Usefulness of the ID-Migraine Screening Tool for Diagnosing Migraines in Patients With Headache and Facial Pain: A Systematic Review and Meta-Analysis
Article information
Abstract
Background and Objectives
A substantial proportion of patients presenting with sinus headaches are ultimately diagnosed with migraines. The ID-Migraine tool has been developed to aid in identifying migraine patients in primary care settings. Therefore, we conducted a meta-analysis of validation studies evaluating the diagnostic accuracy of the ID-Migraine tool.
Methods
Data were extracted regarding true positives, false positives, true negatives, and false negatives, along with study characteristics. Methodological quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool.
Results
In total, 30 studies involving 9,942 participants were included in the final analysis. The pooled diagnostic odds ratio for the ID-Migraine tool was 13.84. The area under the summary receiver operating characteristic curve was 0.846. The pooled sensitivity and specificity were 0.865 and 0.702, respectively, while the negative predictive value was 0.791 and the positive predictive value was 0.799.
Conclusion
The ID-Migraine screening tool is highly accurate and user-friendly for identifying migraines among patients presenting with sinus headaches. By employing this tool, unnecessary diagnostic tests or treatments can be minimized, and patients can be promptly referred to appropriate specialists.
INTRODUCTION
Patients experiencing various facial or headache syndromes accompanied by rhinological symptoms—commonly termed “sinus headache”—frequently consult otolaryngologists, either through self-referral or referrals from other healthcare providers. However, many of these patients are ultimately diagnosed with migraines [1]. Otolaryngologists should be prepared to evaluate and manage these cases, particularly in determining whether facial pain originates from rhinogenic or migrainous causes [2,3]. Additionally, the global age-standardized prevalence of migraines increased by 1.7% between 1990 and 2019, with current prevalence estimates ranging from 8% to 22%. Approximately 1.1 billion people worldwide are estimated to suffer from migraines [4]. Given this significant prevalence, migraines are frequently encountered in clinical practice, making it critical for otolaryngologists to actively participate in managing patients presenting with migraines that include rhinologic symptoms.
The ID-Migraine screening tool was developed to simplify the identification of migraine patients in primary care environments [5]. This tool has been validated across various clinical contexts [6].
We hypothesize that this tool may serve as a practical and efficient initial screening instrument for distinguishing sinogenic headaches from migraines. Therefore, we conducted a meta-analysis of validation studies examining the ID-Migraine tool to assess its diagnostic accuracy specifically for patients with migraines.
METHODS
Study protocol and registration
This systematic review and meta-analysis were performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [7]. The study protocol was pre-registered on the Open Science Framework (https://osf.io/3gsy5/).
Literature search strategy
A comprehensive, systematic literature search was conducted in several electronic databases, including PubMed, Scopus, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials, to identify relevant clinical studies published up to January 2025. The search strategy combined keywords and controlled vocabulary terms related to migraine diagnosis, including “migraine disorders,” “ID-Migraine,” “sensitivity and specificity,” as well as various forms of the terms “diagnose,” “diagnosis,” “diagnostic,” and “differential diagnosis.” A complete list of the specific search terms and queries is provided in Supplementary Table 1 (in the online-only Data Supplement).
To maximize comprehensiveness, the reference lists of all included studies were meticulously examined to identify additional relevant articles potentially missed in the initial search. Two reviewers (DHK and SHH) independently screened the titles, abstracts, and full texts of retrieved studies to exclude those not directly related to the use of the diagnostic tool for migraines. In cases of disagreement, discrepancies were resolved through discussion and consultation with a third reviewer (DWJ), ensuring accuracy and consistency in study selection.
Selection criteria
Studies meeting the following predefined eligibility criteria were included: 1) studies involving patients experiencing headache symptoms; 2) cohort studies or randomized controlled trials; 3) studies published in English; and 4) studies that directly compared the performance of the ID-Migraine tool against a confirmed migraine diagnosis established according to the International Classification of Headache Disorders (ICHD-II or III) criteria [8,9]. These criteria ensured that only relevant, high-quality studies were analyzed.
Studies were excluded if they were review articles, case reports, or commentaries; focused on headache types other than migraines, such as cluster headaches, tension-type headaches, or secondary headache disorders; or lacked sufficient data necessary for meaningful analysis. Additionally, studies with ambiguous diagnostic criteria or without clearly defined methods for confirming migraines were excluded.
A detailed overview of the study selection process, including the number of records screened, excluded, and included in the final analysis, is visually summarized in Fig. 1.
Data organization and risk of bias assessment
Data from selected studies were systematically extracted using a standardized data collection form to ensure accuracy and consistency [10,11]. The primary outcomes analyzed included the diagnostic odds ratio (DOR), summary receiver operating characteristic (sROC) curve, and the area under the curve (AUC). Together, these provided a comprehensive evaluation of the diagnostic performance of the ID-Migraine tool [5,12–40].
The DOR, a single indicator of test effectiveness, was calculated using the formula: (true positives [TPs]/false positives [FPs]) divided by (false negatives [FNs]/true negatives [TNs]). This metric quantifies the odds of a positive test result in migraine patients relative to the odds in individuals without migraines. Random-effects models were employed to compute DOR values with 95% confidence intervals (CIs), accounting for both within-study variability and heterogeneity between studies. DOR values theoretically range from 0 to infinity, with higher values indicating better diagnostic accuracy; a value of 1 indicates no discriminatory power.
The sROC curve was utilized to summarize the balance between sensitivity and specificity, providing a visual representation of the diagnostic accuracy. Curves positioned closer to the top-left corner of the ROC space indicate superior diagnostic performance, reflecting high sensitivity and specificity. To further quantify diagnostic accuracy, the AUC was calculated, with values ranging from 0 to 1. The AUC scores were interpreted as follows: ≥0.90 indicated excellent accuracy, 0.80–0.89 good accuracy, 0.70–0.79 fair accuracy, 0.60–0.69 poor diagnostic value, and <0.60 insufficient accuracy for clinical application [41].
In addition to these statistical analyses, data on the total number of patients in each study, along with TP, TN, FP, and FN values, were extracted to facilitate the calculation of AUC and DOR. Furthermore, the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool was employed to evaluate methodological quality and risk of bias in the included studies, ensuring a rigorous assessment of study reliability and potential limitations [42].
Statistical analysis and outcome measurements
The meta-analysis was performed using R statistical software (version 4.4.2; R Foundation for Statistical Computing), providing a robust and reproducible analytical approach. Statistical computations were conducted using specialized meta-analysis packages within R, enabling comprehensive evaluation of diagnostic performance across the studies.
To assess heterogeneity among studies, the Q statistic was applied, with a corresponding p-value to determine whether variability across studies was due to true differences in effect sizes or random sampling error. Additionally, the I2 statistic was calculated to quantify the percentage of total variation attributable to heterogeneity rather than chance, with higher values indicating greater inconsistency between studies.
Subgroup analyses were performed to explore potential sources of heterogeneity, focusing particularly on variations across imaging modalities and other study characteristics potentially influencing diagnostic performance. These subgroup analyses assessed whether specific methodological factors contributed to variations in sensitivity and specificity estimates.
Forest plots were generated to visually represent key diagnostic parameters, including sensitivity, specificity, and the sROC curve. These graphical summaries aided interpretation of findings by displaying individual and pooled estimates of diagnostic performance, along with their respective confidence intervals.
To evaluate potential publication bias, Deeks’ funnel plot asymmetry test was utilized. This statistical method evaluates whether smaller studies with less favorable results are underrepresented, thus ensuring the reliability of synthesized findings. If significant asymmetry was detected in the funnel plot, additional sensitivity analyses were performed to explore its potential impact on overall conclusions.
RESULTS
Search and study selection
In total, 30 studies involving 9,942 participants were included in this analysis. The key characteristics of these studies are summarized in Table 1, and their bias assessments are provided in Supplementary Table 2 (in the online-only Data Supplement). The p-value obtained from Deeks’ funnel plot asymmetry test was 0.0835, indicating no significant evidence of publication bias (Fig. 2).
Diagnostic accuracy of different cutoff values in ID-Migraine
The pooled DOR for the ID-Migraine tool was 13.84 (95% CI, 10.21–18.76; I2=89.7%) (Fig. 3). The pooled area under the sROC curve was 0.846 (Fig. 4). The pooled sensitivity was 0.865 (95% CI, 0.821–0.900; I2=94.6%), specificity was 0.702 (95% CI, 0.628–0.766; I2=90.8%), the positive predictive value (PPV) was 0.799 (95% CI, 0.738–0.835; I2=92.7%), and the negative predictive value (NPV) was 0.791 (95% CI, 0.738–0.849; I2=95.6%) (Supplementary Fig. 1 in the online-only Data Supplement). However, substantial heterogeneity (I2≥50%) was noted, likely due to variations in cutoff values (scores of 1, 2, and 3). To explore this heterogeneity further, a subgroup analysis was conducted. Results demonstrated significant differences in sensitivity, specificity, and NPV among the three cutoff values (p<0.05), whereas no significant differences were found for DOR or AUC (p>0.05). The cutoff value of 1 demonstrated the highest sensitivity (0.981) but the lowest specificity (0.178) and PPV (0.560). Conversely, the cutoff value of 3 provided the highest specificity (0.817) but the lowest sensitivity (0.702) and NPV (0.652). The cutoff value of 2 presented balanced diagnostic performance, with sensitivity, specificity, PPV, and NPV each ranging from 0.7 to 0.8.
Diagnostic accuracy of individual items in ID-Migraine
The ID-Migraine tool comprises three items: nausea, photophobia, and disability due to headache. The diagnostic accuracy of each individual item was analyzed to determine its relative contribution to the overall performance of the tool. The analysis revealed significant differences in specificity among the three items (p<0.05). The disability item exhibited the highest sensitivity (0.876) but the lowest specificity (0.537). In contrast, the nausea and photophobia items showed similar diagnostic performance in terms of sensitivity, specificity, PPV, and NPV.
DISCUSSION
The DOR of the ID-Migraine tool was 13.84, and the AUC was 0.846, indicating excellent diagnostic accuracy. Although specificity was relatively moderate at 70.2%, the tool demonstrated robust sensitivity at 86.5%, making it suitable as an initial screening instrument. While ID-Migraine analyses typically use a cutoff value of 2, several studies have examined diagnostic performance (sensitivity and specificity) across different cutoff values. These findings were integrated into our analysis, which confirmed that a cutoff of 2 provides significant diagnostic utility. Comparison with a cutoff value of 3 revealed no substantial differences; however, the NPV at this cutoff was relatively low at 0.65, suggesting that approximately 35% of patients classified as ID-Migraine-negative might be incorrectly identified as non-migraine cases. Given this limitation, a cutoff value of 2 is recommended. Further analysis evaluated the diagnostic contribution of each individual symptom within the three-item questionnaire to determine if any single symptom offered greater diagnostic utility. Among these symptoms, “disability” demonstrated a sensitivity of approximately 80%, specificity of around 50%, and both PPV and NPV of approximately 70%. These findings suggest that “disability” might provide greater overall diagnostic value compared to the other symptoms.
The term “sinus headache” is frequently employed, yet it remains a widely misunderstood clinical complaint, leading many patients to seek evaluation by otolaryngologists [43]. Sinus headache is traditionally associated with sinonasal inflammatory disease. However, a significant proportion of individuals who self-diagnose, or who are diagnosed by physicians as having sinus headaches, are ultimately identified as migraine sufferers. For example, a case series involving 2,991 patients who reported sinus headaches revealed that 88% fulfilled the International Headache Society (IHS) criteria for migraines [44]. Additionally, 42% of those later diagnosed with migraines based on IHS criteria were initially misdiagnosed with sinus headaches. Diagnostic challenges are compounded further by frequent symptom overlap between sinusitis and migraines [44]. A study demonstrated that among patients meeting IHS criteria for migraines, 84% reported sinus pressure, 82% had pain localized in sinus areas, 63% presented nasal congestion, and 40% experienced rhinorrhea during their initial evaluation [44]. Moreover, even when radiographic evidence of chronic rhinosinusitis was present, no significant relationship was observed between pain scores and findings obtained through Lund-Kennedy endoscopy or Lund-Mackay computed tomography (CT) scoring systems [45].
Current differential diagnostic protocols advocate for the use of CT as a primary screening tool [1,2]. Utilizing CT as an initial diagnostic instrument could facilitate rapid differential diagnosis, reduce unnecessary antibiotic prescriptions, and enable timely specialist referrals when necessary [2]. However, CT imaging involves potential unnecessary radiation exposure and imposes considerable cost burdens. Additionally, CT might not be feasible as a screening method in all primary care environments.
In situations where headaches are determined not to originate from sinus-related causes through patient history and physical examination, collaboration with other medical specialties is typically required. This study demonstrates that the ID-Migraine tool provides a simple yet effective method for distinguishing sinus headaches from migraines. Given the high prevalence of migraines, if a migraine is suspected using the ID-Migraine tool, empiric treatment with triptans can be initiated as initial therapy [46,47], along with a neurologist referral. This approach allows timely patient feedback regarding treatment efficacy during neurological consultation, thus minimizing delays in obtaining an accurate diagnosis and initiating appropriate care. Furthermore, if migraine is excluded based on patient history assessments, further investigations can be pursued to determine the actual underlying cause of sinus headaches.
This meta-analysis successfully confirms the utility of ID-Migraine as a screening instrument; however, it is subject to several limitations. First, the 30 included studies were conducted across four continents, representing diverse geographic settings and varying study qualities. This variability likely contributed to the substantial heterogeneity observed among the included studies, potentially influencing pooled estimates of sensitivity, specificity, and DOR. Nevertheless, this diversity might have mitigated selection bias by preventing overrepresentation of studies from a limited number of countries. Second, only English-language publications were included, potentially introducing language bias. However, Deeks’ funnel plot asymmetry test indicated no significant evidence of publication bias, suggesting the meta-analysis likely yielded an accurate estimate of the diagnostic tool’s performance. Third, given the relatively low incidence of migraines, the NPV might have been overestimated due to the high proportion of test-negative cases. Consequently, although the NPV (FN/TN) appeared high, the specificity (70.2%) was lower than the sensitivity (86.5%) [48].
In conclusion, the ID-Migraine screening tool demonstrates high diagnostic accuracy and user-friendliness for identifying migraines in patients presenting with sinus headaches. Implementing this tool can prevent unnecessary tests or treatments and facilitate prompt referrals to appropriate specialists.
Supplementary Materials
The online-only Data Supplement is available with this article at https://doi.org/10.18787/jr.2025.00019.
Notes
Ethics Statement
Not applicable
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
Do Hyun Kim and Se Hwan Hwang who are on the editorial board of the Journal of Rhinology were not involved in the editorial evaluation or decision to publish this article. The remaining author has declared no conflicts of interest.
Author Contributions
Conceptualization: Se Hwan Hwang. Data curation: Do Hyun Kim, David W. Jang. Formal analysis: Se Hwan Hwang. Funding acquisition: Do Hyun Kim, Se Hwan Hwang. Investigation: Do Hyun Kim. Methodology: David W. Jang, Se Hwan Hwang. Project administration: Se Hwan Hwang. Resources: David W. Jang. Software: Se Hwan Hwang. Supervision: Se Hwan Hwang. Validation: Do Hyun Kim, David W. Jang. Visualization: Do Hyun Kim, Se Hwan Hwang. Writing—original draft: Do Hyun Kim. Writing—review & editing: David W. Jang, Se Hwan Hwang.
Funding Statement
This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT)(RS-2023-00209494), and a Korean Fund for Regenerative Medicine (KFRM) grant funded by the Korean government (23C0121L1).
Acknowledgments
None
