Underwriting Requirements And Reality
The only risk appraisal resource likely to emerge unscathed is the venerable attending physician statement (APS). All the others are under more scrutiny than ever to either perform or pass into the mists of history.
This article will focus on those underwriting assets most powerfully influenced by this process, one way or the other.
NT-proBNP belongs in the Underwriting Book of Records on three counts: (1) most outstanding cardiac test, (2) fastest growing requirement since HIV-1, and (3) most poorly characterized resource (by people who should know better).
The quantity of misinformation/disinformation about NT-proBNP available for industry consumption is staggering. After reading virtually everything written in English about this test and writing two white papers on the subject, let me assure you that:
• The test is called NT-proBNP—not BNP!
• It is a marker for a pathological process affecting the ventricles that is incited—sooner or later—by virtually all heart disorders—whether silent or symptomatic.
• It is as specific for heart damage as the troponins.
• Elevated levels at some threshold are consistently, independently and highly significant to cardiovascular and all-cause mortality.
• An insurer can do roughly 40 NT-proBNP tests for the out-of-pocket cost of one exercise ECG (sans the baggage associated with the latter).
• It should and will be used as a routine screening requirement at age 55 and older and as a reflexive test in every potentially rateable cardiac context at all ages.
In 1977, only a handful of proactive carriers used NT-proBNP. Four years later, the substantial majority did so, and its use will become universal in due course.
A few years ago, APPS Paramedical Services introduced a novel approach for taking medical histories during paramedical exams—using laptops or the equivalent in lieu of scribbling on paper forms. This accommodates scripted drilldown questioning of all “yes” answers in the same manner as telephone interviews.
The “smart” paramedical hit the street running and has won over the hearts, minds and wallets of a growing roster of major role model caliber insurers. Now that other paramedical providers also offer variations on this theme, it should become the only way this essential process is carried off.
On the matter of paramedicals, the time has come for insurers to insist that providers measure waist circumference (WC) routinely.
Body mass index (BMI) is a flawed marker. It cannot distinguish between central (bad) obesity versus peripheral (insignificant, other than cosmetically) adiposity. By adding WC, we will finally be empowered to make this key determination, and by doing so properly apportion “build” debits.
There is a lively debate going on in clinical medicine about the pros and cons of screening for prostate specific antigen (PSA). While the arguments from both sides make interesting reading, they have no implications for the extent of PSA screening by insurers.
Why? Because as long as there is any significant incidence of clinical PSA screening, insurers must deploy PSA—if only to ward off anti-selection.
The typical individual found to have high-grade prostate cancer as a result of a positive PSA test has a life expectancy spanning multiple contestable periods. Therefore, absent diligent insurer screening, anyone told of a positive test could readily embellish his insurance portfolio before enduring treatment.
Besides, a recent study challenges the longstanding notion that elderly men with prostate cancer “die with, not of” this disease.
When prescription profiles were first introduced, we had many concerns about their implications. All of these issues have been laid to rest, as amply demonstrated by the meteoric increase in prescription history use in the last few years.
When utilized competently, these records should allow underwriters to forego medical records on many common conditions. Prescription profiles also exert a potent sentinel effect, unmasking applicants afflicted with “prescription amnesia.”
In due course, their use will be universal. This is because their combination of relatively low-cost, high-protective value and instantaneous access satisfy the above-mentioned wish list of senior management as well as or better than any asset we have except the Medical Inspection Bureau (MIB).
The Tedious Trio
This gaggle of antiquities consists of resting ECGs, exercise ECGs and chest x-rays.
• The only saving grace for resting ECGs is their comparatively low cost. Balanced against this are increasingly common complaints about poor quality and inherent customer-unfriendliness, as well as the fact that 50 percent of individuals with coronary disease have normal tracings.
• The stress ECG, on the other hand, is poised to (finally) disappear. A majority of chief underwriters who took the 2011 Life Underwriting Requirement Survey concurred with this assessment.
It is interesting in this regard that these underwriters also recognize the potential downstream consequences if we inadvertently compel an elder with an undisclosed or overlooked contra-indication to undergo this rigorous process—with a fatal outcome.
Consider that clinical medicine abandoned routine indiscriminate screening with stress ECGs three decades ago, it would be timely for insurers to follow suit.
• Chest x-rays are almost extinct in underwriting. The fact that an insurer would expose an individual to carcinogenic ionizing radiation to complete a financial transaction is nothing short of appalling.
Predictive Models Based on Lab Tests and Physical Measurements
All three industry laboratories have introduced predictive model risk scoring based on the results of screening blood and urine profiles, plus three physical measurements (build, blood pressure and resting heart rate).
This underwriter has had an overview look at all three and the opportunity to dissect one of them in considerable detail. ExamOne’s RISK IQ model assigns a score to each applicant based on the aforementioned components. That score ranges from 1 to 99; the higher the score, the greater the implied risk and vice versa.
The first thing that grabs your attention here is the impact of high normal and, to a lesser extent, low normal findings. After investigating the legitimacy of these results based on an extensive review of the medical literature, I can assure you that these scores are, as Aussies are fond of saying, “spot on.”
Underwriters flummoxed by scores in the 1990s on proposed insureds with more or less all normal tests and measurements have sent me numerous cases for a second opinion. After I recounted the scientific evidence underpinning the findings, they were intellectually assuaged—but still understandably anxious about having to explain the implications to the producer!
The good news here is that these scores are not used arbitrarily; they are just one of many factors considered in assessing risk. Medical history remains, as ever, the overarching arbiter of insurability.
At a recent study group, we were told that more than a dozen carriers have already implemented this type of predictive modeling. And it goes without saying that all the others will be obliged, sooner or later, to sort out the pros and cons of one or more of these scoring systems. If nothing else, this should be interesting!
Predictive Models Based on Personal Purchase Records
The potential for using this information in an underwriting setting has been touted in some quarters as a way of lowering business acquisition costs while also improving mortality outcomes.
There are two ways we can assess the viability of this approach.
One is to look at the evidence in a vacuum. The other is to take a step back and ask: What are the big picture implications of indulging this scheme?
The latter—and superior—approach gives rise to certain fundamental questions.
Do we want to embrace an inference-based approach to directly or indirectly assessing risk?
These models require us to make a dubious leap of faith (i.e., if one buys any significant quantity of alcohol, therefore one drinks too much, etc.).
Consider this example: Purchasing a piece of exercise equipment is perceived as favorable in this context. I bought an exercise bike four years ago with my credit card, rode it twice, hated it, and gave it to a friend. My son bought a similar device at a garage sale, rides it religiously and has the cardio-respiratory status consistent with his faithful use of this instrument.
Using this model, I would be absurdly “credited” (because there is a record of my purchase) whereas my son would fail to get well-deserved statistical recognition because he bought a used machine for cash!
There are myriad other scenarios where this flawed, inference-dependent model comes up woefully short of reality.
How will your clients and insurance regulators respond to this practice?
Those who fancy arbitrary and intrusive practices should be thrilled. The other 99 percent may be less enamored. Ditto for insurance departments. There are important reasons why race, ethnicity, religion and sexual orientation are off limits as insurability considerations.
The same must now be said for personal purchase records, whether or not a statistical argument can be framed for their efficacy in the vacuum of unconsciousness.
If you agree (like the substantial majority of chief underwriters completing the 2011 survey), you may wish to make your views known, because unfortunate things happen when conscientious people remain silent.
Cognitive and Frailty Testing
The available evidence suggests that both the three-word delayed recall and clock drawing tests have limitations that would ideally exclude them from use in an underwriting context.
We have superior alternatives at hand, including the ten-word delayed recall test and, most notably, the multi-domain state of the art protocol called the Massachusetts Comprehensive Assessment System (MCAS).
Physical frailty is a dominant factor impacting insurability at ages 70 and older. Indeed, one could make the argument that screening for frailty is more important to us overall than testing for cognitive dysfunction.
The Timed Get Up and Go test gets high marks from gerontologists and also happens to be the most widely used frailty screen used by insurers. All things considered, measuring walking speed may be an even better resource and its deployment would be feasible with fixed-site paramedical facilities.
Greater emphasis needs to be placed on questioning elder applicants about their capacity to carry out activities of daily living (ADLs) and perhaps instrumental activities of daily living (IADLs) as well. The evidence for their implications is more robust than for Timed Get Up and Go and they are easily carried off in a paramedical context.
Many subjects touched on here have, to say the least, elements of controversy. Thus, it would be to our collective advantage if they were aired out in ways that contrasting viewpoints were presented for scrutiny.
By Hank C. George, FALU, CLU, FLMI FALU, CLU, FLMI, for the November 2012 issue of Broker World Magazine. Hank is CEO of an underwriting education and consulting firm in Milwaukee, WI.