Time to antibiotic administration in children with febrile neutropenia: Report from a low middle-income country

In the management of childhood cancers, prompt treatment of febrile neutropenia (FN) constitutes an essential aspect, especially in developing countries is where the mortality in childhood cancer secondary to infections, is higher in comparison to high-income countries12. Morbidity and mortality varies widely in children presenting with FN3. A relatively well-looking child might progress to septic shock and multiorgan failure within a matter of hours, emphasizing the importance of immediate attention to every febrile neutropenic child as soon as they present to the hospital14. Although the risk varies depending on the phase of chemotherapy, duration of fever and focus of infection, significant morbidity and mortality can be prevented by timely administration of the first dose of intravenous antibiotic as proven time and again by various studies conducted all over the world14567. Time to antibiotic administration (TTA) has been proposed as a quality of care measure for infectious diseases16. Worldwide consensus for the standard time to antibiotic administration is 60 min, as infection in these patients has a potential for rapid progression689. Multiple factors are seen to be a detriment in achieving this target. The delay can be at various levels, including physician, nurse and other logistical factors.

The objective of this study was to assess the time taken to receive the first dose of antibiotics, i.e. TTA, after triaging in the setting of a low middle-income country referral hospital and evaluate the main causes of delay.

Material & Methods

This was a prospective, cross-sectional study on children with malignancy on treatment in the Paediatric Haematology-Oncology Unit, Advanced Pediatric Centre, Postgraduate Institute of Medical Education and Research, Chandigarh who presented with FN. This study was conducted for nine months between September 2014 and June 2015 after prowring approval from the Institutional Ethics Committee. Children up to the age of 18 yr were included in the study. Patients treated with antibiotics on an ambulatory basis for the episode of FN before admission, patients with fever at the time of the diagnosis of cancer and those who presented with haemodynamic instability were excluded. Only one episode per patient was recorded. Patient enrolment was done by a non-probability sampling technique, wherein the patients who met the practical criteria of being available at a given time were enlisted.

Criteria for assessment of study parameters:

Primary outcome: This was taken as the proportion of patients with FN who received antibiotics within 60 min of triage assessment.

Secondary outcome: This included the length of hospital stay, transfer to paediatric intensive care unit (PICU) within 24 h of admission and death in that particular hospital admission.

Time of day, place of presentation, weekend versus weekday presentation and the presence of focus of infection were the predictor variables evaluated for TTA.

Patients who did not meet the primary outcome of receiving antibiotics within an hour of triage were subjected to a root cause analysis to help identify the cause of delay.

Febrile neutropenia (FN): Presence of a single oral/axillary temperature of ≥38.3°C (101°F) or a temperature of ≥38.0°C (100.4°F) for greater than or equal to one hour along with an absolute neutrophil count <500/mm³, or <1000/mm³ with expected decline to <500/mm³ within the next two days was considered for inclusion of children as FN.

Time to antibiotic administration (TTA): The time from triage assessment to the time of receipt of the first dose of antibiotic was evaluated as TTA.

Root cause analysis: The various causes for delay in antibiotic administration were divided into (i) physician-level causes, (ii) nursing-level causes, and (iii) logistical issues (Annexure 1). The causes of delay were investigated by the study Physician through a predesigned questionnaire, and the causes for delay were noted. This was done within 24 h of admission. The hospital personnel and the patient’s guardians were asked for details of delay as per the questionnaire. Informed consent was obtained from parents/guardians of all participants.

Process flow of patient triage and TTA: A child being treated for malignancy at our centre is typically advised to attend the Pediatric Oncology Clinic (Tuesdays and Fridays), Paediatric Oncology Daycare (PODC) (daily 9:00 AM – 5:00 PM, Sundays 9:00 AM –1:00 PM) and the Haematology-Oncology ward (HO ward) for any concerns such as fever, vomiting, loose stools and lethargy. After 5:00 PM (after 1:00 PM on Sundays), patients have instructions to attend the HO ward and not the Pediatric emergency room. The flow of a walk-in patient with FN is depicted in the Figure. Our hospital has only inpatient beds and no outpatient beds. The resident staffs posted in the HO ward are under instructions to attend to such patients as soon as a possible triage form is filled by the physician on duty (occasionally nurse in the daycare) and requisite blood samples are taken before initiation of antibiotics in children with FN. The time of triage is entered on the triage form.

Close

Figure

Flow of patients with febrile neutropenia.

Export to PPT

The time of charting antibiotics is entered on the patient’s chart by the physician and the time of administration of the antibiotics is entered by the nurse. Typically, a peripheral cannula is inserted for the same as central lines are sparingly used. Antibiotics are to be procured from the pharmacy by the patients’ family and the nurse prepares the antibiotics for administration. Antibiotics are often given before blood count reports if a child is in the intensive phase of therapy as a complete blood count report may take from two to four hours to be received. At the time of doing the study, our laboratory reports were not computerized.

Possible predictors of delayed administration: These included (i) time of presentation, (ii) place of presentation (PODC; HO ward; clinic) and (iii) presence of focus of infection. Time of presentation was subdivided into day (9:00 AM to 5:00 PM); evening (5:00 PM to 10:00 PM) and night (10:00 PM to 9:00 AM). Monday to Saturday 2.00 PM was week day time (Our hospital has a five and a half-day week); Saturday afternoon to Monday morning 9:00 AM was weekend time (Annexure 2). Patients who had fever with normal counts (even if they received parenteral antibiotics) were excluded from the study. Patients were enrolled randomly for the study and the study physician, recorded all clinical details in the study case record form at enrolment. Enrolment to the study was done within 24 h of admission.

Root cause analysis for the delay were also recorded (Annexure 2) as (i) Physician level delays included (a) not informed regarding the patient, (b) busy with other work, (c) awaiting blood results, (d) unaware of the need of antibiotics, (e) decision to defer parenteral antibiotics, (f) difficult intravenous access; (ii) Nursing level delay included (a) not informed regarding patient, (b) delay in preparing antibiotics, (c) busy with other work, (d) no intravenous access on the patient; (iii) Logistical issues considered were: (a) delay in allotment of bed, (b) delay in procurement of antibiotics, (c) time spent in radiology for X-ray, etc.

Statistical analysis: Descriptive analysis included measures of central tendency (mean, median and mode) as well as measures of dispersion [standard deviation, interquartile range (IQR)] for continuous variables and frequencies as well as proportions for categorical variables. The outcome variable of time to receive antibiotics after hospitalization was analyzed as a dichotomous variable (≤60 min vs. >60 min) using the effect size measure of odds ratio (OR). Univariate risk factors for delayed administration of antibiotics (beyond 60 min) were statistically analyzed using non-parametric Mann–Whitney U test and Chi-square test (with Fisher’s modification as required). All statistical tests were two-tailed and limit of significance was taken as <5 per cent. The statistical analysis was performed using SPSS® software version 20 (IBM Corp., Chicago, IL, USA). A sample size of 191 produced a two-sided 95 per cent confidence interval (CI) with a width equal to 0.14 when the sample proportion was 0.35. Sample size calculation was performed using PASS software 2019 vs. 19.03. To account for attrition, the final sample was increased to 211 patients. No assumptions were made.

Results

A total of 211 children with FN were evaluated for time to antibiotic (TTA) administration. Seventy per cent had an underlying diagnosis of acute lymphoblastic leukaemia. The mean age was six years (four months-15 yr), median being five years (IQR: 3-9 yr). The demographic details are given in Table I. Median time from triage to antibiotic administration (TTA) was 60 min [interquartile range (IQR): 30-120 min]. The 95 per cent CI of the proportion of patients who received antibiotics within 60 min was 59.6-72.3 per cent.

Sixty six per cent of patients received their antibiotics within one hour, with 79 per cent having received their antibiotics within two hours. Seventy one patients (34%) had a delay in their antibiotic administration, TTA being >60 min.

Delayed TTA was correlated with the possible predictor variables, which included time and place of presentation, and presence of focus of infection (Table II).

In terms of time of presentation the odds of delayed TTA in participants assessed in the day was (OR=1.37; 95% CI; 0.73-2.57; P=0.29) higher than those assessed during other times, however, not significantly. Similarly, the odds of delayed TTA in patients seen in the evening was higher (OR=1.13; 95% CI; 0.56-2.21; P=0.70) compared to those assessed during the day and night, but again not significantly. Interestingly, most of the patients who presented at night (81%) had a TTA ≤60 min, the odds of delayed TTA (OR=0.40; 95% CI; 0.13-1.08; P=0.053) at night was lower compared to other times, but it was not significant. Odds of delayed TTA were higher over the weekend (OR=1.06; 95% CI; 0.5-2.17; P=0.86).

In terms of place of presentation, the odds of delayed TTA were found to be greater in participants who were assessed in the PODC as compared with those assessed in other areas (Paediatric Oncology Clinic and HO ward), and it was however, not significant. (OR=1.37; 95% CI; 0.74-2.52; P=0.58). Twenty seven per cent of the participants who had a delayed TTA were seen in the PODC.

Children who had a focus of infection at presentation received antibiotics earlier with the odds of delayed TTA being higher (OR=1.72; 95% CI; 0.73-4.41; P=0.40) in participants who had no focus of infection. The association between TTA and various predictor variables is given in Table II.

The mean duration of hospital stay in children with a TTA ≤60 min, was 7.7 days compared to 6.6 days in participants whose TTA was >60 min. (P=0.36). The median duration of hospital stay in children with a TTA ≤60 min, was six days (IQR: 4-9 days) which was similar to those with a TTA of >60 min of six days (IQR: 4-9 days). None of children were admitted to PICU within 24 h of admission. However, three patients needed PICU admissions after 24 h of their hospital stay. None of the study participants succumbed during admission.

Root cause analysis was done to evaluate the cause for delay at physician or nursing level and for other logistical issues. The cause for impediment, which led to a delayed TTA is given in Table III. Twenty episodes (28%) were found to have a delay at more than one level. Fifty two of 71 patients had a delay in TTA owing to physician level wait. Awaiting blood results (40%), busy with another patient (21%) and a delay in the decision for antibiotics (15%) were the most cited reasons. Twenty three of 71 patients had a delay attributable to the nursing level, with 65 per cent of cases being delayed secondary to the time taken for preparing antibiotics. Twenty six per cent of delays were due to the nursing staffs being busy with other work. Eight per cent of delays were observed to be secondary to the nurse not knowing about the patient. Postponement in delivery of medications secondary to logistics was observed in 31/71 patients (32 episodes), representing 44 per cent of all delays. The most common reason seen in 21 (64%) cases was attributable to the time taken for getting an inpatient file and a bed allotted in the ward, the rest being ascribable to the time taken for procuring antibiotics.

Discussion

The cure rates of childhood cancers are a success story of the last century10. Mortality due to sepsis, however, remains a problem in the management of cancers with the potential for rapid progression of infection in children with low counts secondary to therapy/malignancy11 per se. This demands promptness of initial antibiotic administration in a child presenting with FN. Given the many competing demands that exist in an oncology ward, especially in a low-middle-income country where resources (doctor/nurse:patient ratio) are limited, completing any task in a specified time frame can be challenging2. Although simple in principle, achieving high levels of compliance with early antibiotic administration can be a challenge. Early administration of antibiotics has been proposed as a quality indicator for oncology centres caring for febrile neutropenic patients1489. However, few studies have examined the prevalence and predictors of antibiotic delay in FN12.

Two-thirds (66%) of our patients received antibiotics within the stipulated time of 60 min. Adult studies have reported a varying time ranging from 107 to 216 min as the time for administering antibiotics13141516. Studies in children with malignancies have also reported a wide range of TTA, with the goal of a TTA of one hour being achieved in 1-50 per cent patients61718192021222324, with only one study reporting 80 per cent patients to have received antibiotics in the stipulated time of two hours21 and two studies reporting no patient to have received antibiotics in the given hour1720. Median waiting time from emergency room admission to examination has been observed as 75 min, with 210 min to receive antibiotics and 5.5 h to hospital admission15. A quality improvement project to reduce treatment delays in patients with FN revealed three areas needing improvement: (i) inpatient orders, (ii) the admission communication process, and (iii) multidisciplinary staff accountability, which after implementation resulted in nearly 50 per cent reduction in cycle time on the inpatient unit16.

Predictor variables assessed included time of presentation during the day and week, place of evaluation and focus of infection for their effect on a delay in TTA. Patients who were assessed in the night received antibiotics earlier compared to those assessed during any other time of the day. The resident staff on evening/night were under instructions to attend to such patients at the earliest. This along with a possible feeling that a patient coming at night has a greater likelihood of being sick and should not wait could explain the short TTA at night compared to daytime. Delay in the daycare was an unexpected finding. Competing demands on the daycare personnel could be the plausible reason for delay.

Patients admitted over the weekend had a longer TTA with 65.3 per cent facing a delay as compared to 33.3 per cent delay in those admitted on weekdays, though the difference was not significant. This finding was similar to that reported earlier where patients who arrived during the weekend had a higher percentage of composite adverse event outcomes24. The likely reason could be less physician staff on duty over the weekend.

The odds of delayed TTA was lesser by 20 per cent in patients who were assessed in the clinic (Tuesdays and Fridays) versus the rest of the areas (PODC and HO Ward-daily). Fletcher et al24 reported, patients who presented to the emergency room to have a longer TTA (145 min) compared to those who presented to the inpatient (60 min) or clinic (93 min). A survey of Paediatric Oncology centres, reported that >50 per cent of patients who presented to the emergency room had a mean TTA over 60 min whereas among those who presented to the outpatient department and inpatient units, the proportion was <25 per cent17. Children who had a focus of infection received antibiotics earlier. This could be attributable to the fact that finding a focus of infection prompts the physician to administer antibiotics straightaway.

Literature reports a delayed TTA to be associated with prolonged hospitalization and an adverse outcome32526, as also to have no correlation with mortality and hospital stay71127. We did not find a correlation between delay in antibiotic administration and the outcome (duration of hospital stay, PICU admission within 24 h of admission and death). It is important to note that children who presented with hemodynamic compromise were excluded in this study. The major causes of delay were waiting for blood results, time spent in preparing antibiotics, time in getting an inpatient file and allocation of a bed. Studies have reported the availability of laboratory results, physician evaluation and venous access to be major causes of delay182028. Complete blood count testing before physician assessment also delays antibiotic administration in FN patients12. Time in preparing and procuring antibiotics can be circumvented by having an emergency antibiotic box for FN children in the PODC and the HO ward. Administrative issues of bed availability are difficult in the best of centers and an effort can be made to administer antibiotics while awaiting a bed. Possibly, one can recommend initiating antibiotics in anticipation of neutropenia in a ‘sick’ looking child to avoid delay 29.

Limitations of the study include restricted non-randomized enrolment owing to logistic issues. Residents and nurses were aware of the study contributing to possible bias. The causes for delay ascertained were based on memory recall by doctors and nurses. We did not include children who presented with hemodynamic compromise as they were attended to as per the emergency protocol.

Overall, this study shows that in resource-constrained settings, a TTA of ≤60 min was achieved in 66 per cent of our patients, as recommended by the surviving sepsis campaign. Root cause analysis identified factors on which targeted intervention could reduce the TTA. Interventions required to improve the quality measure include a rapid assessment protocol, having an emergency antibiotic trolley/box and an automatic token system to alert the arrival of the patient to the staff on duty, which should be a standard operating protocol in all paediatric cancer units2830313233. We have instituted an emergency antibiotic box and have instructions in the ward to administer antibiotics before allocation of a bed, after this study. Future studies can test the effect of these measures on TTA.