Introduction
Measles is a respiratory infection caused by the measles virus, a member of the Paramyxoviridae virus family. Reference Bellini, Rota and Rota1 Common symptoms of infection include high fevers, cough, coryza, conjunctivitis, sore throat, and a distinctive rash. Reference Dunn, Baldanti, Puchhammer, Panning, Perez and Harvala2 While most cases are self-limited, some patients have complications, including otitis media, laryngotracheobronchitis, pneumonia, and diarrhea. Reference Perry and Halsey3–Reference Hester, Nickel and LeBlanc5 Less common, but serious, complications include encephalitis and subacute sclerosing panencephalitis. Reference Wendorf, Winter and Zipprich6,Reference Fisher, Defres and Solomon7 Before the widespread use of the measles vaccine, measles was a common childhood illness and a major cause of morbidity and mortality. For example, during the prevaccination era, three to four million infections occurred each year in the United States (US) alone, with 48,000 associated hospitalizations and up to 500 deaths. Reference Bloch, Orenstein and Stetler8–Reference Gastanaduy, Rota and Patel10
In 2000, measles was declared to be eliminated from the US. Reference Parker Fiebelkorn, Redd and Gallagher11 Although sporadic outbreaks occurred, they were mostly linked to returning travelers who were infected abroad. Reference Jost, Luzi, Metzler, Miran and Mutsch12,Reference Poterek, Kraemer, Watts, Khan and Perkins13 Recently, the US has experienced multiple prolonged outbreaks, sustained by populations with low vaccination rates. Reference Gastañaduy, Goodson, Panagiotakopoulos, Rota, Orenstein and Patel14 These outbreaks are difficult to control for three major reasons. First, the virus is highly transmissible via respiratory droplets during close contacts and by aerosolized particles that can remain airborne for hours and infect persons who are not directly exposed to the source. Reference Bloch, Orenstein and Ewing15,Reference Remington, Hall, Davis, Herald and Gunn16 Individuals without protective immunity who are exposed to the virus have up to a 95% chance of acquiring measles. Reference Azan, Chuecos-Escalante, Marte and Bhagi17 Second, asymptomatic or minimally symptomatic persons who are incubating measles can spread the virus for up to four days before the classic rash appears, Reference Graber, Andrade, Bost and Gibbs4,Reference Bester18 and their infections may not be diagnosed because the non-specific symptoms during this prodromal period overlap with other, more common, respiratory infections. Furthermore, even patients with a rash can be misdiagnosed as having other exanthematous infections, including rubella, roseola, mononucleosis, and scarlet fever. Reference Azan, Chuecos-Escalante, Marte and Bhagi17 Finally, during the postvaccination era, healthcare professionals may not include measles in their differential diagnoses because they have never encountered a patient with measles. Reference Popofsky and Romero19,Reference Steingart, Thomas, Dykewicz and Redd20
The difficulty in diagnosing measles contributes to disease spread. Multiple outbreaks have been attributable, at least in part, to delays in diagnosing measles. Reference Maltezou and Wicker21–Reference Clegg, Linchangco and Arwady26 Given the dramatic and sustained increase in the number of measles cases in the US in recent years, we sought to estimate the incidence and duration of diagnostic delays for measles and to describe factors associated with delays.
Methods
Data source and study population
We conducted a retrospective study using longitudinal health insurance claims from the Merative MarketScan Research Database, including the Commercial and Medicare databases from 2001–2023 and the Multi-State Medicaid databases from 2014–2021. These data represent one of the largest databases of commercial insurance claims in the US and contain records from inpatient, outpatient, and emergency department (ED) encounters, along with outpatient prescription medications.
Measles cases were identified using ICD-9-CM diagnosis codes 055.XX and ICD-10-CM codes B05.XX. The index diagnosis date/visit was defined as a patient’s first claim containing any of the relevant ICD codes for measles. Since our database lacks laboratory test results to confirm cases of measles, we also required evidence of PCR or serology testing for measles. Specifically, the study population was restricted to patients with claims for PCR or serology testing (CPT 86765, 87798) within 30 days of the index measles diagnosis. Our goal was to identify diagnostic delays in clinical suspicion rather than testing. If a test was ordered before the index diagnosis, we used the test date as the index date, assuming measles was suspected then. Visits between the test and index dates were not attributed to delays in diagnosis. Patients were required to be continuously enrolled in their insurance plan for ≥180 days before the index measles diagnosis.
Statistical analysis
We identified all potential missed diagnostic opportunities by evaluating healthcare visits before the index measles diagnosis during which signs or symptoms of measles were present. We reviewed all ICD-9/10-CM diagnosis codes recorded during healthcare visits in the 180 days before the index measles diagnosis to identify visits with: (1) signs or symptoms of measles, (2) tests ordered for conditions with manifestations similar to measles, or (3) diagnoses with presentations similar to measles. Codes meeting one of these criteria were deemed clinically plausible as potential evidence of measles and defined as symptomatically similar diagnoses (SSDs). Appendix Table 1 presents the complete set of SSDs and corresponding ICD-9/10-CM codes used to identify potential missed diagnostic opportunities.
Missed opportunities to diagnose measles are not directly observable, as signs or symptoms occurring before diagnosis (e.g., fever, cough) may be coincidental and caused by other diseases with similar symptoms (e.g., pneumonia, bronchitis). Thus, to estimate the true number of missed opportunities, we employed a two-step process utilized in prior studies. Reference Miller, Arakkal and Koeneman27–Reference Miller, Koeneman, Arakkal, Cavanaugh and Polgreen31
Step 1: Estimating the number of missed opportunities
We began by implementing a type of case-crossover design to calculate the excess number of SSD visits (i.e., visits with signs or symptoms) each day before the measles diagnosis. We refer to the case period of 14 days before the measles diagnosis as the diagnostic opportunity window, which we define as the time before diagnosis when diagnostic opportunities may occur. The period from 15–180 days before the measles diagnosis represents the crossover (control) period for each enrollee. We selected this 14-day cutoff as the point in time where the frequency of SSD-related visits began to increase from the baseline trend (Figure 1). We also conducted a sensitivity analysis (described below) evaluating alternative crossover periods.
Daily counts of symptomatically similar diagnosis (SSD)-related healthcare visits in the 180 days before the index measles diagnosis.
To estimate the expected number of SSD visits each day before diagnosis, we first computed the number of patients with an SSD-associated visit each day during the crossover period (15–180 days before diagnosis). We fit a linear time trend to these daily visit counts and included an additive qualitative effect for day of the week to capture weekly periodicity. We then extrapolated the trend forward into the diagnostic opportunity window (i.e. 1–14 days before diagnosis) to estimate the expected number of SSD visits. Finally, we computed the number of missed diagnostic opportunities each day before diagnosis as the difference between the observed and expected number of SSD visits (i.e., the excess number of SSD visits) during the diagnostic opportunity window.
Step 2: Estimating diagnostic delay metrics and potential risk factors
We utilized a bootstrapping approach to estimate patient-level measures for the frequency, duration, and potential risk factors for diagnostic delays. Reference Miller, Cavanaugh, Arakkal, Koeneman and Polgreen32 First, we randomly sampled individual patient visits representing a missed opportunity on a given day using the computed number of missed opportunities from the case-crossover analysis (described above). Using these selected visits, we calculated: (1) the number of patients experiencing a missed opportunity, (2) the number of missed opportunities per patient, (3) the number of missed opportunities by healthcare setting, and (4) the diagnostic delay duration.
The bootstrapping procedure was implemented using the following four-step process: (1) draw a bootstrapped sample of patients with replacement, (2) implement the case-crossover analysis computing the number of missed opportunities each day before diagnosis, (3) randomly draw which individual patient visits represented a missed opportunity (we use the uncorrelated algorithm from Miller et al. 2023), Reference Miller, Cavanaugh, Arakkal, Koeneman and Polgreen32 and (4) calculate delay metrics and evaluate the risk factor models (described below). We repeated steps 1 and 2 100 times generating 100 bootstrapped samples. Then for each bootstrapped sample, we repeated steps 3 and 4 100 times, resulting in a total of 10,000 trials. We aggregated results across all trials. We report the median of the bootstrapped estimates along with the 95% percentile-based confidence interval (95%CI).
We conducted an exploratory analysis to identify possible risk factors associated with missed diagnostic opportunities. We evaluated a patient-level Firth logistic regression model to estimate the odds of a patient experiencing at least one missed opportunity as a function of patient and clinical characteristics. Firth logistic regression was employed due to issues with complete/quasi-complete separation induced by sparsity. For each simulation trial, we assigned each patient selected as having at least one missed opportunity an outcome value of 1 (i.e., delayed patient) and each patient not selected a value of 0 (i.e., not a delayed patient). We evaluated the following patient demographic characteristics: age (categorized into <2, 2–4, 5–11, 12–17, 18–35, 36–45, 46–55, 56–65, and >65 years), sex, and rurality (i.e., patient visited a rural health clinic). We also evaluated if treatments for potential symptoms of measles were associated with delays by including an indicator for respiratory antibiotics prescribed during the diagnostic opportunity window. We included the month and year of the index measles diagnosis as categorical variables. Due to sparsity with respect to the outcome, we excluded the years 2001 and 2002. In total, we fit 10,000 models, one for each simulation trial. For each risk factor of interest, we present the median of the bootstrapped effect estimate and the 95% percentile-based confidence interval.
We used the R programming language for all statistical analyses and the delaySim package (github.com/aarmiller/delaySim) to implement the bootstrapping procedure. The codes used for identifying cases along with diagnosis, procedure, and medication codes are available in the GitHub repository: github.com/aarmiller/delay_diagnosis/tree/main/publications/measles.
Sensitivity analyses
We evaluated the impact of the assumed 14-day diagnostic opportunity window. The timing of the diagnostic opportunity window defines the crossover-control period (i.e., 15–180 days) used to estimate the expected number of symptomatic healthcare visits before diagnosis. Thus, we also considered a 21-day diagnostic opportunity window (i.e., 22–180-day control period) and repeated our entire analysis using this window.
Furthermore, since some signs and symptoms present during clinic visits may not be captured by diagnostic codes in insurance claims, the SSD visits identified may underestimate the true number of missed opportunities. Therefore, we repeated all our analyses to compute the frequency and duration of delays using all visits (with or without an SSD) as potential diagnostic opportunities. This analysis essentially provides an upper bound of potential missed opportunities not captured by our SSD list.
Results
We identified 4,651 index cases of measles during 2001–2023 among patients continuously enrolled for at least 180 days before their index diagnosis. Table 1 summarizes baseline characteristics of the study population. The median age of the cohort was 34.0 years and 67.2% of the cohort was female. Most of patients were identified in the Commercial Claims database (94.6%), followed by the Multi-State Medicaid (3.7%) and the Medicare databases (1.7%).
Baseline characteristics of the study population identified between 2001–2023
Table 1 Long description
A table summarizing baseline characteristics of a study population identified between 2001 and 2023. The table has 4,651 patients and includes characteristics such as age at diagnosis, sex, database source, and enrollment time prior to index. The table is divided into several rows and columns. Row 1: Age at diagnosis (years), Number of patients (percent). Row 2: <2, 85 (1.8 percent). Row 3: 2-4, 82 (1.8 percent). Row 4: 5-11, 187 (4.0 percent). Row 5: 12-17, 149 (3.2 percent). Row 6: 18-35, 1,976 (42.5 percent). Row 7: 36-45, 916 (19.7 percent). Row 8: 46-55, 709 (15.2 percent). Row 9: 56-65, 471 (10.1 percent). Row 10: >65, 76 (1.6 percent). Row 11: Mean (SD), 34.8 (16.1). Row 12: Median (IQR), 34.0 (24.0). Row 13: Sex, Number of patients (percent). Row 14: Male, 1,525 (32.8 percent). Row 15: Female, 3,126 (67.2 percent). Row 16: Database source, Number of patients (percent). Row 17: Commercial, 4,398 (94.6 percent). Row 18: Medicare, 80 (1.7 percent). Row 19: Medicaid, 173 (3.7 percent). Row 20: Enrollment Time Prior to Index (years), Number of patients (percent). Row 21: 1, 813 (17.5 percent). Row 22: 2, 1,886 (40.6 percent). Row 23: 3, 2,613 (56.2 percent). Row 24: >3, 2,038 (43.8 percent). Row 25: Mean (SD), 3.7 (3.3). Row 26: Median (IQR), 2.6 (3.7).
Note. SD, standard deviation; IQR, interquartile range.
In the 180 days before the index measles diagnosis, 3,968 (85.3%) patients had at least one healthcare visit for any reason and 1,402 (30.1%) patients had at least one SSD visit. We identified a total of 2,486 SSD visit days in the 180 days before the index diagnosis. Figure 1 depicts the trend in the number of SSD visits over the 180-day period before the index measles diagnosis. Notably, the frequency of SSD-related visits increases substantially beginning a few weeks before the index diagnosis.
During the 2-week diagnostic opportunity window, we identified 542 total SSD visit days from 426 (9.2%) patients representing “potential” missed opportunities. Figure 2(A) depicts the expected trend (in red) of SSD-related visits, estimated based on visits during the control period from 15–180 days before diagnosis. The estimated number of missed diagnostic opportunities is depicted as the shaded area (in orange) between the observed (in black) and the expected (in red) trend during the diagnostic opportunity window (region to the right of the blue line). Figure 2(B) depicts the distribution of the estimated number of missed diagnostic opportunities each day before the index diagnosis.
Trends in symptomatically similar diagnosis (SSD)-related healthcare visits and estimated number of missed diagnostic opportunities each day prior to the index measles diagnosis. (A) Raw SSD visit counts for each day before the index measles diagnosis in black, expected trendline estimated from visits between 15–180 days before the index measles diagnosis in red, and the start of the diagnostic opportunity window beginning 14 days before the index diagnosis in blue. (B) Estimated number of missed diagnostic opportunities based on the excess number of SSD-associated healthcare visits, computed as the difference between the observed and expected values in the orange shaded region of (A).
Figure 2 Long description
Panel A: A line graph shows the number of symptomatically similar diagnosis (SSD) visits against days before the index measles diagnosis. The black line represents raw SSD visit counts, the red line shows the expected trendline estimated from visits between 151-80 days before the index diagnosis, and the blue line marks the start of the diagnostic opportunity window beginning 14 days before the index diagnosis. There is a noticeable spike in SSD visits as the index diagnosis approaches. Panel B: A bar graph illustrates the estimated number of missed diagnostic opportunities, computed as the difference between observed and expected values in the orange shaded region of Panel A. The number of missed opportunities increases significantly as the index diagnosis nears.
Based on the bootstrapping approach, we estimated that 381 (95%CI: 330–440) of the total SSD visit days during the diagnostic opportunity window represent true missed diagnostic opportunities. Table 2 presents a breakdown of the estimated number of missed diagnostic opportunities by medical setting. Approximately 81.0% (95%CI: 77.2–84.8%) of missed opportunities occurred in an outpatient setting, 14.4% (95%CI: 11.1–17.6%) in ED settings, 3.9% (95%CI: 1.6–7.5%) in inpatient settings, and 0.7% (95%CI: 0.0–1.9%) in observational stay settings.
Index diagnosis visits and simulation results for estimated number and percent of missed diagnostic opportunities by setting
Table 2 Long description
A table comparing missed diagnostic opportunities across different medical settings. The table has five rows and five columns. The columns are labeled Setting, Number of potential missed opportunity visit days, Index diagnosis visits, Percent of all index visits, Number of missed opportunities (95% CI), and Percent of all missed opportunities (95% CI). The rows are labeled Outpatient, ED, Observational stay, Inpatient, and Inpatient visit. Row 1: Outpatient, 491, 4608, 95.9 percent, 343 (299-398), 81.0 percent (77.2-84.8 percent). Row 2: ED, 84, 168, 3.5 percent, 62 (45-84), 14.4 percent (11.1-17.6 percent). Row 3: Observational stay, 5, 6, 0.1 percent, 4 (0-8), 0.7 percent (0.0-1.9 percent). Row 4: Inpatient, 25, 25, 0.5 percent, 17 (6-34), 3.9 percent (1.6-7.5 percent). Row 5: Inpatient visit, 9, 25, 0.5 percent, 9 (4-15), -.
Note. a Multiple settings may be associated with an index visit date (i.e., we are unable to order the timing of visits from claims on the same day); thus, the total number exceeds the number of individuals in the study population.
b Multiple settings may occur on a given visit day (i.e., we are unable to order the timing of visits from claims on the same day); thus, the total number exceeds the total estimated potential missed diagnostic opportunities.
c Inpatient visits represent one or more days admitted to an inpatient facility. Our analysis of missed diagnostic opportunities is based on “visit days” in which the patient received a symptomatically similar diagnosis on a given day. We can aggregate potential missed opportunities, index diagnoses, and estimated missed opportunities from the bootstrapping approach into the overall inpatient stay in which they occurred. However, the denominator for computing the percent of missed opportunities that occurred in a given setting represents the total number of “visit days” representing a missed opportunity; thus, a value for aggregate inpatient visits is not computed.
Table 3 presents the estimates for the duration and number of missed diagnostic opportunities that individual patients experienced. Approximately 6.9% (95%CI: 6.1–8.0%) of patients experienced at least one missed diagnostic opportunity. On average these patients experienced 1.18 (95%CI: 1.13–1.26) visits representing missed opportunities and had an average diagnostic delay duration of 4.05 days (95%CI: 3.48–4.51). An estimated 1.7% (95%CI: 1.1–2.2%) of patients experienced a diagnostic delay duration that lasted a week or longer.
Simulation results for estimated number and duration of delayed visits per patienta
Table 3 Long description
The table presents data on the number of missed opportunities per patient and the duration of delayed visits. It has 21 rows and 4 columns. The columns are labeled 'Metric', 'Number of patients (% of all patients)', and '95% CI'. The rows are grouped into two main sections: Number of missed opportunities per patient and Duration of delayed visits. For the number of missed opportunities per patient, the metrics include ≥ 1, ≥ 2, ≥ 3, ≥ 4, and ≥ 5, with corresponding patient numbers and percentages. The mean and median number of missed opportunities per patient are also provided. For the duration of delayed visits, the metrics include durations from 1 day to 14 days, with corresponding patient numbers and percentages. The mean and median delay duration in days are also included. Notable trends include a higher percentage of patients experiencing one missed opportunity and a shorter duration of delayed visits being more common.
Note. a Units in counts (or % of all patients) unless otherwise specified.
b The distribution and mean number of potential missed diagnostic opportunities each patient experienced along with the distribution and mean and median duration of diagnostic delays (in days) are presented. Potential missed diagnostic opportunities represent healthcare visits in which sign/symptoms were present, but measles was not diagnosed. Delay duration was defined as the time between the earliest potential missed diagnostic opportunity a patient experienced and their index diagnosis.
Table 4 presents the results of the exploratory risk factor analysis. Compared with patients who had commercial insurance, patients with Medicare or Medicaid were more likely to experience a delay (i.e., experience at least one missed diagnostic opportunity), although these results were statistically significant. Patients aged less than two were more likely to experience a delay. Finally, receipt of respiratory antibiotics to treat potential symptoms of measles during the diagnostic opportunity window was positively associated with the likelihood of experiencing a diagnostic delay.
Results for Firth logistic regression model estimating the odds of a patient experiencing a delay
Table 4 Long description
A table comparing the odds ratios (OR) and 95 percent confidence intervals (CI) for various risk factors affecting patient diagnostic delays. The table has 25 rows and 2 columns. Column headers are Risk factor and OR (95 percent CI). Row labels include Database source, Age, Female, Rural location, Respiratory antibiotic prescription during delay window, Month, and Year. Each row provides the OR and CI for different risk factors. Notable trends include higher odds ratios for certain age groups and specific months, indicating increased likelihood of diagnostic delays.
Appendix Table 2 summarizes the findings of the sensitivity analysis comparing results of the 2- and 3-week diagnostic opportunity windows. In general, the results were relatively similar. When a 3-week window was used, the mean number of missed opportunities per patient and percent of patients experiencing a delay increased by approximately 0.8% and 5.8%, respectively, while the mean duration of delay increased by around 20.7%.
Appendix Table 3 presents the estimates for the duration and number of missed diagnostic opportunities that individual patients experienced when we included all visits instead of SSD visits only. We estimated that 19.2% (95%CI: 17.3–21.3%) of patients experienced a missed opportunity, with an average of 1.20 (95%CI: 1.15–1.25) missed opportunities per patient and an average delay of 4.81 days (95%CI: 4.32–5.33).
Discussion
Our study found that approximately 6.9% of patients with measles experienced at least one missed diagnostic opportunity involving an SSD-related visit during the two weeks before diagnosis, with an average delay of approximately four days. While this finding suggests diagnostic delays for measles are not common overall, given the highly transmissible nature of the disease, any delays are still an important public health concern. Furthermore, we found that diagnostic delays were associated with age and receipt of respiratory antibiotics.
Numerous case reports and outbreak descriptions have shown that patients with measles may have multiple healthcare visits before they receive the correct diagnosis. Reference Maltezou and Wicker21–Reference Clegg, Linchangco and Arwady26 However, these studies are limited due to the uncertainty surrounding the exact onset date of measles symptoms and because they did not account for coincidental conditions with clinical presentations similar to measles. In contrast, the case-crossover and bootstrapping approach that we employed enabled us to estimate which visits were attributable to measles and which were attributable to unrelated diagnoses.
Fortunately, we estimated that relatively few patients experienced multiple missed opportunities. For example, we estimated that only 0.2% of patients experienced three or more missed opportunities. Even when we considered a visit for any reason during the diagnostic window, only 0.6% of patients experienced three or more missed opportunities before diagnosis. Nevertheless, because measles is so transmissible, even a small number of healthcare visits attributable to missed opportunities will expose healthcare professionals and patients. Reference Steingart, Thomas, Dykewicz and Redd20,Reference Miranda, Falcao and Dias33–Reference Dales and Kizer35 The risk of transmission is increased when patients spend time in busy waiting rooms or clinics because virus aerosols remain airborne for hours. Reference Bloch, Orenstein and Ewing15,Reference Remington, Hall, Davis, Herald and Gunn16 Thus, even a few missed opportunities can amplify the cost and burden of infection prevention and public health efforts. A review of cost estimates associated with outbreaks found that the median cost for healthcare and contact tracing to be $32,805 per case, Reference Pike, Leidner and Gastañaduy36 and the per-case cost has been as high as $142,000. Reference Sundaram, Guterman and Omer37
We identified several factors associated with missed opportunities for diagnosing measles. First, we found that patients under the age of two were more likely to experience a missed diagnosis, possibly because these children cannot describe their symptoms well. Second, receipt of an antibiotic prescription was also associated with delays. Interestingly, antibiotic prescriptions have been associated with diagnostic delays for other infectious diseases including histoplasmosis, Reference Miller, Arakkal and Koeneman27 tuberculosis, Reference Miller, Arakkal and Koeneman28 pertussis, Reference Evans, Arakkal, Cavanaugh, Newland, Polgreen and Miller29 and sepsis. Reference Struble, Arakkal, Cavanaugh, Polgreen and Miller30 The association between antibiotic prescribing and diagnostic delays may involve two inter-related processes: (1) patients who are misdiagnosed may be prescribed antibiotics in response to their symptoms, and (2) patients who receive antibiotics may delay seeking another diagnosis while they wait to see if their symptoms resolve on the antibiotics, increasing the time until diagnosis.
We found that most missed opportunities to diagnose measles occurred in outpatient settings (81.0%), compared with ED (14.4%) and inpatient settings (3.9%). Outpatient settings often have fewer infection control resources than hospitals, making diagnostic delays in these settings particularly concerning. This finding highlights the need for targeted efforts to reduce diagnostic delays and strengthen infection control measures in outpatient care.
Our study has several limitations. First, we lacked access to laboratory data to confirm measles diagnoses. Accordingly, we only included patients who had a diagnostic test for measles. Second, we also lacked clinical notes to verify related symptoms, so we may have missed symptomatic visits not captured by diagnosis codes. To address this limitation, we conducted a sensitivity analysis including all visits as potential missed opportunities. This approach increased the estimated proportion of patients experiencing a missed opportunity (6.9% vs 19.2%), though the average number of missed opportunities per patient was nearly identical (1.18 vs 1.20) and the delay duration was similar (4.05 vs 4.81 days). Second, our dataset included individuals with employer-sponsored insurance or privately managed Medicare or Medicaid, limiting generalizability to uninsured populations. Third, missed opportunities to diagnose measles may decrease after a case is identified in the community, as close contacts (e.g., household members) should be easier to diagnose. However, we lacked sufficient secondary household cases to evaluate this phenomenon. Finally, our risk factor analysis was exploratory, so effect estimates should be interpreted cautiously rather than as definitive causal effects.
Measles cases are increasing in the US. While our findings suggest measles diagnostic delays may be infrequent, they still represent a major public health concern. Because diagnostic delays expose other patients and healthcare workers to a highly transmissible virus, any delays in diagnosing measles will increase the likelihood of additional cases in healthcare settings and in the community, highlighting the importance of reducing these delays. Collectively, in the setting of an outbreak, our results highlight the need to consider early testing for measles among patients with respiratory symptoms to prevent secondary measles cases attributable to diagnostic delays.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2025.10393.
Acknowledgments
None.
Financial support
This work was supported in part by the Centers for Disease Control and Prevention contract numbers U54CK000613 (Prevention Epicenters Program) and U01CK000594 (Modeling Infectious Diseases in Healthcare Network: MInD – Healthcare) and the Agency for Healthcare Research and Quality grant number R01HS027375.
Competing interests
All authors report no conflicts of interest relevant to this article.
