Accuracy of International Classification of Diseases Codes for Identifying Acute Optic Neuritis

Background: The accuracy of International Classification of Diseases (ICD) codes for identifying cases of acute optic neuritis (aON) is not known. A prior study reported 61% accuracy for ICD code plus MRI consistent with aON within 2 months. This study determined accuracy for ICD code plus MRI within 2 months regardless of results. Methods: Retrospective chart review was conducted using a medical record research repository of a tertiary care institution from 1998 to 2019. Subjects with ICD-9/10 codes for ON and an MRI brain and/or orbits within 2 months of earliest (initial) ICD code were included. MRI was classified as positive or negative for aON based on report noting gadolinium-contrast enhancement. Clinical diagnosis at the time of initial code was classified as aON, prior ON, considered ON, alternative diagnosis, or unknown based on review of physician authored clinical notes within 7 days of the initial code. Accuracy of ICD code for aON, acute or prior ON, and acute, prior, or considered ON were calculated for all subjects and stratified based on MRI result. Results: Two hundred fifty-one subjects had MRI results within 2 months of their initial ON ICD code (49 positive MRI [previously reported]; 202 negative MRI). Among those with negative MRI, 32 (16%) had aON, 40 (20%) had prior ON, 19 (9%) considered ON as a diagnosis, 92 (46%) had other confirmed diagnoses, and 19 (9%) had unknown diagnosis at time of code. Considering all subjects, accuracy for ICD code was 25% for acute ON, 41% for acute or prior ON, and 48% for acute, prior, or considered ON. Positive MRI, increased number of ON ICD codes, a code given by an ophthalmologist or neurologist within 2 months, and the presence of a neurology encounter within 2 months were associated with an increased accuracy for clinical aON diagnosis. Conclusions: In the setting of an MRI within 2 months, ICD codes for ON have low accuracy for acute ON and only slightly better accuracy for acute or prior ON. Accuracy is higher for cases with a positive MRI than those with a negative MRI, suggesting positive MRI in conjunction with ICD codes may help more accurately identify cases. Reliance on ICD and Current Procedural Terminology codes alone to identify aON cases may introduce substantial misclassification bias in claims-based research.

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Accuracy of International Classification of Diseases Codes for Identifying Acute Optic Neuritis Elena A. Muro-Fuentes, MD, Sylvia E. Villarreal Navarro, MD, Heather E. Moss, MD, PhD Background: The accuracy of International Classification of Diseases (ICD) codes for identifying cases of acute optic neuritis (aON) is not known. A prior study reported 61% accuracy for ICD code plus MRI consistent with aON within 2 months. This study determined accuracy for ICD code plus MRI within 2 months regardless of results. Methods: Retrospective chart review was conducted using a medical record research repository of a tertiary care institution from 1998 to 2019. Subjects with ICD-9/10 codes for ON and an MRI brain and/or orbits within 2 months of earliest (initial) ICD code were included. MRI was classified as positive or negative for aON based on report noting gadolinium-contrast enhancement. Clinical diagnosis at the time of initial code was classified as aON, prior ON, considered ON, alternative diagnosis, or unknown based on review of physician authored clinical notes within 7 days of the initial code. Accuracy of ICD code for aON, acute or prior ON, and acute, prior, or considered ON were calculated for all subjects and stratified based on MRI result. Results: Two hundred fifty-one subjects had MRI results within 2 months of their initial ON ICD code (49 positive MRI [previously reported]; 202 negative MRI). Among those with negative MRI, 32 (16%) had aON, 40 (20%) had prior ON, 19 (9%) considered ON as a diagnosis, 92 (46%) had other confirmed diagnoses, and 19 (9%) had unknown diagnosis at time of code. Considering all subjects, accuracy for ICD code was 25% for acute ON, 41% for acute or prior ON, and 48% for acute, prior, or considered ON. Positive MRI, increased number of ON ICD codes, a code given by an ophthalmologist or neurologist within 2 months, and the presence of a neurology encounter within 2 months were School of Medicine (EM-F), Saint Louis University, St. Louis, Missouri; and Departments of Ophthalmology (SVN, HEM) and Neurology and Neurological Sciences (HEM), Stanford University, Palo Alto, California. Funding from the National Eye Institute (P30 026877) and unrestricted grant to Stanford Department of Ophthalmology from Research to Prevent Blindness, Inc. The authors report no conflicts of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www. jneuro-ophthalmology.com). Address correspondence to Heather E. Moss, MD, PhD, Spencer Center for Vision Research at Stanford, 2370 Watson Court, Suite 200, MC 5353, Palo Alto, CA 94303; E-mail: [e-mail] Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 associated with an increased accuracy for clinical aON diagnosis. Conclusions: In the setting of an MRI within 2 months, ICD codes for ON have low accuracy for acute ON and only slightly better accuracy for acute or prior ON. Accuracy is higher for cases with a positive MRI than those with a negative MRI, suggesting positive MRI in conjunction with ICD codes may help more accurately identify cases. Reliance on ICD and Current Procedural Terminology codes alone to identify aON cases may introduce substantial misclassification bias in claims-based research. Journal of Neuro-Ophthalmology 2023;43:317–322 doi: 10.1097/WNO.0000000000001805 © 2023 by North American Neuro-Ophthalmology Society O ptic neuritis (ON) is a condition involving inflammation of the optic nerve that can occur in association with neuromyelitis optica spectrum disorder (NMOSD), myelin oligodendrocyte glycoprotein (MOG)–associated disease (MOGAD), and multiple sclerosis (MS) or can be idiopathic (1–3). Incidence rates have been reported from 0.83 to 33 per 100,000 in different populations (4–8). With pathophysiologic, treatment, and prognostic differences among NMO, MOG, and MS associated ON, accurate identification of cases for both clinical and research purposes is crucial. Big data research using medical claims is a potential approach to study ON if cases can be accurately identified using International Classification of Diseases (ICD) codes. However, the accuracy of ICD codes for identifying acute ON is not known. A recent review found 3 studies that reported the positive predictive value (PPV) of algorithms for ON ranging from 25% to 100% using the treating clinician’s diagnosis as ground truth (9–12). Our group previously reported 61% accuracy for acute ON in subjects with ICD code for ON, MRI performance within 2 months of ICD code and MRI report of optic nerve enhancement (13). This follow-up study aimed to determine the accuracy for acute ON of ICD code for acute optic neuritis with an 317 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution MRI performed within 2 months regardless of MRI result and to identify strategies for increasing accuracy. METHODS Subjects This is a retrospective study of people with an ICD code for optic neuritis and MRI at a tertiary care academic institution in the Western United States from 1998 to 2019. Inclusion criteria were adults ($18 years old), with 1 or more ICD-9 code 377 or ICD-10 code H46 for optic neuritis, completed MRI brain and/or orbits with and without contrast (current procedural technology code 70553 or 70543) within 2 months of the earliest optic neuritis ICD code, and provider documentation regarding patient diagnosis at the time of ICD code. Subjects were excluded if the only ICD codes indicated a secondary cause of optic neuritis or did not include the term optic neuritis (e.g., ICD-9 377.00 unspecified papilledema, 377.31 optic papillitis, 377.33 nutritional optic neuropathy, 377.34 toxic optic neuropathy; ICD-10 H46.0 optic papillitis, H46.2 nutritional optic neuropathy, H46.3 toxic optic neuropathy) or if there were insufficient medical records available for review (e.g., no MRI report, no provider documentation corresponding to ICD code). Subjects were identified using the Stanford Research Repository (STARR) cohort selection and chart review tools. This study was approved with waiver of informed consent by the Stanford University Institutional Review Board. Subjects were grouped by MRI result. MRI was classified as positive for acute optic neuritis if there was gadoliniumcontrast enhancement of the optic nerve noted on the MRI report and otherwise classified as negative. Subjects with positive MRI are the same as those in our previously published study (13). Subjects with negative MRI have not been previously reported. Variable Assessment Diagnosis in the opinion of the treating providers was categorized as acute ON, prior ON, consideration of ON, alternative diagnosis (not ON), and unknown diagnosis based on review of records within 7 days of initial ICD code. Patient journey characteristics that could be inferred from medical claims data were extracted. These included the number of ON-ICD codes within 2 months of initial ON ICD code, billing provider specialty associated with the initial ICD code, an ophthalmology or neurology billing provider for at least 1 ON ICD code within 2 months, number of days (,365) between initial ICD code and first encounters (inpatient or outpatient) with an ophthalmology or neurology provider, total number ophthalmology and neurology clinical encounters within 2 months, and treatment with steroids within 2 months of initial code. High-dose steroid treatment was defined as 1 g intravenous methylprednisolone for at least 3 days or 318 1,250 mg oral prednisone for at least 3 days. Low-dose steroid treatment was defined as solely an oral dose less than 1,250 mg. Clinical encounters excluded any visits for a procedural, specific diagnostic testing, or perioperative reason. Any encounters with the same specialty but different provider on the same date were counted as separate clinical encounters. Analysis and Statistical Methods Accuracy for acute ON (number of acute ON/total number of subjects), acute or prior ON (number of acute ON + number of prior ON/total number of subjects), and acute, prior, or considered ON (number of acute ON + number of prior ON + number of considered ON/total number of subjects) was calculated for all subjects and stratified based on MRI result. Although these calculations of true positive/ (true positive + false positive) are technically positive predictive value, we chose to use the term accuracy because the connotation was more in line with the assessment. Patient journey characteristics were compared between the acute optic neuritis and nonacute optic neuritis groups using Mann–Whitney for continuous variables with nonnormal distributions. Receiver operating characteristic (ROC) analysis was used to calculate area under the curve (AUC) to evaluate the classification ability of continuous variables. Categorical measures were compared using the chi-squared test or Fisher exact test for any tables with cell size less than 10. Statistical significance was set at P , 0.05. Statistical analysis was performed using SPSS V28 (IBM Inc, Armonk, NY). RESULTS From an initial cohort of 1,510 potential subjects with at least one ICD-9/10 code for optic neuritis, 1,167 (77%) were excluded due to exclusively nonapplicable ICD code for optic neuritis, lack of MRI current procedural terminology (CPT) code within 2 months of earliest ICD code or lack of available MRI report. Of the remaining 343, 268 (78%) had negative MRIs (no optic nerve enhancement) and 75 (22%) had positive MRIs (optic nerve enhancement). Of these groups, 66 and 26, respectively, were excluded due to insufficient records to confirm the clinical diagnosis, leaving 202 negative MRI and 49 positive MRI subjects (Fig. 1). Of these 251 subjects included in the study, 102 had MRI brain only, 75 had MRI orbit only, and 74 had both. Two hundred twenty subjects had MRI performed within 30 days of optic neuritis ICD code. As previously reported, of the 49 with positive MRI, 30 (61%) had acute optic neuritis, 18 (39%) had other confirmed diagnoses, and 1 (2%) was undiagnosed (12). Newly reported in this article are those with negative MRI; among which, 32 (16%) had acute ON, 40 (20%) had prior ON, 19 (9%) had ON considered, 92 (46%) had other confirmed diagnoses, and 19 (9%) were unknown (Fig. 1). Of the 19 with ON considered at time of initial Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution discussed below (See Supplemental Digital Content, Table 2, http://links.lww.com/WNO/A672). International Classification of Diseases Codes for Identification of Acute Optic Neuritis Cases FIG. 1. Subject identification. CPT, current procedural terminology; ICD, International Classification of Diseases; ON, optic neuritis. ICD code, subsequent records indicated that 5 ultimately had acute ON, 8 had an alternate diagnosis, 1 had no diagnosis, and 5 had insufficient follow-up records. Of those with other confirmed diagnoses, the most common were other optic neuropathies (See Supplemental Digital Content, Table 1, http://links.lww.com/WNO/A672). Considering all subjects, accuracy of ICD code was 25% for acute ON with accuracy being better in subjects with positive MRI than negative MRI within 2 months (Table 1). Accuracy for acute or prior ON was 41%. Accuracy for acute, prior, or considered ON was 48%. Accuracy was similar when considering the subgroup with orbit MRI performed within 30 days of ICD code (n = 137) (Table 2). Accuracy for approaches to acute optic neuritis case selection with a yield of .50% are summarized in Table 3 and In analysis of patient journey features, acute optic neuritis diagnosis (aON) vs all other diagnostic categories (non-aON) was associated with a greater number of ICD codes (P , 0.001 considering all subjects, Mann–Whitney, See Supplemental Digital Content, Table 3, http://links.lww.com/ WNO/A672), with area under the curve of 0.77 (ROC analysis). Accuracy for case selection at the largest Youden Index (corresponding to $5 ICD codes) was 49% of 94 subjects. A maximum accuracy of 60% was achieved with $10 ICD codes, but this reduced number of subjects to 47 (See Supplemental Digital Content, Table 2, http://links. lww.com/WNO/A672). A minority of subjects had initial ICD code documented by an ophthalmologist, neurologist, or either, and this did not differ between aON and non-aON subjects (P = 0.135–0.841, Chi-square, See Supplemental Digital Content, Table 4, http://links.lww.com/WNO/ A672). Thus, considering that this did not improve accuracy of case identification. Although the presence of an ophthalmology-given code within 2 months was not associated with aON (P = 0.10 chi square), presence of a neurology-given code, neurology, or ophthalmology-given codes, or both neurology and ophthalmology–given codes was associated with aON (P , 0.001 chi square, Fisher exact for all, See Supplemental Digital Content, Table 4, http://links.lww.com/ WNO/A672) and consideration improved case selection accuracy to 50% (76 subjects), 34% (155 subjects) and 61% (28 subjects), respectively. Specialty Encounters for Identification of Acute Optic Neuritis Cases aON was associated with the presence of $1 neurology clinical encounter within 2 months but not with $1 ophthalmology clinical encounter (P , 0.001, 0.87, Chisquare, See Supplemental Digital Content, Table 5, http://links.lww.com/WNO/A672) with the accuracy of ICD coding increasing to 39% for 111 subjects with a neurology encounter. aON was neither associated with the number of days between initial ICD code and first neurology nor first ophthalmology encounter (P = 0.14, 0.68, Mann–Whitney, See Supplemental Digital TABLE 1. Accuracy of optic neuritis diagnosis among subjects with $1 International Classification of Diseases code and MRI brain or orbit performance within 2 months stratified by MRI result Diagnosis Overall N = 251 MRI+ N = 49 MRI2 N = 202 Acute optic neuritis Acute + prior optic neuritis Acute + prior + considered optic neuritis 62 (25%) 102 (41%) 121 (48%) 30 (61%) 32 (16%) 72 (36%) 91 (45%) Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 319 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 2. Accuracy of optic neuritis diagnosis among subjects with $1 International Classification of Diseases code and MRI orbit performance within 30 days stratified by MRI result Acute optic neuritis Acute + prior optic neuritis Acute + prior + considered optic neuritis Overall N = 137 MRI+ N = 36 MRI2 N = 101 38 (28%) 42 (31%) 58 (42%) 24 (67%) 14 (14%) 18 (18%) 34 (34%) Content, Table 6, http://links.lww.com/WNO/A672). The number of neurology encounters, but not ophthalmology encounters, within 2 months was associated with aON (P , 0.001, P = 0.89, Mann–Whitney, See Supplemental Digital Content, Table 7, http://links.lww. com/WNO/A672) with ROC-AUC of 0.68 for neurology encounter number. Maximum accuracy for aON cases was 64% for $3 neurology encounters (14 subjects). Number of inpatient encounters with neurology, but not ophthalmology, was associated with aON (P , 0.001, 0.19 Mann– Whitney) with ROC-AUC of 0.66 for inpatient neurology encounter number with best accuracy of 52% for $1 encounter (56 subjects). Treatment for Identification of Acute Optic Neuritis Cases Seventy-eight percent of the aON group received any steroid treatment, compared with 24% of the non-ON group (P , 0.001 chi square). Consideration steroid administration improved case identification to 53% (89 subjects). Accuracy improved to 61% among 72 subjects who received high-dose steroids. CONCLUSIONS The use of ICD codes to identify cases facilitates the use of medical claims databases to study disease risk factors, treatments, and care delivery. However, the accuracy of the ICD codes is critical to ensure accurate conclusions and appropriate clinical applications. Our study assessed the accuracy of ICD codes for acute optic neuritis in those with an MRI within 2 months of initial ICD code and identify factors associated with increased accuracy. We found the accuracy of the ICD codes with an MRI within 2 months to be low at 25%, which further decreases to 16% if limited to negative MRIs and increases to 61% with positive MRIs. This suggests that MRI results are useful for identifying acute ON cases based on ICD codes. However, although performance of an MRI can be assessed from medical claims data, the result cannot. Therefore, we assessed other parameters and found none that improved identification accuracy above 70% and all of which compromised the sample size. Previous articles have shown a wide range of ON diagnostic accuracy using claims data. A Canadian study 320 using health claims data reported 25% accuracy for aON, which is similar to our estimate (11). However, this study only considered patients seen in an MS center. Although an American study documented an accuracy of 100% for ON (12), the data supporting their algorithm of ICD code with angiotensin-converting enzyme (ACE) serum test within 90 days or $3 ICD codes within 60 days was not provided, and only 6 cases were identified. A study from Denmark reported hospital discharge ICD code accuracy for aON in a pediatric population to be 70% compared with a review of records supervised by a pediatric neurologist (10). This study builds on this prior work by considering ON codes in all settings of an American tertiary care medical center and evaluating MRI performance and result as a strategy to accurately identify cases of aON. The lack of granularity of the ICD code for optic neuritis (which does not differentiate acute from prior) likely contributes to the low accuracy of ICD codes for aON. The improved accuracy in patients with positive MRI suggests that MRI results may help to refine case selection beyond the ICD code and is not surprising given the high (94%) sensitivity of MRI enhancement for optic neuritis TABLE 3. Strategies that improve accuracy of identification of acute optic neuritis cases to above 50% Strategy Baseline ($1 ICD code and MRI performance within 2 mo) MRI+ $10 ICD codes $1 neurology given ICD code $1 neurology and $1 ophthalmology given ICD code $3 neurology encounters $1 neurology inpatient encounter Steroid treatment High dose steroid treatment Accuracy (True Positive) Subjects (n) 25% 252 61% 60% 50% 61% 49 47 76 28 64% 52% 14 56 53% 61% 89 72 All strategies assessed within 2 months of initial ICD code. ICD, International Classification of Diseases. Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution (14). However, reliance on MRI may exclude cases in which known demyelinating disease prompts treatment without new imaging. Overall, 130 patients (52%) with an ICD code for optic neuritis and an MRI had neither acute, prior nor a considered diagnosis of optic neuritis. In fact, 110 (44%) had confirmed alternative diagnoses. For the cohort with optic nerve enhancement, our group previously reported that the most predominant alternative etiologies were neoplasm or lesions, followed by other optic neuropathies (13). For the cohort without enhancement, other etiologies were predominantly other optic neuropathies, followed by neoplasms or other lesions. Although our study design cannot point to the root of these errors, one possibility is physician error with ON having previously been shown to have a near 60% rate of misdiagnosis, with the most common alternative diagnoses found to be primary headache disorder, functional visual loss, and other optic neuropathies (15). As illustrated by a recent case series, accurate diagnosis of ON is further complicated not only by the atypical ON subtypes but also by atypical presentations (16). With the discovery of the aquaporin-4 (AQP4) antibody and MOG antibody associated with NMOSD and MOGAD, respectively, subsequent research has helped better characterize each and direct prospective and retrospective studies on these specific patient populations, but this may not yet been adopted throughout physician practice (17–19). Although the presence of an initial ICD code given by a specialty other than ophthalmology or neurology was not associated with decreased accuracy, we found an association between presence of an ophthalmology- or neurology-given code and aON, suggesting that expertise plays an important role in accurate diagnosis. Furthermore, becauseICD codes are designed to maximize efficiency in processing medical claims for reimbursement, there is potential misclassification when a diagnosis remains unknown or despite diagnosis change. Other possible sources of error include administrative errors and transcription errors (20). Our results demonstrate features that could inform an algorithm to identify cases of aON. Some features such as number of codes may be helpful but would bias toward longer follow-up. Considering a neurology encounter or steroid administration in conjunction with a positive MRI are promising approach that will need to be validated across multiple sites and larger cohorts. Our study was limited to a single tertiary care center in the Western United States, and further study is needed to determine accuracy of ON ICD coding in institutions/ regions without reimbursement based on coding. Sample size was limited due to our inclusion criteria requiring an MRI CPT code and report. Thus, we did not assess accuracy for ICD code for optic neuritis overall. There may have been variable classification of MRI outcome depending on MRI sequences performed (brain vs orbit) and interval between MRI and clinical syndrome. It is reinforced that the purpose of this study is not to evaluate the specificity of MRI for optic Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 neuritis, rather it is to evaluate strategies for identifying account ON cases from medical claims. Ground truth defined as clinical diagnosis within 7 days of initial ICD code may limit the number of cases identified. However, a large portion had confirmed alternate diagnoses within this time frame and only a minority remained undiagnosed, suggesting that this was not a primary limiting factor. Ultimately our results suggest that optic neuritis cases cannot be accurately identified by ICD and MRI completion alone and that clinical record review is required. Reliance on ICD codes and MRI performance to identify cases of optic neuritis may lead to result in incorrect conclusions due to misclassification bias (15). STATEMENT OF AUTHORSHIP Conception and design: E. Muro-Fuentes, H. Moss; Acquisition of data: E. Muro-Fuentes, S. Villarreal Navarro; Analysis and interpretation of data: E. Muro-Fuentes, H. Moss. Drafting the manuscript: E. Muro-Fuentes; Revising the manuscript for intellectual content: E. Muro-Fuentes, H. Moss, S. Villarreal Navarro. Final approval of the completed manuscript: E. Muro-Fuentes, H. Moss, S. Villarreal Navarro. REFERENCES 1. Optic Neuritis Study Group. Multiple sclerosis risk after optic neuritis: final optic neuritis treatment trial follow-up. Arch Neurol. 2008;65:727–732. 2. Abel A, McClelland C, Lee MS. Critical review: typical and atypical optic neuritis. Surv Ophthalmol. 2019;64:770–779. 3. Hassan MB, Stern C, Flanagan EP, Pittock SJ, Kunchok A, Foster RC, Jitprapaikulsan J, Hodge DO, Bhatti MT, Chen JJ. Population-based incidence of optic neuritis in the era of aquaporin-4 and myelin oligodendrocyte glycoprotein antibodies. Am J Ophthalmol. 2020;220:110–114. 4. Soelberg K, Jarius S, Skejoe H, Engberg H, Mehlsen J, Nilsson A, Madsen J, Reindl M, Wildemann B, Grauslund J, Kyvik K, Smith T, Lillevang S, Paul F, Weinshenker B, Asgari N. A population-based prospective study of optic neuritis. Mult Scler. 2017;23:1893–1901. 5. Rodriguez M, Siva A, Cross SA, O’Brien PC, Kurland LT. Optic neuritis: a population-based study in Olrnsted County, Minnesota. Neurology. 1995;45:244–250. 6. Woung LC, Lin CH, Tsai CY, Tsai MT, Jou JR, Chou P. Optic neuritis among national health insurance enrollees in taiwan, 2000-2004. Neuroepidemiology. 2007;29:250–254. 7. Lim SA, Wong WL, Fu E, Goh KY, Seah A, Tan C, Tow S, Cullen JF, Wong TY. The incidence of neuro-ophthalmic diseases in Singapore: a prospective study in public hospitals. Ophthal Epidemiol. 2009;16:65–73. 8. Martinez-Lapiscina EH, Fraga-Pumar E, Pastor X, Gomez M, Conesa A, Lozano-Rubi R, Sanchez-Dalmau B, Alonso A, Villoslada P. Is the incidence of optic neuritis rising? Evidence from an epidemiological study in Barcelona (Spain), 20082012. J Neurol. 2014;261:759–767. 9. Hamedani AG, De Lott LB, Deveney T, Moss HE. Validity of international classification of diseases codes for identifying neuro-ophthalmic disease in large data sets: a systematic review. J Neuroophthalmol. 2020;40:514–519. 10. Boesen MS, Magyari M, Born AP, Thygesen LC. Pediatric acquired demyelinating syndromes: a nationwide validation study of the Danish National Patient Register. Clin Epidemiol. 2018;10:391–399. 321 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 11. Marrie RA, Ekuma O, Wijnands JMA, Kingwell E, Zhu F, Zhao Y, Fisk JD, Evans C, Tremlett H. Identifying optic neuritis and transverse myelitis using administrative data. Mult Scler Relat Disord. 2018;25:258–264. 12. Winthrop KL, Chen L, Fraunfelder FW, Ku JH, Varley CD, Suhler E, Hills WL, Gattey D, Baddley JW, Liu L, Grijalva CG, Delzell E, Beukelman T, Patkar NM, Xie F, Herrinton LJ, Fraunfelder FT, Saag KG, Lewis JD, Solomon DH, Curtis JR. Initiation of antiTNF therapy and the risk of optic neuritis: from the safety assessment of biologic ThERapy (SABER) Study. Am J Ophthalmol. 2013;155:183–189.e1. 13. Muro-Fuentes EA, Moss HE. Factors associated with increased emergency department utilization in patients with acute optic neuritis. J Neuroophthalmol. 2021;41:335–341. 14. Kupersmith MJ, Alban T, Zeiffer B, Lefton D. Contrastenhanced MRI in acute optic neuritis: relationship to visual performance. Brain. 2002;125:812–822. 15. Stunkel L, Kung NH, Wilson B, McClelland CM, Van Stavern GP. Incidence and causes of overdiagnosis of optic neuritis. JAMA Ophthalmol. 2018;136:76–81. 322 16. Bouffard MA, Mallery RM, Liao YJ, Torun N. Variation in evolving optic neuritis. J Neuroophthalmol. 2021;41:476–479. 17. Lennon VA, Wingerchuk DM, Kryzer TJ, Pittock SJ, Lucchinetti CF, Fujihara K, Nakashima I, Weinshenker BG. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet. 2004;364:2106–2112. 18. Linington C, Lassmann H. Antibody responses in chronic relapsing experimental allergic encephalomyelitis: correlation of serum demyelinating activity with antibody titre to the myelin/oligodendrocyte glycoprotein (MOG). J Neuroimmunol. 1987;17:61–69. 19. Waters P, Woodhall M, O’Connor KC, Reindl M, Lang B, Sato DK, Jurynczyk M, Tackley G, Rocha J, Takahashi T, Misu T, Nakashima I, Palace J, Fujihara K, Leite MI, Vincent A. MOG cell-based assay detects non-MS patients with inflammatory neurologic disease. Neurol Neuroimmunol Neuroinflamm. 2015;2:e89. 20. O’Malley KJ, Cook KF, Price MD, Wildes KR, Hurdle JF, Ashton CM. Measuring diagnoses: ICD code accuracy. Health Serv Res. 2005;40:1620–1639. Muro-Fuentes et al: J Neuro-Ophthalmol 2023; 43: 317-322 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.