Delayed Diagnoses in Patients With Dizziness in the US Commonwealth of Virginia and the Tidewater Region

Objective: In a region of approximately 1.7 million people (Tidewater, coastal Virginia), identify secondary diagnoses in persons with dizziness. Methods: This cross-sectional study utilizing TriNetX included individuals in the region of interest diagnosed with dizziness between 2010 and 2020. Subsequent diagnoses of vestibular disease or medical conditions possibly associated with dizziness in the same subjects were catalogued. Results: During the study period, 31,670 subjects were identified with diagnoses of dizziness as a symptom; 18,390 subjects were subsequently given a dizziness-related nonvestibular diagnosis, and 930 were given a subsequent vestibular disease diagnosis. The proportion of subjects diagnosed with vestibular disease (3%) after the dizziness diagnosis is far below expected norms (25%–34%) in the general population. There were greater proportions of delayed diagnoses of labyrinth dysfunction (odds ratio [OR], 4.8; P < 0.0001), superior semicircular canal dehiscence (OR, 3.1; P = 0.0023), otolith disease (OR, 3.1; P = 0.0023), among others, and a decreased proportion of delayed diagnosis of benign paroxysmal positional vertigo (OR, 0.56; P < 0.0001). Conclusions: The discrepancy between expected and observed prevalence in our region indicates that vestibular disease is likely underdiagnosed.

Dizziness is a common chief complaint in primary care clinics and emergency departments (1)(2)(3)(4).Evaluating the symptom of dizziness can be challenging for clinicians as the diagnosis relies heavily on the patient-reported history, during which patients often use the nonspecific term dizziness interchangeably with vertigo; moreover, patient responses are often inconsistent (2,5).The broad differential diagnosis of dizziness includes both vestibular and nonvestibular causes (1,2,6).
A recent study of regional prevalence in the Tidewater region of Virginia revealed that dizziness (symptom) was diagnosed in approximately 3% of patients, while vestibular disease (VD) was diagnosed in approximately 1% of patients (7).Given the substantial proportion of dizziness (symptom) diagnoses rendered without an underlying diagnosed medical cause in our region, we sought to identify temporally-associated subsequent diagnoses that could possibly relate to the dizziness symptom.Vestibular and nonvestibular diagnoses were considered, with the assumption that a second diagnosis following the symptom of dizziness indicated a potential cause of the symptom.

METHODS
This study was approved by the Eastern Virginia Medical School institutional review board.Patient data for this study were obtained from the global health research network, TriNetX, containing deidentified clinical data for all patients seen at Sentara Healthcare facilities.Included in this cross-sectional study were adults between the ages 18 and 89 who have been diagnosed with the dizziness symptom (primary study cohort) at a Sentara Healthcare affiliated practice from January 1, 2010 to December 31, 2020.Data regarding sociodemographic information, diagnoses, and comorbidities were collected.
TriNetX was searched by International Classification of Diseases, 10th edition diagnosis codes for dizziness (symptom R42) and a variety of VD diagnoses and nonvestibular diagnoses (Table 1) within the proposed study period.From the population of all dizziness diagnoses (symptom R42), 2 separate cohorts were isolated: 1 cohort of VD diagnoses and 1 cohort of dizziness-related nonvestibular disorder (NVD) diagnoses, each occurring subsequent to a dizziness diagnosis by one or more days.The control group utilized in this study, for comparison to the cohort of VD diagnoses following a dizziness diagnosis, comprised of the total number of VD diagnosis in the region of interest (ROI) during the study period, as recently published in our related study (7).
Chi-squared tests were performed to analyze the observed differences between groups.Odds ratios (ORs) and P values were calculated for the prevalence of VD diagnoses subsequent to dizziness (symptom R42) compared to all diagnoses of VD in the database for the study period (control).An OR was considered significant if the value was not close to 1.We used a 95% confidence interval with a P value <0.05 signifying statistical significance.

Region of Interest (ROI)
Prior work (7) outlined the assumptions and limitations of TriNetX as a source and appraisal of the subject selection methodology.No substantial changes in the population have occurred during the interim to alter rationale for utilizing TriNetX to study the ROI.

Study Sample
Demographic data for this population are outlined in Table 2.

Dizzy Patients Later Receiving a Contributing Nonvestibular Diagnosis
During the 10-year study period, there were 31,670 diagnoses of dizziness (symptom R42) (7).
We identified 18,390 subjects (58.1%) in which a subsequent diagnosis of NVD was rendered, plausibly accounting for the dizziness symptom.These individuals did not receive any VD diagnoses during the study period.NVD diagnosed subsequent to a dizziness diagnosis, and their frequencies, are outlined in Table 3.The most common NVD included essential hypertension (63%), diabetes mellitus type II (25%), anxiety disorder (21%), chronic ischemic heart disease (17%), and depressive episode (16%).Sensorineural hearing loss was diagnosed infrequently (6%).

DISCUSSION
Our research (7) is the first to quantify burdens of dizziness and VD in the region.Delayed or missed diagnoses of dizziness can magnify physical and mental disability (8,9).Most cases of dizziness do not receive a subsequent diagnosis that could account for the symptom.The present study identifies key differences in the population of patients who receive a diagnosis of VD or NVD subsequent to a diagnosis of dizziness and provides a critical comparison against all patients with dizziness and VD.
Concerning VD alone, the conditions diagnosed in greater proportion after assignment of the dizziness symptom (R42) are chronic otologic conditions, typically diagnosed by a specialist.A retrospective analysis of patients referred to a specialized center reported fewer diagnoses of unclear dizziness from 70% to 10% upon referral, validating the assumption that specialists are less likely to apply the dizziness symptom diagnosis (10).Consequently, subjects with dizziness who are subsequently diagnosed with VD (about 3%) plausibly reflect the proportion of individuals accessing specialist care to comprehensively evaluate and manage the dizziness symptom.
Comparing the cohort of all subjects with VD and the cohort of subjects with dizziness receiving a subsequent diagnosis of VD (Fig. 1), we found that patients with BPPV tend to receive this as a primary diagnosis, rather than a diagnosis subsequent to a prior encounter for dizziness.This can likely be explained by the ability to diagnose BPPV using the Dix-Hallpike test, a physical exam positioning maneuver that can be performed in any setting with a bed, sofa, table, or stretcher (11)(12)(13)(14).Other peripheral vestibular disorders, typically diagnosed subsequent to dizziness, require more nuanced physical examination-such as the observation of spontaneous nystagmus, the head impulse test, the head-shake test-or videonystagmography for diagnosis (3,(14)(15)(16).
In accordance with existing literature (17)(18)(19), VD was found to be more common in females compared to males, and there was a slightly greater proportion of females in the group with diagnoses of subsequent VD (70%) compared to the group with subsequent NVD (63%), as demonstrated in Table 2. Still, the frequency of females with subsequent VD diagnoses falls near the reported 66.7% predominance of all VD diagnoses in females, as seen in an epidemiologic survey of 70 million patients published by Hülse et al (17).Frequencies for age, race, and ethnicities did not show any substantial differences between the 2 groups.
Previous studies estimate that vestibular disorders account for 25%-34% of dizziness presentations (20,21).In this study, they were only identified in about 3% of dizziness cases.The incidence of VD diagnoses following a dizziness diagnosis in our ROI falls vastly short of the expected true prevalence of VD in patients with dizziness.In the ROI, perhaps VD is underdiagnosed, or alternatively, a preponderance of dizziness is attributable to NVD.Notably, no subsequent diagnosis was identified in 12,350 of the 31,670 dizzy subjects, or possibly, the symptom resolved.
In cases of acute vertigo, a complete neurological examination and supportive care may be appropriate for primary care settings.Notably, visiting emergency departments for complaints of dizziness can introduce unnecessary costs for patients, due to the overuse of neuroimaging that is insufficient in diagnosing peripheral causes of dizziness (4,22,23).Admittedly, arriving at one diagnosis or several diagnoses to account for the symptom of dizziness is a challenge.After a general medical evaluation, referral to specialized care providers, such as neurologists, cardiologists, psychiatrists, otolaryngologists, or vascular specialists, could increase diagnostic rates of chronic dizziness and vertigo.

Selection Bias Inherent to Study Design.
The database used in this study, TriNetX, introduced limitations in data collection and analysis.To offer large amounts of data from the entire Sentara Healthcare System, the platform compresses data and reports of all frequencies rounded to the  nearest factor of 10 and all percentage values rounded to the nearest whole number.This reporting error should not impact the interpretation of the results due to the large size of the population explored in this study.Additionally, due to the size of the population receiving subsequent diagnoses of NVD, data were compressed by the TriNetX database, reporting only the frequency and percent values for prevalence of disease from 9820 of 18,390 total diagnoses in the cohort.Enough diagnoses are included in the reported percentage values that these should be generalizable to the whole population and are, therefore, reported in this study.Because frequencies for demographics were only reported in percentages, numerical frequency reported in the results was calculated based on the percentage value of the whole population number, and is, therefore, also a rounded value.

Diagnosis
We also consider the limitation that some dizziness diagnoses may have been rendered near the end of the study period and are, therefore, excluded from the cohorts of patients who received a follow-up diagnosis to explain their dizziness.Therefore, this study may report an underestimated number of  patients who received a follow-up diagnosis and may overestimate the number of patients who never received an additional diagnosis following a diagnosis of dizziness.However, the criteria to be included in the subsequent diagnoses cohort only required a minimum of 1 day between diagnoses; therefore, few cases were likely excluded.It was only possible to query TriNetX for this one or more day criteria, the database does not provide the duration of the interval between dizziness diagnoses and subsequent diagnoses for neither individual cases nor an average for the cohort.The TriNetX database did not enable us to collect data on frequency of patient consultation or referral to a specialist in our study population.Additionally, we chose not to investigate resolved dizziness in this study, since physicians do not always remove symptoms from problem lists in medical records, lending to overestimation of prevalence.

CONCLUSIONS
A majority of dizziness symptom diagnoses were followed by plausible (nonvestibular) causes (58%), while very few individuals with the dizziness symptom diagnoses were ultimately diagnosed with VD (3%).Many individuals with the dizziness symptom (39%) did not receive an additional qualifying diagnosis to explain the dizziness symptom.
One may surmise that a majority of individuals either experience spontaneous resolution of dizziness (improbable) or their symptom is incompletely evaluated (probable); comparisons to existing prevalence data (7) in the ROI reveal that VD is either very uncommon (improbable) or underdiagnosed (probable).We consider it important to deepen our understanding of these regional differences, because consequences of delayed or missed diagnoses include lost opportunities for early access to medical management or rehabilitation and prolonged disability.

FIG. 1 .
FIG.1.Disease frequencies from the total population of primary VD diagnoses compared with disease frequencies of VD diagnoses subsequent to a dizziness diagnosis.This bar graph displays a greater proportion of Labyrinthine dysfunction, Meniere disease, vestibular schwannoma, and peripheral vestibulopathy diagnosed subsequent to the dizziness symptom, and a decreased proportion of BPPV diagnosed subsequent to the dizziness symptom, compared to all diagnoses of primary VD.BPPV indicates benign paroxysmal positional vertigo; VD, vestibular disease.

TABLE 1 .
Comprehensive list of vestibular disease and nonvestibular disease diagnoses and associated ICD-10 codes included in the study population ICD-10 indicates International Classification of Diseases, 10th edition; TIA, transient ischemic attack.

TABLE 3 .
Frequencies of nonvestibular diagnoses following initial dizziness diagnosis between 2010 and 2020, grouped by organ system ICD-10 indicates International Classification of Diseases, 10th edition; TIA, transient ischemic attack.

TABLE 4 .
Frequencies of vestibular disease diagnoses following initial dizziness diagnosis between 2010 and 2020 ICD-10 indicates International Classification of Diseases, 10th edition; VD, vestibular disease.