| アブストラクト | PURPOSE: The study aimed to validate and compare coding algorithms for identifying people with migraine within the Japanese claims database. METHODS: This study used the administrative claim database provided by DeSC Healthcare, Inc., that was linked to the results of an online survey administered to adult users of the health app "kencom((R))." The ability of the 12 algorithms to detect migraines using diagnostic records alone or with prescription records was evaluated based on sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs). We used a migraine diagnosis judged based on respondents' self-reported symptoms according to the diagnostic criteria of the International Classification of Headache Disorders, version 3 (ICHD-3), as true. RESULTS: Of the 21,480 individuals, 691 had migraine according to the ICHD-3 criteria. The 12 algorithms had a sensitivity of 5.4-8.8%, specificity of 98.8-99.6%, PPVs of 19.2-32.5%, and NPVs of 96.9-97.0%. Algorithm 9 (migraine diagnostic records more than once AND at least one prescription record for migraine prophylaxis or triptans in the same month as diagnosis) produced the highest PPV, whereas Algorithm 2 (at least one diagnostic record of migraine or tension-type headache) had the highest sensitivity. Similar trends were observed when using the ID-Migraine or 4-item migraine screener, instead of the ICHD-3 criteria, for case ascertainment. CONCLUSION: Strict algorithms, such as Algorithm 9, yielded a higher PPV but a lower sensitivity, and such algorithms may be suitable for studies estimating the relative risk. Conversely, algorithms based on a single diagnostic record, such as Algorithm 2, had a higher sensitivity and may be suitable for studies estimating the prevalence/incidence of disease. Our findings will help select a desirable algorithm for migraine studies using a Japanese claim database. |
|---|
| 組織名 | Medical Affairs, Otsuka Pharmaceutical Co., Ltd., Osaka, Japan.;Dokkyo Medical University, Tochigi, Japan.;Department of Health Informatics, Graduate School of Medicine and School of;Public Health, Kyoto University, Kyoto, Japan. |
|---|
