"70%" - It's Baaack! Kind of...

Bill Comfort, CSA?, CLTC?, LTCCP?

December, 13, 2020

(Edited August 20, 2024 to include updated 2022 study on HIPAA-level risk data)

The Stat Trap:

Relying on any statistic to "prove" to clients that they will likely need extended care is a lazy, ineffective way to engage and motivate clients to take action on the purchase of a meaningful long-term care (LTC) insurance benefit. In my professional practice and as an 18-year instructor with the CLTC? designation training program, I have learned there is no "magic number," no perfect stat that will get clients to take action to plan for their future care including buying LTC insurance.

The human brain is evolutionarily wired to reject data - even "true" data - that conflicts with existing beliefs, like, "I don't think I'll ever need care, and I certainly don't want to talk about it."

More specifically, we do not make these kinds of decisions based on logic; we are hard-wired to make decisions on emotion supported by facts, sometimes wrongly. Said another way, facts, statistics, do not change minds. Ever. And yet, in the LTC insurance industry we are forever chasing a bigger and bigger, ever-more-frightening statistic that will suddenly, magically, motivate prospects to action. It is a fool's game, and as the woeful market penetration of any type of LTC insurance coverage shows, it is ineffective to the point of even driving many prospects away.

If stats worked to motivate the purchase of insurance, every life insurance brochure would print in 80-point, bold type at the top of the first page this unimpeachable statistic: 100% chance of dying - at some point in your life - according to God Himself!

I have spent years tilting at the compliance and lazy-selling windmills that continue to grind out "70% ... 70% ... 70% ..." in the context of "educating" consumers about extended care and LTC insurance. The study that gave us the 70% stat (Kemper, et al, 2005) was so deeply flawed for use in any context related to LTC insurance that its use should be considered a marketing lie. Yes, a lie! It has been my quixotic mission to eliminate the use of "70%" in anything related to LTC insurance.

If you must have a LTC stat, 52% is the best one (Favreault, rev. 2016) - more on both of these statistical studies if you are willing to follow me down the rabbit hole.

The "new 70%"!

Imagine my surprise when I just found a newer study, published in 2019 by the US Department of Health and Human Services, that appears to validate that, indeed, there is a 70% chance that a person will experience a Health Insurance Portability and Accountability Act (HIPAA)-qualifying need for long-term care services: "What is the Lifetime Risk of Needing and Receiving Long-Term Services and Supports?", Richard W. Johnson, Urban Institute, HHS ASEP Research Brief, April 2019. ("Johnson, 2019")

I was gobsmacked ... until I started to dig deeper ...

This new study, Johnson, 2019, appears to be a perfect, credible confirmation that "70%" is a legit stat for use with TQ LTC insurance, but it is fundamentally flawed.

"70%" remains a fundamentally flawed stat, and the LTC insurance industry needs to be very careful, because getting it "right" matters!

To understand how we got here, why Johnson, 2019, is flawed, and why "70%" should still NOT be used in marketing LTC insurance, we need to review what has come before.

A Short History of LTC Statistics:

It all started with The New England Journal of Medicine from a February, 1991, article: "43% of people turning age 65 will spend some time in a nursing home." The LTC insurance industry flogged that one for years, even after its credibility waned as it only considered care in a nursing home including short-term, Medicare-paid rehabilitation.

Through the 1990s into early 2000s, other studies attempted to measure the risk of needing care at home or in assisted living as well as nursing homes, and a bizarre statistical "auction" began; 50% will need any type of care, 65% of couples will have one spouse need any type of care, and on and on and on, until we got to "70%."

Anyone who has followed my writing and teaching knows that I have been a harsh critic of the use of the "70% of people over the age of 65 will need 'some type of care' in their lifetime" in any context even remotely associated with LTC insurance. Not just because it is ineffective, but it is also inaccurate and misleading. The study that set off the "70%" exclamations actually says the risk of needing "some type of care" is 69%. But why not just round up?!

This is the study that got everyone excited about "70%" - as infamously rounded-up from 69% by HHS: "Long-Term Care Over an Uncertain Future: What Can Current Retirees Expect?". Peter Kemper, Ph.D., Harriet Komisar, Ph.D., Lisa Alecxih, Ph.D., Inquiry 42: 335-350, Winter 2005/2006 ("Kemper, 2005")

The problem is Kemper, 2005, clearly uses impairment criteria that is significantly more liberal than HIPAA TQ LTC insurance benefit eligibility criteria. Specifically, Kemper, 2005, used one (1) ADL or four (4) IADLs to get to its 69% conclusion. This is the "flaw" I have always harped on in using 70% for the "risk" especially in the context of TQ-LTC insurance.

The "70%" flaw: Neither one ADL nor four IADLs will trigger benefits in a TQ LTC insurance contract.

52% Remains the "Best" LTC Insurance Risk Stat

Next up, the study - and statistic, 52%, that I think remains most-applicable to TQ LTC insurance marketing if you must use a stat: "Long-Term Services and Supports for Older Americans: Risks and Financing Research Brief," Melissa Favreault, Urban Institute & Judith Dey, HHS, ASPE Research Brief, July, 2015, revised 2016. ("Favreault, 2016")

Favreault, 2016, used two of six ADLs for at least 90-days or a "severe cognitive impairment" to exactly track the HIPAA TQ LTC "chronic illness" impairment/benefit eligibility standard. Endnote 5. clearly states that the six ADLs used are consistent with HIPAA and TQ LTCI; "eating, toileting, transferring, bathing, dressing, and continence." This study repeatedly mentions using criteria exactly consistent with HIPAA-level impairments.

Favreault, 2016, is also consistent with LTC insurance industry claims data that suggests that about 50% of policyholders with a 0-day EP will receive at least $1 in benefits (33% with a 90-day EP will receive at least $1).

(Note that this report, this stat, is also published by the US Department of Health and Human Services! Yet, HHS continues to (improperly) "70%" on its own LTC insurance promotion website. 52% is still a very compelling risk stat if needed to be used and it is the ONLY stat that is appropriate when used in any context related to a discussion or promotion of LTC insurance.)

[This HIPAA TQ-level risk of needing care data analysis was updated in 2022 by the Urban Institute & HHS, co-authored by Judith Dey and, interestingly, Richard Johnson. Their updated HIPAA TQ-level risk: 56%. More info on this updated study by following this link: https://www.dhirubhai.net/pulse/70-ltci-lie-bill-comfort-csa-cltc-/?trackingId=VwXBu6%2FGQ1KexUYdXvc2SQ%3D%3D ]

Kemper, 2005, and Favreault, 2016, both used a "microsimulation" model. For lack of a better layman's analogy, their microsimulations were a type of "Monte Carlo" projection.

The Fatal Flaws in the "NEW 70%":

Johnson, 2019, does not mention that he used a "microsimulation" tool, even though the model used by the previous studies was created by the Urban Institute for which the author works. It appears he made his own "straight-line" extrapolation from the data, which could predict a higher need than the others that take into account multiple, variable factors on a simulated basis.

I would like to know more about the statistical process used in Johnson, 2019, but several fundamental, fatal flaws should condemn its usage for LTC insurance marketing regardless.

It is also telling that Johnson, 2019, completely ignores, makes no reference to Favreault, 2016, despite the fact that Favreault was authored by an Urban Institute colleague, and that it presented a significant advancement in the data analysis from Kemper, 2005 which Johnson does quote.

Johnson, 2019, in attempting to draw a similarity to and credibility from Kemper, 2005, misquotes Kemper by saying that Kemper also used two ADLs to get to its 69%; Kemper clearly did not, which, again, has been the problem in using it for marketing TQ LTCI. I am not sure what significance this misquoting has, if any, but it seems a glaring, sloppy quoting error in a report like this.

Now to the significant data flaws in Johnson, 2019. Nowhere in the report does he mention HIPAA or HIPAA-level impairments other than in an odd Endnote (#4), but in the narrative, he misleadingly, repeatedly appears to suggest he followed HIPAA standards by stating the projections are based on 2 ADLs for at least 90-days or severe cognitive impairment.

Digging deeper reveals that Johnson, 2019, does NOT use HIPAA TQ LTCI-compliant ADLs!

?This is the cheat - or inconsistency - presented deep in the report: "ADLs include getting in and out of bed, dressing, walking across a room, bathing or showering, eating, and using the toilet." [EMPHASIS ADDED]

Johnson, 2019, drops "continence" from the list of ADLs which is not a big deal statistically I think, especially related to LTCI claims, but he adds the non-HIPAA-compliant "walking across a room" (i.e., "mobility" or "ambulating"). Anecdotally, from nearly 30-years of LTC experience, when "mobility" is considered a measure, as part of the ADL trigger, the need or the claim risk increases significantly.

The inclusion of "walking across a room" alone could explain the Johnson, 2019, "new 70%" results vs. the Favreault, 2016, 52%!

The Johnson, 2019, Endnote 4. also raises questions about the value of using this study and its data in relation to marketing TQ LTC insurance. Johnson, 2019, notes that in the surveys of those needing care - or their caregivers/former caregivers that, "the HRS does not ask respondents if they need assistance with various ADLs; it only asks if they have any difficulty with ADLs. To create a threshold more consistent with HIPAA, we added the requirement that individuals must also receive some LTSS."

So, first, there's a lack of specificity regarding the self-reported (or observed) ADL needs, particularly multiple ADL needs, and second, simply adding that those with any reported help with ADLs be actually receiving "LTSS" is not a good proxy as there are many people privately paying for professional "companion" or "homemaker" care who do not yet meet HIPAA TQ LTCI impairment standards, and that could skew such an assumption, inaccurately raising the "risk" percentage.?The potential inaccuracy of relying on caregiver/former caregiver reporting of care needs is compounded by the misuse of "walking across the room" as an ADL.

If you absolutely, positively, must use a statistic to state the possible risk of needing extended care in the context of marketing LTC insurance, the ONLY credible stat is 52% from Favreault, 2016. (I'm talking to you home office marketing and compliance departments!)

For agents and advisors who want to meaningfully, proactively engage their clients who must plan ahead for the possibility of an extended care need, learn how to gain the proper emotional foothold in a client's heart without scary statistics! The CLTC? designation training program that pioneered and champions this approach - which is the only effective approach - would be a great place to start: www.ltc-cltc.com.

Or dig deeper into the many planning and LTC insurance resources on my website: www.ComfortLTC.com.

(c)2020 William E. Comfort, Comfort Long Term Care

?