Dear Sir,
We thank Dr. Sabour for his interest in our article, which gave us an opportunity to clarify our message [Citation1,Citation2]. In his letter Dr Sabour’s first concern was our use of the word ‘validity’. Validity is not an exactly defined term in the literature. It can have slightly different meaning in different contexts but we agree with Dr. Sabour’s definition that, generally speaking, validity refers to whether or not the test measures what it claims to measure. However, Dr. Sabour seems to be referring to a laboratory situation where for example a new laboratory test is compared to the correct reference standard. This is not exactly applicable to our study. We compared diagnoses which are based on human interpretations of the electrocardiogram (ECG) recordings and there is thus no absolutely correct ‘golden standard’ although the coding of the cardiology resident was used as the reference. This is why we deliberately did not calculate the absolute numbers of true positives and false positives etc, as requested by Dr. Sabour. As a whole, our study examined whether or not the diagnoses of ST-elevation myocardial infarction (STEMI), non-ST-elevation infarction (NSTEMI) and non-classifiable myocardial infarction (NCMI) in Finnish electronic hospital discharge register (HDR) are correct when carefully re-examined using standardized criteria. Therefore, we believe that the use of the word ‘validity’ is appropriate in our article, although we acknowledge that in some sentences words like ‘repeatability’ or ‘reproducibility’ could also have been used.
The second concern of Dr. Sabour was about applying Cohen’s kappa coefficient in our study to assess the agreement. He points out in his letter that Cohen’s kappa depends on the prevalence in each category and gives an example from his own work. If we have understood Dr. Sabour’s example correctly (the numbers do not add up in the second 2 × 2 table), this situation is not applicable to our paper. As described in the methods-section of the paper, we have deliberately selected the same amount of patients from every category (NSTEMI, STEMI and NCMI) so the prevalence is same in each category. We compared the hospital discharge register diagnoses to the diagnoses of the cardiology resident and resident’s diagnoses to those of a senior cardiologist. These comparisons showed that if the diagnoses are carefully set correctly they are reliable and repeatable (resident vs cardiologist), whereas if they are not carefully set correctly they are not repeatable and thus not reliable and valid (resident vs hospital discharge register). In all, we have used Cohen’s kappa-coefficients appropriately [Citation3] to assess the reproducibility of STEMI and NSTEMI diagnoses and we are confident that our results give a correct picture on the reliability and validity of these diagnoses in the Finnish electronic hospital discharge register.
References
- Sabour S. Hospital discharge register data on non-ST-elevation and ST-elevation myocardial infarction in Finland; Terminology and statistical issues on validity and agreement to avoid misinterpretation.
- Okkonen M, Havulinna AS, Ukkola O, et al. The validity of hospital discharge register data on non-ST-elevation and ST-elevation myocardial infarction in Finland. Scand Cardiovasc J. 2020;54(2):108–114.
- Cleophas TJ, Droogendijk J, van Ouwerkerk BM. Validating diagnostic tests, correct and incorrect methods, new developments. Curr Clin Pharmacol. 2008;3(2):70–76.