . .

Challenge #1: Quoting New Group Business

One of the first and most obvious impacts of the rating change is that the systems carriers use to sell new business (that is, to quote rates to prospective insurance buyers) will have to be modified to calculate rates following the new rating rules.

This sounds simple on the surface, since there are really just four pieces of information you need to know to calculate the rate for each member: the product(s) they want to see quotes for, their age, their tobacco-use status, and their family's zip code. Add them up, and you have the family's rates. If you're quoting business for a small group, add up the family rates for all the employees being covered, and you have the total cost for the business to cover its employees.

Ah, but the challenges are actually quite complex, and there aren't good answers for all of them.


Challenge #1A: Hardwired Sales & Quoting Systems

It would be nice if all carriers were using sophisticated, cutting-edge quoting technology where the information to be gathered from prospective groups and their employees and the calculations used to generate quotes could be updated through a simple, intuitive graphical-user interface.

Unfortunately, most carriers' quoting systems were built years ago, and many (if not most) of them are hardwired specifically to that carrier's particular quoting methods and market practices. Tobacco-use has not typically been captured except in the health questionnaire, which now has to go away since rating based upon pre-existing conditions is no longer allowed. The logic to generate rates has typically been at the subscriber/family level, not at the member level--and that logic tends to be encapsulated in a lot of complex procedural computer code that has to be totally rewritten, and hardwired into reports that have to be written, and stored in rigidly-structure database tables that have to be rewritten. In the small group market, most of the systems were designed to support insurance agents requesting quotes for their small business customers, not the small business owners themselves, so they vary greatly depending upon each carrier's relationship with their broker network.

Here it is almost the end of February, and all this custom code has to be analyze, redesigned, rewritten, tested, and put into production by October 1st. Not an easy task.


Challenge #1B: What Do You Do With the Street Quote?

For most carriers, the small group sales process has two distinct steps, with two different types of information being gathered and quotes being produced. In both cases, these steps were being driven by insurance agents/brokers quoting business of behalf of their small business clients:

1. The "Street Quote": This is a quick and dirty quote intended to give the agent a quick look at the different rates his or her client might be able to get for various types of small group plans and to compare rates across multiple carriers. They are simply ballpark numbers, based upon some very limited information about the company such as number of employees and their average ages. It's a way for the agent to "run the numbers" and lay out some options to pursue without knowing all the specific details about each employee. 
2. The Final Quote: Because a group's final rates have (up until 2014, at least) depended upon the health status of the group's employees, a detailed group application has typically been required before an insurer could create a final quote and formally offer coverage to a prospective group. That means submitting detailed information--birthdate, SSNs, and answers to lots of health questions--about each employee. Agents would use the street quote to narrow down the options to just one or two plans, and only then go through the work of applying for a final quote, and usually then with just a single carrier that the prospect had decided to go with.

The new rating methodology actually has a good bit more precision than the old, allowing differences in age and tobacco use be reflect more directly in the premium each member is charges than in the past, when they were sort of lumped together and the differences averaged away in the rating algorithms. The downside to this precision, though, is that the rates are going to vary much more sharply from one group of employees to the next, increasing the likelihood that any sort of street quote isn't going to be terribly accurate compared to the final quote.

But, the sales process isn't likely to change just because the rating rules have changed. Small business owners will still want to compare their options across multiple carriers without having to gather tons of information from their employees , and agents will still want to be able to run quick quotes, too, without having to fill out a lot of forms (be they paper or online). Will the street quote remain? If so, how will it need to change in order to be more accurate, since the member-by-member rating might greatly skew the actual rates available from one business to another?

Challenge #1C: Integrating Vendors, Brokers, and Partners

As if the first two challenges we're pressing enough, many carriers rely up vendors for delivering their sales systems, and they have networks of large brokers and other partners who may have their own quoting and sales systems that let agents run quotes on their own and submit the final applications to the carrier electronically. It's one thing for carriers to get their own house in order; to keep business going in the future, they will also have to work with vendors, brokers, and other partners to make sure those external systems are updated properly, too. And all by October.

The Silver Lining

Some carriers have been aggressively looking ahead and have already thought through all these challenges, and a few have already begun implementing the changes that the regulations require. The majority, however, are really just getting started, and for good reason: the rating rules themselves weren't finalized until last Friday. It's hard to blame anyone for not having already ripped out and started rewriting a bunch of complex but well-tested and functioning system code.

But, one silver lining to the cloud is that these changes are affecting insurance carriers all across the country, so no one company is on its own in figuring out what needs to happen. The short timelines make best practices and information sharing all the more important, and we'll look at some of the best practices for solving these challenges in upcoming posts.

In the long run, the outlook is not as bleak as it appears. These changes are all pushing carriers to standardize their operations so that each does business more like their peers in the industry. While this might at first seem like a bad thing--those peers are competitors, after all--the long term effect is likely to be increased efficiency and overall smoother sales and quoting operations. It's just going to take some elbow grease to get there.

The Exchange Landscape by State and Type

It's now looking clear how each state plans to tackle their Exchanges in 2014--either as a standalone State-Based Exchange (SBE), as an Exchange run in partnership with the Federal government, or as a full Federally-Facilitated Exchange (FFE), whether by choice or by default because they took no action to establish an exchange on their own.

I put together a map to give a visual snapshotof the landscape.

Per-Member Rating: The Long Cascade

From the perspective of an insurance carrier’s operations, one of the most significant provisions of the Affordable Care Act (ACA) is the new individual and small group rating methodology defined by the Department of Health and Human Services (HHS) under their proposed Health Insurance Market Rules. The change may seem simple—rating individual families and small group plans on the member- rather than subscriber-level—but it creates huge downstream ripples within a carrier’s operations.

We’ll take a look at a lot of those ripples in the future, but let’s start by reviewing the proposed change in rating methodology.

How It (Often) Works Today

Let’s start with rating for individual direct-to-consumer plans. Rating methodologies vary from carrier to carrier, and some states prescribe specific methods, but here’s a typical one.

Rate = (Base Rate * Age Factor * Tier Factor * Health Risk Factor)

In this model, you start with the base rate defined for the product (say, $238 per month) and then adjust it based upon the age of the subscriber, the type of coverage being purchased (e.g. individual coverage or family coverage) and then by a factor based upon the health history of the family—that painful list of dozens of questions beginning “Has anyone applying for coverage been treated for [insert long list of dreadful diseases and conditions here].”

So, suppose you have the following family of four:

            Bill (age 36)
            Sandra (35)
            Mark (10)
            Sarah (8)

           

Bill selects a PPO plan with a base rate of $238.25 per month and fills out an application, including completing the health questionnaire. Because he is 36, Bill is assigned a subscriber age factor of .79, and because he chooses family coverage his contract has a tier factor of 2.68. Based upon some health conditions Sandra and her son Mark have, the carrier’s underwriters assign a health risk factor of 1.1. 

So, Bill’s final rate for family coverage is:

$238.25 * .79 * 2.68 * 1.1

Which equals $554.87 per month.

Note that in this case Sandra's age and the fact that she and Bill have two children do not figure into the equation. The rate is based off the oldest adult (Bill) as well as the fact that they selected family coverage, and the rate would be the same whether they have one child or four. The rate also doesn’t change if they have another child and add coverage for the baby.

How It Will Work in 2014

Under the proposed HHS rule, the rates for Bill’s family will be calculated a little differently starting in 2014. One of the big ACA changes is that rates can no longer be based upon members’ health history, so that health risk factor will go away. Now, only age, tobacco use, and “geolocation” (basically, the county in which the family lives) can be taken into account.

Here’s the kicker. The family rate must now be determined by adding up the rate for each individual family member. The rule also stipulates that no more than three children under the age of 21 can be taken into account. (Any children after the first three can be added for no additional cost.)

So, let’s rate Bill’s family under the new methodology. Bill’s monthly premium will be the sum of the rates for each of the members in his family. Each member’s rate will be calculated as follows:

member rate = (base rate) * (age factor) * (tobacco factor) * (geo factor)

Instead of filling out a lengthy health questionnaire, Bill simply has to indicate whether each person he is covering uses tobacco. In his family’s case, only Bill does. So here’s the calculation:

Bill’s rate = $152.62 * 1.22 * 1.35 = $250.34
Sandra’s rate = $152.62 * 1.18 * 1 = $179.36
Mark’s rate = $152.62 * .635 * 1 = $96.52
Sarah’s rate = $152.62 * .635 * 1 = $96.52
Total Family Rate = $620.62

Now, suppose Bill and Sandra have a third child, Nellie. Under the original rating method, they just add Nellie to their family coverage and their premium stays the same. Under the new rating methodology, Nellie gets rated separately and Bill’s coverage goes up by $96.52.

If they have a fourth child, under the new method, the premium would not increase since only the first 3 family members under 21 are counted.

Small Group Rating

The impact on small group rating is very similar, for the proposed rule does away with former “composite rating” approaches commonly used with groups.

Under composite rating, every employee in the company who purchases a plan pays the same as his or her fellow employees, varying only by the coverage tier selected (e.g. employee only, employee + spouse, employee + children, or employee + family). The rates themselves are determined by the characteristics of the group as a whole, such as average employee age, average employee health risk, geography, and size of the group. The rates for the year are determined during the sales process at the start of the benefit year. If a new employee is hired mid-year, he or she can buy coverage for the same price that his or her new co-workers pay.

That is all going to change under the new proposed rules. Starting in 2014, a group’s total rate is going to be the sum of the rates determined for each employee, and the rate for each employee will be the sum of the rates for family members being covered, just as we saw for Bill in the individual example above. A small business, in other words, will be just a collection of individual families, with each member in the group rated separately.

In other words, if you have a small business with 12 employees, go through the same pricing exercise for each employee like we did for Bill’s family above. Sum those 12 employees together, and you have your total group rate. Hire a new employee mid-year, and that new employee will be rated separately and the total group price will go by whatever that new employee is priced to pay.

I’ll let others debate the merits of the “family unit” approach versus the “build up” approach in terms of fairness, accuracy, and efficiency. From an operational perspective, though, it greatly changes things. We'll take a look next at a few of the challenges it poses to various operational areas within an insurance carrier.

The Big Shift and Why It Brings Hope

Joseph Stiglitz has a nice thought-provoking piece in the latest Vanity Fair on the underlying causes of the Great Depression and analogues to our current economic situation.  The crux of his argument is that, just as in the 1920s and 1930s, the economic struggles of today are not the result simply of problems in the financial industry.  Instead, they are part of a larger economic shift taking place as the United States moves from being a manufacturing economy to being a service economy, just as the Great Depression was ultimately the result of a fundamental shift from an agricultural to an industrial economy.

He makes a compelling argument, and one that, admittedly, is open to interpretation for how to proceed from here.  Stiglitz, for his part, advocates a series of massive government investment programs to help goose the transition along, like more funding for education, basic research, and cleaner, more efficient energy production and a return to much tighter regulation of the financial industry.  But, whether one agrees with those prescriptions or not, the fundamental analysis of our current situation--the transition to a new economy--can be an inspiring one.

Inspiring how?  Because it gives hope.  Because if there's a transition underway, then there's another side we can get to.  It's a way to break out of the rather vicious Sophie's choice of a declining manufacturing economy.  Is it really such a bad thing (from a macro perspective) that our manufacturing jobs are rapidly moving to China and Southeast Asia?  If your fundamental premise of an economy is that it needs to make big, heavy, physical things, then the picture doesn't look so good.  There are only two scenarios there:

1. We will continue to see our prosperity slip away overseas as one manufacturing job after another leaves, or

2. We somehow turn it around and reestablish the perfect, glorious economy of our golden days.

And what would #2 look like?  The old 1950s and 1960s Eden where you got a job with a big factory right out of high school and could stay there the rest of your working life (and maybe your kids could get on, too)?   A world where, when you get on up in years and retire you have  a reliable pension and can live out the rest of your years in the exact same house you've been in since you got started?

Is this the best we have to offer, the real American Dream?  It doesn't sound all that terrible, but it doesn't exactly sound like the stuff of dreams, either.

The transition to a "service economy" offers some hope--or, at least it does if you define service as a "knowledge economy" or "professional service" type occupation and not more rote services like answering customer support lines or cleaning motel rooms.  And I think Stiglitz is talking about the former.

If he's correct, it means that there's a green, unexplored field out there for continual expansion and improvement and innovation, a pretty good chance--though, my no means, any sort of guarantee--for the United States to make the turn and not only to stay a prosperous nation but to grow and lead the pack.

The true test, it seems to me, will be how we handle the transition and whether we can repeat our past performance when, in the middle part of the 20th Century, we successfully transitioned from being one of the world's leading agricultural economies to being the leading industrial one.  Time will tell, but we have a lot of assets on our side.

Returning to the New Frontier

This blog started with a good idea and, even with six years hindsight, still a very good initial post.  Then it sagged to the ground, while the other side of my interests--food writing and food history--really took off.  But, it's a new year, and I'm feeling in a very forward looking mood.

So, I'm dusting off The New Frontier.  I gave it a fancy new logo today.  Not a bad way to kick off the year.

And, my New Year's Resolution for 2012 is to look forward.  Not to forget the past, or ignore it.  If anything, we should continue to delve into it as deeply as we can.  But not to live in the past but rather to gain ideas and inspiration and guidance for how to proceed in the future.  And looking at the future seems very important to me right now, for I feel that we are in many ways at a turning point, where we need new ideas and new energy to move forward.

So, here's to 2012.  And the future, too.

Why not Excel?

I read Rory Blyth's blog post "Excel as a Database" (which is now available here) not on the web but in old-fashioned hardback form in Joel Spolsky's collection The Best Software Writing I. I flipped to it right off after seeing the title in the table of contents because I thought, "Ah! I bet this guy has written about a topic that I've often thought about . . . why people use Excel as a database." But, unfortunately, it took the exact opposite direction I expected and was in fact a rather mean-spirit lampooning of people who build complicated collections of data in Excel. The tenor of the piece is essentially, "anyone who creates a database in Excel is an idiot, and they would only use a spreadsheet like this because they are too stupid to understand a thing about computers." I expected a thoughtful piece on why people find it so useful to treat Excel like a database; instead, I got techy bile.

But maybe instead of sneering at semi-technical users who create convoluted, difficult spreadsheets, perhaps there are a lot of lessons we can learn from Excel to understand why it is such a pervasive and popular tool. And saying "because business and marketing people are a bunch of morons" is cathartic but not very helpful.

I was also disappointed in John Gruber's post "The Location Field is the New Command Line" not because I disagree with his points nor because I think it's a bad post. I don't and it isn't. It's just not what I expected from reading the title. I expected a meditation on the Location field of a web browser itself, while Gruber actually writes about web apps vs. desktops apps.

Now, as a little background: I approach this subject as a database guy, having served many years as a hardcore database developer, DBA, and BI developer. I've designed the data models for complex enterprise applications, written more lines of stored procedures than I care to remember, built large data warehouses, and created complex frameworks for data analysis. For much of my career I was in the business of taking "spreadmarts" created by legions of database-challenged but Excel-happy business people and transforming them into structured databases. Many times I've been handed Rory's proverbial crammed-full Excel file and been told "do something with it." So, by all means I should despise Excel.

But I don't. I think it's one of the most wonderful applications ever written.

Also, over the past few years I've been involved in migrating enterprise applications to the Web and, in fact, now work for a 100% browser-based software-as-a-service company. So, I appreciate the many complexities of the web model vs. the desktop model. John Gruber's title, however, hints at another interesting question that his article doesn't address: why has the URL/Location bar persisted across all the evolving variations of web browsers and web applications?

And what, you may ask, do Excel and web browser's Location field have to do with each other?

Quite a lot, actually. They are both very effective and very under-appreciated modes of user interface.

Let's start with Excel. Why is it so popular? Well, in part because you don't have to know how to create complex databases and write complex SQL queries in order to create a basic database. You don't have to write code and compile it. You don't have to build out all the frameworks of a database application in order to get a job done. Look at it from the marketing analyst's side of things. Let's call him Timmy.

Timmy has a job to do. The boss has given Timmy an assignment: I need to know how many of our existing customers have purchased Product A, how many have purchased Product B, and who the best candidates are for cross-selling based on company revenue and budget.

Why does it work so well? A few possibilities:

The UI and the business logic and the data layers are not (contrary to good n-tier design principles) separated at all. Not a good model for building a complex Enterprise app, but for Timmy's 4-hour job it's just the trick. He can click on a cell and immediately see where the formula came from. There's no debugger required (Excel does have a built-in formula debugger, but I doubt many people actually use it--there's no need).

He also doesn't have to compile, run, and execute. He sees the application right away. It's also remarkably easy to replicate data. How do you copy one table to another in a database (INSERT INTO T1(F1,F2) SELECT . . .)? How do you do it in Excel? Drag mouse, CTRL-C, click, CTRL-V. And you can instantly see the result without having to run a query.

And for old Timmy, data entry is a breeze. How many hours have you spent as a programmer creating forms and grids with the sole purpose of allowing data to get easily into a database? Sometimes I think I've spent the better part of my life wrestling with data grid controls and trying to do basic things like putting drop-down combo boxes in cells to allow list-based data entry. Or, when things got a little better, wrestling with complex 3rd party data grids that had all that stuff built in if you could just figure out the object model enough to make it work.

Excel is hardly perfect, and too many people don't take advantage of some of the built in data-validation features, but, boy, is it fast to get data in. It even has that magic auto-complete feature that will guess cell content based on previous keystrokes. And you can copy and paste entire lines.

A gentle reminder to my fellow database developers: there are a lot of smart people out there who don't know anything about database development. And they don't need to. They have other things to focus their energies on.

On the subject of the location bar: how many times do you find yourself cutting and pasting a URL into the Location bar? I do it all the time. And I love it.

If you click a link from another site and land on a page eighty layers deep into a new website and want to get back to the top (maybe you follow a link to an old blog post, like it, and want to read the blogger's most recent posts), you can try to follow links on the page and navigate back using the web designer's predefined routes. Or, you can seize control and whack off all but the first part of the URL in the Location field and, voila!, back to the top-level home page.

You can also guess URLs. I want to go to Old Navy's website and look for clothes. Hmm, let me guess: "OldNavy.com"? Yep. That's it. How easy was that? No search engine, no directory, no looking anything up or barely even thinking about it. I've got Google's toolbar installed in my browser, but I rarely use it. And how about this: I don't have to know whether I want to see a web page or view a Word doc on the file system or launch an application. I just type or paste in a URL and, "bing", the browser magically shifts to display the resource I want.

What's the common thread here? Unstructured, ad-hoc, immediate, easy to copy and paste. Control in the hands of the user and not the application designer. It flies in the face of everything an enterprise software developer would recommend, but in terms of serving the immediate needs of end users it's ideal.

There are a few lessons here somewhere, but I'm not exactly sure what they are. One is that if you are designing Enterprise software, do everything you can to support ad-hoc data input and retrieval modes--including exporting to and import from Excel. Yes, that's right: enable the dumping and retrieving of data from Excel. In fact, with clever use of VBA and hidden properties, I bet there are some really powerful user interface options you could work out that leverage Excel as the primary user interface to an application.

And, most of all, don't restrict the power of the user to efficiently adapt and interact with your application in ways you never expected they would. That is the real holy grail of application development.

Great User Interface Design

Not too long ago the snack machine at my company was replaced by a new model. No big deal. It still had the same basic assortment of chips, candy bars, and peanut butter crackers. But, there was one big difference: the numbering for the items had changed.

In the old machine, everything had a two-digit number kept in simple sequence: 23 for Doritos, 24 for the Snicker's bar next door. The new machine had a different number scheme, and a three-digit one to boot. Why on earth start numbering the snack cells all 100? What a waste of finger power.

Ah . . . but not so fast. With the old machines, every now and again I would look to buy the crackers in slot 19, see they were priced at 55 cents, and type 55 on the keypad (transposing the price for the slot number). And instead, item #55--Pringles, yuck!--would come spinning out of the machine.

Not possible with the new machine. No prices began with one, so if you even hit the 1 key first you got an error from the little screen.

Ingenious! And good user interface design, too, because it anticipates the user's mistakes and, rather than saying in the classic socially-maladjusted programmer way, "too bad, you screwed up, live with the Pringles", it says, "Hey, buddy--are you sure about that?" and allows you to correct the error of your ways.

And that's good interface design.