OUP user menu

Paediatric utilization of an emergency department in Italy

Claudia Pileggi , Gianluca Raffaele , Italo F. Angelillo
DOI: http://dx.doi.org/10.1093/eurpub/ckl019 565-569 First published online: 14 February 2006

Abstract

Background: The purpose of this study was to determine the frequency and characteristics of paediatric attendance as a source of medically non-urgent problems at an accident and emergency department (ED) of a public non-teaching hospital in Crotone (Italy). Methods: For each patient aged 16 years or younger, there were collected information on demographics and socioeconomic characteristics, medical history, route of referral, clinical complaints that they presented at the moment of their presentation at the ED, duration of presenting problems prior to arrival, hour of arrival, day of the week of arrival, and reason for attending the ED. Data about the consultation process and the final decision made were also recorded. Results: Of a total of 980 patients included in the study, 27.6% had conditions that were definitely non-urgent. Multiple logistic regression analysis showed that the visit was non-urgent in younger population, in females, and in those attending the ED on the weekend. The results of the second multivariable regression analysis model indicate that patients who did not receive medical or surgical examination at the ED, with problems of longer duration prior to arrival at the ED, with non-traumatic injuries, and who did not require inpatient hospital admission were more likely to use the ED as a source of non-urgent care. The most frequent presenting problems for patient visits to ED were injury, respiratory diseases, and digestive symptoms. Conclusion: A closer cooperation within the health care organization system to provide a service responsive to the real needs of patients is essential.

  • emergency department
  • hospital
  • Italy
  • non-urgent visits
  • paediatric

Accident and emergency departments (EDs) have been created in hospitals with the primary function of providing immediate care for patients with life-threatening medical conditions, trauma, or injuries, but not to treat minor illnesses or provide primary care. In recent years, the ED has increasingly become a major provider of health care and this overcrowding has become problematic. A large proportion of the demand for ED utilization often may be attributed to visits for medical problems that do not require emergency treatment.

In the context of limited inpatient hospital resources, it is acknowledged that the phenomenon of medically non-urgent ED visits, which can be managed alternatively and appropriately in general practice, has raised serious concerns among health-care planners, both because of its magnitude and because the appropriate utilization of hospital care can significantly improve health outcomes. The use of ED by paediatric patients has become an important problem that must be solved, since medical, social, economic, and psychological factors mainly influence the parent's decision to visit the ED rather than manage their children at home either with or without additional support, or prior to making an unscheduled visit to an ED their demands could have been satisfactorily met by an appropriate visit at a different health care level.

There is a paucity of literature investigating the frequency of paediatric patients who present to the ED for medically non-urgent visits and it is imperative to generate data on this topic. Therefore, in this survey we sought to determine the frequency and characteristics of paediatric attendance at a hospital ED as a source of medically non-urgent problems in Italy.

Materials and methods

This cross-sectional study was undertaken between April and July 2004 in the ED of one public non-teaching hospital with 468 beds located in the area of Crotone (Italy). The catchment area of Crotone encompasses ∼60 000 inhabitants and this is the only hospital in the area of 1717 km2 with an ED.

All paediatric patients aged 16 years or younger attending the ED within 10 randomly selected week periods were studied. A medical interviewer, who was involved in health care, collected the following information for all patients agreeing to participate: demographics and socioeconomic characteristics (age, gender, birth order, parent(s)/guardian(s) employment status), medical history, route of referral, clinical complaints that they presented at the moment of their presentation at the ED, duration of presenting problems prior to arrival, hour of arrival, day of the week of arrival, and reason for attending the ED. When a child had working parents/guardians, the highest occupation was used. A total of 20 patients were excluded because they were judged to have severe impairments, requiring immediate medical or surgical attention, that interfered with completion of the verbal assessment tools and no one was available to provide the information. For each patient, the attending physician completed a form about the consultation process (investigations, medical or surgical examinations, treatment) and the final decision made.

The following five levels of definition of urgent care was developed before the beginning of the study according to the Guidelines for the Canadian Pediatric Triage and Acuity Scale1: conditions that are threats to life or limb (or imminent risk of deterioration) with immediate aggressive interventions (resuscitation); conditions that are potential threats to life, limb, or function with rapid medical intervention or delegated acts (emergent); conditions that could possibly progress to a serious status requiring emergency intervention, perhaps those associated with significant discomfort or dysfunction at work or activities of daily living (urgent); conditions relating to age, distress, or potential for deterioration or complications that would benefit from interventions or reassurance within 1–2 h (less urgent); conditions that are acute and non-urgent as well as conditions which may be part of a chronic problem with or without evidence of deterioration. The investigation or interventions for some of these illnesses or injuries could be delayed or even referred to other areas of the hospital or health care system (non-urgent).

Statistical analysis

In the primary analysis, t-test and the χ2-test were performed to assess differences between urgent and non-urgent care on several continuous and categorical explanatory variables, respectively. We performed multiple logistic regression analysis with forward elimination. The first outcome of interest was to determine the profile of the patients who attend the ED for non-urgent care and this model included the following explanatory variables: age (continuous), sex (1 = male, 2 = female), birth order (1 = 1, 2 = 2–3, 3 = >3), highest parents' occupation stated (1 = unemployed/retired, 2 = artisan/lower managerial, 3 = high professional and managerial), chronic disease (0 = no, 1 = yes), hour of arrival at the ED (1 = 8.00 am to 4.59 pm, 2 = 5.00 to 9.59 pm, 3 = 10.00 pm to 7.59 am), and day of the week of attending the ED (0 = weekend, 1 = weekday). The second outcome investigated the characteristics that were related to the utilization of the ED as a source of non-urgent care and in this model the variables included were the following: the person who referred the patient to the ED (1 = physician, 2 = self-referral/relatives), reason for attending the ED (1 = non-traumatic injuries, 2 = traumatic injuries), duration of presenting problem prior to arrival at the ED, in hours (1 = <1, 2 = 1–23, 3 = ≥24), number of investigations performed at the ED (continuous), medical/surgical examination performed at the ED (0 = no, 1 = yes), and the result of the visit at the ED (1 = hospital admission, 2 = other). The significance level for variables entering the models was set at 0.2 and that for removing from the models at 0.4. Adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated. Stata software was used to analyse the data.2

Results

A total of 980 patients, 16 years of age or younger, who registered for care in the ED, participated in the study. table 1 presents the principal characteristics of the sample and the distribution of non-urgent visits according to several explanatory characteristics. The mean age was 6.7 years (range 1 month to 16 years), two-thirds attended on weekend, almost half were brought in during daytime hours (8 am to 5 pm), only 7.7% were referred by health care professionals, approximately one-third sought care for problems that had been present for <1 h, and for more than half the most common general reason was injury.

View this table:
Table 1

Selected characteristics of the study population and distribution of urgent and non-urgent visits according to explanatory variables

CharacteristicsNa%UrgentNon-urgent
N%N%
Sex
    Male57959.144376.513623.5
    Female40140.926766.613433.4
χ2= 11.69, 1 df, P = 0.001
Age (years)
    0–224424.914157.810342.2
    3–520721.113766.27033.8
    6–920821.216177.44722.6
    10–1217117.513880.73319.3
    >1215015.313388.71711.3
t-test = 7.86, 978 df, P < 0.0001
Birth's order
    14254530070.612529.4
    233835.824773.19126.9
    ≥318219.213674.74625.3
χ2= 1.26, 2 df, P = 0.53
Stated chronic disease
    No92795.466771.926028.1
    Yes454.63782.2817.8
χ2= 2.27, 1 df, P = 0.13
Highest parent(s)' occupation
    Artisan/lower managerial54259.84017414126
    High professional and managerial28531.519869.58730.5
    Unemployed/retired798.749623038
χ2= 5.69, 2 df, P = 0.06
Day of the week of attending the ED
    Weekday32232.949074.516825.5
    Weekend65867.122068.310231.7
χ2 = 4.09, 1 df, P = 0.04
Arrival time at the ED
    8.00 am to 4.59 pm46547.532970.713629.3
    5.00 to 9.59 pm3633728177.48222.6
    10.00 pm to 7.59 am15215.510065.85219.2
χ2 = 8.52, 2 df, P = 0.01
Reason for attending the ED
    Non-traumatic injuries47548.524351.223248.8
    Traumatic injuries50551.546792.5387.5
χ2 = 209.34, 1 df, P < 0.0001
Referral to the ED
    Physician757.75877.31722.7
    Self/relatives90192.36497225228
χ2 = 0.98, 1 df, P = 0.32
Duration of presenting problem prior to arrival at the ED (hours)
    <125228.222790.1259.9
    1–2342347.331073.311326.7
    ≥2422024.511150.510949.5
χ2 = 92.61, 2 df, P < 0.0001
Investigation received at the ED
    No53354.48769.13930.9
    Yes44745.634883.96716.1
χ2 = 129.11, 1 df, P < 0.0001
Medical/surgical examination received at the ED
    No36937.731485.15514.9
    Yes61162.339664.821535.2
χ2 = 47.42, 1 df, P < 0.0001
Result of the visit at the ED
    Hospital admission15115.414394.785.3
    Other82984.656768.426231.6
χ2 = 44.28, 1 df, P < 0.0001
  • a: Numbers that do not add up to 980 are due to missing data for that variable

According to the judgment of the observers, 27.6% of the total number of patients surveyed at the ED had conditions that were definitely non-urgent. The univariate analysis indicated that of the patients' demographics and socioeconomic characteristics, attending to non-urgent care at the ED was significantly associated with the patient's age (t-test = 7.86, 978 df, P < 0.0001) and sex (χ2 = 11.69, 1 df, P = 0.001), being more frequently in those younger and in females. The number of accesses for non-urgent care was significantly higher for those patients attending the ED in the late evening or in the early morning (χ2 = 8.52, 2 df, P = 0.01), in the weekend (χ2 = 4.09, 1 df, P = 0.04), in those who attend the ED for non-traumatic injuries (χ2 = 209.34, 1 df, P < 0.0001), and in those with problems of longer duration (χ2 = 92.61, 2 df, P < 0.0001). The proportion of patients whose visit was considered non-urgent varied according to the number of investigations received in the ED, since the frequency of those receiving at least one investigation (χ2 = 129.11, 1 df, P < 0.0001) or a medical/surgical examination (χ2 = 47.42, 1 df, P < 0.0001) was significantly lower in non-urgent patients. Moreover, non-urgent care was significantly higher for those patients not requiring a hospital admission (χ2 = 44.28, 1 df, P < 0.0001). When the multivariate regression analysis was performed, the results basically did not change. In the first multivariate model in table 2 we included the demographic and social characteristics of patients attending the ED, in order to describe which of these predicted non-urgent care. The visit was non-urgent in younger (OR = 0.88, 95% CI = 0.85–0.91), in females (OR = 1.71, 95% CI = 1.26–2.32), and in those attending the ED on the weekend (OR = 0.71, 95% CI = 0.52–0.97). In the second multivariable regression analysis model, characteristics describing the ED utilization were added in order to explore which aspects were directly related to the utilization of the ED as a source of non-urgent care. Non-urgent care was significantly associated with the number of consultation activities performed in the ED and the outcome at discharge, since patients who were non-urgent did not receive medical or surgical examination (OR = 0.33, 95% CI = 0.23–0.48) and did not require inpatient hospital admission and were able to return to their original residence (OR = 19.15, 95% CI = 8.07–45.43). Moreover, non-urgent patients were more likely to have non-traumatic injuries (OR = 0.12, 95% CI = 0.08–0.2) and problems of longer duration prior to arrival at the ED (OR = 1.43, 95% CI = 1.04–1.96) (Model 2 in table 2).

View this table:
Table 2

Logistic regression model results

VariableORSE95% CIP
Model 1. Outcome: profile of the patients who attend the ED for non-urgent care
    Log-likelihood = −492.6, χ2 = 77.35, P < 0.0001
    Age0.880.020.85–0.91<0.0001
    Sex1.710.271.26–2.320.012
    Day of the week of attending the ED0.710.110.52–0.970.034
    Stated chronic disease0.580.230.26–1.280.175
    Birth's order0.90.090.74–1.110.334
Model 2. Outcome: utilization of the ED as a source of non-urgent care
    Log-likelihood = −346.24, χ2 = 360.74, P < 0.0001
    Result of the visit at the ED19.158.448.07–45.43<0.0001
    Medical/surgical examination received at the ED0.330.060.23–0.48<0.0001
    Reason for attending the ED0.120.030.08–0.2<0.0001
    Duration of presenting problem prior to arrival at the ED1.430.231.04–1.960.027
    Referral to the ED1.410.510.69–2.880.344

The most frequent presenting problems for patient visits to the ED were injury (51.5%), respiratory diseases (18.7%), and digestive symptoms (15.3%).

Discussion

We obtained a comprehensive picture of the characteristics of paediatric attendance at a hospital ED in Italy as a source of medically non-urgent problems. We found that 27.6% of ED attendance aged 16 years or younger presented as non-urgent, meaning that interventions could have been provided by a more appropriate health care source. Comparisons can be made with paediatric data obtained in other countries, but in doing so the difficulty owing to differences in the methodology and in the age groups of the population, in the organization and provision of health care, in the attitudes and in the behaviour of the physician and the patient should be taken into account (table 3). The prevalence of non-urgent visits reported in this study is comparable to previous observations in other hospitals. In the United States, similar results have been reported in a paediatric ED of an urban children's hospital with 21% of 426 adolescents aged 13–18 years identified as presenting with non-urgent conditions,3 and in a cohort of infants, born in the State of Missouri, 29% made at least one non-urgent visit.4 Moreover, our data was considerably lower than the values of non-urgent visits found in the United States, ranging from 61.6% in females for the group 11 years of age or younger and 47.5% in males in the 11–14 age group, observed in a sample of 418 ED5; the values of 46%6 and 51.6%,7 respectively, for children 0–18 years and with a mean age of 9.4 years; the values of 65% in children with a mean age of 6.2 in two urban hospitals8; and the values of 70% for children 0–18 years who attended EDs affiliated with paediatric residency training programmes.9 Finally, Hotvedt et al.10 in a study including children aged 0–10 years admitted to emergency rooms in seven general hospitals in a Norwegian county concluded that 94.7% did not need surgical or intensive medical treatment within 8 h of their arrival at the hospital.

View this table:
Table 3

Comparison of data for non-urgent paediatric utilization of emergency department from various countries

AuthorCountryAgeNon-urgent visits (%)
Isaacman and Davis (1993)USA≤1870
Melzer-Lange and Lye (1996)USA13–1821
Ziv et al. (1998)USA<1157.3–61.6
11–1447.5–57.5
15–1752.2–59.3
Hotvedt et al. (1999)Norway≤1094.7
Phelps et al. (2000)USA≤1665
Sharma et al. (2000)USA<129
Mistry et al. (2005)USA9.4 (mean)51.6
Zimmer et al. (2005)USA≤1846
Present studyItaly≤1627.6

As in the elderly population, the paediatric patient may be very vulnerable to inappropriately utilizing hospital facilities. A pertinent example provided in this study was the scenario indicating the profile of the patients visiting the ED for non-urgent care. Indeed, one of the most interesting observations of the multivariable analyses was that young age, being a female, attending the ED in the weekend, non-traumatic injuries, problems of longer duration, and problems that did not need a hospital admission were some of the predictors of patients to be with non-urgent problems. This inappropriate use is not a new phenomenon.11 However, emerging from the results of this study is that the majority of cases allocated to this scenario may be considered preventable and it illustrates that paediatric patients are important targets for efforts aimed at reducing the over dependence of emergency services for both crisis intervention and routine care. It is interesting to find that the result that younger children were associated with higher likelihood to have an ED visit for non-urgent problems corroborate previous results.12 The results regarding the very low rate of hospitalization, the common medical diagnoses that included respiratory and abdominal symptoms, and attending the ED for non-traumatic injuries suggest that many patients use the EDs for primary health care. Reports in the literature indicate similar presenting problems in the already mentioned study conducted by Ziv et al.,5 in children attending paediatric accident and EDs in the UK13,14 and in paediatric (<18 years) attendees to an urban ED in Malaysia.15 Therefore, this could also be attributable to the lack of linkage between institutional and community services, and the lack of paediatric primary care among these patients may result in their over dependence on EDs. This is also supported by the finding that the day a patient attended the ED affected the frequency of medically non-urgent problems. In fact, the patients in the weekend were more likely to be with non-urgent problems, despite a presumably consistent day-to-day burden of disease. The reduction in clinical primary care services on weekends may explain, in part, such pattern and this is of particular interest since primary care paediatricians play a vital role in medical emergencies in public health care systems, and more programmes should be adopted to provide acute prehospital care. The reasons for ED attendance are complex and are related, in the case of paediatric patients, to parents' perceptions of the purpose of ED. This is supported by the result that more than 90% of the patients examined were self-referred or referred by relatives to the ED, although such patients did not show a significantly higher proportion of non-urgent visits than those referred by physicians. This picture strongly suggests that efforts to minimize such unnecessary ED use are important to reduce the spread of resources. It may therefore be appropriate that health educational interventions focus on both the optimal approach and management by the parents and the social services and community resources.

In conclusion, considering the extreme importance that patients receive services at the appropriate level of care and in view of the high direct and indirect costs associated with inappropriate hospital use, it is essential to have a closer cooperation within the health care organization system to provide a service responsive to the real needs of patients.

Key points

  • Our objective was to determine the frequency and characteristics of paediatric attendance as a source of medically non-urgent problems at an emergency department (ED).

  • The visit was non-urgent in younger population, females, and in those attending the ED on the weekend.

  • Adolescents who did not receive medical or surgical examination at the ED, with problems of longer duration prior to arrival at the ED, with non-traumatic injuries, and who did not require inpatient hospital admission were more likely to use the ED as a source of non-urgent care.

  • The most frequent presenting problems for visits to ED were injury, respiratory diseases, and digestive symptoms.

  • A closer cooperation within the health care organization system to provide a service responsive to the real needs of patients is essential.

References

View Abstract