THE LATEST TRENDS IN YOUR FAVORITE TOPIC: HEALTHCARE

If you’re frustrated with healthcare these days, you’re not alone. That value-added service PEOs either insure or administer has continued to grow in cost and complexity. In this article, we’ll cover top trends as seen through the lens of benefits consulting. Specifically, we’ll deal with cost dynamics and how PEOs can chart new paths to sustainable growth.

ECONOMIC IMPACT

The US spends 18% of GDP on healthcare. That’s astounding considering we have the highest GDP in the world at $33 trillion (T) annually. 18% of $33T is $5.9T. That’s our spend on healthcare. Germany has the world’s third-largest economy and comes in second with spending only 12% of GDP (a ratio 1/3 less than us). To put their 12% in context, Germany’s entire GDP is $5.5T annually, or less than what the US spends on healthcare alone!

Yes, we spend a lot. About 50% of what we spend is on publicly funded programs like Medicare, Medicaid and military and veterans’ benefits. The other 50% are dollars we put into privately funded solutions through employers and the individual market. Over 80% of the privately funded healthcare is provided through employer-sponsored benefits, or about 40% of all money allocated to healthcare in the US. That’s approximately $2.5T, or large enough to exceed the annual GDP of countries like Russia and Mexico. That’s just the value of healthcare benefits provided through US employers. That’s heady stuff.

You wouldn’t be surprised to learn that growth in our healthcare costs is expected to outpace growth in our GDP. By 2034, healthcare spending is forecasted to reach nearly 21% of GDP or $9T. That’s over a 50% increase in just 8 more years.

There are numerous factors at play. Before we talk about how technology is driving cost, let’s discuss how changes in workforce are driving consumption.

WORKFORCE CHANGES

According to the US census bureau, the fastest-growing age group in the labor force is workers aged 55 and older (a group to which I, admittedly, belong). Workers in that age segment are approximately 25% of the workforce, growing at a rate 50% higher than average. Our data suggest that workers in that segment consume healthcare at a rate 2x that of workers 35-54 and nearly 4x that of workers under 34. Chronic diseases like diabetes, heart disease and neurological disorders manifest more frequently in that group because lifestyle and genetic factors have had ample time to express themselves. Increased age also brings increased exposure to cancer and autoimmune conditions. So, suffice it to say that condition-based demand for healthcare is rising apart from manufactured demand via new technology and advertising.

Let’s now cover where money is being spent on healthcare delivery. Viewing in broad categories, inpatient hospital care is 15%-20% of total spend but declining. Outpatient hospital care is growing quickly at 25%-30% of total spend with many expensive drug therapies being administered in that setting. Hospital ER adds another 5%, bringing hospital-based care to 45%-55% of expenditure. Physician and surgical costs are another 20-25% of total spend, and retail Rx is the single fastest growing segment of cost at 20-25% and increasing rapidly. It’s easiest to think of roughly half of the cost concentrated in hospital-based services and half in the combination of physician and retail RX. The two fastest growing segments are outpatient hospital and retail Rx.

Hospital consolidation and renegotiation of insurance contracts are driving higher hospital fees while advancements in care continue to shift hospital services to an outpatient setting. Hospitals are acquiring physician services and steering more patients to hospital outpatient. Specialty pharmaceuticals are commonly administered in a hospital-outpatient setting for treatment of cancer, autoimmune diseases and genetic illnesses.

Meanwhile, the development of new, specialty-retail drugs is driving a wave of consumption through new, highly effective treatments for rheumatoid arthritis, plaque psoriasis, psoriatic arthritis, Crohn’s disease, MS, lupus, cystic fibrosis and a host of other genetic illnesses. Cancer treatments have also advanced with new oral medications that change the way the human body confronts the disease. The emergence of new treatments offers incredible promise of more effective care across the spectrum, but that comes with an increased price tag.

THE 2 PERCENT

We routinely run analyses of how members enrolled through PEO plans use healthcare. Those analyses reveal consistently reliable statistics. One of the key findings is that a very small number of covered members drives an increasingly high % of total consumption. Simplistically, 2 of every 100 covered members generate half of the insurance claims by cost. The remaining 98 generate the other half. To compare those ratios, they would be 50/2 = 25 for the high-utilizing 2 members, and 50/98 = 0.51 for the balance. That means the high-cost population is roughly 50 times more expensive than the rest (25/0.51=49). That means insurance plans, even very large ones, are susceptible to wide swings in performance from just a small change in how the high utilizers impact a policy year. Those 2 members spread throughout a worksite-employer population have a different impact.

For example, 4 employers with 25 members each, 100 in total, would have 2 members generating half of their medical cost (2 of 100). If those 2 members fell across 2 different worksite employers, half of the employers in the sample (2 of 4) would be underfunded. So, just 2 of 100 members could create underfunding with half of the worksites! When viewing that phenomenon across a large sample, approximately 30% of employers contribute 70% of total cost (underfunded worksites), while 70% of employers contribute the remaining 30%. From a ratio standpoint, that’s 70/30 = 2.33 compared to 30/70 = 0.43, with the high-consuming employers manifesting nearly 5.5x the claims of the others (2.33/0.43=5.4). That means the 30% of worksites with the highest claims average about 5-6x the cost of the rest of the population.

Further, the “50% from 2%” phenomenon also appears within Rx data. Specialty Rx, treating the rare diseases referenced earlier, comprise only 2% of scripts filled in retail Rx. Because of the high cost associated with those drugs, they generate 50% of total retail Rx spending. Those claims and attendant costs will only grow in significance with the release of additional specialty drugs. The 50/2 and 70/30 ratios appear commonly in our data sets and they often inform how plan performance is evaluated at renewal.

In looking a bit deeper, our data show that retail Rx treatment of autoimmune disorders is the most commonly occurring claim that exceeds $50,000 ($50k) in a 12-month period. Examples of drugs in that class include Skyrizi, Stelara, Cosentyx, Tremfya, Dupixent and Humira. Some are available in a less-costly biosimilar, but still at significant cost. The average cost of those drugs in our data is $100k/yr. They represent 66% of all retail Rx claims over $50k in a year, and those drugs have emerged significantly post COVID. Unlike certain types of drug treatments, they are largely not delaying or eliminating costly hospitalization and are simply adding to spending. Retail Rx for cancer comes in second at just 6% of claims over $50k in 12 months. All retail-Rx claims in the cancer category averaged $150k annually and that doesn’t capture more expensive IV cancer drugs administered in an outpatient setting. Advanced therapies like antineoplastic immunotherapy are easier on the body than traditional chemotherapy but come at an average annual cost of $500k- billed as medical claims and not Rx. Retail Rx treatment of liver and kidney disease average $150k/yr, with treatment of endocrine disorders and MS averaging just under $100k/yr. Blood disorders represented only 1% of high-cost Rx but averaged $500k/yr in treatment cost for the drugs only. To reiterate, these costs are for the retail Rx scripts only and include no other treatment costs.

The costs from retail Rx alone are staggering and, although some can be predicted from risk-management models, the most common-occurring claims often appear without warning: the high-cost Rx is the first notable treatment someone receives after testing. Aside from high-cost Rx, GLP-1s, not high-dollar claims themselves, are being prescribed often enough to add approximately 10% to total Rx spending or 2% to overall plan costs. All of this is driving annual forecasts of medical/rx trend to a consensus of 9%/yr but with significant volatility related to high-dollar claims. Inside the 9% trend is an approximate 8% medical trend coupled with an approximate 12% Rx trend. Also, there’s a hidden stat within the 50/2 ratio mentioned earlier: 1 in 1,000 covered members generate 10% of claims cost. Those claimants carry a cost index of 10/0.1 = 100 compared to the 25 cost index for high utilizers in general and 0.51 for everyone else. A very slight variance in the 1-in-a-1000 group has a very measurable impact to annual plan performance. One notable story within our data set was a baby born with a genetic illness carrying unfortunate, grave, complications. The baby received a newly approved treatment for the illness and, after intravenous therapy and a protracted hospitalization, was cured- not stabilized but cured. That’s amazing. We live in amazing times. The claim was $4 million (M).

The trend in healthcare cost appears to not be cyclical but structural and compounding. We need to appreciate that significant investment is being made in treatment of conditions that are classically defined as “rare.” The National Organization for Rare Disorders defines a rare disease as one affecting less than 200,000 people in the US. The broader rare-disease community defines it more loosely as less than 1.5M people. With a US population of 350M, even 1.5M people with an illness represent just 0.4% of the population. Using that broader definition, the number of catalogued diseases sits at approximately 10,000. That’s right, 10,000 distinct rare diseases. Given the sheer number of diseases, the odds of someone in the US having at least one of the diseases considered “rare” is 1 in 10. That doesn’t seem very rare. The point is simply that there is ample opportunity for new investors to pour billions into the development of treatments for a fractional number of eligible patients, driving up the cost per treatment. Given there are so many rare conditions eligible for treatment, you can appreciate the challenge we all face.

To recap, we have an aging workforce, hospital and physician-group consolidation with price inflation, growing specialty drugs, increased treatment of rare conditions (often chronic), increased cancer diagnosis with advanced treatments and more significant high-dollar claims. Some argue that healthcare use is up in general in the years following COVID simply because of delayed diagnosis and treatment, possibly combined with suboptimal habits during COVID. Our data shows that claim trend has roundly outpaced post-renewal premium yield, putting policyholders on the back foot when it comes to negotiating renewals. Many health policies are early in the process of achieving suitable premium given new claims, and overall performance reflects that. That issue is impacting all health-insurance markets and all carriers. So, how does all this impact PEOs offering access to health insurance?

THE PEO IMPACT

Risk-prediction models used in health underwriting are directionally useful and, based upon testing and actuarial feedback, are most effective at making near-term cost predictions for actively treated chronic conditions. Generally, there is a tremendous amount of uncertainty in the months following point-in-time predictions. That comes from a combination of a few things. First, data is lagged. Timing is always an issue. Second, data is often incomplete. Models using aggregator data for new-business evaluation could be missing a few pieces (from inaccurate member data and low data robustness), and renewal models contend with an increasing number of masked conditions (conditions too concerning to be revealed by insurers even with appropriate protections and documentation). Third, the data set itself is insufficient, defined by ICD-10s, CPI-4s and NDCs. Those are the codes showing diagnoses, procedures and drugs. The data set was created to reimburse providers for services and was adapted for clinical intervention and cost prediction. The data set is missing electronic medical and health records that capture a much larger picture of member health. We can see a test was run, but we can’t see the results. We need to wait for any ensuing treatment to forecast near-term cost using lagged data, so far from ideal. We are missing lab results, vitals, clinical notes, and physician observations and treatment recommendations. The absence of complete data stymies significant predictive gains from AI, too. AI can improve how people evaluate options and buy insurance. AI in underwriting can and will improve data integrity, speed, clarity, reporting, guided decision making and user interface. But AI will struggle to markedly improve cost prediction until we expand the data set. Fourth, precursors to many claims are not present in the data. People can have little to no expensive treatment before being treated with an Rx at $100k/yr. High-cost claims can pop up with little warning. Lastly, predictive models were developed from large data sets to then apply to large data sets. We can predict average incidence and condition cost for an entire health plan, but we are applying models to small groups where the forecasting is directionally useful but subject to substantial volatility in the months after prediction.

Let’s revisit one earlier point about prediction models being best at predicting cost for immediate, high-cost risk from actively treated chronic conditions. They cannot assure us that groups presenting well will continue well. The traditional approach of assigning groups most-favored status based upon point-in-time risk metrics no longer meshes with emerging data on actual performance. Data suggest that, within an entire health plan, the best assumption we can make is that most clients will present with average risk. Given that average cost is higher now than in years past, average premium needs to keep pace. Benefits consultants are squarely focused on identifying strategies to increase premium.

One last important consideration is determining the best markets for health insurance. Our data suggest that member costs across groups, defined by enrollment count, are relatively homogenous. If anything, there is a slight bias toward higher claims per member within the largest worksites. That’s based on larger groups, given their size, carrying a greater statistical likelihood of picking up high-dollar claims, and the fact that a subset of high-dollar claims will be very high, even by large-group standards. More high dollar claims increases the possibility of an extremely high claim, and competition for larger groups typically drives their premium lower: a double whammy. Our data suggest that the best opportunity to achieve competitive, yet appropriate, premium lies with small groups (under 50 eligible employees in most states), including groups new to benefits. Our enterprise-wide data substantiate that smaller groups represent the safest growth path for now. That conclusion runs counter to traditional thinking about safety in numbers, but reveals an important maxim that safety often comes from seeking shelter rather than trusting yourself to consistently predict the weather.

How do PEOs seek shelter while charting paths to sustainable growth? We believe that comes from carefully defining company objectives, accepting realities in the healthcare market today, and setting strategies for achieving quality growth despite market challenges. That would impact both how PEOs issue their new-business proposals and how they retain their current clients. It also comes from recontextualizing benefits and the role that masters play in a much larger benefits arena. We need to meet worksite employers where they are with flexible, cost-effective, dynamic, and appropriate insurance solutions. That likely means supplementing masters with complementary options and, for PEOs without masters, leaning into diversification. Benefits consultants can help.

I’m fond of asking our PEO clients, in the context of health benefits, how they define “winning.” That’s a deceptively difficult question to answer. I urge you to consider how you define winning and pressure test that answer with data. With reflection, we often find we know more than we appreciate but less than we should. Accepting our new reality is key to success, and as we change the way we approach healthcare solutions, we can adapt and overcome. We got this.