Category: Research

Sep
29

Causation and Correlation and the fallacy thereof

Posted by Joe Paduda on Tuesday, September 29th, 2015 | Comments

There’s a big difference between “causation” and “correlation”.  Just because A is around when B happens does not mean A causes B.  It could be that A’s friend C is actually the instigator, and A and C happen to hang out together a lot.  Or let’s say B is graffiti at the subway. A can only get out of school on Saturdays which is the only day B occurs because that’s when the graffiti school has their practical exercises.

Or, to be a bit more complex, when Justin Bieber was born, cholesterol scores began to decrease.  Until, that is, Facebook was invented, which reversed the decrease, thereby proving the Beebs is not as powerful as Facebook!

justinbeiber

Did you know that the more mozzarella consumed in the US, the more civil engineering degrees that are awarded?

Did you know that the age of Miss America strongly correlates with the number of murders by steam, hot objects, or hot vapours?

And the more the US spends on science research, the higher the number of suicides by suffocation strangulation, or hanging… (damn science!)

For those consumed by marriage in Kentucky, know that the marriage rate is closely tied to the frequency of death by falling out of a fishing boat.  Perhaps a fishing honeymoon involving moonshine is a time-honored tradition in Kentucky…

Now you know…

But seriously folks, even esteemed medical journals sometimes screw this up – HealthNewsReview highlighted one such faux pas last week in the British Medical Journal of all places.

According to HNR, a BMJ piece about SSRIs and crime stated:

‘Use of selective serotonin reuptake inhibitors (SSRIs) increases the rate of violent crime among young adults.’

The piece SHOULD have said “there is a correlation between SSRIs and an increased risk of violent crime among young adults.”

See the difference?  A statistical correlation does not mean one causes the other. In fact, it could be that violent youths are more likely to be prescribed SSRIs, or the crimes studied were committed in an area where psychiatrists prescribe SSRIs a lot more often than in other areas, or any one of dozens of other reasons.

I bring this to your attention, dear reader, because there’s far too much sloppy reporting out there that confuses correlation with causation, often by folks that should know better.  

What does this mean for you?

Think critically.  Always.  

Sep
4

Work comp medical spend and other fun facts

Posted by Joe Paduda on Friday, September 4th, 2015 | Comments

In my ongoing effort to serve the public good, here are current facts and figures related to how many dollars are spent on medical care in worker’s comp. 

Total medical dollars

In 2013 workers comp medical spend totaled $31.5 billion.  Source – NASI’s Workers’ Compensation 2013 Report.  NASI is the definitive source for this data; their primary sources are NCCI for the 38 states where NCCI is the rating agency and state rating agencies/bureaus for the other 13.

Worker’s comp medical trend rate

NCCI’s Annual State of the Line presentation at the firm’s Annual Issues Symposium provides the earliest – and most complete – insight into medical inflation.  For 2014, initial indication was medical severity for lost time claims increased 4 percent.

A couple of caveats – this is for lost time claims only; while LT claims account for the vast majority of medical spend, medical only claims account for perhaps 15% of spend.  In addition, NCCI’s data does not include self-insureds; about a quarter of comp benefits are self-insured.

Pharmacy spend

Work comp pharmacy spend accounts for somewhere around $5.5 – $6 billion.

Sources

  • Internal HSA data from research projects (my consulting work)
  • NCCI – by their estimate drugs accounted for 18% of all medical expenses in 2011; note that this is based on total incurred cost, or for the layperson, their estimate of what the total including already-paid and future drug costs. Therefore this isn’t actual annual “spend”. And, the data comes from 2011 reports.  There’s a lot more to this, but suffice it to say the $ range above is solid.
  • note – some claimants are submitting their work comp scripts to their group health plans.  While this won’t affect “spend”, it does impact the addressable market.

There are a lot of other sub-categories out there – and just as much confusion about what services, codes, provider types, or locations of service belong in what buckets.

If you are attempting to categorize spend, make very sure you understand your sources’ definitions.  E.g., script count; how do you define a “script”?

  • Is it the prescriber’s prescription written out for a patient?  If so, understand that most “scripts” include 2+ drugs.
  • Is it each individual medication prescribed?  If so, understand that some “scripts” are for 3 days’ supply, others for 90.
  • Is it for a certain number of days’ supply?

Different stakeholders use different definitions – and not just for pharma.  How is “surgery” captured, and what is included?  CPT codes? Facility fees?  Associated office visits? Bills submitted by providers with a surgical specialty?

I could go on, but hopefully you get the (cloudy) picture by now.  If not, your bad.

What does this mean for you?

Work comp data is dirty, inconsistently categorized, and there are no single sources for all categories/spend types.

If you want to really understand the space, get granular, precisely understand definitions, and do NOT make any assumptions that other non-primary sources have got it right.  

Aug
26

Steroid injections – they kinda sort work some of the time…

Posted by Joe Paduda on Wednesday, August 26th, 2015 | Comments

Thanks to Steve Feinberg, M.D. for forwarding a study on epidural steroid injections.

Here’s the brief findings:

Epidural corticosteroid injections for radiculopathy [pain radiating from the spine] were associated with immediate reductions in pain and function. However, benefits were small and not sustained, and there was no effect on long-term surgery risk. Limited evidence suggested no effectiveness for spinal stenosis.

In a follow up, Dr Feinberg provided this:

I have a 68 year old physician colleague who is highly functional both at work and recreationally. He has rather severe cervical and lumbar degenerative disease and stenosis and a very damaged left knee. He has undergone a number of injections (more than would be allowed via EBM) and takes Vicodin 10/325 3 times a day and uses some oxycodone for “breakthrough” pain. He lives on 5 acres and takes care of 10 horses and the property. He told me that working on his property makes him hurt more but that he is not going to stop being active just because of the pain/discomfort. He has been on the same opioid dose for years and has no obvious negative side-effects. He told me that without his medications, he would have trouble practicing as a physician and he certainly would not be active on his property.

Dr Feinberg closed with:

“I ask myself everyday if so little works, what are we left with to treat?”

A colleague of the good doctor provided this as well: “Could it be that Osler’s words from over a century ago continue to direct our best efforts? “The job of the physician is to entertain the patient while nature takes its course?”

I bring this to your attention as a reminder to all that medicine can be as much art as science, that we often don’t know what works for whom why and when and how.

However, make no mistake that treatment can and should be guided by evidence-based clinical guidelines. There should be a way to navigate the care management and authorization process to allow Dr Feinberg’s colleague access to the treatment that works for him, just as there should be a high standard for approval of “non-standard” care that puts patients at risk.

I’d close with this note – there is far too much use of procedures similar to ESIs, and far too little challenging of that use.

What does this mean for you?

Promote EBM, and ensure your authorization processes work well.

 

Jul
30

So what’s up with health care costs?

Posted by Joe Paduda on Thursday, July 30th, 2015 | Comments

Actuaries are projecting health care costs will increase 5.8% annually over the next 9 years. Others think that increases will be significantly smaller.

While the 5.8% is a bit higher than we’ve seen of late, it is a heckuva lot lower than the average for the last three decades.

Currently health care is responsible for 17.4% of US GDP; if the inflation rate prediction holds true and other economic sectors also grow as projected, health care will account for one out of every five dollars in ten years (19.6% to be precise).

We do know that the prediction will prove to be somewhat wrong, and economic growth for the next quarter is hard enough to predict, making a ten-year projection the proverbial dartboard in a dark room.  So, what’s with the discrepancy between predictions?

The actuaries responsible for the 5.8% figure believe the soft economy over the last few years has been the primary driver of low health care cost inflation.  Their thinking is that now the the economy is back to steady and significant growth, demand and prices will both heat up.

The counter-argument attributes the recent happy days of low medical cost inflation to structural changes in the health care delivery system. Their view is these changes, while overwhelmed somewhat by the big increase in the insured population due to PPACA, will help keep cost growth low as they become increasingly commonplace.

At this point, we just don’t know; there is anecdotal evidence that medical homes work and don’t; that ACOs are a success and a failure; that behavioral changes are working and are non-existent. That is far from surprising; we are still pretty early into this process, a process which is massively changing almost one-fifth of our nation’s economy.

What does this mean for you?

The key message is costs will continue to increase, with health insurance cost increases somewhat mitigated by higher deductibles and copays.

What’s also very clear is the health plans that are able to deliver lower costs and sufficient outcomes will do very well.

 

 

Jul
15

Drug testing (partially) explained

Posted by Joe Paduda on Wednesday, July 15th, 2015 | Comments

Once more we will delve into the minutiae of an issue…this time into testing patients who are prescribed opioids to ascertain if they are taking the prescribed medications, and if there is evidence they are consuming other licit and/or illicit drugs.

(full disclosure – Millennium Health, the largest toxicology testing company, is a consulting client)

All guidelines suggest/encourage/require testing of patients prescribed opioids.  

There are two types of urine drug tests – qualitative, where a cup test simply indicates if a drug is or is not present, and quantitative, which is much more accurate and must be done in a lab.

I’m surprised at the continued use of qualitative tests as they are notoriously unreliable; research indicates the cup test failed to show benzodiazepines were present for 28% of specimens, and cocaine for fully half of specimens evaluated – and false negatives and positives for other drugs are much higher than one would expect.  These “false negatives” are obviously misleading; the usefulness of cup tests is further compromised by how easy they are to fool. (there are about a gazillion web pages that provide info on passing a cup test…)

Say you are prescribed Oxycontin, but haven’t been taking the pills.  You’re scheduled for an office visit, have sold your pills, and don’t want to get caught.  You can rent pills to pass a pill count, and if you’re asked to pee in a cup, you can shave one of the pills into the cup, thereby adding the chemicals that will show you are compliant.
Voila!  you’re clean!

Except, if your sample gets sent to a lab for quantitative testing.  It is much harder to fool good lab testing because the testing equipment:

  • uses much lower cutoff levels for drugs, thereby finding more positives than cups do;
  • tests for metabolites – the chemicals created by your body after it processes the drugs: metabolites show you’ve actually taken the drug
  • checks for certain chemical markers that can indicate if the urine is fake or from another person (or, in some cases, another animal)

There’s much more to this; warning, if you start looking around on the web, you’ll find some incredible stories and myths and tales about folks allegedly passing tests; great for entertainment but very easy to become mesmerized for hours.

I recently reviewed data from a very large sample, specifically looking for data about cup results vs lab (quantitative) results.  The analysis was rather disturbing…Cup tests missed:

  • 45% of opiates (cup reported no opiates, lab reported opiates)
  • 44% of benzodiazepines
  • 28% of marijuana

Cup tests also indicate drugs are present when the lab tests show they are not, false positives occurred in:

  • 27% of reported opiates
  • 69% of antidepressants
  • 100% of PCP

What does this mean for you?

Be very careful about basing decisions on cup tests – even if they show there aren’t any anomalies or “unexpected” results.

Jun
19

My favorite day of the year

Posted by Joe Paduda on Friday, June 19th, 2015 | Comments

This year Father’s Day and the Summer Solstice fall on the same day – making for a very long day here in upstate NY with lots of daylight so I can loll around while being waited on (well, maybe not that last part).

While I was busy inundating your inbox with posts on the profitability – or lack thereof – of workers’ comp, a bunch of other stuff happened.

Another shot in the subrogation/third party liability battle was fired by Kentucky’s Medicaid program.  According to WorkCompCentral’s Ben Miller, hundreds of letters have been sent to work comp insurers in an attempt to ascertain if specific individuals’ medical care is due to a work comp injury.  The rationale is clear; Medicaid doesn’t want to pay for medical care it doesn’t have to.  As a taxpayer I completely support this.  Where it could get really sticky involves settled claims; if the work comp insurer/employer has settled the claim, my assumption (always dangerous) is the settlement requires the claimant to use those funds to pay for injury-related medical care.

What if the claimant doesn’t have any of the settlement dollars left?  If the claimant doesn’t pay, is the work comp insurer/employer liable? Who’s going to be stuck with the bill; the claimant?  the provider? Medicaid? another insurer?

Oh boy.

A terrific article in Harvard Business Review on what private equity investors do when they buy companies notes three distinct types of “engineering”; financial, governance, and operational.  Lots of insight, data, and examples make this a must read for anyone considering a transaction, or trying to understand how PE firms work.

Activity in the oil patch is slowing down, but claims counts are not going up.  Reuters quotes a Travelers insurance exec who’s a bit surprised about this; I have a call into Travelers to see if we can get more insight into the issue, and will share whatever I learn.

The new, updated Washington Guidelines for Prescribing Opioids for Pain are out; a product of the Agency Medical Directors (AMD), the new guidelines address opioid usage for many different conditions, cover special population issues, and update and expand a variety of treatment- and risk-assessment-related topics.  With five years’ experience under its belt, the AMD have learned a lot, lessons that other jurisdictions would be well-served to consider.

Finally, for many families in Charleston – and elsewhere – this Father’s Day is anything but joyful.  If I may be so bold, I’d suggest we strive to be part of the solution.

Jun
17

Workers’ comp profitability, Part 3

Posted by Joe Paduda on Wednesday, June 17th, 2015 | Comments

We’ve seen that work comp’s “profitability” isn’t very good, whether measured (inappropriately) as operating gain or (appropriately) as return on net worth/return on equity (ROE).

Today we’ll dig deeper into the data; below is a chart provided courtesy of CWCI, it is NAIC’s 2004-2013 Profitability Report, comparing average rates of return on net worth among California and US WC, property & casualty (P&C) insurers and all industries.

2004 2005 2006 2007 2008 2009  2010 2011 2012 2013 04-13Avg
Calif WC 12.6% 14.2% 16.4% 12.1% 7.0% 4.6% 5.2% 7.4% 3.9% 3.0% 8.6%
Calif All Lines 14.8% 14.5% 17.1% 11.9% 6.0% 9.4% 9.7% 8.4% 7.4% 7.6% 10.7%
US WC 10.1% 9.6% 10.0% 9.0% 5.1% 4.2% 3.9% 6.2% 5.9% 7.2% 7.1%
US All Lines 10.0% 5.3% 14.4% 12.5% 2.4% 6.3% 8.0% 4.9% 5.8% 9.0% 7.9%
NAIC P&C 8.0% 8.3% 12.2% 9.7% 2.2% 5.7% 6.0% 3.4% 5.2% 8.0% 6.9%
Fortune All Ind 13.9% 14.9% 15.4% 15.2% 13.1% 10.5% 12.7% 14.3% 13.4% 16.6% 14.0%

First up, look at the last row, Fortune’s All Industry average is higher than the US WC results every year for the last decade.

Over the last decade, WC’s returns have been just half the All Industry average.

Next, kindly allow me to direct your attention to the bolded red numbers – California WC insurers’ return on net worth for 2013 and the national average for the same year.  Fellow WC geeks will recall 2013 was identified by ProPubica/NPR as WC insurers’ “most profitable year in over a decade, bringing in a hefty 18 percent return.”

Oh were it only so.

(We dissected PP/NPR’s interpretation of profitability yesterday)

PP/NPR’s series of “reports” on workers’ comp allege that this “hefty” return is due in large part to reductions in benefits for workers pushed by employers and insurers and “reforms” that have taken away workers’ ability to choose their doctors – among other changes.  The reporters specifically cited big problems in California, where insurers’ doctors “deny” care without seeing the patient, where benefits have been slashed and workers made to suffer due to ill-conceived “reforms”.

This is a classic example of writers looking for “facts” that support their pre-conceived bias.  NAIC’s data shows just how wrong reporters Grabell and Berkes are; if the “reforms” in California were so one-sided, so employee-unfriendly, designed to benefit insurers at the expense of injured workers, those reforms have clearly NOT delivered the intended financial results.

By way of reference, historically the target ROE for US companies is in the 12-15 percent range, making the US WC insurance industry’s 7.1% return over the last decade look shabby indeed.

Remind me again why anyone would want to be a workers comp insurer???

 

 

Jun
10

Health care cost drivers, or, Here’s where you’re getting screwed

Posted by Joe Paduda on Wednesday, June 10th, 2015 | Comments

Forgive the vulgarity, but it seems apt when considering two articles just published in the venerable journal Health Affairs.

First, as physician practices consolidate, markets become more concentrated.   A study indicates orthopedic fees paid by private insurers are measurably higher in those markets with higher concentration.  As “Physician groups are growing larger in size and fewer in number”, expect this trend will affect other, currently-less-concentrated markets, thereby driving up the price of orthopedic services.

While the research by Alex Sun and Laurence Baker focused narrowly on knee arthroplasty, it’s likely an examination of other orthopedic procedures would yield the same finding.

A couple key quotes that should resonate among workers’ comp payers:

  • Our results suggest that the potential for reduced costs [due to larger physician groups] may be outweighed by providers’ ability to negotiate higher physician fees.
  • the ACA encourages further concentration to some degree by incentivizing physician groups to form ACOs to provide care. Again, our results suggest that the potential benefits of the formation of ACOs must be balanced against the potential for these organizations to negotiate higher physician fees.

I’d suggest that if private insurers are paying higher rates, workers comp payers are likely paying way higher rates.

Which is an excellent segue to the companion article on hospital markups (hat tip to Richard Krasner for getting to this a day before I did). The authors identified the 50 hospitals with the highest charge to cost ratios; this is a simple analysis comparing their chargemaster, or published price list, to Medicare’s assessment of their allowable costs. Here are a couple enlightening excerpts:

While most public and private health insurers do not use hospital charges to set their payment rates, uninsured patients are commonly asked to pay the full charges, and out-of-network patients and casualty and workers’ compensation insurers are often expected to pay a large portion of the full charges [emphasis added]

forty-nine (98 percent) are for profit, compared to 30 percent in the overall sample; one for-profit hospital system (Community Health Systems) operates half of the fifty hospitals with the highest markups (Exhibit 3). Hospital Corporation of America operates more than one-quarter of them.

Florida has 20 of the fifty hospitals with the highest markups; this is also a state with a fee schedule based on a percentage of “usual and customary” charges.

A notable finding; “markup varies substantially across medical services in the same hospital, and an overall hospital-level charge-to-cost ratio might not reflect the extent of markup for a specific patient. For example, among the fifty hospitals analyzed in this study, the average charge-to-cost ratio for anesthesiology is 112, for diagnostic radiology it is 15, and for nursery it is 3.”

What does this mean for you?

External forces are dramatically reshaping the health care delivery landscape; winners will be those payers who successfully adapt to those changes, not those who ignore them.

Jun
8

WC Rx Survey – early results are in…

Posted by Joe Paduda on Monday, June 8th, 2015 | Comments

We’ve finished collecting the data for CompPharma’s 12th (!) Annual Survey of Prescription Drug Management in Workers’ Comp;  working on compiling and analyzing the data now and expect to get the report out next week.

The survey uses both quantitative and qualitative questions; this enables us to track changes over time to key metrics including network penetration, mail order usage, generic efficiency, and compound drug growth.

The qualitative responses are really helpful in gaining an understanding of what’s keeping payers up at night, where they are seeing success, and how they view key issues.

Here are a few initial findings…

  • Drug management is viewed as more or much more important than other medical issues by 80% of respondents
  • 75% said drug costs will become more important over the next year
  • 2/3rds indicated compounds are the most concerning new issue in WC pharmacy

There’s a lot more analysis to be done as we dig into the qualitative responses. One key area will be the cost and quality control programs payers have implemented over the last 18 months and the results of those programs.

Once the final report is done and proofed, I’ll put up a link.

Public versions of the previous surveys are available free for download (no registration required) here.

 

Jun
4

A work comp exec’s MUST read

Posted by Joe Paduda on Thursday, June 4th, 2015 | Comments

The health care “system’s” problems are even worse for worker’s comp.

That’s the conclusion I reached after I finally got around to finishing “Overkill“, Atul Gawande’s latest piece on the clustermess that is the American health care system.

The top takeaway is this – there is huge over-diagnosis of medical “problems” due to an over-reliance on fancy diagnostic technology, technology that far-too-often identifies physical abnormalities that have little to no effect on one’s health or functionality.

An excerpt makes the point:

Studies of adults with no back pain find that half or more have degenerative disk disease on imaging. Disk disease is a turtle—an abnormality that generally causes no harm. It’s different when a diseased disk compresses the spinal cord or nerve root enough to cause specific symptoms, such as pain or weakness along the affected nerve’s territory, typically the leg or the arm. In those situations, surgery is proved to be more effective than nonsurgical treatment. For someone without such symptoms, though, there is no evidence that surgery helps to reduce pain or to prevent problems. One study found that between 1997 and 2005 national health-care expenditures for back-pain patients increased by nearly two-thirds, yet population surveys revealed no improvement in the level of back pain reported by patients. [emphasis added]

More specific to workers’ comp, the good folks at Liberty Mutual’s Institute for Research found claimants with back pain with:

early or non-indicated MRIs led to a cascade of medical services in the six-month period post-MRI that included electromyography, nerve conduction testing, advanced imaging, injections or surgery.

These procedures often occurred soon after the MRI and were 17 to nearly 55 times more likely to occur than in similar claims without MRI.

“Being a highly sensitive test, MRI will quite often reveal common age-related changes that have no correlation to the anatomical source of the lower back pain,” said one of the researchers, Glenn S. Pransky, MD, Center for Disability Research, which is part of the Liberty Mutual Research Institute for Safety, in a statement.

According to Pransky, evidence-based practice guidelines for lower back pain recommend against early MRI except for “red flag” indications such as severe trauma, infection or cancer.

Dr. William Gaines, associate national medical director, Liberty Mutual Commercial Insurance Claims, said that the National Committee for Quality Assurance and the American Board of Internal Medicine have emphasized for years that overuse of imaging does not represent good care for low back pain.

What does this mean for you?

Our health care system is very, very good at finding physiological and anatomical “problems”.  Unfortunately, it is also very good at assuming those findings actually indicate an underlying and significant pathology.