|Fast track and the pediatric emergency department: resource utilization and patient outcomes
|Hampers L C, Cha S, Gutglass D J, Binns H J, Krug S E
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
Use of on-site fast track (FT) in the management of patients arriving at the emergency department (ED) with non-urgent presentations. In accordance with hospital policy, all patients were triaged by a qualified registered nurse (RN) at a desk near the ED entrance. Triage RNs were nurses with more than six months of ED experience who had completed a specialised training curriculum. Standardised triage protocols incorporating both objective and subjective findings (which were developed and implemented several years prior to the establishment of the FT) were used to assign patients to one of four categories ('emergent', 'urgent high', 'urgent low', or 'nonurgent'). Nonurgent patients were defined as those with an uncomplicated history and reassuring physical examination for whom a substantial delay (up to 8 hours) in medical attention was considered unlikely to result in serious harm. The FT was located in the hospital roughly 150 feet from the ED. Staffing and resources in the FT consisted of a board-certified or board-eligible paediatrician, an RN, a registration clerk, and four examining rooms. The ED staffing consisted of paediatric, emergency medicine, and family practice residents supervised by board-certified paediatric emergency medicine (PEM) attendings and fellows. Evening and weekend care was provided in the FT; at all other times these low-acuity patients were seen in the ED.
Type of intervention
Diagnosis, treatment and organisational issues.
Economic study type
Nonurgent paediatric patients presenting to EDs for acute care. Patients included were: triaged as 'nonurgent'; aged 2 months to 10 years; not chronically ill; and had fever, or complaint of vomiting, diarrhoea, or decreased oral intake.
Hospital. The economic analysis was carried out in the USA.
Dates to which data relate
Effectiveness and resource use data corresponded to patients managed in the study institution from September to December 1997. The price year was not explicitly specified.
Source of effectiveness data
The evidence for the final outcomes was based on a single study.
Link between effectiveness and cost data
Costing was performed prospectively on the same patient sample as that used in the effectiveness analysis.
Power calculations were not used to determine the sample size. The study sample consisted of 1,036 patient visits, for which physician providers completed study forms. There were 557 patients with a mean (SD) age of 37 (30) months in the ED group and 479 patients with a mean (SD) age of 33 (28) months in the FT group. Review of daily ED records revealed that roughly 90% of eligible patient visits were appropriately included. A total of 15 attending physicians staffed both the ED and the FT (FT paediatricians), to serve as their own controls. A total of 25% of the ED patients were seen solely by these paediatricians and the remaining 75% were seen by the PEM-supervised house staff. The FT paediatricians saw 88% of the FT patients (the rest were seen by 'moonlighting' paediatric subspecialty fellows, who did not see patients in the ED).
This was a prospective cohort study, carried out at an urban, university-affiliated paediatric ED with a total annual patient volume of some 39,000 patients. The mean (SD) duration of the follow-up was for 7 (2.8) days after the ED visit. In total, about 10% of eligible patients were inappropriately excluded from the study mainly because of clerical staff failure to attach a study form to the ED record and failure of physicians to complete the study form. Regarding the loss to follow-up, it was reported that telephone follow-up interviews were completed for 480 of 755 families with working telephone numbers (64%); resulting in a fairly low overall follow-up rate of 46%. Physicians, blinded to the cohort nature of the design, consented to a study of 'resource utilisation patterns', with assurance that information on individual providers would remain coded and anonymous. Triage requirements for nonurgent patients did not vary based on time of day, and the triage nurses were blinded to the study investigation. A registration clerk attached a data form to all patient charts after triage. The study forms required the physicians to identify nonurgent patients fulfilling the study inclusion criteria. Following the visit, the primary investigators reviewed the medical records of included patients. Inappropriately excluded patients did not differ from included patients on any demographic or clinical characteristics. In the two settings, the patients lost to follow-up were comparable with patients with successful follow-up in terms of age, triage temperature, race/ethnicity, and insurance status.
Analysis of effectiveness
The principle used in the analysis of effectiveness was treatment completers only. The clinical outcome measures were rates of admission, use of intravenous fluids, and intensity of diagnostic investigations. The chief outcomes of interest in the 7-day follow-up telephone interviews were subjective patient improvement, unscheduled follow-up care, and family satisfaction. Extra comparisons were made between the management decisions of the paediatricians who staffed both the ED and the FT in the two corresponding settings (they served as their own controls to isolate the effect of the FT environment on individual physician's medical decision making) and with those of the PEM-supervised house staff in the ED. The patients in the two settings did not differ in age, clinical condition, race, or commercial insurance status.
The effectiveness results were as follows:
The FT And ED patients did not differ at follow-up: 90% versus 88% had improved conditions (p=0.53), 18% versus 15% had received unscheduled follow-up care (p=0.44), and 94% of the families in both groups were satisfied with the visit (p=0.98).
Rate of admission was 1 (0.2%) case in the FT group versus 13 cases (2.3%) in the ED group (OR 12; 95% CI: 1.6 - 241).
Intravenous fluid bolus was given to 3 (0.6%) of the FT patients and 25 (4.5%) of the ED patients (OR 7.7; 95% CI: 2.2 - 32).
24% of the FT patients versus 41% of the ED patients had tests performed.
When compared with patients managed by FT paediatricians in the ED, PEM-supervised house staff involvement was not associated with any significant differences in care.
However, the care of both groups of ED patients differed from that of FT patients seen by the FT paediatricians. Both groups of ED patients were more likely to be admitted and receive intravenous fluids than FT patients.
It was noteworthy that despite differences in assessment and management, no effect on the outcomes of patients contacted for follow-up could be detected. There was no difference in the rate of unscheduled follow-up visits ('bounce-backs') either to the study ED or elsewhere. In addition, family satisfaction was uniformly high. Comparing the FT physicians' outcomes in FT with their outcomes in the ED shows that even the same providers seem to have behaved differently in the two settings. These findings suggest that there are elements intrinsic to the ED environment that prompt a higher rate of test ordering.
Measure of benefits used in the economic analysis
No summary benefit measure was identified in the economic analysis, and only individual clinical outcomes, reported in the effectiveness results, were included.
Costs were not discounted due to the short time frame of the cost analysis. The quantities of lengths of stay (LOS) and intensity of diagnostic investigations were reported separately from the costs. Some cost items were reported separately. Cost analysis covered the costs of diagnostic investigations. The perspective adopted in the cost analysis was not explicitly specified. Charge data were used instead of true costs. LOS data were reported but were not valued in monetary terms. The price year was not explicitly specified.
Statistical analysis of costs
The Mann-Whitney U test was used to compare the study groups in terms of charges for diagnostic testing. The Student's t test was used to compare the groups in terms of mean length of stay.
Indirect costs were not considered.
No sensitivity analysis was carried out.
Estimated benefits used in the economic analysis
See effectiveness results above.
The patient mean (SD) test charges were $27 (63) for the FT and $52 (91) for the ED (p<0.01). Both groups of ED patients (those who were treated by the FT paediatricians and those who were treated by the PEM-supervised house staff) were more likely than the FT patients to have higher test charges and longer lengths of stay. In the two settings, the patients lost to follow-up were comparable with patients with successful follow-up in terms of test charges (p=0.99), and length of stay (p=0.11).
Synthesis of costs and benefits
Costs and benefits were not combined.
Compared with those in the main ED, the study patients seen in the FT had fewer tests ordered and had briefer lengths of stay. Differences in patient ages, vital signs, or demographic characteristics did not explain these findings. No difference in final outcomes or satisfaction was detected among the families contacted for follow-up.
CRD COMMENTARY - Selection of comparators
The strategy of using the main ED, as the traditional approach, was regarded as the comparator. You, as a database user, should consider whether this is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness
The internal validity of the effectiveness results cannot be guaranteed due to the non-randomised nature of the study design (which resulted from logistical factors), lack of power analysis, and the fact that the effectiveness analysis was based on treatment completers only. As acknowledged by the authors, despite the fact that the two study groups had comparable clinical profiles the following confounding factors may have played a role in determining the effectiveness outcomes: subjective distinctions could explain some of the differences in test orderings (particularly among the overnight house staff), independent of objective clinical characteristics; the subjective parameters (such as the perception of nursing regarding parental fatigue and anxiety) in the application of the triage criteria may have played a role in influencing triage destinations during late-night hours, culminating in 'up-triaging' and subsequent exclusion of such patients (biasing the findings toward the null hypothesis); it is conceivable that other independent variations existed within the two study cohorts (elements of the history, social and family conditions, etc.) that impacted the study dependent variables of interest; the fairly low overall follow-up rate (46% of the total sample) may have resulted in sampling bias. The study sample appears to have been representative of the study population (with a fairly narrow range of presentations such as fever, vomiting, diarrhoea, or decreased oral intake).
Validity of estimate of measure of benefit
The authors did not derive a measure of health benefit. The analysis was therefore one of cost-consequences design.
Validity of estimate of costs
The following features enhanced the validity of the cost results: some quantities were reported separately from the costs; adequate details of cost estimation were given; and statistical analysis was performed on resource use and cost data. However, the following features may have adversely affected the validity: the costs were based on charge data rather than true costs; the price year and the perspective adopted in the cost analysis were not explicitly specified; LOS data were not converted into monetary values thus preventing calculation of the total cost; the effects of the alternative procedures on indirect costs (productivity loss) were not addressed; the cost results may not be generalisable outside the study setting.
In view of the limitations of the study design and the lack of sensitivity analysis, some degree of caution may need to be exercised in interpreting the study results. The issue of generalisability to other settings or countries was partially addressed by acknowledging that the low response rate may limit the generalisability of the study findings. Some comparisons were made with other studies. The issue of whether the study sample was representative of the study population was addressed by pointing out that the study included a fairly narrow range of presentations (such as fever, vomiting, diarrhoea, or decreased oral intake), the aim being to minimise patient-to-patient variability when comparing physician decision-making process; furthermore, it was believed that this range of complaints presented clinicians with a large degree of discretion regarding test ordering.
Implications of the study
The authors believed that the study findings should encourage busy EDs to continue their development of the fast track paradigm. For nonurgent paediatric patients presenting to EDs for acute care, fast track may provide a substantially more efficient decision-making environment without decreasing the quality of care. In an attempt to explain the differences in resource utilisation of the same FT paediatricians, the following series of hypotheses was put forward in informal interviews (which presumably can be further investigated in future studies):
(1) as the sole providers for a large volume of patients, the FT paediatricians reported feeling pressured to see patients quickly and were thus discouraged from ordering extensive and unnecessary investigations; they appeared to be insulated from such pressures in the main ED because of diffusion of responsibility for patient flow;
(2) the FT paradigm creates a 'bias toward wellness' in the judgement of the physicians;
(3) subtle clinical clues may vary in the two settings; conceivably, paediatric patients may appear fussy or cranky in the noisy and harsh milieu of the main ED, whereas similar patients may be calm in the quieter FT environment;
(4) there are differences in the intensity of human and technological monitoring of patients in the two settings (the routine availability of pulse oximetry in the ED, for instance);
(5) the queuing rules and other differences in clerical and nursing staffing may have contributed to the more rapid management of FT patients.
Source of funding
Supported in part by a Special Projects Grant from the Ambulatory Pediatric Association.
Hampers L C, Cha S, Gutglass D J, Binns H J, Krug S E. Fast track and the pediatric emergency department: resource utilization and patient outcomes. Academic Emergency Medicine 1999; 6(11): 1153-1159
Subject indexing assigned by NLM
Chi-Square Distribution; Child; Child, Preschool; Clinical Competence; Cohort Studies; Emergency Service, Hospital /organization & Emergency Treatment /methods; Evaluation Studies as Topic; Female; Health Resources /utilization; Humans; Infant; Male; Outcome Assessment (Health Care); Pediatrics /statistics & Prospective Studies; Statistics, Nonparametric; Time and Motion Studies; Triage /organization & United States; Urban Population; administration; administration /utilization; numerical data
Date bibliographic record published
Date abstract record published