To evaluate variation in resource use for children with acute respiratory tract illness (ARTI) by vaccination status.
We conducted a retrospective cohort study of children 0 to 16 years of age at 5 children’s hospitals with 1 of 4 ARTI diagnoses (pneumonia, croup, asthma, and bronchiolitis) between July 2014 and June 2016. The predictor variable was provider-documented up-to-date (UTD) vaccination status (yes or no). Outcomes were receipt of each of the following tests or treatments (yes or no): complete blood cell count, blood cultures, C-reactive protein (CRP) level testing, viral testing, influenza testing, pertussis testing, chest radiographs, neck radiographs, antibiotics, and corticosteroids. We generated multivariable logistic regression models to examine the associations between our predictor and outcomes.
Of the 2302 participants included in analysis, 568 (25%) were diagnosed with pneumonia, 343 (15%) were diagnosed with croup, 653 (28%) were diagnosed with asthma, and 738 (32%) were diagnosed with bronchiolitis. Most (92%) vaccination statuses were documented as UTD. Across conditions, children whose vaccination status was documented as not UTD had higher adjusted odds of receiving a complete blood cell count, blood culture, CRP level testing, and influenza testing (P < .001). Children with pneumonia whose vaccination status was documented as not UTD had higher adjusted odds of receiving CRP level testing and influenza testing (P < .001). Children with croup whose vaccination status was documented as not UTD had higher adjusted odds of receiving blood cultures (P < .001).
Children with ARTI whose vaccination status was documented as not UTD had higher odds of undergoing laboratory testing compared with children whose vaccination status was documented as UTD.
Children who are undervaccinated are at risk of vaccine-preventable diseases, including acute respiratory tract illnesses. Additionally, they have higher rates of emergency department visits and inpatient hospitalizations compared with children who are fully vaccinated.
There is variation in care for children who are undervaccinated compared with children who are fully vaccinated. Children who are undervaccinated and receive hospital-based care for acute respiratory tract illnesses undergo more laboratory testing than children who are fully vaccinated.
Acute respiratory tract illnesses (ARTIs) account for significant morbidity in children and are a main cause of pediatric emergency department (ED) visits and hospitalizations annually.1–6 Despite their high frequency, significant variation in the care of children with ARTIs has been observed.7–11 The reason for this variation remains incompletely understood.
One potential explanation for this variability in the care of children with ARTIs is a child’s vaccination status. Children who are undervaccinated for vaccine-preventable ARTIs, such as pneumonia, influenza, and pertussis, are at a higher risk of developing these ARTIs.12–14 In addition, children who are undervaccinated have higher all-cause hospital admissions and more frequent ED visits.15–17 Because of their higher risk and/or more frequent use of health care in hospital-based care settings, providers may conduct more testing and treatment of undervaccinated children presenting with ARTIs.
Our primary objective for this study was to evaluate variation in ED and inpatient resource use for children with ARTI by vaccination status. We hypothesized that undervaccinated children presenting to the ED or hospital with ARTI would receive more laboratory and radiographic testing and medication treatment.
Methods
Study Population
We conducted a retrospective cohort study of children who presented with ARTI between July 2014 and June 2016 to 5 freestanding academic children’s hospitals participating in the Pediatric Research in Inpatient Settings network. ARTI was defined as community-acquired pneumonia (CAP), croup, asthma, and bronchiolitis. Children were included in the study if they were between 2 weeks and 16 years old, if the family spoke English or Spanish, and if they had 1 of the 4 aforementioned ARTI conditions. We focused on ARTI because (1) ARTI is a common cause of pediatric hospital admission and (2) a child’s vaccination status is relevant to the diagnosis of ARTI because many vaccine-preventable diseases (VPDs), such as influenza, pertussis, and pneumococcal disease, are in the differential diagnosis. Subjects were excluded if they (1) had an underlying condition that would alter the routine childhood vaccination schedule (including requiring additional vaccines), such as immune deficiencies, HIV, or asplenia, per the recommendations of the Advisory Committee on Immunization Practices18 or (2) had a chronic medical condition that would alter the standard of care for ARTI, including history of cardiac disease, anatomic airway abnormalities, cystic fibrosis, neuromuscular disease, bronchopulmonary dysplasia, or chronic lung disease. All children in the study who were diagnosed with CAP, asthma, and bronchiolitis were admitted to the hospital. Children with croup were either seen in the ED and admitted or seen in the ED and discharged. We opted to include children with croup who were discharged from the ED given the lower rates of admission for croup. If a child had multiple hospitalizations during the study period, only the first hospitalization was included in this analysis. This study was approved by the Western Institutional Review Board.
Variables
The primary predictor variable was provider-documented up-to-date (UTD) vaccination status (yes or no) at the time of hospital presentation. This variable was abstracted from the electronic medical record (EMR) from either the ED or admission note. Previous studies have revealed that provider-documented UTD status may not be reflective of true vaccination status.19–21 However, we selected it as our predictor variable because it represented the information that providers had at the point of care when making clinical decisions during the hospital visit. Subjects were excluded if there was no provider documentation of vaccination status in the EMR. Of note, all hospitals in the study had a specific vaccination question included in their EMR note templates for ED notes and admission history and physical notes. At the time of this study, none of the 5 hospitals received vaccination data from local population-based immunization information systems (IISs). Thus, provider-documented UTD status was the only vaccination data included in the EMR.
The outcome variables of interest were receipt of laboratory tests, radiographic studies, and medication treatment of ARTI. We used the Pediatric Health Information System (PHIS) database to obtain outcome data. PHIS is a de-identified administrative database that contains information on diagnoses, laboratory investigations, radiographic studies, and medications for 47 children’s hospitals across the United States. For each participant, we examined the PHIS to identify laboratory tests, radiographic studies, and medications they received during the current hospitalization. We a priori selected laboratory tests, radiographic studies, and medications that were clinically relevant to ARTI: complete blood cell (CBC) count, blood culture, C-reactive protein (CRP) level testing, respiratory viral testing, influenza testing, pertussis testing, chest radiographs, neck radiographs, antibiotic use, and corticosteroid use. Clinical relevance was determined from clinical practice guidelines, a literature review, and consensus among study authors.7,11,22–24 For respiratory viral testing, we included testing for any of the following respiratory viruses: adenovirus, metapneumovirus, respiratory syncytial virus, rhinovirus, parainfluenza, and influenza. We also included nonspecific viral tests, such as viral cultures and unspecified viral polymerase chain reaction tests, and nucleic acid amplification tests. We examined influenza and pertussis tests as separate variables because both of these tests were specifically related to VPDs.
We queried the PHIS for at least 1 occurrence (yes or no) of each of the aforementioned tests and treatments during the included visit. We examined the mean number of diagnostic laboratory tests and radiographic studies each participant received during hospitalization (range 0–8) as well as the number and type of antibiotics each participant received. Antibiotics were classified by type (ie, penicillin, cephalosporins, macrolides, etc) and categorized into 2 groups: narrow and broad spectrum. Narrow-spectrum antibiotics included penicillin and macrolides or a combination of penicillin and macrolide. Broad-spectrum antibiotics included cephalosporins, fluoroquinolones, aminoglycosides, antistaphylococcal antibiotics (such as vancomycin), or a combination therapy that included at least 1 of these agents.25,26
Covariates of interest included child age at presentation, sex, race and/or ethnicity, insurance, chronic disease status, admission to the ICU, and month of presentation. We categorized age into 4 groups (0–18 months, 19–35 months, 3–6 years, and >6 years) to correspond to the age categories typically used in population-level evaluations of vaccination coverage,27 although participants >6 years were aggregated because of small sample size. Chronic disease status was determined by using the Pediatric Medical Complexity Algorithm (PMCA).28 A chart review was performed for all subjects with complex chronic disease to ensure they met study eligibility criteria. The month of presentation was included to account for seasonal trends and categorized into influenza season (October–March) and noninfluenza season (April–September).29
Statistical Analysis
Univariate descriptive statistics were used to examine the cohort of subjects. We used χ2 tests and bivariate logistic regression to assess the relationship between (1) covariates and the predictor variable, (2) covariates and the outcome variables, and (3) our predictor and outcome variables. To test the independent associations between the predictor and outcome variables, we used multivariable logistic regression models that included covariates with a significant (P < .05) relationship with the predictor variable and outcome variables. We additionally adjusted models for length of stay (LOS) in days as a continuous variable to account for time in the hospital as a possible confounder for receiving more testing. Because variability in practice for these conditions could influence site-specific care, we included hospital site as a fixed effects variable, with clustering of SEs by site.
We generated multivariable generalized linear regression models to examine the relationship between the number of diagnostic tests received (0–8) by each individual by UTD status. We generated multivariable logistic regression models to examine binary use of each test and treatment. We examined models for the entire cohort of subjects as well as models for each of the 4 respiratory conditions. To account for multiple comparisons, we used a Bonferroni-corrected critical P = .006.
Sensitivity Analysis
None of the hospitals in our study had IIS data integrated in their EMR. Because of time and budget constraints, we were unable to obtain individual-level vaccine records for all participants, but we were able to obtain IIS vaccination data for the subjects included in the study from 1 study site. For this subset of subjects, we conducted a sensitivity analysis to compare IIS UTD status against resource use. Children’s vaccination status was considered UTD if the child had received all age-appropriate vaccine doses, including for influenza, per the Advisory Committee on Immunization Practices guidelines. We generated multivariable logistic regression models to examine binary use of each test and treatment by IIS UTD status. We examined models of the entire cohort of subjects as well as each of the 4 respiratory conditions. All models were adjusted for race and/or ethnicity, PMCA, season of presentation, and LOS.
Results
Of 2380 children identified with ARTI, 46 (2%) were excluded because of underlying conditions identified by an EMR review, and 32 (1%) had no documentation of vaccine status in their EMR (Fig 1). Of the remaining 2302 participants included in analysis, the mean age was 3.5 years (SD 3.7), and 59% were boys (Table 1). There were 568 (25%) participants who presented with CAP, 343 (15%) who presented with croup, 653 (28%) who presented with asthma, and 738 (32%) who presented with bronchiolitis. Most (92%) had an UTD vaccination status by provider documentation. Race and/or ethnicity, PMCA, season of presentation, site, and diagnosis were all significantly related to provider-documented UTD status.
Demographic Variables by Provider-Documented Vaccination Status
. | All Participants (N = 2302), n (%) . | UTD (n = 2123), n (%) . | Not UTD (n = 179), n (%) . | Unadjusted Odds Ratio (95% CI) . | P . |
---|---|---|---|---|---|
Age | .34 | ||||
≤18 mo | 1028 (45) | 943 (44) | 85 (47) | Reference | |
19–35 mo | 348 (15) | 320 (15) | 28 (16) | 1.0 (0.7–1.6) | |
3–6 y | 534 (23) | 490 (23) | 44 (25) | 1.0 (0.7–1.5) | |
≥7 y | 392 (17) | 370 (17) | 22 (12) | 1.5 (0.9–2.5) | |
Sex | .43 | ||||
Male | 1364 (59) | 1253 (59) | 111 (62) | 0.9 (0.6–1.2) | |
Race and/or ethnicity | .03 | ||||
White | 910 (40) | 836 (39) | 74 (34) | Reference | |
African American | 512 (22) | 485 (23) | 27 (15) | 1.6 (1.0–2.5) | |
Hispanic | 556 (24) | 500 (24) | 56 (31) | 0.8 (0.5–1.1) | |
Other | 307 (13) | 285 (13) | 22 (12) | 1.1 (0.7–1.9) | |
Missing | 17 (1) | 17 (1) | 0 (0) | — | |
Insurance | .67 | ||||
Private | 999 (43) | 924 (44) | 75 (42) | Reference | |
Public | 1303 (57) | 1199 (56) | 104 (58) | 0.7 (0.7–1.3) | |
PMCA | .03 | ||||
Nonchronic | 1278 (56) | 1162 (55) | 116 (65) | Reference | |
Noncomplex chronic | 920 (40) | 865 (41) | 55 (31) | 1.6 (1.1–2.2) | |
Complex chronic | 101 (4) | 93 (4) | 8 (4) | 1.2 (0.5–2.4) | |
Missing | 3 (0) | 3 (0) | 0 (0) | — | |
ICU admission | 145 (6) | 130 (6) | 15 (8) | 0.7 (0.4–1.2) | .24 |
Seasonality | .002 | ||||
April to September | 806 (35) | 762 (36) | 44 (25) | Reference | |
October to March | 1496 (65) | 1361 (64) | 135 (75) | 0.6 (0.4–0.8) | |
Hospital | <.001 | ||||
1 | 528 (23) | 513 (24) | 15 (8) | Reference | |
2 | 488 (21) | 403 (19) | 85 (47) | 0.1 (0.1–0.2) | |
3 | 468 (20) | 431 (20) | 37 (21) | 0.3 (0.2–0.6) | |
4 | 278 (12) | 250 (12) | 28 (16) | 0.3 (0.1–0.5) | |
5 | 540 (23) | 526 (25) | 14 (8) | 1.1 (0.5–2.3) | |
Diagnosis | .002 | ||||
CAP | 568 (25) | 504 (24) | 64 (36) | Reference | |
Croup | 343 (15) | 314 (15) | 29 (16) | 1.4 (0.9–2.2) | |
Asthma | 653 (28) | 617 (29) | 36 (20) | 2.2 (1.4–3.3) | |
Bronchiolitis | 738 (32) | 688 (32) | 50 (28) | 1.7 (1.2–2.6) |
. | All Participants (N = 2302), n (%) . | UTD (n = 2123), n (%) . | Not UTD (n = 179), n (%) . | Unadjusted Odds Ratio (95% CI) . | P . |
---|---|---|---|---|---|
Age | .34 | ||||
≤18 mo | 1028 (45) | 943 (44) | 85 (47) | Reference | |
19–35 mo | 348 (15) | 320 (15) | 28 (16) | 1.0 (0.7–1.6) | |
3–6 y | 534 (23) | 490 (23) | 44 (25) | 1.0 (0.7–1.5) | |
≥7 y | 392 (17) | 370 (17) | 22 (12) | 1.5 (0.9–2.5) | |
Sex | .43 | ||||
Male | 1364 (59) | 1253 (59) | 111 (62) | 0.9 (0.6–1.2) | |
Race and/or ethnicity | .03 | ||||
White | 910 (40) | 836 (39) | 74 (34) | Reference | |
African American | 512 (22) | 485 (23) | 27 (15) | 1.6 (1.0–2.5) | |
Hispanic | 556 (24) | 500 (24) | 56 (31) | 0.8 (0.5–1.1) | |
Other | 307 (13) | 285 (13) | 22 (12) | 1.1 (0.7–1.9) | |
Missing | 17 (1) | 17 (1) | 0 (0) | — | |
Insurance | .67 | ||||
Private | 999 (43) | 924 (44) | 75 (42) | Reference | |
Public | 1303 (57) | 1199 (56) | 104 (58) | 0.7 (0.7–1.3) | |
PMCA | .03 | ||||
Nonchronic | 1278 (56) | 1162 (55) | 116 (65) | Reference | |
Noncomplex chronic | 920 (40) | 865 (41) | 55 (31) | 1.6 (1.1–2.2) | |
Complex chronic | 101 (4) | 93 (4) | 8 (4) | 1.2 (0.5–2.4) | |
Missing | 3 (0) | 3 (0) | 0 (0) | — | |
ICU admission | 145 (6) | 130 (6) | 15 (8) | 0.7 (0.4–1.2) | .24 |
Seasonality | .002 | ||||
April to September | 806 (35) | 762 (36) | 44 (25) | Reference | |
October to March | 1496 (65) | 1361 (64) | 135 (75) | 0.6 (0.4–0.8) | |
Hospital | <.001 | ||||
1 | 528 (23) | 513 (24) | 15 (8) | Reference | |
2 | 488 (21) | 403 (19) | 85 (47) | 0.1 (0.1–0.2) | |
3 | 468 (20) | 431 (20) | 37 (21) | 0.3 (0.2–0.6) | |
4 | 278 (12) | 250 (12) | 28 (16) | 0.3 (0.1–0.5) | |
5 | 540 (23) | 526 (25) | 14 (8) | 1.1 (0.5–2.3) | |
Diagnosis | .002 | ||||
CAP | 568 (25) | 504 (24) | 64 (36) | Reference | |
Croup | 343 (15) | 314 (15) | 29 (16) | 1.4 (0.9–2.2) | |
Asthma | 653 (28) | 617 (29) | 36 (20) | 2.2 (1.4–3.3) | |
Bronchiolitis | 738 (32) | 688 (32) | 50 (28) | 1.7 (1.2–2.6) |
—, unable to calculate odds ratio because n = 0
The mean number of diagnostic tests received was 1.4 (SD 1.5; range 0–7). There was a significantly higher mean number of diagnostic tests ordered for children whose vaccination status was documented as not UTD compared with those whose vaccination status was UTD (unadjusted: not UTD: 1.9 [95% confidence interval (CI) 1.6–2.1] versus UTD: 1.3 [95% CI 1.3–1.4]; P < .001; adjusted: not UTD: 0.8 [95% CI 0.5–1.0] versus UTD: 0.4 [95% CI 0.2–0.5]; P = .005).
When evaluating each test or treatment individually, children whose vaccination status was documented as not UTD (versus UTD) had higher odds of receiving laboratory testing (Table 2). In analyses across all 4 conditions adjusted for race and/or ethnicity, PMCA, season of presentation, site, and LOS, children whose vaccination status was documented as not UTD (versus UTD) had higher odds of receiving a CBC count (adjusted odds ratio [aOR] 1.7; 95% CI 1.3–2.2; P < .001), blood culture (aOR 1.7; 95% CI 1.5–1.9; P < .001), CRP level testing (aOR 3.1; 95% CI 2.7–3.5; P < .001), and influenza-specific testing (aOR 1.6; 95% CI 1.2–2.1; P = .001). There were not significantly higher odds of receiving radiographic tests or medications for children whose vaccination status was not UTD compared with those whose vaccination status was UTD (Table 3).
Relationship of UTD Vaccination Status With Laboratory Testing for Children With ARTI
Outcome . | Vaccine Status . | n (%) Received . | Unadjusted Odds Ratio (95%CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
CBC count | UTD | 400 (19) | Reference | .02 | Reference | <.001 |
Not UTD | 47 (26) | 1.5 (1.1–2.2) | 1.7 (1.3–2.2) | |||
Blood culture | UTD | 337 (16) | Reference | .07 | Reference | <.001 |
Not UTD | 38 (21) | 1.4 (1.0–2.1) | 1.7 (1.5–1.9) | |||
CRP | UTD | 110 (5) | Reference | <.001 | Reference | <.001 |
Not UTD | 22 (12) | 2.6 (1.6–4.2) | 3.1 (2.7–3.5) | |||
Respiratory viral testing | UTD | 682 (32) | Reference | .001 | Reference | .39 |
Not UTD | 79 (44) | 1.7 (1.2–2.3) | 1.2 (0.8–1.8) | |||
Influenza-specific testing | UTD | 193 (9) | Reference | .36 | Reference | .001 |
Not UTD | 20 (11) | 1.3 (0.8–2.0) | 1.6 (1.2–2.1) | |||
Pertussis-specific testing | UTD | 140 (7) | Reference | <.001 | Reference | .80 |
Not UTD | 26 (15) | 2.4 (1.5–3.8) | 1.1 (0.5–2.2) | |||
CAP | ||||||
CBC count | UTD | 265 (53) | Reference | .53 | Reference | .12 |
Not UTD | 31 (48) | 0.8 (0.5–1.4) | 1.2 (0.9–1.6) | |||
Blood culture | UTD | 208 (41) | Reference | .56 | Reference | .22 |
Not UTD | 24 (38) | 0.8 (0.5–1.5) | 1.2 (0.9–1.7) | |||
CRP | UTD | 80 (16) | Reference | .002 | Reference | .001 |
Not UTD | 20 (31) | 2.4 (1.3–4.3) | 2.9 (1.6–5.4) | |||
Respiratory viral testing | UTD | 202 (40) | Reference | .13 | Reference | .78 |
Not UTD | 32 (50) | 1.5 (0.9–2.5) | 1.1 (0.6–1.9) | |||
Influenza-specific testing | UTD | 67 (13) | Reference | .87 | Reference | .001 |
Not UTD | 9 (14) | 1.1 (0.5–2.2) | 1.8 (1.3–2.5) | |||
Pertussis-specific testing | UTD | 54 (11) | Reference | .002 | Reference | .69 |
Not UTD | 16 (25) | 2.8 (1.5–5.2) | 1.2 (0.5–2.6) | |||
Croup | ||||||
CBC count | UTD | 20 (6) | Reference | .04 | Reference | .68 |
Not UTD | 5 (17) | 3.1 (1.1–8.9) | 1.3 (0.4–4.1) | |||
Blood culture | UTD | 6 (2) | Reference | .11 | Reference | <.001 |
Not UTD | 2 (7) | 3.8 (0.7–19.8) | 3.0 (2.2–4.0) | |||
CRP | UTD | 6 (2) | Reference | .58 | Reference | .24 |
Not UTD | 1 (3) | 1.8 (0.2–15.8) | 1.4 (0.8–2.4) | |||
Respiratory viral testing | UTD | 44 (14) | Reference | <.001 | Reference | .23 |
Not UTD | 12 (41) | 4.3 (1.9–9.7) | 2.3 (0.6–9.0) | |||
Influenza-specific testing | UTD | 9 (3) | Reference | .86 | Reference | .99 |
Not UTD | 1 (3) | 1.2 (0.1–9.9) | 1.0 (0.1–31.9) | |||
Pertussis-specific testing | UTD | 12 (4) | Reference | .004 | Reference | .08 |
Not UTD | 5 (17) | 5.2 (1.7–16.1) | 2.1 (0.9–4.9) | |||
Asthma | ||||||
CBC count | UTD | 28 (5) | Reference | .34 | Reference | .07 |
Not UTD | 3 (8) | 1.9 (0.6–6.6) | 2.6 (0.9–7.1) | |||
Blood culture | UTD | 12 (2) | Reference | .73 | Reference | .27 |
Not UTD | 1 (3) | 1.4 (0.2–11.4) | 3.1 (0.4–22.7) | |||
CRP | UTD | 7 (1) | Reference | .52 | Reference | .86 |
Not UTD | 0 (0) | — | — | |||
Respiratory viral testing | UTD | 136 (22) | Reference | .71 | Reference | .001 |
Not UTD | 7 (19) | 0.9 (0.4–2.0) | 0.5 (0.4–0.8) | |||
Influenza-specific testing | UTD | 18 (3) | Reference | .96 | Reference | .65 |
Not UTD | 1 (3) | 1.0 (0.1–7.3) | 1.8 (0.1–24.3) | |||
Pertussis-specific testing | UTD | 18 (3) | Reference | .96 | Reference | .01 |
Not UTD | 1 (3) | 1.0 (0.1–7.3) | 0.3 (0.1–0.8) | |||
Bronchiolitis | ||||||
CBC count | UTD | 87 (13) | Reference | .51 | Reference | .61 |
Not UTD | 8 (16) | 1.3 (0.6–2.9) | 1.4 (0.3–6.0) | |||
Blood culture | UTD | 111 (16) | Reference | .28 | Reference | .20 |
Not UTD | 11 (22) | 1.5 (0.7–3.0) | 1.8 (0.7–4.3) | |||
CRP | UTD | 17 (2) | Reference | .84 | Reference | .77 |
Not UTD | 1 (2) | 0.8 (0.1–6.2) | 1.4 (0.1–16.0) | |||
Respiratory viral testing | UTD | 300 (44) | Reference | .09 | Reference | .30 |
Not UTD | 28 (56) | 1.6 (0.9–2.9) | 1.7 (0.6–4.8) | |||
Influenza-specific testing | UTD | 99 (14) | Reference | .49 | Reference | .04 |
Not UTD | 9 (18) | 1.3 (0.6–2.8) | 1.7 (1.0–2.9) | |||
Pertussis-specific testing | UTD | 56 (8) | Reference | .97 | Reference | .48 |
Not UTD | 4 (8) | 1.0 (0.3–2.8) | 0.6 (0.1–2.6) |
Outcome . | Vaccine Status . | n (%) Received . | Unadjusted Odds Ratio (95%CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
CBC count | UTD | 400 (19) | Reference | .02 | Reference | <.001 |
Not UTD | 47 (26) | 1.5 (1.1–2.2) | 1.7 (1.3–2.2) | |||
Blood culture | UTD | 337 (16) | Reference | .07 | Reference | <.001 |
Not UTD | 38 (21) | 1.4 (1.0–2.1) | 1.7 (1.5–1.9) | |||
CRP | UTD | 110 (5) | Reference | <.001 | Reference | <.001 |
Not UTD | 22 (12) | 2.6 (1.6–4.2) | 3.1 (2.7–3.5) | |||
Respiratory viral testing | UTD | 682 (32) | Reference | .001 | Reference | .39 |
Not UTD | 79 (44) | 1.7 (1.2–2.3) | 1.2 (0.8–1.8) | |||
Influenza-specific testing | UTD | 193 (9) | Reference | .36 | Reference | .001 |
Not UTD | 20 (11) | 1.3 (0.8–2.0) | 1.6 (1.2–2.1) | |||
Pertussis-specific testing | UTD | 140 (7) | Reference | <.001 | Reference | .80 |
Not UTD | 26 (15) | 2.4 (1.5–3.8) | 1.1 (0.5–2.2) | |||
CAP | ||||||
CBC count | UTD | 265 (53) | Reference | .53 | Reference | .12 |
Not UTD | 31 (48) | 0.8 (0.5–1.4) | 1.2 (0.9–1.6) | |||
Blood culture | UTD | 208 (41) | Reference | .56 | Reference | .22 |
Not UTD | 24 (38) | 0.8 (0.5–1.5) | 1.2 (0.9–1.7) | |||
CRP | UTD | 80 (16) | Reference | .002 | Reference | .001 |
Not UTD | 20 (31) | 2.4 (1.3–4.3) | 2.9 (1.6–5.4) | |||
Respiratory viral testing | UTD | 202 (40) | Reference | .13 | Reference | .78 |
Not UTD | 32 (50) | 1.5 (0.9–2.5) | 1.1 (0.6–1.9) | |||
Influenza-specific testing | UTD | 67 (13) | Reference | .87 | Reference | .001 |
Not UTD | 9 (14) | 1.1 (0.5–2.2) | 1.8 (1.3–2.5) | |||
Pertussis-specific testing | UTD | 54 (11) | Reference | .002 | Reference | .69 |
Not UTD | 16 (25) | 2.8 (1.5–5.2) | 1.2 (0.5–2.6) | |||
Croup | ||||||
CBC count | UTD | 20 (6) | Reference | .04 | Reference | .68 |
Not UTD | 5 (17) | 3.1 (1.1–8.9) | 1.3 (0.4–4.1) | |||
Blood culture | UTD | 6 (2) | Reference | .11 | Reference | <.001 |
Not UTD | 2 (7) | 3.8 (0.7–19.8) | 3.0 (2.2–4.0) | |||
CRP | UTD | 6 (2) | Reference | .58 | Reference | .24 |
Not UTD | 1 (3) | 1.8 (0.2–15.8) | 1.4 (0.8–2.4) | |||
Respiratory viral testing | UTD | 44 (14) | Reference | <.001 | Reference | .23 |
Not UTD | 12 (41) | 4.3 (1.9–9.7) | 2.3 (0.6–9.0) | |||
Influenza-specific testing | UTD | 9 (3) | Reference | .86 | Reference | .99 |
Not UTD | 1 (3) | 1.2 (0.1–9.9) | 1.0 (0.1–31.9) | |||
Pertussis-specific testing | UTD | 12 (4) | Reference | .004 | Reference | .08 |
Not UTD | 5 (17) | 5.2 (1.7–16.1) | 2.1 (0.9–4.9) | |||
Asthma | ||||||
CBC count | UTD | 28 (5) | Reference | .34 | Reference | .07 |
Not UTD | 3 (8) | 1.9 (0.6–6.6) | 2.6 (0.9–7.1) | |||
Blood culture | UTD | 12 (2) | Reference | .73 | Reference | .27 |
Not UTD | 1 (3) | 1.4 (0.2–11.4) | 3.1 (0.4–22.7) | |||
CRP | UTD | 7 (1) | Reference | .52 | Reference | .86 |
Not UTD | 0 (0) | — | — | |||
Respiratory viral testing | UTD | 136 (22) | Reference | .71 | Reference | .001 |
Not UTD | 7 (19) | 0.9 (0.4–2.0) | 0.5 (0.4–0.8) | |||
Influenza-specific testing | UTD | 18 (3) | Reference | .96 | Reference | .65 |
Not UTD | 1 (3) | 1.0 (0.1–7.3) | 1.8 (0.1–24.3) | |||
Pertussis-specific testing | UTD | 18 (3) | Reference | .96 | Reference | .01 |
Not UTD | 1 (3) | 1.0 (0.1–7.3) | 0.3 (0.1–0.8) | |||
Bronchiolitis | ||||||
CBC count | UTD | 87 (13) | Reference | .51 | Reference | .61 |
Not UTD | 8 (16) | 1.3 (0.6–2.9) | 1.4 (0.3–6.0) | |||
Blood culture | UTD | 111 (16) | Reference | .28 | Reference | .20 |
Not UTD | 11 (22) | 1.5 (0.7–3.0) | 1.8 (0.7–4.3) | |||
CRP | UTD | 17 (2) | Reference | .84 | Reference | .77 |
Not UTD | 1 (2) | 0.8 (0.1–6.2) | 1.4 (0.1–16.0) | |||
Respiratory viral testing | UTD | 300 (44) | Reference | .09 | Reference | .30 |
Not UTD | 28 (56) | 1.6 (0.9–2.9) | 1.7 (0.6–4.8) | |||
Influenza-specific testing | UTD | 99 (14) | Reference | .49 | Reference | .04 |
Not UTD | 9 (18) | 1.3 (0.6–2.8) | 1.7 (1.0–2.9) | |||
Pertussis-specific testing | UTD | 56 (8) | Reference | .97 | Reference | .48 |
Not UTD | 4 (8) | 1.0 (0.3–2.8) | 0.6 (0.1–2.6) |
This relationship was evaluated by using logistic regression adjusted for race and/or ethnicity, level of medical complexity, season of admission, LOS, and hospital site. —, unable to calculate odds ratio because n = 0.
Bonferroni-corrected critical P < .006.
Relationship of UTD Vaccination Status With Use of Radiographic Studies and Medications for Children With ARTI
Outcome . | Vaccine Status . | n (%) Received . | Unadjusted Odds Ratio (95%CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
Chest radiograph | UTD | 900 (42) | Reference | .01 | Reference | .04 |
Not UTD | 93 (52) | 1.5 (1.1–2.0) | 1.4 (1.0–2.1) | |||
Neck radiograph | UTD | 80(4) | Reference | .25 | Reference | .41 |
Not UTD | 10 (6) | 1.5 (0.8–3.0) | 1.4 (0.7–2.8) | |||
Corticosteroids | UTD | 1091 (51) | Reference | .01 | Reference | .08 |
Not UTD | 75 (42) | 0.7 (0.5–0.9) | 0.7 (0.5–1.0) | |||
Antibiotics | UTD | 709 (33) | Reference | .002 | Reference | .09 |
Not UTD | 80 (45) | 1.6 (1.2–2.2) | 1.5 (0.9–2.3) | |||
CAP | ||||||
Chest radiograph | UTD | 398 (79) | Reference | .88 | Reference | .14 |
Not UTD | 50 (78) | 1.0 (0.5–1.8) | 0.7 (0.4–1.1) | |||
Neck radiograph | UTD | 3 (1) | Reference | .40 | Reference | .07 |
Not UTD | 1 (2) | 2.7 (0.3–25.9) | 5.0 (0.9–29.2) | |||
Corticosteroids | UTD | 132 (6) | Reference | .07 | Reference | .01 |
Not UTD | 10 (16) | 0.5 (0.3–1.1) | 0.4 (0.2–0.8) | |||
Antibiotics | UTD | 475 (94) | Reference | .05 | Reference | — |
Not UTD | 64 (100) | — | — | |||
Croup | ||||||
Chest radiograph | UTD | 53 (17) | Reference | .16 | Reference | .23 |
Not UTD | 8 (28) | 1.9 (0.8–4.5) | 1.6 (0.8–3.2) | |||
Neck radiograph | UTD | 73 (23) | Reference | .60 | Reference | .92 |
Not UTD | 8 (28) | 1.3 (0.5–3.0) | 1.1 (0.4–2.7) | |||
Corticosteroids | UTD | 260 (83) | Reference | .35 | Reference | .09 |
Not UTD | 22 (76) | 0.7 (0.3–1.6) | 0.5 (0.2–1.1) | |||
Antibiotics | UTD | 25 (8) | Reference | .10 | Reference | .28 |
Not UTD | 5 (17) | 2.4 (0.8–6.8) | 1.8 (0.6–4.9) | |||
Asthma | ||||||
Chest radiograph | UTD | 201 (33) | Reference | .43 | Reference | .58 |
Not UTD | 14 (39) | 1.3 (0.7–2.6) | 1.2 (0.6–2.7) | |||
Neck radiograph | UTD | 1 (0) | Reference | .04 | Reference | .38 |
Not UTD | 1 (3) | 17.6 (1.1–287.3) | 20.3 (0.1–1766) | |||
Corticosteroids | UTD | 599 (97) | Reference | .96 | Reference | .35 |
Not UTD | 35 (97) | 1.1 (0.1–8.1) | 0.5 (0.1–2.1) | |||
Antibiotics | UTD | 63 (10) | Reference | .37 | Reference | <.001 |
Not UTD | 2 (6) | 0.5 (0.1–2.2) | 0.3 (0.2–0.5) | |||
Bronchiolitis | ||||||
Chest radiograph | UTD | 248 (36) | Reference | .41 | Reference | .36 |
Not UTD | 21 (42) | 1.3 (0.7–2.3) | 1.4 (0.7–2.9) | |||
Neck radiograph | UTD | 3 (0) | Reference | .73 | Reference | — |
Not UTD | 0 (0) | — | — | |||
Corticosteroids | UTD | 100 (15) | Reference | .78 | Reference | .67 |
Not UTD | 8 (16) | 1.1 (0.5–2.5) | 0.9 (0.5–1.6) | |||
Antibiotics | UTD | 146 (21) | Reference | .59 | Reference | .32 |
Not UTD | 9 (18) | 0.8 (0.4–1.7) | 0.7 (0.4–1.4) |
Outcome . | Vaccine Status . | n (%) Received . | Unadjusted Odds Ratio (95%CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
Chest radiograph | UTD | 900 (42) | Reference | .01 | Reference | .04 |
Not UTD | 93 (52) | 1.5 (1.1–2.0) | 1.4 (1.0–2.1) | |||
Neck radiograph | UTD | 80(4) | Reference | .25 | Reference | .41 |
Not UTD | 10 (6) | 1.5 (0.8–3.0) | 1.4 (0.7–2.8) | |||
Corticosteroids | UTD | 1091 (51) | Reference | .01 | Reference | .08 |
Not UTD | 75 (42) | 0.7 (0.5–0.9) | 0.7 (0.5–1.0) | |||
Antibiotics | UTD | 709 (33) | Reference | .002 | Reference | .09 |
Not UTD | 80 (45) | 1.6 (1.2–2.2) | 1.5 (0.9–2.3) | |||
CAP | ||||||
Chest radiograph | UTD | 398 (79) | Reference | .88 | Reference | .14 |
Not UTD | 50 (78) | 1.0 (0.5–1.8) | 0.7 (0.4–1.1) | |||
Neck radiograph | UTD | 3 (1) | Reference | .40 | Reference | .07 |
Not UTD | 1 (2) | 2.7 (0.3–25.9) | 5.0 (0.9–29.2) | |||
Corticosteroids | UTD | 132 (6) | Reference | .07 | Reference | .01 |
Not UTD | 10 (16) | 0.5 (0.3–1.1) | 0.4 (0.2–0.8) | |||
Antibiotics | UTD | 475 (94) | Reference | .05 | Reference | — |
Not UTD | 64 (100) | — | — | |||
Croup | ||||||
Chest radiograph | UTD | 53 (17) | Reference | .16 | Reference | .23 |
Not UTD | 8 (28) | 1.9 (0.8–4.5) | 1.6 (0.8–3.2) | |||
Neck radiograph | UTD | 73 (23) | Reference | .60 | Reference | .92 |
Not UTD | 8 (28) | 1.3 (0.5–3.0) | 1.1 (0.4–2.7) | |||
Corticosteroids | UTD | 260 (83) | Reference | .35 | Reference | .09 |
Not UTD | 22 (76) | 0.7 (0.3–1.6) | 0.5 (0.2–1.1) | |||
Antibiotics | UTD | 25 (8) | Reference | .10 | Reference | .28 |
Not UTD | 5 (17) | 2.4 (0.8–6.8) | 1.8 (0.6–4.9) | |||
Asthma | ||||||
Chest radiograph | UTD | 201 (33) | Reference | .43 | Reference | .58 |
Not UTD | 14 (39) | 1.3 (0.7–2.6) | 1.2 (0.6–2.7) | |||
Neck radiograph | UTD | 1 (0) | Reference | .04 | Reference | .38 |
Not UTD | 1 (3) | 17.6 (1.1–287.3) | 20.3 (0.1–1766) | |||
Corticosteroids | UTD | 599 (97) | Reference | .96 | Reference | .35 |
Not UTD | 35 (97) | 1.1 (0.1–8.1) | 0.5 (0.1–2.1) | |||
Antibiotics | UTD | 63 (10) | Reference | .37 | Reference | <.001 |
Not UTD | 2 (6) | 0.5 (0.1–2.2) | 0.3 (0.2–0.5) | |||
Bronchiolitis | ||||||
Chest radiograph | UTD | 248 (36) | Reference | .41 | Reference | .36 |
Not UTD | 21 (42) | 1.3 (0.7–2.3) | 1.4 (0.7–2.9) | |||
Neck radiograph | UTD | 3 (0) | Reference | .73 | Reference | — |
Not UTD | 0 (0) | — | — | |||
Corticosteroids | UTD | 100 (15) | Reference | .78 | Reference | .67 |
Not UTD | 8 (16) | 1.1 (0.5–2.5) | 0.9 (0.5–1.6) | |||
Antibiotics | UTD | 146 (21) | Reference | .59 | Reference | .32 |
Not UTD | 9 (18) | 0.8 (0.4–1.7) | 0.7 (0.4–1.4) |
This relationship was evaluated by using logistic regression adjusted for race and/or ethnicity, level of medical complexity, season of admission, LOS, and hospital site. —, not applicable.
Bonferroni-corrected critical P < .006.
In adjusted condition-specific models, children with CAP and croup whose vaccination status was documented as not UTD (versus UTD) had higher odds of receiving some laboratory tests (CAP: CRP level testing and influenza testing; croup: blood culture) (Table 2). For asthma, there were lower odds of receiving nonspecific viral testing and antibiotics for children without documented UTD vaccination status compared with those with documented UTD vaccination status (Tables 2 and 3). There were no differences in resource use by vaccination status for children with bronchiolitis (Tables 2 and 3). There was also no significant difference in adjusted odds of receipt of narrow- or broad-spectrum antibiotics by vaccination status for subjects with all conditions and those with CAP (Table 4).
Relationship Between UTD Vaccination Status and Antibiotic Use for Children With ARTIs
Type of Antibiotic . | UTD, n (%) . | Not UTD, n (%) . | Odds Ratio (95% CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
None | 1420 (67) | 99 (55) | Reference | .008 | Reference | .38 |
Narrow | 428 (20) | 51 (28) | 1.6 (1.1–2.2) | 1.3 (0.7–2.5) | ||
Broad | 282 (13) | 29 (16) | 1.3 (0.8–1.9) | 1.3 (0.7–2.3) | ||
CAP | ||||||
None | 29 (6) | 0 (0) | Reference | .12 | Reference | .50 |
Narrow | 278 (55) | 40 (63) | 1.4 (0.8–2.3) | 1.1 (0.4–2.6) | ||
Broad | 198 (39) | 38 (24) | 0.9 (0.5–1.6) | 1.0 (0.4–2.7) |
Type of Antibiotic . | UTD, n (%) . | Not UTD, n (%) . | Odds Ratio (95% CI) . | P . | aOR (95%CI) . | Pa . |
---|---|---|---|---|---|---|
All conditions | ||||||
None | 1420 (67) | 99 (55) | Reference | .008 | Reference | .38 |
Narrow | 428 (20) | 51 (28) | 1.6 (1.1–2.2) | 1.3 (0.7–2.5) | ||
Broad | 282 (13) | 29 (16) | 1.3 (0.8–1.9) | 1.3 (0.7–2.3) | ||
CAP | ||||||
None | 29 (6) | 0 (0) | Reference | .12 | Reference | .50 |
Narrow | 278 (55) | 40 (63) | 1.4 (0.8–2.3) | 1.1 (0.4–2.6) | ||
Broad | 198 (39) | 38 (24) | 0.9 (0.5–1.6) | 1.0 (0.4–2.7) |
This relationship was evaluated by using logistic regression models adjusted for race and/or ethnicity, level of medical complexity, season of admission, LOS, and site.
Bonferroni-corrected critical P < .006.
Sensitivity Analysis
In the sensitivity analysis used to examine the relationship between binary use of each test and treatment of children whose vaccination status was UTD by IIS records for 1 site, we identified no significant differences in the adjusted odds of receiving any laboratory tests, radiographic studies, or medications by vaccination status (Supplemental Table 5).
Discussion
In a large multisite retrospective cohort of children with ARTI, we examined variation in resource use by provider-documented child vaccination status at hospital presentation. Overall, we identified that children with ARTI whose vaccination status was documented as not UTD had higher odds of undergoing diagnostic testing compared with children whose vaccination status was documented as UTD. In particular, we found higher odds of testing for children with CAP and croup whose vaccination status was documented as not UTD compared with children with the same conditions whose vaccination status was documented as UTD.
It is noteworthy that many of the tests in which we identified differences by vaccination status are nonspecific, such as CBC counts and CRP level tests, and it is unknown how the results influence subsequent medical decision-making for children admitted with ARTI. In previous work, Glanz et al16 identified higher rates of ED visits and inpatient admissions for children who were undervaccinated compared with those who were fully vaccinated, suggesting there are differences in how children who are undervaccinated seek care for acute illnesses. Our results suggest that when children with ARTI seek care in hospital settings and are documented as not having an UTD vaccination status, they experience higher resource use of nonspecific diagnostic testing.
The routine childhood vaccination schedule includes vaccines against Bordetella pertussis, Haemophilus influenzae type b, Streptococcus pneumoniae, and influenza, all of which can be pathogens in ARTI.30 Because the differential diagnoses for ARTI can include these pathogens, the medical decision-making of a provider treating a child with ARTI whose vaccination status is not UTD may be different from the medical decision-making of a provider when the child’s vaccination status is considered to be UTD. Our finding that differences in resource use by vaccination status were predominantly related to testing versus treatment reveals that providers may experience more diagnostic uncertainty in children with ARTI whose vaccination status is not UTD. This uncertainty may lead to more nonspecific testing to help rule out, for example, serious bacterial illness. However, if diagnostic uncertainty existed in a child whose vaccination status was not UTD, specific testing for certain VPDs would be most appropriate. With the exception of influenza testing, we did not observe any differences in testing for specific VPDs by vaccination status. Secondly, this observed increase in nonspecific testing did not appear to result in increased odds of receiving treatment. For example, we identified no difference in the use of broad-spectrum antibiotics in children with CAP who were undervaccinated. Therefore, an important takeaway from our results is for hospital-based physicians to consider whether nonspecific testing contributes to their management of children who they have identified as undervaccinated.
There have been substantial quality improvement efforts to decrease nonspecific diagnostic tests, particularly for ARTI.31–36 Our study results reveal that vaccination status is an important variable to consider when implementing these efforts. For instance, improving vaccination rates within a population may be an important component to efforts to help hospital-based providers be more parsimonious with nonspecific diagnostic testing. Understanding medical decision-making by providers in hospital settings regarding the testing and treatment of undervaccinated children is key to our ability to provide this high-risk population with appropriate, high-quality health care.
There has been increasing recognition of hospitalization as a missed opportunity to vaccinate children whose vaccination status is not UTD, particularly in the era of population-based IISs, in which a child’s vaccination status may be accessible to hospital-based providers.19,37–40 This study demonstrates that vaccination is relevant to hospital-based pediatric care because we identified variation in diagnostic testing in the hospital by vaccination status. In areas with high rates of undervaccination, there may be spillover effects for resource use in hospitals and health systems because more children without UTD vaccination status will be seen in hospital care settings. Better integration of IISs into hospital EMRs may help hospital-based providers more accurately assess the risk of VPDs for children who are undervaccinated and present with ARTI. Access to high-quality, accurate vaccination data at the point of care will allow us to focus future efforts on safely and effectively reducing nonspecific diagnostic testing in this population. Hospital-based quality improvement efforts have shown improvement in vaccination rates for high-risk children who are hospitalized, and hospitals should continue to provide catch-up vaccinations during hospitalization when clinically appropriate.37,38
One limitation of this study is the use of provider-documented vaccination status as our predictor variable rather than the child’s actual vaccination status. Provider-documented vaccination status is often inaccurate when compared with statewide vaccine registries.19 We hypothesize that provider-documented vaccination status in this study was likely obtained via parent report; however, it is possible that providers used their local population-based IIS registries or outpatient vaccination records to obtain this information. In our sensitivity analysis, using IIS UTD status, we saw similar point estimates to provider-documented UTD status, suggesting provider-documented UTD status most likely reflects the information acted on by the provider in medical decision-making at the point of care. However, we lack sufficient power to detect significant differences in this subsample. Additionally, by categorizing UTD as binary (yes or no), we were unable to evaluate the effect of missing specific vaccines related to ARTI. The high rate of provider-documented UTD status that we observed also limited the power of our study to detect differences, despite a large sample size.
This study was a retrospective cohort study; thus we cannot make any causal inferences between vaccination status and subsequent testing and treatment. We adjusted for the possible confounders that were measured in this study. However, we are unable to adjust for other potential causes of bias. There was a high percentage of children documented with UTD vaccination status in this study, and there is a possibility that not being documented with an UTD vaccination status serves as a proxy for other important patient-level factors that could influence clinical care.
All of the hospitals in this study were academic children’s hospitals. Although geographically dispersed, they may not be representative of all children’s hospitals or community-based settings where many children are hospitalized for ARTI.41 Additionally, although we included variables such as ICU admission, PMCA, and LOS to account for potential differences in acuity of illnesses, these proxy measures are imperfect, and it is unknown how other clinical characteristics interact with vaccination status to influence the care children receive for ARTI.
Conclusions
Children with ARTI who presented to the hospital and whose vaccination status was documented as not UTD had higher odds of receiving laboratory tests compared with children whose vaccination status was documented as UTD. Most additional laboratory tests done were not specific for VPDs, and we identified no increased odds of receiving treatment. Not only does undervaccination place children at risk for VPDs but also there is variation in care for common ARTIs, such as croup and pneumonia, by vaccination status, with higher use of diagnostic testing for children who are undervaccinated.
Acknowledgment
We thank Dr Mangione-Smith for her contributions to the project.
Dr Bryan conceptualized the study design, conducted the data analysis, and wrote the first draft of the manuscript; Drs Hofstetter, deHart, Simon, and Opel contributed to the study design and interpretation of study data and critically reviewed the manuscript to provide key intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: Some of the data for this article were obtained through support from National Heart, Lung, and Blood Institute grant 1R01HL121067-01 (principal investigator: Rita Mangione-Smith, MD, MPH). Funded by the National Institutes of Health (NIH).
- aOR
adjusted oddsratio
- ARTI
acute respiratory tract illness
- CAP
community-acquired pneumonia
- CBC
complete bloodcell
- CI
confidence interval
- CRP
C-reactive protein
- ED
emergency department
- EMR
electronic medical record
- IIS
immunization information system
- LOS
length ofstay
- PHIS
Pediatric Health Information System
- PMCA
Pediatric Medical Complexity Algorithm
- UTD
up-to-date
- VPD
vaccine-preventable disease
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: Dr Hofsetter previously received research support from Pfizer Independent Grants for Learning and Change; the other authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.