Patient, Provider, and Health Care System Characteristics Associated With Overuse in Bronchiolitis

Abstract

BACKGROUND AND OBJECTIVES:

The American Academy of Pediatrics recommends against the routine use of β-agonists, corticosteroids, antibiotics, chest radiographs, and viral testing in bronchiolitis, but use of these modalities continues. Our objective for this study was to determine the patient, provider, and health care system characteristics that are associated with receipt of low-value services.

METHODS:

Using the Virginia All-Payers Claims Database, we conducted a retrospective cross-sectional study of children aged 0 to 23 months with bronchiolitis (code J21, International Classification of Diseases, 10th Revision) in 2018. We recorded medications within 3 days and chest radiography or viral testing within 1 day of diagnosis. Using Poisson regression, we identified characteristics associated with each type of overuse.

RESULTS:

Fifty-six percent of children with bronchiolitis received ≥1 form of overuse, including 9% corticosteroids, 17% antibiotics, 20% β-agonists, 26% respiratory syncytial virus testing, and 18% chest radiographs. Commercially insured children were more likely than publicly insured children to receive a low-value service (adjusted prevalence ratio [aPR] 1.21; 95% confidence interval [CI]: 1.15–1.30; P < .0001). Children in emergency settings were more likely to receive a low-value service (aPR 1.24; 95% CI: 1.15–1.33; P < .0001) compared with children in inpatient settings. Children seen in rural locations were more likely than children seen in cities to receive a low-value service (aPR 1.19; 95% CI: 1.11–1.29; P < .0001).

CONCLUSIONS:

Overuse in bronchiolitis remains common and occurs frequently in emergency and outpatient settings and rural locations. Quality improvement initiatives aimed at reducing overuse should include these clinical environments.

Bronchiolitis is the leading cause of hospitalization among infants and young children, with an annual cost of 1.7 billion dollars in the United States.¹ In 2014, the American Academy of Pediatrics (AAP) published updated clinical practice guidelines for children 1 to 23 months of age that recommended against routine use of β-agonists, corticosteroids, antibiotics, chest radiography, and viral testing.² Despite these guidelines, overuse of these low-value services continues.³

Overuse in bronchiolitis can cause patients harm and lead to unnecessary medical costs.⁴ For example, β-agonists can lead to increased oxygen consumption⁵; corticosteroids can cause irritability and immunosuppression; antibiotics can cause allergic reactions, alterations in the microbiome,⁶ and increasing antimicrobial resistance; and chest radiography may increase the risk of certain malignancies and lead to over diagnosis and increased cost.⁷ Sometimes a single unnecessary test or treatment can lead to a cascade of overuse with downstream harms and costs. For example, a chest radiograph can lead to the misdiagnosis of bacterial pneumonia and the prescription of antibiotics.⁸

Previous studies have shown that there is wide variation in bronchiolitis care.^8–16 Most studies have not included overuse in the outpatient setting^8–19 or were limited to a single insurance type.²⁰ Many occurred before the 2014 AAP guidelines.^8–15,17 We aimed to conduct a study of commercially and publicly insured children in inpatient, emergency, urgent care, and outpatient settings after the 2014 AAP guidelines were published. Our goals were to determine the magnitude of overuse and the patient, provider, and health care characteristics associated with overuse in bronchiolitis.

METHODS

Data Source

Data were obtained from the 2018 Virginia All-Payers Claim Database (VA-APCD). The VA-APCD contains deidentified information on patient claims, including visit dates, diagnostic and treatment codes, Health Cost Guidelines groupings, and the National Provider Identifier of the ordering provider.

Inclusion Criteria

Children aged 0 to 23 months with a diagnosis of bronchiolitis (International Classification of Diseases, 10th Revision code J21) who were continually enrolled in the 2018 VA-APCD were included.

Overuse Measures

Overuse measures were derived from Choosing Wisely recommendations and current AAP guidelines. Similar measures have been used in previous studies of claims databases.^20,21 For 3 medication measures (corticosteroids, antibiotics, and β-agonists), overuse was included if a prescription was claimed within 3 days after an initial bronchiolitis diagnosis. Of 2 service-based overuse measures (testing for respiratory syncytial virus [RSV] and chest radiographs), a service was included if it was claimed within 1 day after diagnosis. A composite measure was also created for counting the number of overuse measures recorded per unique child with bronchiolitis. The complete list of drugs and procedure codes are provided in Supplemental Table 3.

Patient, Provider, and Health Care System Characteristics

Patient-level covariates extracted from the VA-APCD included age at diagnosis, race, sex, and insurance type. Each provider National Provider Identifier was linked to the National Plan and Provider Enumeration System database,²² in which each provider is associated with a National Uniform Claim Committee domain of care (ie, emergency medicine, family medicine). The domains were classified into 5 groups: family medicine, pediatrics, pediatric specialist, emergency medicine, and other (Supplemental Table 4). We determined rurality of place of care (rural, suburban or town, urban) using the geographic codes created by the Education Demographic and Geographic Estimates Program from the National Center for Education Statistics.²³ The clinical setting was obtained through the Health Cost Guidelines grouping and classified as emergency department or urgent care, inpatient, or outpatient.

Statistical Analysis

Patient-, provider-, and practice-level characteristics were summarized as frequencies and percentages (Table 1). Generalized estimating equations with a Poisson distribution and log-link function were used to examine the associations between covariates and the overuse measurements. For each measure, we modeled the rate of the overuse service being provided. Each model included patient-level characteristics (age, sex, and insurance type), provider-level characteristics (specialty), and clinic-level characteristics (rurality, clinical setting). Because of the high degree of missingness, race was not included in the final model. To estimate prevalence rates, an offset term was included in each model, valuing 1 for the binary overuse measures and 5 for the composite measure. Hierarchical nesting of patients within providers was accounted for by specifying a compound symmetric dependence structure. For each model, adjusted prevalence ratios (aPRs), 95% confidence intervals (CIs), and P values were reported. All analyses were performed at a statistical significance level of .05 by using SAS version 9.4 statistical software (SAS Institute, Inc, Cary, NC).

TABLE 1.

Descriptive Statistics for Overuse in Bronchiolitis

Characteristics	Frequency (%)
Age, mo
0–11	4766 of 8308 (57)
12–23	3542 of 8308 (43)
Sex
Female	3511 of 8308 (42)
Male	4797 of 8308 (58)
Race
Asian American	207 of 8308 (2)
Black	1479 of 8308 (18)
White	1815 of 8308 (22)
Other or unknown	4807 of 8308 (58)
Insurance
Commercial	2025 of 8308 (24)
Public	6283 of 8308 (76)
Corticosteroids
Yes	737 of 8308 (9)
No	7571 of 8308 (91)
RSV testing
Yes	2183 of 8308 (26)
No	6125 of 8308 (74)
Antibiotics
Yes	1452 of 8308 (17)
No	6856 of 8308 (83)
β-agonists
Yes	1632 of 8308 (20)
No	6676 of 8308 (80)
Chest radiographs
Yes	1454 of 8308 (18)
No	6854 of 8308 (82)
No. overuse types experienced (per child)
0	3634 of 8308 (44)
1	2624 of 8308 (32)
2	1470 of 8308 (18)
3	442 of 8308 (5)
4	122 of 8308 (1.5)
5	16 of 8308 (0.2)
Provider specialty
Pediatrics	5377 of 8308 (65)
Emergency medicine	1456 of 8308 (18)
Family medicine	450 of 8308 (5)
Pediatric specialty	117 of 8308 (1)
Other	908 of 8308 (11)
Practice location
City	2158 of 8272 (26)
Rural	3384 of 8272 (41)
Suburban or town	2730 of 8272 (33)
Clinical setting
Emergency or urgent care	2420 of 8308 (29)
Inpatient	314 of 8308 (4)
Outpatient	5574 of 8308 (67)

Ethical Approval

Ethical approval was given by the Virginia Commonwealth University Institutional Review Board with an exempt determination (HM20019581).

RESULTS

We identified 177 344 children aged 0 to 23 months in the VA-APCD during the study period, of whom 8308 (5%) were diagnosed with bronchiolitis (Table 1). Most children in the cohort were 0 to 11 months (57%), male (58%), and publicly insured (76%). Most of the diagnosing providers were pediatric providers (65%). The most common location of care was rural (41%), followed by suburban (33%) and urban (26%). Most diagnoses were made in outpatient settings (67%), although many were made in emergency or urgent care settings (29%), including stand-alone urgent care facilities or hospital-affiliated urgent care facilities. Overall, 56% of children diagnosed with bronchiolitis received at least 1 low-value service. Nine percent of children with bronchiolitis were prescribed corticosteroids, 17% were prescribed antibiotics, 20% were prescribed β-agonists, 26% were tested for RSV, and 18% underwent chest radiographs.

Children with commercial insurance were more likely than those with public insurance to receive ≥1 low-value service (aPR 1.21; 95% CI: 1.15–1.30; P < .0001). Children seen in rural locations were more likely than children seen in cities to receive a low-value service (aPR 1.19; 95% CI: 1.11–1.29; P < .0001) (Table 2). Children treated by emergency and family medicine providers were more likely than those treated by pediatric providers to receive a low-value service (aPR 1.65 [95% CI: 1.49–1.82; P < .0001] and aPR 1.48 [95% CI: 1.31–1.67; P < .0001], respectively). Children in emergency (including urgent care) settings were more likely to receive a low-value service (aPR 1.24; 95% CI: 1.15–1.33; P < .0001) compared with children in inpatient settings.

TABLE 2.

aPRs, 95% CIs, and P Values for Various Measures of Overuse With Bronchiolitis

Covariate	Corticosteroids			β-Agonlsts			Antibiotics			RSV Testing			Chest Radiographs			≥1Forms of Overuse
	aPR	95% CI	P	aPR	95% CI	P	aPR	95% CI	P	aPR	95% CI	P	aPR	95% CI	P	aPR	95% CI	P
Agea 0–11 mo	0.66*	0.57–0.76*	<.0001*	1	0.92–1.09	.98	0.7*	0.64–0.78*	<.0001*	1.15*	1.06–1.24*	<.001*	0.98	0.90–1.06	.56	0.93*	0.88–0.97*	.002*
Female sexb	0.88	0.77–1.00	.06	0.97	0.89–1.06	.52	1.02	0.93–1.12	.69	1.04	0.97–1.11	.3	0.91*	0.84–0.98*	.01*	0.98	0.93–1.03	.36
Public Insurancec	0.77*	0.64–0.92*	.003*	0.84*	0.75–0.94*	.002*	0.71*	0.63–0.79*	<.0001*	0.82*	0.74–0.91*	<.0001*	1.01	0.90–1.14	.85	0.82*	0.77–0.87*	<.0001*
Locatlond
Rural	1.49*	1.15–1.93*	.003*	1.4*	1.20–1.64*	<.0001*	1.28*	1.10–1.48*	.003*	1.12	0.99–1.28	.07	0.89	0.77–1.02	.08	1.19*	1.11–1.29*	<.0001*
Suburban	0.84	0.62–1.13	.24	1.27*	1.07–1.51*	.006*	0.92	0.78–1.09	.36	0.93	0.80–1.08	.35	0.72*	0.62–0.85*	<.0001*	0.95	0.86–1.03	.22
Specialtye
Emergency	2.05*	1.55–2.70*	<.0001*	1.11	0.87–1.40	.4	0.98	0.79–1.22	.87	1.58*	1.35–1.85*	<.0001*	2.95*	2.42–3.60*	<.0001*	1.65*	1.49–1.82*	<.0001*
Family	3.92*	3.08–4.99*	<.0001*	1.03	0.83–1.27	.8	1.38*	1.14–1.66*	<.001*	1.14	0.93–1.40	.22	1.75*	1.35–2.25*	<.0001*	1.48*	1.31–1.67*	<.0001*
Other	2.08*	1.59–2.73*	<.0001*	1.05	0.87–1.27	.59	1.09	0.90–1.30	.38	1.25*	1.06–1.40*	.008*	2.34*	1.90–2.90*	<.0001*	1.37*	1.24–1.50*	<.0001*
Settingf
Emergency or urgent care	0.91	0.73–1.15	.44	0.53*	0.43–0.65*	<.0001*	0.84*	0.72–0.99*	.04*	1.37*	1.20–1.56*	<.0001*	2.69*	2.18–3.31*	<.0001*	1.14*	1.05–1.24*	.003*
Inpatient	0.89	0.60–1.33	.58	0.45*	0.32–0.63*	<.0001*	0.4*	0.26–0.62*	<.0001*	0.8	0.61–1.04	.1	3.15*	2.47–4.02*	<.0001*	0.92	0.79–1.08	.31

^aReference group: 12–23 mo.

^bReference group: male sex.

^cReference group: commercially insured.

^dReference group: city.

^eReference group: pediatrics.

^fReference group: outpatient.

^*Prevalence ratios significant at P < .05.

Regarding specific low-value services, commercially insured patients were more likely than publicly insured patients to receive corticosteroids, β-agonists, antibiotics, and viral testing (Table 2). There was no difference in receipt of chest radiographs by insurance status. Family medicine providers were most likely to prescribe corticosteroids and antibiotics, whereas emergency medicine providers were most likely to order chest radiographs and viral testing. There was no significant difference in the prescribing of b-agonists by provider type. Providers in rural locations were more likely than providers in other locations to order every type of overuse besides chest radiography and viral testing. β-agonists and antibiotics were most likely to be prescribed in the outpatient setting. Viral testing was most likely to be ordered in emergency and urgent care settings. Chest radiographs were most likely to be ordered in the inpatient setting. We conducted a sensitivity analysis with the composite measure, excluding the 4% of children who were seen as inpatients, and the results did not substantively change.

DISCUSSION

We found that 4 years after the most recent AAP guidelines, overuse of low-value services in bronchiolitis remains common. More than half of the children in this cohort (56%) received ≥1 low-value service. It should be noted that some have suggested that target levels for these types of overuse should not necessarily be 0.^3,24–26 For example, antibiotics may be justified if the provider suspects a concomitant bacterial infection. Recent estimates for achievable benchmarks of care derived from the results of quality improvement initiatives range from 7% to 19% for β-agonist use, 0% to 6% for corticosteroid use 1% to 19% for antibiotic use, 1% to 2% for RSV testing, and 4% to 32% for chest radiography.^3,24–26 These benchmarks were largely derived for hospitalized patients. In Spain, quality improvement initiatives have been able to reduce corticosteroid, β-agonist, and antibiotic use to 4%, 16%, and 10%, respectively, in outpatient settings.²⁷

The most common low-value service was RSV testing. The isolation of RSV may bring additional certainty to a provider’s diagnosis and increase parental satisfaction. However, there is no specific antiviral treatment of RSV, and testing has not been shown to lead to improved outcomes.² The authors of one study found that RSV polymerase chain reaction testing increased costs by ~385 per patient.²⁸ It should be noted that although we did not record testing that was specific for influenza, we included multiplex viral panels that may have included influenza. These panels can sometimes be used for cohorting purposes when a child is admitted to the hospital. A viral panel may also be used to determine the etiology of a fever of unknown origin or in the evaluation of a child who is medically complex or immunocompromised. Further research could exclude these specific clinical scenarios, although it would be difficult to distinguish viral panels that are solely ordered for cohorting purposes with claims data alone.

The second most ordered service was b-agonists. The 20% of children who received b-agonists in our study is probably an underestimate because we measured prescriptions and did not have access to medications given in the clinic. β-agonists have been shown to be ineffective at improving oxygen saturation, admission rate, and length of stay.^2,29,30 In 2006, the AAP clinical practice guidelines listed b-agonists as optional: “A carefully monitored trial of a-adrenergic or b-adrenergic medication is an option. Inhaled bronchodilators should be continued only if there is a documented positive clinical response to the trial using an objective means of evaluation (option).”³¹ In 2014, the wording was updated to reflect more recent evidence: “Clinicians should not administer albuterol (or salbutamol) to infants and children with a diagnosis of bronchiolitis (Evidence Quality: B; Recommendation Strength: Strong Recommendation).”² It is unclear if providers continue to prescribe b-agonists because of their reluctance to give up historical practices or the commission bias in which the provider prefers “doing something” over “doing nothing.”³²

Children aged <1 year were more likely to receive RSV testing compared with children aged 12 to 23 months. One explanation for this phenomenon is that providers may feel the need to obtain a conclusive RSV test result in this age group to reassure themselves or parents against the possibility of a more serious bacterial infection. Other providers may feel a heightened concern for apnea in children with RSV, although this association has been disproven.³³ In contrast, older children were more likely to receive corticosteroids and antibiotics compared with younger children. This may be due to providers’ concern for medication-related adverse events in younger ages or may indicate greater comfort in treating children along a clinical spectrum approaching asthma and bacterial pneumonia. Interestingly, there was no significant difference in the prescription of b-agonists by age, although it can be more challenging to distinguish between bronchiolitis and asthma in older age groups.

We found that commercially insured patients were more likely than publicly insured patients to receive all types of low-value services, except for chest radiographs. This finding is particularly noteworthy given that commercially insured patients tend to have higher out-of-pocket costs.³⁴ The authors of another study found that commercially insured children with bronchiolitis were slightly more likely than publicly insured children to receive viral testing and b-agonists but less likely to receive corticosteroids and antibiotics.²¹ One reason for the differences we observed may be that the comprehensiveness of the VA-APCD allowed us to control for provider type, location, and setting, which can confound differences in payer type. There are several possible reasons why commercially insured children may receive more low-value services. One reason is that in fee-for-service plans, providers are paid higher reimbursement rates for certain services by commercial insurers compared with Medicaid,³⁵ although this would not explain the difference we observed with prescription medications. Another explanation is that parents of commercially insured children may request low-value services more frequently. One study showed that parents of commercially insured children were less likely to request antibiotics,³⁶ but this may not hold true across other types of low-value services.

Family medicine providers and emergency medicine providers were more likely to order low-value services compared with pediatric providers. One explanation is that pediatric providers tend to see proportionally more children compared with family medicine providers, a pattern that has become even more pronounced over the last decade.³⁷ Another explanation is that pediatric guideline dissemination may occur more rapidly among pediatric providers compared with providers who see primarily adults.³⁸

Children seen in rural locations made up the largest group of children (41%). These children were more likely to experience ≥1 type of overuse, even after we controlled for provider specialty. Rural hospitals tend to be smaller,³⁹ and studies have shown that hospitals that have fewer patients with bronchiolitis tend to overuse low-value services in higher proportions.^40,41 Providers in rural locations may also be further from tertiary care and feel compelled to attempt interventions before transport of sicker children to urban centers.⁴²

Children seen in emergency and urgent care settings were most likely to receive ≥1 low-value service. Our findings are similar to the findings of a survey of providers showing that individuals practicing in the emergency department were more likely to order b-agonists compared with other providers.⁴³ Two-thirds of the children with bronchiolitis in this study were seen as outpatients. Children in outpatient settings were more likely to receive b-agonists and antibiotics compared with children seen in emergency and inpatient settings. This is significant because, with few exceptions,^27,44–47 most studies of overuse in bronchiolitis have targeted hospitals and emergency departments.^8–19 International studies have shown that reaching benchmarks in the outpatient setting is possible.²⁷ In the United States, it may be more challenging to conduct quality improvement efforts outside of urban academic children’s hospitals because of decentralized and unique electronic health records and lack of collaborative relationships with unaffiliated clinics. The finding that there tended to be less overuse in inpatient settings could suggest that an individual child receives low-value care in a clinic or emergency department before being admitted. For example, if an inpatient provider knows that a β-agonist was trialed in a clinic without benefit to the patient, the provider may not feel the need to trial a β-agonist while that child is admitted in the hospital. Similarly, if a viral test was ordered in the emergency department, a provider in the inpatient setting would not need to repeat the test after the child was admitted. Conversely, when a child is sent from a clinic to the emergency department for higher-acuity care, the provider working in the emergency department may feel compelled to trial additional therapies.

In addition to the missingness of the race variable mentioned above, there were several limitations to this study. First, in this study, we used claims data to determine the diagnosis of bronchiolitis. Diagnosis codes can be imprecise and generated by individuals with limited medical training.⁴⁸ Second, we excluded children without continuous coverage. This may have differentially excluded immigrant children who may not qualify for Medicaid or the Children’s Health Insurance Program. Third, the VA-APCD does not contain patients who have military insurance or certain commercial insurances or who are uninsured. We do not know if these patients experienced more or less overuse. Fourth, there may be differential coding misclassification away from bronchiolitis and toward more severe conditions, such as respiratory failure, in higher-acuity settings, such as emergency departments, and hospitals driven by reimbursement at the facility level. Fifth, we did not adjust for acuity in the models. However, we do not think that this was a major constraint of the study. Because we could not distinguish between children admitted to an ICU and those admitted to a general ward, even children who were critically ill would be classified as inpatients. Children in the inpatient setting experienced lower overall overuse, although they tend to be sicker and more medically complex than children in the outpatient setting. Furthermore, some might argue that overuse in children who are sicker is even more problematic because of their potential for medication-related adverse events and medication interactions.

CONCLUSIONS

We found that overuse associated with bronchiolitis frequently occurred in emergency and outpatient settings and rural locations. This suggests that quality improvement initiatives aimed at reducing overuse in bronchiolitis should include these clinical settings.

WHAT’S KNOWN ON THIS SUBJECT:

The American Academy of Pediatrics recommends against routine use of b-agonists, corticosteroids, antibiotics, chest radiographs, and viral testing in bronchiolitis, but use of these modalities continues. Little is known about provider, patient, and system characteristics associated with overuse in bronchiolitis.

WHAT THIS STUDY ADDS:

Commercially insured children were more likely than publicly insured children to receive low-value services. Emergency and family medicine providers ordered low-value services more frequently than pediatric providers. Overuse in bronchiolitis frequently occurred in emergency and outpatient settings and rural locations.

ABBREVIATIONS

AAP

American Academy of Pediatrics

aPR

adjusted prevalence ratio

CI

confidence interval

RSV

respiratory syncytial virus

VA-APCD

Virginia All-Payers Claims Database