Registry studies revisited: indispensable in improving care.

Mkoma and colleagues have written a very interesting paper on the use of reperfusion therapy and time delay with regard to immigration status in patients with acute ischaemic stroke in Denmark [1]. These findings were derived from large nationwide registries. This provides several clear benefits over smaller studies designed to answer a specific clinical question, but also introduces different hurdles which were tackled very well in the present study.

To answer important questions, as in this paper, in which outcomes between large societal groups are compared, by definition very large numbers are needed, not feasible in ‘regular’ clinical studies. Large nationwide registries such as the Danish Stroke Registry can be very useful to answer such questions. An often‐heard limitation with performing analyses using quality registries is that real‐life data in these large registries are not as complete or precise as in clinical studies designed for one specific question (and with enough financial recourse to ensure data completeness). However, the fact that registry data, collected in daily clinical practice, may not be as complete as data in a specific clinical study would be does not preclude a valid and precise answer. Moreover, real‐life data provide better ‘real‐life’ estimates than smaller clinical studies would do. ‘Real‐life data’ ensure that, if a clear contrast is found, for example regarding outcome between certain groups, this by definition will be ‘true’.

A possible threat for large nationwide registry studies or quality of care audits is that designers aim at collecting too many variables. We all know that if the registration burden for healthcare workers, who are asked to collect these data during their clinical daily practice, becomes too high, this will almost by definition reduce the willingness to participate. To ensure viability, designers of registries or audits should strongly limit the number of variables and focus to register solely the variables needed to assess the quality of care, for example treatment times in reperfusion therapies in acute ischaemic stroke. Also, the feasibility of a registry increases if sufficient resources are available. Registration and follow‐up of all acute stroke patients is very time consuming and if resources are available to collect this will strongly enhance the success rate. In Denmark, and in Scandinavia in general, but also in the UK, governments have acknowledged the importance of high‐quality audits to improve healthcare, resulting in a large history and very good track record of nationwide acute stroke audits such as the Danish Stroke Registry or the Sentinel Stroke National Audit Programme [2, 3].

To answer their question, the authors also did a very good job in combining different registries and existing databases in their country. They merged data of the Danish Stroke Registry, the Danish Prehospital Database and Statistics Denmark. To enable this, they were allowed to use personal identification numbers of Danish citizens. This has to be praised. In other countries, such as in the Netherlands, this is not allowed based on the novel European laws for data protection and privacy. This creates problems in assessing the quality of care provided, for instance for stroke patients who are transferred a comprehensive stroke centre for endovascular thrombectomy [4]. Surprisingly, this is an opposite effect of why these identification numbers were designed in the first place, more specifically to fast trace individuals in medical emergencies for example. Moreover, in the Netherlands patient associations as well as the healthcare inspection authorities have unambiguously endorsed the use of individual identification numbers for this purpose of improving quality of healthcare. Denmark seems to be ahead regarding this issue and other countries should be encouraged to follow this example.

In this specific study, the authors came to very important conclusions: in Denmark immigration status is associated with a lower chance of receiving reperfusion therapy and in delayed treatment times in acute ischaemic stroke. This is a clear example of identifying an important problem from a societal perspective, and this was only possible by using and combining large nationwide registries. The findings may result in more detailed research in specific causes and, for example, targeted education for general improvement of quality of care—the ultimate goal of high‐quality nationwide registry studies such as presented here.

AUTHOR CONTRIBUTION

Paul J. Nederkoorn: Writing—original draft (lead).

CONFLICT OF INTEREST

Paul J Nederkoorn is chair of the Dutch Acute Stroke Audit (DASA).