I’m on an email group whose members are heavy into the money side of healthcare. They are fighting back against our exorbitantly expensive healthcare system. The members of this group know a lot about the macro-economics of healthcare—overall spend, spend by major categories, some cross-national comparisons—and the finance of healthcare—carve-outs, loss ratios, those god-awful PBMs, and so on.
But they don’t understand the micro-economics of healthcare. My trigger was this statement:
I wonder if any any of the group know of good quality $ data on the cost savings to health plans from diabetes chronic care programs, So, linking HbA1c improvement (clinical) tp hard cost savings for health plans (which is their only interest). We all know that story.
Improving A1Cs across a population of diabetics does not lower healthcare costs, in the short term or the long term. This group’s grand muse, Al, has told you that chronic disease costs are not as big a driver of the overall spend as other conditions, but that message does not seem to be taking hold of the collective consciousness.
I am determined to teach open-minded people the micro-economic realities of healthcare.
This science is 47 years old. The July 31, 1975 issue of NEJM is where this began. I will back up all my assertions with links to articles from the standard peer-reviewed medical literature. This is not only a U.S. science. These approaches are the basis for the coverage decisions of every other major developed country in the world. The National Institute for Health and Care Excellence (NICE) in the UK is the most transparent about how they decide what they’ll pay for healthcare interventions, but all other countries have an agency with a similar mission. In contrast, it is illegal for U.S. Medicare to consider cost-effectiveness in its coverage determinations.
There are many facets to the chronic care of diabetes so it’s a little more complicated than other chronic diseases. I’ll start with a simpler disease, hypertension. I’ll cover diabetes in later posts, if Brian approves of course.
The methodology for these cost-effectiveness studies were standardized in 1996. An update appeared in 2016—in a book called Cost-Effectiveness in Health and Medicine. I won’t get into the deep weeds on methodology, unless people have questions. The outcome of a cost-effectiveness analysis study is the incremental cost-effectiveness ratio (ICER). The unit of measurement of an ICER is the dollars spent/saved per quality-adjusted life year (QALY).
The cost trade-off for all chronic diseases is that there are costs on the front end for services like doctors’ visits, tests, and medications that may or may not pay off with lower treatment costs decades later. For hypertension in particular, the associated bad diseases of heart attacks and strokes may not appear for 50 years.
One of the disconnects I’ve observed in people who don’t understand the inputs and outputs of chronic disease care is that they over-estimate the incidence of the long-term outcomes and the effectiveness of treatments. Just because you have high blood pressure that is well treated on medication does not mean you will never have a stroke or heart attack.
Here’s the evidence:
One of the earliest comprehensive cost-effectiveness analyses was in a 1990 Annals of Internal Medicine study. https://www.acpjournals.org/doi/abs/10.7326/0003-4819-112-3-192 One of the authors was Harold Sox, MD, who was one of the early leaders of the evidence-based medicine movement. They concluded that “the absolute benefits of screening are quite small. Under our base-case assumptions, the average screenee (someone screened once for hypertension) can expect to save between 1 and 20 days of quality-adjusted life.” This study assumed about 20% of the screened population would have hypertension, so the increased life expectancy per hypertensive patient would be about 5 days to 3 months. The ICER for a 40-year-old man was $16,280/QALY, for a 40-year-old woman $23,216.
These positive numbers mean that the cost of the screening and treatment outweighed future savings. Therefore, this service increases the total cost of care over the short term and the long term.
In the same year in JAMA, a study looked at the ICERs for different drug class treatments over a lifetime of treatment. https://jamanetwork.com/journals/jama/article-abstract/380290. They calculated the total costs and outcomes for the U.S. population with hypertension. For the cheapest drug, propranolol, the total cost of therapy was $85.8 billion, the cost of saved events was $39.7 billion, for a net increase in costs of $46.1 billion. This resulted in a cost per year of life saved of $10,900. More expensive drugs had higher ICERs. They did not quality adjust in this study.
There have been studies is specific sub-populations. This one is from the UKPDS type 2 diabetes study.
https://pubmed.ncbi.nlm.nih.gov/9732339/ They compared tight blood pressure control (<180/105) vs. loose control (<150/85). Life expectancy improved by 9.7 months, 5 if future events are discounted at 6%. They found treating hypertension in diabetics to be relatively inexpensive, but not cost saving at £1049/QALY. Diabetics are more effected by the vascular damage of hypertension than non-diabetics, therefore one would expect that treatment would have a larger effect on life expectancy and would therefore be more cost-effective.
Some studies looked at cost-effectiveness in a slightly different way. This study looked at the net cost per coronary event prevented. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC286247/. It didn’t explicitly report change in life expectancy. It calculated the cost to prevent a coronary event was $18,300 for “intensive antihypertensive” treatment.
The treatment of hypertension doesn’t have to be just pills. This study looked at the added value of case management in vulnerable populations. https://pubmed.ncbi.nlm.nih.gov/29239999/. It estimated an increased life expectancy of .04 QALYs at a cost of $52,850/QALY.
This 2011 study appropriately pointed out that some of the cost-effectiveness studies for hypertension assumed brand drug pricing and estimated that the true cost-effectiveness was better at generic prices.
https://www.healthaffairs.org/doi/full/10.1377/hlthaff.2010.0431. It calculated $52,983/QALY with brand-name drug and $7,753/QALY for generic. Again, there the total cost goes up with treatment.
There are a few dissenting studies.
In 1977, there was a study published before the first consensus statement on cost-effectiveness methodology was published that found a very high impact of treating hypertension.
https://www.nejm.org/doi/10.1056/NEJM197703312961307. They calculated an increased life expectancy for treating hypertension of 2.3 to 5.0 years for women, 1.4 to 8.1 for men. As you can see from the other examples, this is the clear outlier on the effectiveness of treating hypertension. In spite of this huge estimated positive impact of treatment, they estimated no cost savings.
For some reason, the CDC wrote 2 studies that found small savings. In 2002, the CDC published a study in JAMA that calculated the ICER for “intensified hypertension control” in type 2 diabetics as cost saving. It decreased the total spend by about 2%. https://pubmed.ncbi.nlm.nih.gov/12020335/
The CDC also published an estimate of the benefit of “Intensified hypertension control” based on UKPDS study findings. This is odd given that the British researchers who conducted the UKPDS trial found that treating hypertension at British drug and clinic prices didn’t cost much, but it was not cost saving.
https://pubmed.ncbi.nlm.nih.gov/12020335/. The CDC estimated that “intensified hypertension control” increased life expectancy by 0.40 to 0.47 QALYs and was cost saving.
I’ll finish with a 2015 study in NEJM that asked the question, “what if we magically found all untreated people with hypertension in the U.S. and treated them (in other words, no costs for outreach and clinic visits to find these people)?” Cost-Effectiveness of Hypertension Therapy According to 2014 Guidelines | NEJM. Their results were more nuanced. They estimated that treating hypertension was cost-saving for men and women ages 60-74, and cost saving for men with more severe hypertension (>160/100) at ages 33-74. For women with more severe disease, the cost was $26,000/QALY ages 33-44, but cost saving for other ages. For men and women with an untreated pressure of >140/90, the overall ICER for men was $40,000/QALY, for women it ranged from $7,000 to $181,000/QALY. And remember, these cost estimates are biased low because they did not add costs that would be required to discover these people to treat them.
Now let’s dig a little deeper into their estimates. For men, they assumed that 2.4 million new cases of stage 1 hypertension in males ages 35-59 would be discovered and treated. The number of CVD events in this population would decrease from 195,000 to 179,000 (8% decrease) over 10 years, CVD deaths from 49,100 to 46,200 (6% decrease), CVD costs from $32.5 billion to $31.9 billion (2% decrease), and QALYS from 44,162 to 44,187 (.05% increase). For females, the numbers are CVD events 101,000 to 100,000 (1% decrease), CVD deaths 20,600 to 20,400 (1% decrease), CVD costs $20.3 billion to $20.2 billion (.4% increase), and QALYs 44,995 to 44,998 (.01% increase).
Summary
The overall summary of 40+ years of cost-effectiveness studies of hypertension treatment is that it helps patients a little from slightly fewer heart attacks and strokes, and increases the total cost of healthcare. It increases life expectancy in higher-risk patients by less than a year, a few weeks to a few months in lower risk patients.
Thanks for sharing this. As always “strong work”
Larry