
==== Front
Eur J Neurol
Eur J Neurol
10.1111/(ISSN)1468-1331
ENE
European Journal of Neurology
1351-5101
1468-1331
John Wiley and Sons Inc. Hoboken

39005191
10.1111/ene.16413
ENE16413
EJoN-24-0684.R1
Original Article
Stroke
Diagnostic accuracy and radiological validation of intracerebral hemorrhage diagnosis in the Swedish Stroke Register (Riksstroke)
DIAGNOSTIC ACCURACY OF ICH IN RIKSSTROKE
Sultani et al.
Sultani Gabriella 1 2
Hillal Amir 2 3
Ramgren Birgitta 2 3
Apostolaki‐Hansson Trine https://orcid.org/0000-0003-4063-5476
1 2 trine.apostolaki-hansson@med.lu.se

Norrving Bo 1 2
Wasselius Johan 2 3
Ullberg Teresa 1 2
1 Department of Neurology Skåne University Hospital Malmö Sweden
2 Department of Clinical Sciences Lund University Lund Sweden
3 Medical Imaging and Physiology Skåne University Hospital Lund Sweden
* Correspondence
Trine Apostolaki‐Hansson, Department of Neurology, Skåne University Hospital, Malmö, Sweden.
Email: trine.apostolaki-hansson@med.lu.se

15 7 2024
10 2024
31 10 10.1111/ene.v31.10 e1641322 5 2024
08 4 2024
03 7 2024
© 2024 The Author(s). European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology.
https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Abstract

Background and purpose

National quality registries for stroke care operate under the assumption that the included patients are correctly diagnosed. We aimed to validate the clinical diagnosis of spontaneous intracerebral hemorrhage (ICH) in Riksstroke (RS) by evaluating radiological data from a large, unselected ICH population.

Methods

We conducted a retrospective, multicenter study including all ICH patients registered in RS between 2016 and 2020 residing in Skåne County in Sweden (1.41 million inhabitants). Radiological data from first imaging were evaluated for the presence of spontaneous ICH. Other types of bleeds were registered if a spontaneous ICH was not identified on imaging. The radiological evaluation was independently performed by one radiology fellow and one senior neuroradiologist.

Results

Between 2016 and 2020, 1784 ICH cases were registered in RS, of which 1655 (92.8%) had a radiological diagnosis consistent with spontaneous ICH. In the 129 (7.2%) remaining cases, the radiological diagnosis was instead traumatic bleed (n = 80), subarachnoid hemorrhage (n = 15), brain tumor bleed (n = 14), ischemic lesion with hemorrhagic transformation (n = 14), ischemic lesion (n = 3), or no bleed at all (n = 3). There was a higher degree of incorrect coding in the older age groups.

Conclusion

At radiological evaluation, 92.8% of ICH diagnoses in RS were consistent with spontaneous ICH, yielding a high rate of agreement that strengthens the validity of the diagnostic accuracy in the register, justifying the use of high coverage quality register data for epidemiological purposes. The most common coding error was traumatic bleeds that were classified as spontaneous ICH.

diagnostic accuracy
intracerebral hemorrhage
radiological validation
stroke register
Crafoordska Stiftelsen 10.13039/501100003173 20200548 VINNOVA 10.13039/501100001858 2021‐02617 source-schema-version-number2.0
cover-dateOctober 2024
details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.4.8 mode:remove_FC converted:20.09.2024
Sultani G , Hillal A , Ramgren B , et al. Diagnostic accuracy and radiological validation of intracerebral hemorrhage diagnosis in the Swedish Stroke Register (Riksstroke). Eur J Neurol. 2024;31 :e16413. doi:10.1111/ene.16413
==== Body
pmcINTRODUCTION

Intracerebral hemorrhage (ICH) accounts for 13% of acute stroke cases in Sweden and up to 30% worldwide [1, 2]. The 90‐day mortality rate is approximately 40%, and only one in five patients regains functional independence [3]. Further research is needed to highlight existing inadequacies in the management of ICH and to identify strategies to address these areas of improvement.

National quality registries provide an excellent data source to monitor health care and evaluate the introduction and implementation of treatment protocols, provided that they have high coverage and accuracy [4]. Several examples exist where national register data have provided valuable information about the safety and efficacy of novel treatments regarding ischemic stroke [5, 6, 7]. Although new evidence‐based treatments have been scarce in regard to ICH, improved therapies are emerging [8]. To perform reliable evaluations of health care performance on a nationwide level, national quality registries must obtain high coverage, and key variables should be independently validated to ensure the internal and external validity of data [9]. Riksstroke (RS), the Swedish national quality register for stroke care, has high coverage (>90%) of all hospital admissions for stroke and has been extensively used for research purposes [10], but the accuracy of ICH diagnoses has yet to be independently validated.

The aim of this study was to assess the accuracy of registered ICH diagnoses in RS compared to radiological evaluation as gold standard and to classify wrongly coded patients into diagnostic categories.

METHODS

Study design

We conducted a diagnostic accuracy study based on data from RS where the coding of ICH in the register was validated against the radiological diagnosis. The article follows the standards of reporting in accordance with STARD (Standards for Reporting Diagnostic Accuracy Studies) 2015 [11]. We studied ICH patients residing in Skåne County (1.4 million inhabitants) [12], including one comprehensive and nine primary stroke centers.

Participants

Patients were included if they were aged ≥18 years, residing in Skåne County, and registered with an ICH diagnosis (International Classification of Diseases, 10th Revision [ICD‐10] I61) in RS during 2016–2020. Patients were excluded if their initial head computed tomography (CT) scan was unavailable in the picture archiving and communication system (PACS) or if the CT scan was not possible to evaluate due to substantial technical artifacts.

Spontaneous ICH was defined as ICH caused by deep perforating vasculopathy, cerebral amyloid angiopathy (CAA), vascular malformations, intracranial venous thrombosis, brain surgery, or other rare conditions [13]. Intracerebral hemorrhage due to an underlying brain tumor or metastases, as well as hemorrhagic transformation of an ischemic lesion, is by definition spontaneous ICH. However, conforming to the stroke registration guidance manual for RS [14], they were not included in this study. Hemorrhagic transformation of an ischemic lesion is not registered as ICH in RS, as the primary diagnosis is considered to be ischemic stroke [14]. For this reason, these two conditions will not be included when referring to spontaneous ICH henceforth in this study. Furthermore, by definition, spontaneous aneurysmatic/nonaneurysmatic subarachnoid hemorrhage (SAH) is not defined as an ICH and is therefore not included in this group. In cases of nonaneurysmatic SAH, CAA may have been an underlying pathology, although this was not further analyzed in patients without accompanying ICH identified on imaging. Traumatic bleeds were defined as isolated or combinations of SAH, epidural hematoma, subdural hematoma, or ICH, with a history of head trauma, presence of neck/trauma on CT, or presence of a skull fracture. Cerebral contusions were defined as punctate/large hyperdense foci/hematomas involving the grey matter and subcortical white matter. Usual locations of cerebral contusions are the anterior cranial fossa, inferior frontal lobes, and anterior–inferior temporal lobes, in addition to other locations depending on the direction of the head strike [15]. Traumatic SAH was defined as sulcal SAH occurring at the site of impact (coup) or opposite the site of impact (contrecoup).

Cases where traumatic or spontaneous ICH could not be distinguished were classified as spontaneous ICH (n = 15).

Baseline data variables

To ensure external validity, aggregated data from the entire Swedish stroke population in RS were compared with the baseline and clinical characteristics of patients residing in Skåne County. All patient characteristics were collected from RS except for oral anticoagulation treatment at the time of ICH diagnosis, which was collected from the Swedish Prescribed Drugs Register.

Data sources

Riksstroke

RS is the Swedish quality register for stroke care. All Swedish hospitals (n = 72) admitting patients with stroke contribute to data collection, resulting in high coverage (>90%) of Swedish hospital‐admitted stroke cases. The setting in which the ICD‐10 classification was applied in this study was if the patient included in the registry was treated in a hospital setting or had sought health care [10]. The register contains data on patient characteristics, provided care, and patient outcomes. The data are collected by licensed health care workers through the completion of RS's acute phase questionnaire. All patients registered in RS are informed of registration and handling of their patient data for research purposes. Approximately 2600 ICH cases are annually registered in RS [1].

Radiological database

Skåne County has a common PACS where radiological data from all hospitals in the region are gathered. By using the PACS system, access was granted to all available imaging data from hospitals providing stroke care within the region.

Radiological evaluation

Radiological data from the initial head CT scan following ICH were evaluated to verify the presence of a spontaneous ICH. If not present, other types of bleeds were categorized. The radiological evaluation was performed by one radiology fellow with 1 year of neuroradiology experience who examined and evaluated the CT scans for all included patients (n = 1794). For interrater agreement assessment, the initial 500 scans were also independently evaluated by a senior neuroradiologist with >20 years of experience. Prior to the image evaluation, both raters made consensus reading of >50 other ICH cases to ensure an even standard of assessment. Any disagreements were resolved by consensus. To distinguish between spontaneous ICH and other types of bleeds, the assessors looked for typical radiological patterns and signs of causes other than spontaneous ICH [16]. Other available imaging (e.g., CT angiography [CTA], magnetic resonance imaging [MRI], or magnetic resonance angiography [MRA]) and the radiological referral text in PACS were also used for the radiological evaluation of ICH when required. Medical records, other than the radiological referral text in PACS, were not reviewed in this study.

Statistical analysis

IBM SPSS statistics version 28 was used for all statistical analyses. Cohen kappa was used to assess interrater agreement for the presence of spontaneous ICH. The data were analyzed using standard descriptive statistics. After reviewing the radiological data, the ICH diagnosis registered in RS was defined as true positive (TP) if it was consistent with the radiological evaluation of ICH, or false positive (FP) if this was not the case. The positive predictive value (PPV = TP / [TP + FP]) was expressed as a percentage. We calculated the PPV for the total study population and stratified patients into the following age categories: <65 years, 65–74 years, 75–84 years, and ≥85 years. The proportions of patients incorrectly coded in RS were compared between the different age groups using chi‐squared test. Probability values (two‐sided) < 0.05 were considered statistically significant.

RESULTS

Participants

The inclusion criteria were fulfilled in 1794 patients. In total, 10 patients were subsequently excluded from the study due either to missing the initial head CT (n = 8) or to the presence of substantial image artifacts (n = 2), resulting in a final study population of 1784 patients.

Baseline and clinical characteristics for the study population and the entire Swedish ICH population between 2016 and 2020 are shown in Table 1. Median age for both groups was 76 years, and male sex represented approximately 55% of the total patient population. Most clinical parameters were similar between the two groups, except for admission to a stroke unit, which was 62% in Skåne County compared to 56% in the entire Swedish ICH population, and admission to an intensive care unit, which was correspondingly 13% and 21%.

TABLE 1 Baseline characteristics of patients with ICH in the Skåne County population and in the entire Swedish ICH population during the study period (2016–2020).

Baseline characteristics	Skåne County population, n = 1786, n (%)	Swedish population, n = 13,980, n (%)	
Age, years, median (IQR)	76 (67–84)	76 (66–83)	
Gender, male	1002 (56.1)	7724 (55.3)	
Hypertension	1058 (59.2)	8190 (58.6)	
Diabetes	334 (18.7)	2422 (17.3)	
Previous stroke	439 (24.6)	3158 (22.6)	
Level of consciousness	
Alert	986 (55.2)	8446 (60.4)	
Drowsy	448 (25.1)	3135 (22.4)	
Comatose	336 (18.8)	2399 (17.2)	
OAC at onset	398 (22.3)	2943 (21.1)	
Admitted to regular care unit	185 (10.4)	1717 (12.3)	
Admitted to stroke unit	1109 (62.1)	7825 (56.0)	
Admitted to intensive care unit	239 (13.4)	2960 (21.2)	
Admitted to neurosurgical ward	45 (2.5)	412 (2.9)	
Note: The proportion of missing data varied between 0% and 1.1% for all variables.

Abbreviations: ICH, intracerebral hemorrhage; IQR, interquartile range; OAC, oral anticoagulation.

Radiological evaluation

In total, 1784 ICH cases were evaluated, of which 1655 cases were radiologically consistent with the diagnosis of spontaneous ICH. In all cases, first imaging was conducted using CT scans; none of the initial scans was performed using MRI. This resulted in a PPV of 92.8%. For the remaining 129 cases (7.2%), the radiological assessment was consistent with traumatic bleed (n = 80), SAH (n = 15), tumor bleed (n = 14), ischemic lesion with hemorrhagic transformation (n = 14), ischemic lesion (n = 3), or no bleed at all (n = 3; Figure 1). Disagreement between the two independent raters occurred in four of the 500 cases, yielding an interrater agreement of 99.2% (kappa = 0.96). All cases of disagreement were subsequently resolved by consensus.

FIGURE 1 Donut chart demonstrating the proportions of patients incorrectly coded in Riksstroke (RS), as well as a specification of the different diagnoses and their proportions in this group.

Diagnoses after radiological evaluation for the total study population, in addition to diagnosis stratification by age, are shown in Table 2. The proportion of patients incorrectly coded in RS as having a spontaneous ICH was 12.1% in patients aged ≥85 years compared to 5.0%–6.6% in younger age groups (p < 0.001; Figure 2). Specifically, traumatic bleeds that were incorrectly diagnosed as spontaneous ICH were more prevalent in patients aged ≥85 years. Regarding tumor bleeds, the tumor was known and MRI‐verified prior to the ICH diagnosis in 13 of the 14 cases that were wrongly classified as spontaneous ICH in RS.

TABLE 2 Table of diagnoses upon radiological evaluation in the total study population and in different age groups within the study population.

Diagnosis	All ages, n = 1784, n (%)	<65 years, n = 378, n (%)	65–74 years, n = 424, n (%)	75–84 years, n = 584, n (%)	≥85 years, n = 398, n (%)	
Spontaneous ICH	1655 (92.8)	353 (93.4)	403 (95.0)	549 (94.0)	350 (87.9)	
Traumatic bleed	80 (4.5)	8 (2.1)	12 (2.8)	24 (4.1)	36 (9.0)	
Subarachnoid hemorrhage	15 (0.8)	7 (1.9)	1 (0.2)	2 (0.3)	5 (1.3)	
Tumor bleed	14 (0.8)	5 (1.3)	4 (0.9)	3 (0.5)	2 (0.5)	
Ischemic lesion with hemorrhagic transformation	14 (0.8)	3 (0.8)	3 (0.7)	5 (0.9)	3 (0.8)	
Ischemic lesion	3 (0.2)	1 (0.3)	1 (0.2)	0 (0.0)	1 (0.3)	
No bleeding	3 (0.2)	1 (0.3)	0 (0.0)	1 (0.2)	1 (0.3)	
Abbreviation: ICH, intracerebral hemorrhage.

FIGURE 2 Pie charts demonstrating the proportions of incorrectly coded cases in Riksstroke (RS) in the entire study population as well as for the four different age categories.

DISCUSSION

In our study population of 1784 individuals, the assessment of radiological data was consistent with RSs diagnosis of spontaneous ICH in 92.8% of cases, yielding a PPV of 93%.

Previous studies on the validation of ICH diagnoses using administrative or register data have reported PPVs ranging from 66% to 97% [17, 18, 19, 20, 21, 22], although these studies were largely based on the review of medical records and not on the assessment of radiological images. Several of these studies were not solely focused on ICH, but on the validation of stroke diagnoses in general, including ischemic stroke and/or SAH [18, 19, 20, 21, 22]. In several of the studies [17, 18, 20, 21, 22], the number of ICH cases included was less than in ours. The study conducted by Gaist et al. was the only abovementioned study that included a comparable number of ICH cases, yet they reported a PPV of 73% for the diagnosis of ICH in the Health Improvement Network, a health care database from the United Kingdom [19]. A Norwegian study conducted by Oie et al. reported a PPV of 96.9% for the primary diagnosis of ICH in inpatient admissions registered in the Norwegian Patient Register (NPR) using information provided by electronic medical records (diagnostic imaging, radiology reports, and laboratory tests) [17]. As only radiology reports were screened, the determination of the consistency of the diagnosis of ICH in the NPR with the assessment of radiological data was not performed. Thus, certain patients may have been falsely coded due to misdiagnosis.

In our population, we identified that traumatic bleeds constituted more than half of all misdiagnosed cases and that faulty coding was more common in the elderly (≥85 years old). The finding that older patients have a higher rate of traumatic bleeds misdiagnosed/miscoded as spontaneous ICH may potentially be explained by the higher prevalence of traumatic bleeds and intracranial hematomas in older age groups. Another possible explanation is the finding that elderly stroke patients more often receive care outside neurological wards [23], where physicians may be less accustomed to diagnostic coding of ICH. Traumatic bleeds may be incorrectly coded as spontaneous ICH in RS, because the ICD‐10 code for traumatic ICH (S06.3) reads “focal brain injury” and not “traumatic ICH.” This could presumably pose a challenge for physicians to identify the correct diagnostic code, and they may instead apply the code for spontaneous unspecified ICH (I61.9), a code that should only be used in cases of spontaneous ICH [24]. With the additional training of health care personnel in diagnostic coding, there is potential to improve the diagnostic accuracy of stroke, and thus further improve the validity of the quality register.

Differentiating between traumatic versus nontraumatic ICH was ambiguous in 15 patients in whom a history of head trauma was reported but CT revealed only an isolated ICH and no imaging findings of traumatic injuries. In these instances, patients were diagnosed with spontaneous ICH. It was unclear and indistinguishable based on the CT findings alone whether these patients first had an ICH followed by a traumatic fall or vice versa. Furthermore, tumor bleeds constituted only a small minority of cases (<1%). Although a small number overall, the proportion of previously known tumors in patients with tumor bleeds was high (13/14 cases). Because the stroke registration guidance manual for RS clearly states that tumor bleeds should not be registered in RS, we expected the number of previously known tumors to be smaller.

The number of patients diagnosed with ICH due to vascular malformations, such as arteriovenous malformations or cavernomas, was lower than expected. One potential reason for this is that only a minority of patients in our study population underwent radiological examination with CTA and/or MRI/MRA. The increased use of these imaging modalities in the diagnostic workup of ICH patients would improve the causal classification of ICH. The causal classification of ICH is of importance because the management, recurrence risk, and prognosis may differ based on the underlying etiology [25]. Because CTA and/or MRI are not extensively used in the diagnostic workup of ICH in Sweden, the underlying etiology of spontaneous ICH may be unknown in a number of patients included in this study. Due to the potential absence of an adequately comprehensive diagnostic workup to identify an underlying cause of ICH, there is a small possibility that misclassification of spontaneous ICH may have occurred if, for example, an underlying tumor was not initially identified on CT. Nonetheless, we presume that this number is uninfluential to our results, because patients are to be included in RS regardless of the exact underlying cause of spontaneous ICH (except for tumor bleeds and the hemorrhagic transformation of an ischemic lesion). The annual RS register report has shown an upward trend when it comes to diagnostic workup with MRI and CTA in ICH patients. According to the annual report from 2020, the proportion of ICH patients examined with CTA and/or MRI was 32% and 16%, respectively. The proportion of patients undergoing these examinations was higher in the younger population, correspondingly 55% and 37% in patients <55 years old [26]. Compared to other national cohorts, the use of MRI following ICH during the acute phase is low in Sweden. In a patient cohort of 3572 patients registered in the Swiss Stroke Register, approximately 40% of patients underwent MRI within 90 days of ICH [27]. RS does not include data on MRI investigations performed after patient discharge. Therefore, MRI use in Sweden is probably underestimated in this study.

Establishing national‐ and regional‐level systems for assessing stroke clinical services, providing peer support for quality improvement, and making audit data routinely available to the general public are among the goals of the Stroke Action Plan for Europe [28]. However, national quality registers are currently available in fewer than half of the European countries [29]. To provide reliable data, it is essential that definitions and terminology are standardized and that data for main stroke subtypes are validated separately.

Strengths

The strengths of this study include the large sample size based on an unselected patient cohort. The patient cohort represents a large Swedish region, encompassing approximately 13% of the total Swedish population, which contributes to the generalizability of this study [30]. Second, the number of missing CT examinations was low, and the coverage of stroke cases in RS during the study period was high. The study covers 10 hospitals ranging in size from university hospitals to county and local hospitals. Therefore, in addition to the large number of patients, the study population can be considered representative of the entire Swedish ICH population, which is also seen in Table 1, as the baseline and clinical characteristics of the two populations are similar.

Limitations

As this study validated the diagnosis of ICH registered in the Swedish Stroke Register, RS, results from this study may not be generalizable to other stroke registers. Selection bias is a second limitation of our study, because any ICH patient not included in RS is also missed in our analysis. In a recent study on case ascertainment in Swedish hospital‐based stroke registries, including RS and the Lund Stroke Register, by Aked et al. [31], it was demonstrated that, despite best efforts, there would always be stroke cases that are missed. Fifteen percent of the patient population in the aforementioned study had ICH; the remaining had ischemic stroke (84%) or underdetermined stroke cause (1%). In the small ICH population included in this study, case ascertainment in RS was 83% (n = 50/60). In the total patient population, patients who were not identified by RS more often had less severe strokes or a higher early case fatality. As the ICH population in this study was small, the generalizability of these results to the entire country is low. Thus, even though the number of missing ICH cases in RS is estimated to be low given the high coverage of RS [1], it is not possible to estimate the number of false negative cases, that is, patients with ICH who were not registered in RS, in this study.

CONCLUSIONS

In conclusion, our study shows a high rate of agreement between ICH diagnoses registered in RS and the radiological evaluation. Our results strengthen the validity of the diagnosis of ICH in RS and support the use of the register for research and quality evaluation. Due to its high coverage and high validity, RS is a register that can be appropriately used for epidemiological studies. Although our results confirm a high diagnostic accuracy, they also identify areas for improvement, such as the coding of traumatic bleeds and bleeds in patients with underlying brain tumors. The current trend seen in the increased use of CTA and/or MRI may further improve the causal classification of ICH.

AUTHOR CONTRIBUTIONS

Gabriella Sultani: Conceptualization; writing – original draft; formal analysis; writing – review and editing. Amir Hillal: Data curation. Birgitta Ramgren: Data curation; writing – review and editing. Trine Apostolaki‐Hansson: Writing – review and editing; supervision. Bo Norrving: Conceptualization; supervision; writing – review and editing. Johan Wasselius: Conceptualization; supervision; writing – review and editing. Teresa Ullberg: Conceptualization; supervision; writing – review and editing.

FUNDING INFORMATION

This study was funded by grants from the Crafoord Foundation (20200548), VINNOVA (2021‐02617), SUS Stiftelser & Fonder, Södra Sjukvårdsregionens doktorandanslag, and ALF Region Skåne.

CONFLICT OF INTEREST STATEMENT

J.W. is a founder and shareholder in Uman Sense. T.U. has received honoraria for an expert group assignment from AstraZeneca. B.N. has received honoraria for serving on a data and safety monitoring board for the HOVID trial (Simbec‐Orion). The other authors have no conflicts of interest to disclose.

ETHICS STATEMENT

The Swedish Ethical Review Authority approved the study (#2020–06800) and waived individual informed consent.

ACKNOWLEDGMENTS

We would like to acknowledge Fredrik Jönsson at Riksstroke for preparing the dataset.

DATA AVAILABILITY STATEMENT

An anonymized dataset supporting the conclusions of this article may be provided upon reasonable request.
==== Refs
REFERENCES

1 Riksstroke . The Swedish Stroke Register. Årsrapport Stroke TIA 2022. Accessed November 29, 2023. https://www.riksstroke.org/wp‐content/uploads/2023/12/Riksstroke_Arsrapport_2022.pdf
2 van Asch CJ , Luitse MJ , Rinkel GJ , van der Tweel I , Algra A , Klijn CJ . Incidence, case fatality, and functional outcome of intracerebral haemorrhage over time, according to age, sex, and ethnic origin: a systematic review and meta‐analysis. Lancet Neurol. 2010;9 (2 ):167‐176. doi:10.1016/S1474-4422(09)70340-0 20056489
3 Sennfalt S , Norrving B , Petersson J , Ullberg T . Long‐term survival and function after stroke: a longitudinal observational study from the Swedish stroke register. Stroke. 2019;50 (1 ):53‐61. doi:10.1161/STROKEAHA.118.022913 30580719
4 Yu AYX , Bravata DM , Norrving B , Reeves MJ , Liu L , Kilkenny MF . Measuring stroke quality: methodological considerations in selecting, defining, and analyzing quality measures. Stroke. 2022;53 (10 ):3214‐3221. doi:10.1161/STROKEAHA.122.036485 35876016
5 Ullberg T , von Euler M , Wasselius J , Wester P , Arnberg F . Survival and functional outcome following endovascular thrombectomy for anterior circulation acute ischemic stroke caused by large vessel occlusion in Sweden 2017–2019‐a nationwide, prospective, observational study. Interv Neuroradiol. 2023;29 (1 ):94‐101. doi:10.1177/15910199211073019 35044270
6 Ullberg T , von Euler M , Wester P , et al. Endovascular thrombectomy for anterior circulation stroke beyond 6 hours of onset in Sweden 2015 to 2020: rates and outcomes in a nationwide register‐based study. J Neurointerv Surg. 2023;15 (4 ):330‐335. doi:10.1136/neurintsurg-2022-018670 35301261
7 Hall E , Ullberg T , Andsberg G , Wasselius J . Incidence of intracranial hemorrhagic complications after anterior circulation endovascular thrombectomy in relation to occlusion site: a nationwide observational register study. J Neurointerv Surg. 2023:jnis‐2023‐020768. doi:10.1136/jnis-2023-020768
8 Anderson CS . Stroke advances in 2023: a new horizon for the management of intracerebral haemorrhage. Lancet Neurol. 2024;23 (1 ):12‐13. doi:10.1016/S1474-4422(23)00451-9 38101883
9 Rannikmae K , Ngoh K , Bush K , et al. Accuracy of identifying incident stroke cases from linked health care data in UK biobank. Neurology. 2020;95 (6 ):e697‐e707. doi:10.1212/WNL.0000000000009924 32616677
10 Riksstroke . The Swedish Stroke Register. Accessed December 7, 2023. https://www.riksstroke.org/sve/
11 Korevaar DA , Cohen JF , Reitsma JB , et al. Updating standards for reporting diagnostic accuracy: the development of STARD 2015. Res Integr Peer Rev. 2016;1 :7. doi:10.1186/s41073-016-0014-7 29451535
12 The Swedish Statistics Agency . Population of Sweden, Swedish Statistics Agency. Accessed March 22, 2023. https://www.scb.se/hitta‐statistik/sverige‐i‐siffror/manniskorna‐i‐sverige/sveriges‐befolkning/#lan_och_kommuner
13 Raposo N , Zanon Zotin MC , Seiffge DJ , et al. A causal classification system for intracerebral hemorrhage subtypes. Ann Neurol. 2023;93 (1 ):16‐28. doi:10.1002/ana.26519 36197294
14 Riksstroke . The Swedish Stroke Register. “Vägledning för Stroke registrering i Riksstroke, 2023”. Accessed October 13, 2023. https://www.riksstroke.org/wp‐content/uploads/2023/01/Vagledning_stroke_‐2023‐1.pdf
15 Parizel PM , Philips CD . Traumatic Neuroemergency: imaging patients with traumatic brain injury—an introduction. In: Hodler J , Kubik‐Huch RA , von Schulthess GK , eds. Diseases of the Brain, Head and Neck, Spine 2020–2023: Diagnostic Imaging. Springer International Publishing; 2020:77‐92.
16 Hillal A , Ullberg T , Ramgren B , Wasselius J . Computed tomography in acute intracerebral hemorrhage: neuroimaging predictors of hematoma expansion and outcome. Insights Imaging. 2022;13 (1 ):180. doi:10.1186/s13244-022-01309-1 36417131
17 Oie LR , Madsbu MA , Giannadakis C , et al. Validation of intracranial hemorrhage in the Norwegian patient registry. Brain Behav. 2018;8 (2 ):e00900. doi:10.1002/brb3.900 29484261
18 Sedova P , Brown RD Jr , Zvolsky M , et al. Validation of stroke diagnosis in the National Registry of hospitalized patients in The Czech Republic. J Stroke Cerebrovasc Dis. 2015;24 (9 ):2032‐2038. doi:10.1016/j.jstrokecerebrovasdis.2015.04.019 26139454
19 Gaist D , Wallander MA , Gonzalez‐Perez A , Garcia‐Rodriguez LA . Incidence of hemorrhagic stroke in the general population: validation of data from the health improvement network. Pharmacoepidemiol Drug Saf. 2013;22 (2 ):176‐182. doi:10.1002/pds.3391 23229888
20 Tirschwell DL , Longstreth WT Jr . Validating administrative data in stroke research. Stroke. 2002;33 (10 ):2465‐2470. doi:10.1161/01.str.0000032240.28636.bd 12364739
21 Johnsen SP , Overvad K , Sorensen HT , Tjonneland A , Husted SE . Predictive value of stroke and transient ischemic attack discharge diagnoses in the Danish National Registry of patients. J Clin Epidemiol. 2002;55 (6 ):602‐607. doi:10.1016/s0895-4356(02)00391-8 12063102
22 Tolonen H , Salomaa V , Torppa J , et al. The validation of the Finnish hospital discharge register and causes of death register data on stroke diagnoses. Eur J Cardiovasc Prev Rehabil. 2007;14 (3 ):380‐385. doi:10.1097/01.hjr.0000239466.26132.f2 17568236
23 Weber R , Krogias C , Eyding J , et al. Age and sex differences in ischemic stroke treatment in a Nationwide analysis of 1.11 million hospitalized cases. Stroke. 2019;50 (12 ):3494‐3502. doi:10.1161/STROKEAHA.119.026723 31623547
24 ICD‐10 Version:2019 . International Statistical Classification of Diseases and Related Health Problems 10th Revision. Accessed October 13, 2023. https://icd.who.int/browse10/2019/en/
25 Fandler‐Hofler S , Obergottsberger L , Ambler G , et al. Association of the presence and pattern of MRI markers of cerebral small vessel disease with recurrent intracerebral hemorrhage. Neurology. 2023;101 (8 ):e794‐e804. doi:10.1212/WNL.0000000000207510 37349111
26 Riksstroke . The Swedish Stroke Register “Årsrapport Stroke TIA 2020”. Accessed December 1, 2023. https://www.riksstroke.org/wp‐content/uploads/2021/12/Riksstroke_Arsrapport_2020.pdf
27 Goeldlin MB , Mueller M , Siepen BM , et al. CADMUS: a novel MRI‐based classification of spontaneous intracerebral hemorrhage associated with cerebral small vessel disease. Neurology. 2024;102 (1 ):e207977. doi:10.1212/WNL.0000000000207977 38165372
28 Norrving B , Barrick J , Davalos A , et al. Action plan for stroke in Europe 2018‐2030. Eur Stroke J. 2018;3 (4 ):309‐336. doi:10.1177/2396987318808719 31236480
29 (SAP‐E) SAPfE . Key performance indicators. 2024. Accessed May 21, 2024. https://actionplan.eso‐stroke.org/kpi/kpi‐4
30 The Swedish Statistics Agency . Population Statistics, Swedish Statistics Agency. Accessed September 29, 2023. https://www.scb.se/hitta‐statistik/statistik‐efter‐amne/befolkning/befolkningens‐sammansattning/befolkningsstatistik/#_Tabellerochdiagram
31 Aked J , Delavaran H , Norrving B , Lindgren A . Completeness of case ascertainment in Swedish hospital‐based stroke registers. Acta Neurol Scand. 2020;141 (2 ):148‐155.31664726
