Another Way to Look at Health Spending

If we are to control health spending growth, we obviously have to understand the sources of the increases.  The Bureau of Labor Statistics at the Department of Labor in conjunction with other federal statistical groups has undertaken a project to create a disease-centered analysis of health spending.  (An interesting question, as always about our grossly inefficient government, is why we have at least four or five groups in different agencies doing economic analyses of health spending, instead of doing it all in one place.)  A recent article in a BLS publication summarizes the current results of this analysis.  (BLS Report)   The theory behind the disease-centered approach is that it makes it easier to tease out the effects of disease prevalence and of changes in the prices and methods used to treat diseases.  The researchers use price change data and the Medical Expenditure Panel Survey results to construct their index.

An important aspect of looking at health spending in this manner is the ability to track substitutions over time in the method of treating diseases.  For example, the use of inpatient care has generally been declining and the use of drugs increasing.  Some drugs are getting cheaper and some are very expensive.  The researchers compared a traditional medical price index to a disease-oriented one over a 17 year period ending in 2016.  The traditional price index rose 59% in that time, while the disease-based one grew 36%.  The divergence was not steady, early in the period there were years when the disease index actually increased faster than the traditional one.  So while the cost of a particular service may be rising fairly rapidly, the cost of a basket of services to treat a specific disease is increasing more slowly.  This is due to fewer services being used to treat a disease and to substitution of cheaper services.

From 2003 to 2013, across all diseases, there has been a 9% reduction in physician visits used to treat disease, a 27% decrease in outpatient services and a 15% decline in inpatient use.  This is startling.  Among diseases there is obviously variation.  A few actually see increases, for example there was a 55% rise in outpatient visits for skin diseases.  But generally there were declines, most notably in a couple of common disease areas.  Endocrine and metabolic diseases, which includes diabetes, experienced a 23% reduction in physician visits, a 46% decline in outpatient use and a 39% decrease in inpatient visits.  Circulatory system diseases, which includes high blood pressure and lipid disorders, had 32%, 46% and 42% reductions in physician, outpatient and inpatient visits, respectively.  This must largely be due to greater use of more effective drugs.

For several disease categories, the largest driver of spending growth is in fact prevalence.  For a few others it is use of more services per patient and for a few more it is increases in the cost per service.  Now it is possible that prevalence figures have risen over time due in part to better documentation and coding (or even, dare I say it, upcoding behavior for reimbursement purposes), but most of the prevalence growth probably is real, to some extent due to the aging of the population, but mostly to widespread poor health behaviors.  We have over 70 million people in our country with high blood pressure and another 60 million with high lipid levels.  On the one hand this analysis confirms that utilization is not the primary problem we face in regard to health spending.  And while unit price increases play a role, the biggest issue actually appears to be our disease prevalence.  Time for a few sticks to go along with the carrots that aim at increasing better health behaviors.