Using Real Health Spending Data 101

Real Health Spending Data

When someone makes a statement about health care that I find questionable (e.g., “Americans are fat and that is the reason we have high health costs”), I immediately ask for the data supporting the statement.   Most of the time… I am greeted with stunned silence.  Looking at real health spending data provides me with information I can use to increase my knowledge base.  It helps me recognize the unsubstantiated messages like the one above that are constantly being thrown at me by the media and HICUP (BB Brigade acronym for those who are engaged in the business of health care in the USA) members. As “engaged” health consumers, we need the data to separate the fact from the fiction.

One of the most important factors that affect the price of the health insurance premium is the composition (distribution) of the risk pool.  In my last blog post,  I explained the difference between average (mean) and median statistics for a hypothetical risk pool of 20 people and introduced the concept of dividing up the risk pool into Low Spender, Medium Spender, and High Spender categories.

Let’s now apply some real health spending data found in the Medical Expenditure Panel Survey to my risk pool categories. This survey collects personal health spending data for civilian Americans (all ages) not in institutions (no nursing home or mental institution residents, or incarcerated prison inmates). If you remember from a past blog post,  personal health spending is the major component of national health spending per capita.  While excluding some Americans, the data in this survey is still a good measure of our nationwide health insurance risk pool. Using 2013 data from the survey, I generated the table below showing the median and mean total health services expenses grouped by age.

Medical Expenditure Panel Survey risk pool 2013
There is a lot of information in this one table that can be digested for future reference.  My general observations from the 2013 data are as follows:

  • For people under the age of 65, the median spending was $1059 per person and average (mean) spending was $4282.
  • For people over the age of 65 (Medicare beneficiaries), the median spending was $4830 and the average (mean) spending was $10,125.
  • Median spending is lower than the average (mean) for all age groups (i.e. health spending is concentrated in a small subgroup of heavy spenders).
  • Average (mean) and median figures are “per person with an expense” and therefore exclude people who have $0 in expenses. This serves to make the published values larger than one would expect if one had included the 15.6% (more than 49 million) of the population with $0 in expenses. If I am reading this data correctly, there is something fishy about this exclusion of people who have zero health expense.
  • Expenses are higher with increasing age (except for the under 5 group where the trend departs). This is probably the result of very high costs associated with keeping a small group of premature and sickly newborns alive.  The fact that the highest spending occurs in the over 65 crowd is to be expected since the potential for sickness increases with age.
  • Medicare pays over $68 billion for the health care of people under 65 (remember that Medicare also insures people with disabilities, end stage renal disease and “Lou Gehrig’s” disease).

If I looked at this table longer, I could come up with many more observations and some interesting food for thought.  I am sure you can add to my list above with your own insights.

Applying the Real Data to Low, Medium, and High Spender Categories 

Taking the 2013 data from the Medical Expenditure Panel Survey, I subdivided the total population (315.7 million) into Low Spenders, Medium Spenders, and High Spenders with cut-offs defined by percentage of population.  I chose the lowest 50% (Low Spender), the next 45% (Medium Spender), and the last 5% (High Spender).  If you would like to see another way of presenting the distribution of health spending data from the Medical Expenditure Panel Survey, visit NIHCM’s look at Health Care’s Big Spenders: The Characteristics Behind the Curve. Remember “raw” data can be presented in a number of ways and, depending upon how it is grouped, can convey different messages.  Grouping data in more than one way helps the presenter better understand the distribution of the data and the most important message that the data represents.

Using my groupings for low, medium, and high spenders, the average (mean) annual health spending for each of my categories is tabulated below:

average health spending for low, medium, and high categories
This data can be graphically shown in the bar graph below.

health spending distribution by population 2013

In the figure above, 315+ million people are represented and the numbers at the tops of each bar represent the mean (average) annual spending per person for the given category. The bottom 50% of the population (who I am calling Low Spenders) spent $253 on average.  The median spending given in the table above ($1389) is the largest amount spent by a person in the Low Spenders category.

One serious illness (or premature birth) can place anyone into the 5% High Spender group after decades of being in the Low Spender group. These groups are not set in stone and a given person can move from one category to the next as one’s health changes. If this calamity were to happen to you or someone in your family, you had better have insurance or a spare $43,000 (or more) in the bank!   It does happen to 5% of the population in any given year.

The Importance of This Real Health Spending Data

This real health spending data tells us what Americans are actually spending for health care and raises a lot of questions.  If over (the median spending figure above does not include more than 49 million Americans with $0 health spending) 50% of the population spent less than $1389 in 2013, then why were health insurance premiums so high ($5,884 for single coverage employer-sponsored health insurance according to the Kaiser Family Foundation annual survey)?  Does the spending of the top 5% High Spenders impact the premium cost that much?  I wonder given the fact that Americans aged 65 and older constituted 14% of the American population and consumed 32% of the health spending ($449,714/$1,400,523 from Table 1 above), the spending by High Spenders therefore comes disproportionately out of Medicare and not employer-sponsored health insurance risk pools.

The Bottom Line

As “engaged” health consumers, we need real health spending data for the entire population of the United States to be able to separate the fact from the fiction.  Armed with this data and the information I can derive from it, I am better equipped to understand when I am being “marketed” with unsubstantiated statements.

Using real health spending data, I have grouped Americans into Low Spender (bottom 50% of population), Medium Spender (next 45%), and High Spender (the remaining 5%) categories.  You will be seeing these categories in future bog posts where I discuss health spending for ALL Americans (i.e., the United States risk pool).


Leave a Reply

Your email address will not be published.