BACKGROUND:

Contaminated blood cultures pose a significant burden. We sought to determine the impact of contaminated peripheral blood cultures on patients, families, and the health care system.

METHODS:

In this retrospective case-control study from January 1, 2014, to December 31, 2017, we compared the hospital course, return visits and/or admissions, charges, and length of stay of patients with contaminated peripheral blood cultures (case patients) with those of patients with negative cultures (controls). Patients were categorized into those evaluated and discharged from the emergency department (ED) (ED patients) and those who were hospitalized (inpatients).

RESULTS:

A total of 104 ED case patients were matched with 208 ED control patients. A total of 343 case inpatients were matched with 686 inpatient controls. There was no significant difference between case and control patient demographics, ED, or hospital course at presentation. Fifty-five percent of discharged ED patients returned to the hospital for evaluation and/or admission versus 4% of controls. There was a significant (P < .0001) increase in repeat blood cultures (43% vs 1%), consultations obtained (21% vs 2%), cerebrospinal fluid studies (10% vs 0%), and antibiotic administration (27% vs 1%) in ED patients compared with controls. Each ED patient requiring revisit to the hospital incurred, on average, $4660 in additional charges. There was a significant (P < .04) increase in repeat blood cultures (57% vs 7%), consultations obtained (35% vs 28%), broadening of antibiotic coverage (18% vs 11%), median length of stay (75 vs 64 hours), and median laboratory charges ($3723 vs $3296) in case inpatients compared with controls.

CONCLUSIONS:

Contaminated blood cultures result in increased readmissions, testing and/or procedures, length of stay, and hospital charges in children.

Blood cultures are important to identify bacteremia in children and guide appropriate antibiotic therapy. However, blood cultures are often contaminated during the collection process by commensals on the skin, oral flora, or the environment.1,2  Contaminated blood cultures result in additional testing, unnecessary antibiotic treatment, increased length of stay (LOS), and increased hospital charges in adults.36  Pediatric studies are limited by small study populations, and have not compared the impact of contaminated blood cultures to a control population with negative cultures.710  This comparison is important to truly determine the workup and charges incurred because of contaminated cultures versus the disease process that resulted in the emergency department (ED) or hospital visit. The few pediatric studies available also do not differentiate the impact of contaminated cultures in ED patients compared with those already hospitalized at the time of the culture. The impact of contaminated blood cultures in ED patients is easily measured because patients have been discharged from the hospital by the time of culture positivity. Therefore, return visits and additional workup and treatment can be attributed to the contaminated blood culture. It is more challenging to measure the impact on hospitalized children (inpatients) because they are sicker and are often already on antibiotics at the time of culture positivity. It is difficult to determine if additional workup or continuation of antibiotics is due to contaminated cultures or the disease process that prompted the admission.

Our goal for this study was to measure the impact of contaminated blood cultures by comparing the hospital course, LOS, return visits and/or readmissions, and total charges of ED patients and inpatients with contaminated cultures with those of ED patients and inpatients with negative cultures.

This retrospective case-control study was conducted at a freestanding tertiary care children’s hospital with 390 beds, 17 000 annual admissions, and >49 000 annual emergency visits. The study period was from January 1, 2014, to December 31, 2017.

We compared patients with contaminated peripheral blood cultures (case patients) with those with negative peripheral blood cultures (controls) presenting to the ED and/or admitted to the inpatient unit or PICU. Patients admitted to the NICU were excluded.

Patients were categorized into those evaluated and discharged from the ED (ED patients) and those hospitalized to the inpatient units (inpatients). Each ED patient with a contaminated blood culture was matched with 2 ED patients with negative cultures by using age, ED triage acuity (emergency severity index11 ), and month of visit as match criteria. Each inpatient with a contaminated blood culture was matched with 2 inpatients with negative cultures by using age, All Patient Refined Diagnosis-Related Groups (APR-DRGs) inpatient severity of illness, discharge service, and admission month as match criteria. Discharge service was used as a proxy to match patients with diseases involving similar organ systems.

We obtained patient information on all peripheral blood cultures from the hospital clinical microbiology laboratory database. Blood cultures are processed in our laboratory for 5 days on a continuously monitored blood culture system. When a culture has a positive result, gram-stain, as well as identification by matrix-assisted laser desorption ionization time-of-flight mass spectrometry, was performed, allowing for early identification of pathogens from 2014 to 2017.12  Matrix-assisted laser desorption ionization time-of-flight mass spectrometry was replaced by the Verigene gram-positive blood culture nucleic acid test in May 2017. All cultures with positive results are plated to appropriate agar and susceptibilities performed when appropriate.

We defined contaminated blood cultures as the growth of organisms normally considered as commensals by our laboratory standards and the Centers for Disease Control and Prevention National Health Safety Network common commensal list.13  Patients were excluded (1) if they had true-positive cultures; (2) if they had data quality issues, including missing data fields or discrepancies in data obtained from the different databases; or (3) if a chart review revealed that the medical team treated a commensal organism as a pathogen on the basis of the clinical context of the patient. In addition, we also excluded return visits within 6 months of the initial visit for each patient.

Information on patient demographics, ED triage acuity at presentation, APR-DRG severity of illness, date of visit and/or admission, date of discharge, discharge service, and LOS was obtained from the inpatient and ED electronic medical records. Charge data were obtained from the health information management system. Data from the clinical laboratory, ED, and the inpatient and health information management databases were merged by using the unique accession number for each blood culture and other patient identifiers.

Chief complaint at presentation, past medical history, ED or hospital course (laboratory studies obtained, procedures performed, consultations obtained, imaging studies obtained, and antibiotics administered), discharge diagnosis, return to ED and/or readmission, information regarding follow-up phone call, and difficulties in reaching families about contaminated blood cultures were obtained by chart review. Chief complaint at presentation and discharge diagnosis were categorized into groups during the chart review. For inpatients, we hypothesized that additional workup due to contaminated cultures would likely be performed within the first 48 hours after the culture was obtained. Accordingly, we compared workup and treatment in the first 48 hours after the initial culture was obtained between case and control patients.

The chart review was performed by 5 study physicians, and the data were entered into a study instrument in Research Electronic Data Capture, a secure Web-based data capture application.14  A majority of charts (60%) were reviewed by the first author (M.F.). M.F. and M.S. (principal investigator) audited at least 15% of the charts reviewed by team members to ensure the accuracy of data collection.

The study was approved by our institutional human research protection office.

Data were presented as the median with interquartile range (IQR) for continuous variables and number with percentage for categorical variables. Pearson χ2 tests of independence or Fisher’s exact tests were used to compare categorical variables. Continuous variables were analyzed with Wilcoxon rank tests. Data were analyzed by using SAS version 9.4 (SAS Institute, Inc, Cary, NC). A P value <.05 was considered significant.

Peripheral blood cultures were obtained in 10 071 patients from January 1 2014, to December 31, 2017. After exclusions, 8182 patients with peripheral blood cultures were included (Fig 1). A total of 447 patients had contaminated blood cultures (5.5%) with growth of 540 commensal organisms. Most patients (75%) had blood cultures obtained in the ED. Most patients (81%) had a single organism growing in their blood cultures. Most contaminants were coagulase-negative staphylococci (64%) and viridans group streptococci (18%) (Supplemental Table 5). Contaminants were recovered from 227 (51%) aerobic cultures, 145 (32%) anaerobic cultures, and 75 (17%) from both aerobic and anaerobic cultures. The median time to positivity was 20 hours (IQR 17–24 hours) in aerobic cultures and 24 hours (IQR 20–40 hours) in anaerobic cultures.

FIGURE 1

Study cohort. A case is a patient with a contaminated blood culture. A control is a patient with a negative culture. ED exclusions were as follows: a blood culture from an ED patient (case) was obtained from a central line and was excluded, along with 2 matched controls; one control ED patient grew a contaminant on a repeat culture and was excluded. Inpatient exclusions were as follows: 14 inpatient blood culture (case) results were true-positives due to growth in >1 culture (n = 10) or were treated as true-positive by the medical team (n = 4); 2 case patients were excluded for other reasons (central line, NICU patient), and all matching controls were excluded; 6 control patients had a pathogen growing in blood cultures obtained at an outside hospital before presentation at Saint Louis Children’s Hospital and were excluded; 1 control patient was included twice, and the duplicate was excluded; and the blood culture for another control patient was obtained from a central line and was excluded.

FIGURE 1

Study cohort. A case is a patient with a contaminated blood culture. A control is a patient with a negative culture. ED exclusions were as follows: a blood culture from an ED patient (case) was obtained from a central line and was excluded, along with 2 matched controls; one control ED patient grew a contaminant on a repeat culture and was excluded. Inpatient exclusions were as follows: 14 inpatient blood culture (case) results were true-positives due to growth in >1 culture (n = 10) or were treated as true-positive by the medical team (n = 4); 2 case patients were excluded for other reasons (central line, NICU patient), and all matching controls were excluded; 6 control patients had a pathogen growing in blood cultures obtained at an outside hospital before presentation at Saint Louis Children’s Hospital and were excluded; 1 control patient was included twice, and the duplicate was excluded; and the blood culture for another control patient was obtained from a central line and was excluded.

Close modal

A total of 1766 patients were evaluated and discharged from the ED (ED patients). Of these, 104 patients (5.9%) had contaminated blood cultures (case patients) and were matched with 208 ED patients with negative cultures (controls) by using our match criteria. After chart review, 1 case patient and 3 controls were excluded (Fig 1). Final analysis included 103 case patients and 205 controls.

There was no significant difference in age, sex, ethnicity, ED triage acuity, month of visit, number of chronic medical issues, chief complaint category at presentation, or ED LOS between case patients and controls (Supplemental Table 6). There was no significant difference in the initial laboratory or imaging studies obtained, procedures performed, consultations ordered, or intravenous (IV) or intramuscular (IM) antibiotics administered at the time the blood culture was obtained between case and control patients (Table 1).

TABLE 1

Laboratory Studies Obtained, Procedures Performed, Consultations Obtained, and Antibiotics Administered in ED Patients

Case Patients (n = 103), n (%)Control Patients (n = 205), n (%)POdds Ratio (95% CI)
ED course at the time of initial blood culture     
 Laboratory studies     
  CBC 99 (96) 199 (97) .7 0.8 (0.2–2.7) 
  UA or urine culture 69 (67) 126 (61) .4 1.3 (0.8–1.7) 
  CSF studies 2 (2) 7 (3) .7 0.6 (0.1–2.8) 
  Viral studies 52 (50) 102 (50) >.99 1.0 (0.6–1.7) 
 Imaging 34 (33) 63 (31) .7 1.1 (0.7–1.8) 
 Procedures     
  IV placed 72 (70) 161 (79) .1 0.6 (0.4–1.1) 
  Lumbar puncture 2 (2) 7 (3) .7 0.6 (0.1–2.8) 
 Consultations 26 (25) 45 (22) .6 1.2 (0.7–2.1) 
 IV antibiotics started 18 (17) 31 (15) .6 1.2 (0.6–2.3) 
 Total No. returned to ED or readmitted within 5 d 57 (55) 8 (4) <.0001 30.2 (13.5–67.7) 
 Return or revisit due to contaminated blood culture     
  Reevaluated in ED and discharged from the hospital 26 (25) N/A N/A N/A 
  Admitted to hospital 28 (27) N/A N/A N/A 
ED and/or hospital course after return to ED and/or readmission within 5 d     
 Laboratory studies     
  CBC 28 (27) 6 (3) <.0001 12.4 (4.9–31.1) 
  Blood culture 44 (43) 3 (1) <.0001 52.2 (15.7–174.3) 
  UA or urine culture 7 (7) 1 (0) .002 14.9 (1.8–122.6) 
  CSF studies 10 (10) 0 (0) <.0001 N/A 
  Viral studies 12 (12) 3 (1) .0002 8.9 (2.4–32.2) 
 Imaging 3 (3) 3 (1) .4 2.0 (0.4–10.1) 
 Procedures     
  IV placed 34 (33) 4 (2) <.0001 24.8 (8.5–72.3) 
  Lumbar puncture 10 (10) 0 (0) <.0001 N/A 
 Consultations 22 (21) 5 (2) <.0001 10.8 (3.9–29.4) 
 IV antibiotics started 28 (27) 3 (1) <.0001 26.4 (7.8–89.2) 
Case Patients (n = 103), n (%)Control Patients (n = 205), n (%)POdds Ratio (95% CI)
ED course at the time of initial blood culture     
 Laboratory studies     
  CBC 99 (96) 199 (97) .7 0.8 (0.2–2.7) 
  UA or urine culture 69 (67) 126 (61) .4 1.3 (0.8–1.7) 
  CSF studies 2 (2) 7 (3) .7 0.6 (0.1–2.8) 
  Viral studies 52 (50) 102 (50) >.99 1.0 (0.6–1.7) 
 Imaging 34 (33) 63 (31) .7 1.1 (0.7–1.8) 
 Procedures     
  IV placed 72 (70) 161 (79) .1 0.6 (0.4–1.1) 
  Lumbar puncture 2 (2) 7 (3) .7 0.6 (0.1–2.8) 
 Consultations 26 (25) 45 (22) .6 1.2 (0.7–2.1) 
 IV antibiotics started 18 (17) 31 (15) .6 1.2 (0.6–2.3) 
 Total No. returned to ED or readmitted within 5 d 57 (55) 8 (4) <.0001 30.2 (13.5–67.7) 
 Return or revisit due to contaminated blood culture     
  Reevaluated in ED and discharged from the hospital 26 (25) N/A N/A N/A 
  Admitted to hospital 28 (27) N/A N/A N/A 
ED and/or hospital course after return to ED and/or readmission within 5 d     
 Laboratory studies     
  CBC 28 (27) 6 (3) <.0001 12.4 (4.9–31.1) 
  Blood culture 44 (43) 3 (1) <.0001 52.2 (15.7–174.3) 
  UA or urine culture 7 (7) 1 (0) .002 14.9 (1.8–122.6) 
  CSF studies 10 (10) 0 (0) <.0001 N/A 
  Viral studies 12 (12) 3 (1) .0002 8.9 (2.4–32.2) 
 Imaging 3 (3) 3 (1) .4 2.0 (0.4–10.1) 
 Procedures     
  IV placed 34 (33) 4 (2) <.0001 24.8 (8.5–72.3) 
  Lumbar puncture 10 (10) 0 (0) <.0001 N/A 
 Consultations 22 (21) 5 (2) <.0001 10.8 (3.9–29.4) 
 IV antibiotics started 28 (27) 3 (1) <.0001 26.4 (7.8–89.2) 

CBC, complete blood cell count; CI, confidence interval; CSF, cerebrospinal fluid; N/A, not applicable; UA, urine analysis.

Of the 103 patients with contaminated cultures, 57 (55%) case patients returned to the ED and/or were admitted within 5 days of the initial visit, versus 8 (4%) control patients (Table 1). Fifty-four (52%) patients returned because of contaminated blood cultures, and 3 patients returned because of reasons unrelated to their contaminated blood culture. Follow-up on the other patients are listed in Table 2.

TABLE 2

Follow-up and Difficulties Reaching Families of Patients With Contaminated Cultures

n (%)
Follow-up on patients with contaminated blood cultures (n = 103)  
 Returned to ED and/or readmitted 54 (52) 
 Phone evaluation 17 (17) 
 PMD notified and asked to follow-up on patient 12 (12) 
 Asked to follow-up at an outside ED 8 (8) 
 Chart reviewed with no follow-up indicated 8 (8) 
 Unable to reach family 4 (4) 
Difficulties in reaching family (n = 83)a  
 Contacted easily (1–2 phone calls) 56 (67) 
 Contacted after multiple phone calls 7 (8) 
 Unable to reach family; PMD contacted, and they agreed to follow-up with the patient 7 (8) 
 Other difficulties in reaching family 8 (10) 
 Unable to reach family; police sent to home to contact the family 4 (5) 
 Mom refused to return, DFS notified; mom and patient returned to ED after case worker talked to mom 1 (1) 
n (%)
Follow-up on patients with contaminated blood cultures (n = 103)  
 Returned to ED and/or readmitted 54 (52) 
 Phone evaluation 17 (17) 
 PMD notified and asked to follow-up on patient 12 (12) 
 Asked to follow-up at an outside ED 8 (8) 
 Chart reviewed with no follow-up indicated 8 (8) 
 Unable to reach family 4 (4) 
Difficulties in reaching family (n = 83)a  
 Contacted easily (1–2 phone calls) 56 (67) 
 Contacted after multiple phone calls 7 (8) 
 Unable to reach family; PMD contacted, and they agreed to follow-up with the patient 7 (8) 
 Other difficulties in reaching family 8 (10) 
 Unable to reach family; police sent to home to contact the family 4 (5) 
 Mom refused to return, DFS notified; mom and patient returned to ED after case worker talked to mom 1 (1) 
a

Of the 103 patients with contaminated blood cultures, the provider called the PMD office instead of the family for 12 patients or reviewed the chart and felt that no further action was needed for 8 patients. These patients are excluded from the calculations for difficulties reaching the family. DFS, division of family services.

ED physicians contacted 83 families of patients with contaminated cultures. For the remaining patients, providers either called the primary medical doctor’s (PMD’s) office (n = 12) or felt that no further action was needed (n = 8). Difficulties in reaching the families are listed in Table 2. It is significant to note that police were sent to the homes of 4 (5%) patients because of the inability to contact the family. The parent of 1 patient refused to bring the child back to the ED for evaluation, which then required social work consultation and notification to child protective services (Table 2).

Patients with contaminated blood cultures who returned to the ED and/or were admitted had a significantly greater number of laboratory studies and consultations obtained, procedures performed, and IV or IM antibiotics started compared with controls (Table 1).

Of the 54 ED patients who returned to the ED and/or were admitted because of contaminated blood cultures, 1 patient had additional complications due to a preexisting medical condition unrelated to the contaminated blood culture, which resulted in a long LOS. This patient visit was excluded from the analysis of LOS and charge data. For the 53 patients, the total LOS due to these visits and/or admissions was 56.5 days. Charge data were available for 51 of 53 patients, and charges totaled $237 681, averaging to $4660 per patient with a contaminated blood culture requiring revisit to the hospital system.

A total of 6416 patients were admitted to the inpatient units (inpatients) either through the ED or as direct admissions. A total of 343 patients (5.3%) had contaminated blood cultures (case patients) and were matched with 686 inpatients (controls) with negative cultures by using our match criteria. A chart review was used to identify 16 case patients and 40 controls who were excluded (Fig 1). After the exclusions, 327 case patients and 646 controls were included in the final analysis.

There was no significant difference in age, sex, ethnicity, APR-DRG severity of illness, month of admission, number of chronic medical issues, discharge service, discharge diagnosis category, or number of children admitted or transferred to the ICU between case and control patients (Supplemental Table 7). There was no significant difference in the initial laboratory or imaging studies obtained, procedures performed, consultations obtained, or IV or IM antibiotics administered at the time the blood culture was obtained between case and control patients (Table 3).

TABLE 3

Laboratory Studies Obtained, Procedures Performed, Consultations Obtained, and Antibiotics Administered in Inpatients

Case Patients (n = 327), n (%)Control Patients (n = 646), n (%)POdds Ratio (95% CI)
Hospital course at time of blood culture     
 Laboratory studies     
  CBC 282 (86) 559 (87) .9 1.0 (0.7–1.4) 
  UA or urine culture 180 (55) 345 (53) .6 1.1 (0.8–1.4) 
  CSF studies 86 (26) 159 (25) .6 1.1 (0.8–1.5) 
  Other 295 (90) 583 (90) .9 1.0 (0.6–1.6) 
 Imaging 179 (55) 349 (54) .8 1.0 (0.8–1.4) 
 Procedures     
  IV placed 263 (80) 536 (83) .3 0.8 (0.6–1.2) 
  Lumbar puncture 91 (28) 158 (24) .3 1.2 (0.9–1.6) 
 Consultations 73 (22) 177 (27) .1 0.8 (0.6–1.0) 
 IV or IM antibiotics started 189 (58) 372 (58) .9 1.0 (0.8–1.3) 
 Total No. returned to ED or readmitted within 5 d 15 (5) 17 (3) .1 1.8 (0.9–3.6) 
Hospital course during the 48 h after blood culture was obtained     
 Laboratory studies     
  CBC 91 (28) 186 (29) .8 1.0 (0.7–1.3) 
  Blood culture 188 (57) 47 (7) <.0001 17 (12–25) 
  UA or urine culture 38 (12) 73 (11) .9 1.0 (0.7–1.6) 
  CSF studies 10 (3) 27 (4) .4 0.7 (0.3–1.5) 
  Other 189 (58) 330 (51) .05 1.3 (1.003–1.7) 
 Imaging 129 (39) 259 (40) .8 1.0 (0.7–1.3) 
 Procedures     
  IV placed 35 (11) 52 (8) .2 1.4 (0.9–2.1) 
  Lumbar puncture 9 (3) 29 (4) .2 0.6 (0.3–1.3) 
 Consultations 115 (35) 184 (28) .03 1.4 (1.03–1.8) 
 IV antibiotics started 34 (10) 46 (7) .1 1.5 (1.0–2.4) 
 Already on IV antibioticsa 206 (63) 421 (65) .5 0.9 (0.7–1.2) 
 Antibiotic coverage broadened 59 (18) 69 (11) .001 1.8 (1.3–2.7) 
Case Patients (n = 327), n (%)Control Patients (n = 646), n (%)POdds Ratio (95% CI)
Hospital course at time of blood culture     
 Laboratory studies     
  CBC 282 (86) 559 (87) .9 1.0 (0.7–1.4) 
  UA or urine culture 180 (55) 345 (53) .6 1.1 (0.8–1.4) 
  CSF studies 86 (26) 159 (25) .6 1.1 (0.8–1.5) 
  Other 295 (90) 583 (90) .9 1.0 (0.6–1.6) 
 Imaging 179 (55) 349 (54) .8 1.0 (0.8–1.4) 
 Procedures     
  IV placed 263 (80) 536 (83) .3 0.8 (0.6–1.2) 
  Lumbar puncture 91 (28) 158 (24) .3 1.2 (0.9–1.6) 
 Consultations 73 (22) 177 (27) .1 0.8 (0.6–1.0) 
 IV or IM antibiotics started 189 (58) 372 (58) .9 1.0 (0.8–1.3) 
 Total No. returned to ED or readmitted within 5 d 15 (5) 17 (3) .1 1.8 (0.9–3.6) 
Hospital course during the 48 h after blood culture was obtained     
 Laboratory studies     
  CBC 91 (28) 186 (29) .8 1.0 (0.7–1.3) 
  Blood culture 188 (57) 47 (7) <.0001 17 (12–25) 
  UA or urine culture 38 (12) 73 (11) .9 1.0 (0.7–1.6) 
  CSF studies 10 (3) 27 (4) .4 0.7 (0.3–1.5) 
  Other 189 (58) 330 (51) .05 1.3 (1.003–1.7) 
 Imaging 129 (39) 259 (40) .8 1.0 (0.7–1.3) 
 Procedures     
  IV placed 35 (11) 52 (8) .2 1.4 (0.9–2.1) 
  Lumbar puncture 9 (3) 29 (4) .2 0.6 (0.3–1.3) 
 Consultations 115 (35) 184 (28) .03 1.4 (1.03–1.8) 
 IV antibiotics started 34 (10) 46 (7) .1 1.5 (1.0–2.4) 
 Already on IV antibioticsa 206 (63) 421 (65) .5 0.9 (0.7–1.2) 
 Antibiotic coverage broadened 59 (18) 69 (11) .001 1.8 (1.3–2.7) 

CBC, complete blood cell count; CI, confidence interval; CSF, cerebrospinal fluid; UA, urine analysis.

a

Patients who were already started on antibiotics at time of blood culture.

A significantly greater number of patients with contaminated blood cultures had repeat blood cultures, consultations obtained, and broadening of antibiotic coverage during the 48 hours after the initial blood culture was obtained compared with control patients (Table 3).

Inpatients with contaminated blood cultures (case patients) had a significantly longer LOS (median LOS was 11 hours longer) and higher laboratory charges compared with control patients. This would lead to ∼122.6 additional days of hospital stay due to contaminated cultures. There was no significant increase in total charges for case versus control patients (Table 4).

TABLE 4

LOS and Charges for Inpatients

Case Patients (n = 327)Control Patients (n = 646)P
LOS, h, median (IQR) 75 (49–149) 64 (44–132) .006 
Total charges, $, median (IQR) 15 508 (9169–34 364) 13 731 (8560–30 924) .1 
Breakdown of total charges, $, median (IQR)    
 Laboratory charges 3723 (2387–5594) 3296 (2065–5437) .04 
 Pharmacy charges 668 (265–2202) 560 (167–2142) .08 
 Unit charges 4736 (2444–12 088) 4296 (2368–10 904) .1 
 Other charges 4685 (2918–13 243) 4305 (2687–11 884) .2 
Case Patients (n = 327)Control Patients (n = 646)P
LOS, h, median (IQR) 75 (49–149) 64 (44–132) .006 
Total charges, $, median (IQR) 15 508 (9169–34 364) 13 731 (8560–30 924) .1 
Breakdown of total charges, $, median (IQR)    
 Laboratory charges 3723 (2387–5594) 3296 (2065–5437) .04 
 Pharmacy charges 668 (265–2202) 560 (167–2142) .08 
 Unit charges 4736 (2444–12 088) 4296 (2368–10 904) .1 
 Other charges 4685 (2918–13 243) 4305 (2687–11 884) .2 

There was no significant difference in the number of inpatients who returned to the ED or were readmitted within 5 days of discharge between case and control patients (Table 3).

In total, 6 inpatients either returned to the ED or were readmitted because of contaminated blood cultures. For the 6 patients, the total LOS due to these visits and/or admissions was 5.1 days, with a total charge of $20 993. This averages to $3499 per inpatient with a contaminated blood culture requiring revisit to the hospital system.

Our study reveals that contaminated blood cultures result in increased return ED visits and/or readmissions, invasive testing, antibiotic use, and charges, posing a significant burden for children, families, and the health care system. Even with the increasing availability of rapid diagnostics (typically polymerase chain reaction based) to identify positive blood culture results, pediatric patients with positive blood culture results are often called back in immediately for evaluation when the gram-stain result is known, before polymerase chain reaction results may be available. This is because the limited reserve of young children with bacteremia necessitates prompt intervention and treatment in the case of a true pathogen. The impact of contaminated cultures is thus particularly significant in patients evaluated and discharged from the ED as compared with inpatients.

Studies have revealed that procedural pain early in life, such as repeated heel sticks, can lead to increased pain responses during subsequent procedures.1517  Needle pokes and IV cannula placement are considered by children to be a source of significant pain during their hospitalization.18,19  In our institution, contaminated blood cultures subjected children to unnecessary painful needle pokes for repeat blood cultures, laboratory studies, IV cannula placements, and lumbar punctures.

The gut microbiome plays an important role in human health and disease development. Disruption of the microbiome through antibiotics, especially in early childhood, has been shown to increase the risk of obesity, celiac disease, asthma, allergies, and antibiotic resistance later in childhood.2023  Contaminated blood cultures resulted in initiation of unnecessary antibiotics and broadening of antibiotic coverage. This poses the risk of disrupting the gut microbiome and disease development in the future.

Return visits and/or readmissions resulted in a charge of $4660 per ED patient and a charge of $3499 per inpatient with contaminated cultures. Total charges per contaminated culture reported by Hall et al7  (N = 149 contaminated cultures) was $2800 per contaminated culture. This included both inpatient charges for additional days of hospital admission (90 days) and return visits (25 visits). Although we do not have charge data for additional days of hospital admission for our inpatients, it seems likely that those numbers are much lower than what we found for ED return visits or readmissions. These charges can be a significant financial burden to families, insurance companies, and, ultimately, hospitals if bundled payments become the model for reimbursement in the future.

Contaminated blood cultures resulted in an additional 60 ED visits and/or inpatient hospitalizations, with a total of 61.6 days spent in either the ED or inpatient unit. In addition, it also led to an additional 122.6 days in the hospital for the 327 inpatients with contaminated cultures. This adds additional costs to families in terms of lost days from work, child care for siblings while a parent is at the hospital, costs of transportation to the hospital, and, in many cases, time lost from work for extended family members who take turns to be with the child at the hospital. In addition, there is the emotional cost and stress to families when they are called at home with notification of a positive culture result and the anxiety when their child is subjected to painful procedures because of a contaminated blood culture. There is also the fear of losing their job because of additional days spent at the hospital with their child.

There are additional opportunity costs to the physicians who had to call families of patients with contaminated cultures who were already discharged. Some children were followed-up at their primary care physician’s office or clinic, and these costs are not included in this analysis. In some instances, social work was consulted, and police had to be sent to the homes of children with contaminated cultures. A home visit by the police can be a source of significant stress for families.

The overall impact is higher in ED patients because they are already discharged at the time of culture positivity, leading to phone calls to families, return visits and/or admissions to the hospital, or visits to the primary care offices. Inpatients observed by the hospital team often do not require an extensive additional workup. Most inpatients are already on IV antibiotics. It is not surprising that the major impacts noted in inpatients with contaminated cultures, compared with patients with negative cultures, were repeat blood cultures and broadening of antibiotic coverage. This was reflected in the increase in laboratory charges in patients with contaminated blood cultures. There was no significant increase in total charges in the case versus control group. This may be partly due to the wide spread of charge data, with a slightly higher proportion of control patients with charges in the 75th to 100th percentile, which likely made it difficult to detect a significant difference in total charges between case and control patients.

A limitation of our study is that it was conducted in a single center and these results may not be generalizable to other institutions. In addition, there was often inadequate documentation of the evaluation performed because of contaminated blood cultures in inpatients. We only attributed workup to contaminated cultures if there was clear documentation by the providers. We also missed the workup, treatment, and charges incurred by children with contaminated cultures (or controls) at outside EDs and the PMD offices. In addition, some of the contaminants were treated as pathogens by the medical team and incurred invasive testing, procedures, and treatments, which might have been unnecessary. Thus, our results on the increased resource use due to the workup and treatments triggered by contaminated cultures are an underestimate. We were also unable to calculate the charges specifically attributable to the additional LOS for the inpatients.

Our study is a case-control study, which is inherently subject to selection bias. We minimized this with multiple matching criteria and a 1:2 case to control patient ratio. The similarity in demographics and clinical presentation at the time of blood culture between case and control patients argues against selection bias.

This study highlights the importance of critically evaluating the need for blood cultures in children. In many instances, such as in children with uncomplicated skin and soft tissue infection or community-acquired pneumonia, blood cultures have limited utility and often do not change management.2426  In these instances, the benefit of the blood culture in detecting bacteremia is offset by the significant burden posed by a potential contaminated culture. This is especially a concern because rates of contaminated cultures are often twofold to fourfold higher than true-positive culture results.27,28 

On the basis of our study results, we have successfully implemented a quality improvement initiative to reduce rates of contaminated blood cultures at our hospital (study in progress). We are also planning a quality improvement initiative to reduce the number of unnecessary blood cultures obtained in children at our hospital.

We sincerely thank Angela Niesen BS, MPH, for obtaining charge data from the health information management system. We also sincerely thank Timothy Casper MD for assistance with chart review.

Dr Srinivasan is the principal investigator and conceptualized and designed the study, collected, analyzed, and interpreted the data, and drafted the initial manuscript; Dr Farrell participated in the study design and data collection, analysis, and interpretation; Drs Bram, Messer, and Mathew participated in the study design and data collection, analysis, and interpretation; Dr Hayes and Ms Gu participated in the study design and analysis and interpretation of data; and all authors critically reviewed the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

FUNDING: Supported by the I2 WUSM ICS and NCI Cancer Center Support Grant P30 CA091842, Siteman Comprehensive Cancer Center, supporting the REDCap clinical data capture service as a research resource at WUSM.

1
Hall
KK
,
Lyman
JA
.
Updated review of blood culture contamination
.
Clin Microbiol Rev
.
2006
;
19
(
4
):
788
802
2
Dawson
S
.
Blood culture contaminants
.
J Hosp Infect
.
2014
;
87
(
1
):
1
10
3
Self
WH
,
Talbot
TR
,
Paul
BR
,
Collins
SP
,
Ward
MJ
.
Cost analysis of strategies to reduce blood culture contamination in the emergency department: sterile collection kits and phlebotomy teams
.
Infect Control Hosp Epidemiol
.
2014
;
35
(
8
):
1021
1028
4
Alahmadi
YM
,
Aldeyab
MA
,
McElnay
JC
, et al
.
Clinical and economic impact of contaminated blood cultures within the hospital setting
.
J Hosp Infect
.
2011
;
77
(
3
):
233
236
5
Bates
DW
,
Goldman
L
,
Lee
TH
.
Contaminant blood cultures and resource utilization. The true consequences of false-positive results
.
JAMA
.
1991
;
265
(
3
):
365
369
6
Gander
RM
,
Byrd
L
,
DeCrescenzo
M
,
Hirany
S
,
Bowen
M
,
Baughman
J
.
Impact of blood cultures drawn by phlebotomy on contamination rates and health care costs in a hospital emergency department
.
J Clin Microbiol
.
2009
;
47
(
4
):
1021
1024
7
Hall
RT
,
Domenico
HJ
,
Self
WH
,
Hain
PD
.
Reducing the blood culture contamination rate in a pediatric emergency department and subsequent cost savings
.
Pediatrics
.
2013
;
131
(
1
).
8
Weddle
G
,
Jackson
MA
,
Selvarangan
R
.
Reducing blood culture contamination in a pediatric emergency department
.
Pediatr Emerg Care
.
2011
;
27
(
3
):
179
181
9
Murofushi
Y
,
Furuichi
M
,
Shoji
K
, et al
.
Adverse economic impact associated with blood culture contamination in a pediatric emergency department
.
Pediatr Infect Dis J
.
2018
;
37
(
8
):
755
758
10
Segal
GS
,
Chamberlain
JM
.
Resource utilization and contaminated blood cultures in children at risk for occult bacteremia
.
Arch Pediatr Adolesc Med
.
2000
;
154
(
5
):
469
473
11
Gilboy
N
,
Tanabe
P
,
Travers
D
,
Rosenau
AM
,
Eitel
D
.
Emergency Severity Index, Version 4: Implementation Handbook
.
Rockville, MD
:
Agency for Healthcare Research and Quality
;
2005
12
Faron
ML
,
Buchan
BW
,
Ledeboer
NA
.
Matrix-assisted laser desorption ionization-time of flight mass spectrometry for use with positive blood cultures: methodology, performance, and optimization
.
J Clin Microbiol
.
2017
;
55
(
12
):
3328
3338
13
Centers for Disease Control and Prevention, National Healthcare Safety Network
.
Master organism list/common commensals
.
14
Harris
PA
,
Taylor
R
,
Thielke
R
,
Payne
J
,
Gonzalez
N
,
Conde
JG
.
Research Electronic Data Capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support
.
J Biomed Inform
.
2009
;
42
(
2
):
377
381
15
Taddio
A
,
Shah
V
,
Gilbert-MacLeod
C
,
Katz
J
.
Conditioning and hyperalgesia in newborns exposed to repeated heel lances
.
JAMA
.
2002
;
288
(
7
):
857
861
16
Taddio
A
,
Goldbach
M
,
Ipp
M
,
Stevens
B
,
Koren
G
.
Effect of neonatal circumcision on pain responses during vaccination in boys
.
Lancet
.
1995
;
345
(
8945
):
291
292
17
Peters
JWB
,
Koot
HM
,
de Boer
JB
, et al
.
Major surgery within the first 3 months of life and subsequent biobehavioral pain responses to immunization at later age: a case comparison study
.
Pediatrics
.
2003
;
111
(
1
):
129
135
18
Friedrichsdorf
SJ
,
Postier
A
,
Eull
D
, et al
.
Pain outcomes in a US children’s hospital: a prospective cross-sectional survey
.
Hosp Pediatr
.
2015
;
5
(
1
):
18
26
19
Cummings
EA
,
Reid
GJ
,
Finley
GA
,
McGrath
PJ
,
Ritchie
JA
.
Prevalence and source of pain in pediatric inpatients
.
Pain
.
1996
;
68
(
1
):
25
31
20
Gibson
MK
,
Crofts
TS
,
Dantas
G
.
Antibiotics and the developing infant gut microbiota and resistome
.
Curr Opin Microbiol
.
2015
;
27
:
51
56
21
Gasparrini
AJ
,
Crofts
TS
,
Gibson
MK
,
Tarr
PI
,
Warner
BB
,
Dantas
G
.
Antibiotic perturbation of the preterm infant gut microbiome and resistome
.
Gut Microbes
.
2016
;
7
(
5
):
443
449
22
Ni
J
,
Friedman
H
,
Boyd
BC
, et al
.
Early antibiotic exposure and development of asthma and allergic rhinitis in childhood
.
BMC Pediatr
.
2019
;
19
(
1
):
225
23
Dydensborg Sander
S
,
Nybo Andersen
AM
,
Murray
JA
,
Karlstad
Ø
,
Husby
S
,
Størdal
K
.
Association between antibiotics in the first year of life and celiac disease
.
Gastroenterology
.
2019
;
156
(
8
):
2217
2229
24
Malone
JR
,
Durica
SR
,
Thompson
DM
,
Bogie
A
,
Naifeh
M
.
Blood cultures in the evaluation of uncomplicated skin and soft tissue infections
.
Pediatrics
.
2013
;
132
(
3
):
454
459
25
Kwon
JH
,
Kim
JH
,
Lee
JY
, et al
.
Low utility of blood culture in pediatric community-acquired pneumonia: an observational study on 2705 patients admitted to the emergency department
.
Medicine (Baltimore)
.
2017
;
96
(
22
):
e7028
26
Trenchs
V
,
Hernandez-Bou
S
,
Bianchi
C
,
Arnan
M
,
Gene
A
,
Luaces
C
.
Blood cultures are not useful in the evaluation of children with uncomplicated superficial skin and soft tissue infections
.
Pediatr Infect Dis J
.
2015
;
34
(
9
):
924
927
27
Theodosiou
AA
,
Mashumba
F
,
Flatt
A
.
Excluding clinically significant bacteremia by 24 hours in otherwise well febrile children younger than 16 years: a study of more than 50,000 blood cultures
.
Pediatr Infect Dis J
.
2019
;
38
(
9
):
e203
e208
28
Sard
B
,
Bailey
MC
,
Vinci
R
.
An analysis of pediatric blood cultures in the postpneumococcal conjugate vaccine era in a community hospital emergency department
.
Pediatr Emerg Care
.
2006
;
22
(
5
):
295
300

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The 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.

Supplementary data