Rankings comparing healthcare providers are often misleading

The upper part of the Erasmus MC building surrounded by greenery.
Portrait photo of Nèwel Salet with a neon purple background.

Can you compare healthcare outcomes and how do you go about it? This is an important thread running through the thesis of Nèwel Salet (Erasmus School of Health Policy & Management). Remarkably, there is very little difference in the quality of doctors. The PhD candidate is regularly annoyed by 'judgment articles' comparing healthcare providers. "In the media, individual pieces of the puzzle are regularly presented as the puzzle."

To get as much usable data as possible, the researcher looked at the variation in care outcomes for four common treatments (including knee replacement and bladder and bowel cancer treatment). He included criteria such as risk of readmission, risk of ICU admission and risk of death. In this comparison, he looked at the majority of Dutch hospitals and doctors performing these treatments, using anonymised data.

Rankings are misleading

Everyone who has to go to hospital for an operation or treatment wants the best possible treatment. Websites like ZorgkaartNederland.nl and Ziekenhuischeck.nl cleverly capitalise on this. For each type of treatment, you can see precisely which hospital (ostensibly) comes out on top. On the former website, you can even find scores for individual doctors and specialists. Nèwel Salet says, "Our research shows that such rankings are misleading. You cannot say that one doctor is better than another based on this care map."

Doctors are working in an operating room in blue coats.
Natanael Melchor (Unsplash)

According to the researcher, the key is to take a ‘multi-level’ look. In other words, look at the different levels, from the doctor and their team, to the level of the hospital, and even the healthcare system as a whole. "Those different levels are not usually taken into account. But it’s the only way to look specifically at where in the chain the reason for the deviation lies; for example, with an individual doctor or hospital. So when you look at it this way, it turns out that you cannot make accurate distinctions at the level of doctors", Salet explains. 

Comparing hospitals is difficult

If you look at hospitals, there do appear to be differences for different treatments. "Even so, you can’t say: ‘cardiology is poor at Hospital A, while it’s good at Hospital B’. That’s not how it works. Moreover, such scores are often not calculated properly. A cardiology department may do well in some areas and not so well in others. Say this one scores poorly on readmissions only; you need to pay more attention to that aspect specifically."

For a relatively simple procedure such as a knee replacement, there appears to be little variation between hospitals. In other words, for that procedure, you’re almost equally well off anywhere in the Netherlands. For more complex procedures, such as after a heart attack, the variation between hospitals is greater. "If you want to address those differences, it’s not useful to look over specialists’ shoulders. You have to look at a specialist group as a whole for the different indicators such as readmission or mortality risk."

Overview photo of the Erasmus MC hospital.
Rob van Esch

Judgment articles in the media

Salet stresses that his study is not intended to rank hospitals, but to find starting points for organising care better. He understands the tendency to want to inform patients about quality of care. Yet media reports comparing hospitals regularly cause frustration. "They tend to be articles that pass judgment. Of course, it would be nice to be able to say: for this condition, you should not seek treatment here. But then you would have to work with reliable data and make fair comparisons. In the media, individual pieces of the puzzle are regularly presented as the whole puzzle."

In another study from the thesis, the PhD candidate looked at patient data (e.g. age and gender) and whether this affects healthcare outcomes. Unlike in many other countries, socioeconomic status does not appear to be an influence. "The most striking finding was that the risk of readmission is higher by a factor of 10 to 25 if you have been admitted in the previous year. This, of course, is important for doctors to know. So from a research perspective, this data should ideally be included in risk estimates."

The fundamentals of value-driven care are lacking

The cover of the thesis shows an illustration of a hospital hovering above the ground like a castle under construction. This serves as a metaphor for ‘value-driven care’, a popular idea to optimise care and keep rising healthcare costs in check. "At the moment, the foundation is largely missing", Salet explains. "Much work is still needed to make value-driven care a reality. That starts with good data collection, but it is also about collaboration, training and funding. With this research, we show what quality of care is being provided and where differences in outcomes come from. Understanding that is a crucial first step."

PhD student
Related content
Towards Building a more Solid Foundation for Value-Driven Care
Logo eshpm

Compare @count study programme

  • @title

    • Duration: @duration
Compare study programmes