Preference sensitive health care: the causes of variations

There is a case-mix part (25% according to Wennberg) such as inguinal hernia, cataracts, metrorrhagia or knee osteoarthritis, for which modern medicine has an effective surgical response, although in the application of the technique there is often a margin for the doctor’s interpretation, another margin for the subjectivity of the patient, such as pain perception or adaptation to the lack of visual acuity, as well as a very important factor: the decision of the patient himself. There are men who prefer to wear a brace to hernia surgery and women who prefer to live with their uterus, provided that the degree of the discomfort and metrorrhagy allows them to.

After this introduction, let’s see the Variations in Health Care, the good, the bad and the inexplicable report by John Appleby and his collaborators, published by King's Fund in 2011, which states that variations in hospitalization rates are pervasive and persistent, and even affect common interventions known to be effective such as hip replacement for advanced osteoarthritis cases.

                Distribution rates of hip replacement in England 2009/10

Note that although adjusted for age and sex, the rate of hospitalization for hip replacement (with lighter blue line) in the graphic, are observed in 18 PCT (Primary Care Trust) that show a value of 60 interventions per 100,000 inhabitants per year, while at the other extreme, there are 20 showing a rate of 140. That is, citizens of the latter communities have a 2.3 times higher probability of being operated on for prosthetic hip replacement than the citizens of the former. The same report shows that these regional differences are extended to other elective procedures such as knee, cholecystectomies or coronary angioplasty.

Whichever way you look, this type of variation in the use of health care resources for procedures seemingly well identified and standardized, is universal, and this is seen when comparisons between systems (countries) are being made, but don’t disappear when, as in the English case, intra-communities and intersystem analysis are being made. If I may, I’ll next show a Spanish example from the Variaciones de la Práctica Médica (VPM) initiative that has the support of all autonomous regions. I have in hand the final report of the VPM research project, led by Julian Librero: Variations in the use of knee arthroplasty in the National Health System, from which I have chosen this graph:

I think in the above graph it should be noted that, excluding outliers health areas (in light blue), there are Spanish territories where inhabitants have a probability of knee prosthesis intervention 6.4 times higher than inhabitants from the areas with the lowest rates. And this occurs in a health system that is considered one of the most fair and accessible in the world.

Where is the problem? What explains these extraordinary variations?

To answer these questions, there’s nothing better that to resort to a 1977 investigation led by a young Wennberg. The following table, published in Tracking Medicine by the same author (Oxford University Press 2010), there is a summary of the results of this work. It’s the comparison of two small communities: Middlebury in Vermont vs. Randolph in New Hampshire. As seen in the table, the two communities have the same socioeconomic characteristics, the same prevalence of chronic diseases, and the same doctor accessibility. However, the Middlebury citizens are hospitalized 67% more than those of Randolph, and undergo surgery 63% more.

This work is a benchmark for questions about the causes of variations in the use of health resources. According to Wennberg, if when demographic, socioeconomic and accessibility factors have been adjusted, such overwhelming differences still remain, one must draw the attention of the variations towards the medical practice and the availability of healthcare resources provided by doctors in each area.



Jordi Varela
Editor