Mariëtte Hooiveld
Publicatie
Datum
10-07-2025
Surveillance of severe acute respiratory infections associated with SARS-CoV-2, influenza virus and RSV using ICD-10 codes: a case definition accuracy study across five European countries, 2021 to 2023.
Sanchez Ruiz, M.A., Marques, D.F., Lomholt, F.K., Vestergaard, L.S., Monge, S., Lozano Álvarez, M., Aspelund, G., Thordardottir, M., Dziugyte, A., Cauchi, J.P., Boere, T.M., Veldhuijzen, I.K., Seppälä, E., Bøås, H., Paulsen, T.H., Machado, A., Rodrigues, A.P., Hooiveld, M., 13, Alves de Sousa, L., Torres, A., Carvalho, C. Nunes. B. Surveillance of severe acute respiratory infections associated with SARS-CoV-2, influenza virus and RSV using ICD-10 codes: a case definition accuracy study across five European countries, 2021 to 2023. Eurosurveillance: 2025. 30(27), art. nr. 2400748.
Lees online
Background
Surveillance of severe acute respiratory infections (SARI) using ICD-10 codes from electronic health records (EHR) lacks consensus on optimal case-defining codes.
Aim
We determined codes that maximise sensitivity (Se) and positive predictive value (PPV) for SARI associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus and respiratory syncytial virus (RSV) in Denmark, Iceland, Malta, Norway and Spain.
Methods
We included hospitalisations from week 21/2021 to 39/2023, with ICD-10 diagnostic codes for respiratory disease (three-character codes J00–J99) or COVID-19 (U07.1, U07.2, country-specific codes for Denmark). We assessed Se and PPV of individual codes against laboratory results. Based on Se and PPV rank-sum, we selected the top 10 codes and combined them into 10 sets per pathogen. We identified sets that maximised the clinical utility index (CUI = Se × PPV), categorised as excellent (≥ 0.81), good (0.64–0.80), satisfactory (0.49–0.63) and poor (< 0.49).
Results
We assessed 395,163 hospitalisations for SARI-SARS-CoV-2, 313,418 for SARI-influenza and 192,936 for SARI-RSV, all tested. For SARI-SARS-CoV-2, code U07.1 (B34.2A, B97.2A for Denmark) had excellent utility in Denmark, Malta, Norway, Spain (≥ 0.82), and good utility in Iceland (0.79). For SARI-influenza, J09, J10 and J11 performed excellently in Denmark, Norway, Spain (≥ 0.83), satisfactorily in Malta (0.52), and poorly in Iceland (0.43). For SARI-RSV, J12, J20 and J21 achieved highest CUI but had poor utility (0.17–0.34).
Conclusions
COVID-19- and influenza-specific three-character ICD-10 codes accurately identified SARI associated with SARS-CoV-2 and influenza virus. For SARI-RSV, four-character codes should be explored. We recommend context-specific assessments in countries adopting EHR-based surveillance.
Surveillance of severe acute respiratory infections (SARI) using ICD-10 codes from electronic health records (EHR) lacks consensus on optimal case-defining codes.
Aim
We determined codes that maximise sensitivity (Se) and positive predictive value (PPV) for SARI associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus and respiratory syncytial virus (RSV) in Denmark, Iceland, Malta, Norway and Spain.
Methods
We included hospitalisations from week 21/2021 to 39/2023, with ICD-10 diagnostic codes for respiratory disease (three-character codes J00–J99) or COVID-19 (U07.1, U07.2, country-specific codes for Denmark). We assessed Se and PPV of individual codes against laboratory results. Based on Se and PPV rank-sum, we selected the top 10 codes and combined them into 10 sets per pathogen. We identified sets that maximised the clinical utility index (CUI = Se × PPV), categorised as excellent (≥ 0.81), good (0.64–0.80), satisfactory (0.49–0.63) and poor (< 0.49).
Results
We assessed 395,163 hospitalisations for SARI-SARS-CoV-2, 313,418 for SARI-influenza and 192,936 for SARI-RSV, all tested. For SARI-SARS-CoV-2, code U07.1 (B34.2A, B97.2A for Denmark) had excellent utility in Denmark, Malta, Norway, Spain (≥ 0.82), and good utility in Iceland (0.79). For SARI-influenza, J09, J10 and J11 performed excellently in Denmark, Norway, Spain (≥ 0.83), satisfactorily in Malta (0.52), and poorly in Iceland (0.43). For SARI-RSV, J12, J20 and J21 achieved highest CUI but had poor utility (0.17–0.34).
Conclusions
COVID-19- and influenza-specific three-character ICD-10 codes accurately identified SARI associated with SARS-CoV-2 and influenza virus. For SARI-RSV, four-character codes should be explored. We recommend context-specific assessments in countries adopting EHR-based surveillance.
Background
Surveillance of severe acute respiratory infections (SARI) using ICD-10 codes from electronic health records (EHR) lacks consensus on optimal case-defining codes.
Aim
We determined codes that maximise sensitivity (Se) and positive predictive value (PPV) for SARI associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus and respiratory syncytial virus (RSV) in Denmark, Iceland, Malta, Norway and Spain.
Methods
We included hospitalisations from week 21/2021 to 39/2023, with ICD-10 diagnostic codes for respiratory disease (three-character codes J00–J99) or COVID-19 (U07.1, U07.2, country-specific codes for Denmark). We assessed Se and PPV of individual codes against laboratory results. Based on Se and PPV rank-sum, we selected the top 10 codes and combined them into 10 sets per pathogen. We identified sets that maximised the clinical utility index (CUI = Se × PPV), categorised as excellent (≥ 0.81), good (0.64–0.80), satisfactory (0.49–0.63) and poor (< 0.49).
Results
We assessed 395,163 hospitalisations for SARI-SARS-CoV-2, 313,418 for SARI-influenza and 192,936 for SARI-RSV, all tested. For SARI-SARS-CoV-2, code U07.1 (B34.2A, B97.2A for Denmark) had excellent utility in Denmark, Malta, Norway, Spain (≥ 0.82), and good utility in Iceland (0.79). For SARI-influenza, J09, J10 and J11 performed excellently in Denmark, Norway, Spain (≥ 0.83), satisfactorily in Malta (0.52), and poorly in Iceland (0.43). For SARI-RSV, J12, J20 and J21 achieved highest CUI but had poor utility (0.17–0.34).
Conclusions
COVID-19- and influenza-specific three-character ICD-10 codes accurately identified SARI associated with SARS-CoV-2 and influenza virus. For SARI-RSV, four-character codes should be explored. We recommend context-specific assessments in countries adopting EHR-based surveillance.
Surveillance of severe acute respiratory infections (SARI) using ICD-10 codes from electronic health records (EHR) lacks consensus on optimal case-defining codes.
Aim
We determined codes that maximise sensitivity (Se) and positive predictive value (PPV) for SARI associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), influenza virus and respiratory syncytial virus (RSV) in Denmark, Iceland, Malta, Norway and Spain.
Methods
We included hospitalisations from week 21/2021 to 39/2023, with ICD-10 diagnostic codes for respiratory disease (three-character codes J00–J99) or COVID-19 (U07.1, U07.2, country-specific codes for Denmark). We assessed Se and PPV of individual codes against laboratory results. Based on Se and PPV rank-sum, we selected the top 10 codes and combined them into 10 sets per pathogen. We identified sets that maximised the clinical utility index (CUI = Se × PPV), categorised as excellent (≥ 0.81), good (0.64–0.80), satisfactory (0.49–0.63) and poor (< 0.49).
Results
We assessed 395,163 hospitalisations for SARI-SARS-CoV-2, 313,418 for SARI-influenza and 192,936 for SARI-RSV, all tested. For SARI-SARS-CoV-2, code U07.1 (B34.2A, B97.2A for Denmark) had excellent utility in Denmark, Malta, Norway, Spain (≥ 0.82), and good utility in Iceland (0.79). For SARI-influenza, J09, J10 and J11 performed excellently in Denmark, Norway, Spain (≥ 0.83), satisfactorily in Malta (0.52), and poorly in Iceland (0.43). For SARI-RSV, J12, J20 and J21 achieved highest CUI but had poor utility (0.17–0.34).
Conclusions
COVID-19- and influenza-specific three-character ICD-10 codes accurately identified SARI associated with SARS-CoV-2 and influenza virus. For SARI-RSV, four-character codes should be explored. We recommend context-specific assessments in countries adopting EHR-based surveillance.
Gegevensverzameling