Positive predictive value of ICD-10 diagnosis codes for brain abscess in the Danish National Patient Registry

Abstract

PurposeTo evaluate the positive predictive value (PPV) of ICD-10 diagnosis codes for brain abscess in the Danish National Patient Registry (DNPR).Patients and methodsWe examined medical records of all patients with a first-time diagnosis code of brain abscess in the DNPR from 2007 to 2016. Patients were categorized with either confirmed or disproved brain abscess using a priori defined criteria. We computed the PPV as the proportion of confirmed diagnoses.ResultsWe identified 709 patients, of whom the medical records could not be retrieved for 15 patients, leaving 694 for further analyses. Of these, 444 had a confirmed brain abscess and 250 had another proven diagnosis, including 47 cases of intracranial empyema. The overall PPV was 64% (95% CI: 60–68) ranging from 24% to 96% among the different codes evaluated. By including only patients with either 1) both a diagnosis and surgical procedure code for brain abscess or 2) patients admitted to hospital with certain primary diagnosis codes (DG060[C,E,F] or DG079B) without newly diagnosed central nervous system (CNS) cancer, spondylodiscitis/intraspinal abscess, or procedure codes for evacuation of intracranial empyema, the PPV increased to 84% (95% CI 80–87) and 89% (395/444) of all confirmed cases were identified.ConclusionThe overall PPV of diagnosis codes for brain abscess in the DNPR was moderate. However, by exclusion of newly diagnosed CNS tumors, spondylodiscitis/intraspinal abscess, and intracranial empyemas, the PPV was high and therefore suitable for future registry-based studies of brain abscess.

Hereditary haemorrhagic telangiectasia and pulmonary arteriovenous malformations in brain abscess patients: a nationwide, population-based matched cohort study

“…Although we used validated diagnosis codes with a positive predictive value of 84% [9], risk of misclassification of brain abscess patients may bias results in either way depending on the association with PAVM/HHT in misclassified or missed cases. Further limitations include lack of access to clinical information and risk of detection bias, which may overestimate the prevalence of PAVM/ HHT in patients with brain abscess compared with population controls.…”

“…A unique 10digit civil registration number allows for unambiguous linkage of healthcare and administrative registries at the individual level. The Danish National Patient Registry was accessed to identify all persons with first-time brain abscess in Denmark from 1982 through 2016 using a validated algorithm of hospital diagnosis codes [9,10]. The first date of diagnosis of brain abscess was considered the index date.…”

Incidence and mortality of brain abscess in Denmark: a nationwide population-based study

“…To accurately identify patients with a first-time diagnosis code of brain abscess in the Danish National Patient Registry, we applied a previously validated algorithm of ICD-10 codes with a positive predictive value of 84% and by which 89% of all verified cases were included ( Supplementary Table S1) [17]. Patients were considered to be diagnosed with brain abscess if they had either a diagnosis (primary or secondary diagnosis code) of brain abscess and procedure code for excision or aspiration, or had been admitted with a primary diagnosis code for brain abscess without newly diagnosed central nervous system cancer, spondylodiscitis/intraspinal abscess or a procedure code for evacuation of intracranial empyema.…”

Culture on Selective Media and Amplicon-Based Sequencing of 16S rRNA from Spontaneous Brain Abscess—the View from the Diagnostic Laboratory

“…We examined patient records during a 76-month period and identified 41 cases of aspirated spontaneous brain abscess, corresponding to an incidence rate of 0.51 per 100,000 person-years. This rate is less than the estimated nationwide rate of brain abscess of 0.78 ( 3 ), but the present study is restricted to patients subjected to surgical drainage and excludes cases with recent neurosurgery. Despite examination of aspirated material by culture, microscopic pathology, and amplicon sequencing, 3 of 41 cases (7%) were unexplained; three other cases were caused by Aspergillus or Toxoplasma , while 35 cases (85%) were bacterial infections.…”

“…Brain abscess is a rare but grave infection with a 1-year mortality of approximately 20% ( 1 , 2 ). Examination of medical records of Danish patients with first-time International Classification of Diseases 10th Revision code for brain abscess from 2007 to 2016 revealed an incidence rate of 0.78 per 100,000 person-years ( 3 ). Brain abscess formation may occur after accidental or neurosurgical head trauma, by contiguous spread from parameningeal foci, or by hematogenous spread of microorganisms from distant or cryptogenic foci ( 4 , 5 ).…”