Impact of Delays in Diagnosis on Healthcare Costs Associated With Blastomycosis, Coccidioidomycosis, and Histoplasmosis in a Commercially Insured Population

Abstract

Among patients with blastomycosis (n = 281), coccidioidomycosis (n = 1920), and histoplasmosis (n = 2180), 62% experienced diagnostic delays (mean 29 days). Patients who experienced delays incurred average excess healthcare costs of up to $15 648 (95% confidence interval: $8600–$22 695) compared with those without a delay. Earlier diagnosis may help reduce excess costs.

Patients with blastomycosis, coccidioidomycosis, and histoplasmosis frequently experience diagnostic delays. In his study, patients who experienced delays incurred average excess healthcare costs of up to $15,648 compared with those without a delay.

Blastomycosis, coccidioidomycosis, and histoplasmosis are infections caused primarily by inhalation of environmental fungi in certain geographic areas [1]. Infection can range from asymptomatic to life-threatening disseminated disease, but pulmonary symptoms are common. Laboratory tests are required to distinguish blastomycosis, coccidioidomycosis, and histoplasmosis from other causes of community-acquired pneumonia [1]. Patients with these fungal diseases often visit healthcare providers multiple times before being correctly diagnosed, resulting in unnecessary testing, medical procedures, and antibiotic prescriptions. Previous studies evaluated healthcare use and costs associated with diagnostic delays in coccidioidomycosis at single centers in Tucson and Phoenix, Arizona [2, 3]. However, this association has not been examined for coccidioidomycosis in other areas or for blastomycosis or histoplasmosis. Using a large commercial health insurance claims database, we sought to quantify direct healthcare costs associated with delays in diagnosis for patients with blastomycosis, coccidioidomycosis, and histoplasmosis.

METHODS

Data Source

We used the Merative^TM MarketScan® Commercial/Medicare Database, which contains health insurance claims data submitted by large employers and health plans for >49 million employees, dependents, and retirees with employer-sponsored health insurance throughout the United States during 2017–2022. The data include information about hospitalizations, outpatient visits, outpatient prescriptions, and costs, and have been widely used in public health and health services research [4].

Study Population

We identified patients with noncapitated health plans and ≥1 International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code for blastomycosis (B40), coccidioidomycosis (B38), or histoplasmosis (B39) during January 1, 2017 to July 31, 2022. For each of the 3 cohorts, the index date was the date this ICD-10-CM code was first used during the study period. We required that the diagnosis not be listed on a laboratory or imaging claim alone, to reduce the possibility of selecting patients for whom the healthcare provider was trying to rule out the disease. We also required continuous enrollment in the database in the 180 days before to 365 days after the index date. To capture incident disease, we excluded patients who had an ICD-10-CM code(s) for blastomycosis, coccidioidomycosis, or histoplasmosis in the 180 days before the index date. For patients with histoplasmosis, we excluded those with a histoplasmosis diagnosis code assigned by an eye care provider, as these likely represent patients with presumed ocular histoplasmosis [5].

We examined demographic characteristics; underlying conditions (on or in the 90 days before the index date); signs, symptoms, and respiratory illnesses that are frequent misdiagnoses among patients with blastomycosis, coccidioidomycosis, or histoplasmosis (referred to as “compatible symptoms” for simplicity) in the 90 days before the index date; and disease form (ie, pulmonary, disseminated, other or unspecified) (Supplementary Table 1).

Features of interest in the 90 days before the index date included cost of outpatient visits and hospitalizations for compatible symptoms, and cost of outpatient systemic antibiotics. Features of interest on or in the 365 days after the index date included cost of blastomycosis, coccidioidomycosis, or histoplasmosis-related outpatient visits and hospitalizations, and cost of outpatient systemic antifungals. All costs are the sum of insurer payments plus patient out-of-pocket payments.

Analysis

Time from healthcare seeking to diagnosis (“delay time”) was defined as the number of days between the first visit for compatible symptoms and the index date. We summed all costs per patient and adjusted costs to the January 2024 monthly inflation-adjusted costs using the U.S. Bureau Labor of Statistics’ medical care in U.S. city average index (https://www.bls.gov/data/). We examined delay time and total costs for coccidioidomycosis and histoplasmosis separately and all 3 diseases combined, stratified by demographic and clinical characteristics. We did not present blastomycosis results separately due to small sample size.

Because the association between diagnostic delay and cost may be confounded by certain characteristics (eg, patient age and underlying medical conditions), we created probability weights to equalize the sample size and patient composition across the preindex date time frame. To address both the heteroscedasticity and the high frequency of zero-value delay time observations, we used quantile binning [6] to assign equal numbers of observations to each of 5 categories of delay times greater than zero. Zero delay time (eg, patient did not have a preindex date visit for compatible symptoms) received its own category. We then calculated propensity scores of these bins using multinomial logistic regression. Next, we used overlapping probability weighting to reduce potential confounding, balancing the populations within each time bin. We used 2 outcome models to evaluate the relationship between binned delay time and cost with corresponding 95% confidence intervals. The first model was a zero-truncated linear regression, which excluded patients with a zero-day delay to evaluate the association between costs and delays greater than zero days. The second model was a categorical regression, which compared costs between patients with delays of 1–30 days, 31–60 days, and 61–90 days and the reference group (zero-day delay), which account for differences between patients with various delay times and those with no delay. Analyses were conducted using the Merative™ MarketScan® Treatment Pathways analysis tool, SAS version 9.4 (Cary, North Carolina), and R program version 4.4.0 (Vienna, Austria).

Patient Consent Statement

This study did not include factors necessitating patient consent.

RESULTS

The analytic cohort comprised 4381 patients (n = 281 blastomycosis, n = 1920 coccidioidomycosis, and n = 2180 histoplasmosis) (Supplementary Table 2). In total, 62.2% of patients experienced any diagnostic delay (blastomycosis, 55.9%; coccidioidomycosis, 71.3%, histoplasmosis, 55.1%). Mean delays were 26 days (standard deviation [SD]: 31) for blastomycosis, 30 days (SD: 30) for coccidioidomycosis, and 28 days (SD: 33) for histoplasmosis. Overall mean per-patient costs were $37 217 (SD: $109 681) for blastomycosis, $18 484 (SD: $64 721) for coccidioidomycosis, and $19 609 (SD: $68 794) for histoplasmosis. Patients with underlying conditions had longer delays (38 [SD: 32] vs 25 [SD: 31] days) and higher average costs ($34 955 [SD: $95 758] vs $13 359 [SD: $53 575]) than those without underlying conditions. Among all patients and across all time points, 97.9% had ≥1 outpatient visit, 22.4% were hospitalized, 42.4% received outpatient antibiotics, and 39.0% received outpatient antifungals; mean per-patient costs were $4714 (SD: $10 980) for outpatient visits, $147 362 (SD: $180 651) for hospitalizations, $80 for outpatient antibiotics, and $2021 for outpatient antifungals (Supplementary Table 3).

The zero-truncated linear model showed an average increase per day of delay time of $131 (95% CI: $55–$207) for blastomycosis, coccidioidomycosis, and histoplasmosis combined, $175 (95% CI: $36–$314) for coccidioidomycosis, and $38 (95% CI: -$148–$224) for histoplasmosis. Only the results for all 3 diseases combined and coccidioidomycosis alone were statistically significant.

The categorical model showed that among all 3 diseases, patients with delays of 1–30 days, 31–60 days, and 61–90 days incurred average increases of $12 433 (95% CI: $7147–$19 564), $13 790 (95% CI: $6267–$21 313), and $15 648 (95% CI: $8600–$22 695) respectively, compared with the reference group (Figure 1). Among patients with coccidioidomycosis, those with delays of 1–30 days, 31–60 days, and 61–90 days incurred average increases of $4930 (95% CI: −$4226–$14 085), $4914 (95% CI: −$5092–$14 919), and $15 355 (95% CI: $5728–$24 982), respectively. The comparisons for patients in the 1–30 and 31–60-day delay groups were nonsignificant. Among patients with histoplasmosis, those with delays of 1–30 days, 31–60 days, and 61–90 days incurred average increases of $15 551 (95% CI: $7115–$23 986), $15 223 (95% CI: $6408–$24 039), and $14 614 (95% CI: $7152–$22 077), respectively.

Figure 1.

Excess direct medical costs (adjusted to January 2024 US dollars) by diagnostic delay time category among patients with blastomycosis, coccidioidomycosis, and histoplasmosis. † Nonsignificant estimates. Negative values represent mean cost that is less than the reference of no delay.

DISCUSSION

In this commercial health insurance claims data analysis, healthcare costs for patients with blastomycosis, coccidioidomycosis, and histoplasmosis and diagnostic delays were significantly higher (excess costs >$15 000 per patient) than those without delays, similar to previous studies of patients with coccidioidomycosis in Arizona [2, 3, 7]. Costs increased with each day of delay, for coccidioidomycosis alone as well as for blastomycosis, coccidioidomycosis, and histoplasmosis combined.

Diagnostic delays were frequent, affecting >70% of patients with coccidioidomycosis, comparable to a previous study [8], and >55% of patients with histoplasmosis and blastomycosis. Less frequent delays among patients with histoplasmosis and blastomycosis could be related to incidental diagnosis of histoplasmosis [9] and the potentially higher severity of blastomycosis, supported by its higher hospitalization rate (Supplementary Table 3). The average delay times we observed were consistent with previous studies (coccidioidomycosis: 23–38 days, histoplasmosis: 23–40 days, and blastomycosis: ≥ 24 days) [2, 8–13]. In one study, approximately 20% of patients with coccidioidomycosis experienced a diagnostic delay ≥90 days [3]; however, we chose a shorter preindex period to estimate costs more conservatively and reduce the possibility of capturing unrelated costs.

Our study's main limitation is potential disease and symptom misclassification inherent in medical coding data. Other limitations include the inability to analyze blastomycosis separately or stratify by hospitalization status or severity in the main analysis. Furthermore, our cost estimates represent only direct medical costs; true costs accounting for indirect costs are likely to be substantially higher. Finally, these results may not represent people without commercial health insurance.

This analysis demonstrates substantial excess medical costs associated with diagnostic delays for blastomycosis, coccidioidomycosis, and histoplasmosis, supporting the need for increased healthcare provider suspicion among patients with compatible symptoms. Increased awareness of and earlier testing may help reduce excess costs and likely improve patient outcomes.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.