Admission rates for children to hospital have increased over the last 15 years even though there has been a reduction in the number of inpatient beds.1 Pressure on accident and emergency (A&E) departments has resulted in long waiting times and inappropriate exposure of children to adult emergencies.2 In addition, junior staff working in A&E departments lack sufficient paediatric experience.3 There has been an increase in parental demand for rapid diagnosis and management of sick children by experienced professionals in an appropriate setting.4 5 Moreover, there is a need to deal with parental concerns, such as perceived lack of control and poor knowledge of common symptoms.6

There is wide recognition of the need to minimise hospital admissions and manage acute child illness more appropriately via specialised ambulatory care, such as short stay units and hospital-at-home services.7 8 Such innovative proposals have a long history and are in line with the National Service Framework for Children.2 9 10 A recent systematic review showed that ambulatory care is becoming an acceptable alternative to inpatient admission, but that more research on impact is required.7

In response to the changing context, Northwick Park Hospital (NPH) in the UK developed a purpose-built paediatric assessment centre, the Paediatric Ambulatory Care Unit (PACU), close to the inpatient wards and outpatients department. PACU opened in October 1998 and during the study period had an average yearly attendance of 4190, with increasing throughput. Patients were referred to the unit for observation, diagnosis and management for periods of between 4 and 6 h. A decision would then be made to admit, discharge or refer. The unit was staffed by an administrator, full time consultant, middle grade doctor and two specialist paediatric nurses per shift. Opening hours were 9 am–10 pm 7 days a week, with the last patient accepted not later than 7.30 pm. Patients arriving after this time were seen in A&E which did not have a separate children’s A&E section. Once PACU was closed at night, the same PACU middle grade paediatric doctors (but not PACU nurses) would work in A&E along with adult trained staff. There were no paediatric nurses in A&E. Any families attending A&E during the day would be referred to PACU as appropriate.

The PACU model was aimed at providing a user-centred service with specialist multi-disciplinary care within a modernised environment directed towards a better patient journey, reduced admissions, timely diagnosis for parents and referrers, enhanced information delivery satisfaction, confidence and understanding of parents when dealing with their sick child. We conducted an evaluation 4 years after PACU opened, focusing in particular on staff and parental views, as well as on comparison with A&E, areas that have attracted only limited study.7 11 12 We have previously reported evidence suggesting this model can reduce overnight admissions.13 Here, we report the impact on the patient journey and the perceptions of parents, staff and referrers.

METHODS

A multi-method design including quantitative and qualitative techniques was used. This approach is appropriate for service evaluation due to its ability to assess a range of service objectives from different perspectives.14 15 Firstly, we aimed to invite all eligible parents attending PACU over a 3-month period (May–July 2003) to complete a questionnaire administered by nursing staff. The questionnaire was developed by our team and validated through a pilot study. It was designed to assess satisfaction, information provision, waiting times and perceived anxiety levels. To provide some comparative data, over a 1-month period we invited parents attending A&E during the night (when PACU was closed) to complete an appropriately modified questionnaire. Parents who did not understand English were excluded.

In addition, independent health researchers (JG, SP) carried out a telephone survey of main referrers (GPs), semi-structured interviews with a range of PACU staff, and direct observation of patient journeys. Qualitative samples were purposively selected.14 Informed consent was sought in all instances and the guidelines of the Research Governance Framework (RGF) closely followed. Harrow Research Ethics Committee approved the study (led by MB) in August 2002. The project also formed part of a wider evaluation of redesigned services at the trust (led by JG). The overall time-frame was August 2002 to January 2005.

Quantitative data were analysed using SPSS. Means are presented where data were normal and medians where skewed. χ² Tests were performed for comparing percentages (unless otherwise stated), t test for means and the Mann-Whitney U test for medians. Differences were considered statistically significant where p<0.05. Qualitative data were categorised into themes and a content analysis performed.14

RESULTS

Parental perceptions

During the questionnaire phase, 148 parents attending PACU, out of 455 eligible respondents, were offered questionnaires at the end of the care process. The response rate was 70% (n = 104). In the A&E survey, 41 out of 60 parents responded to the questionnaire (response rate: 68%). Not all parents were given a questionnaire as staff and parents sometimes had little time. Demographic characteristics are shown in table 1. The main differences were found in comparisons with A&E.

Key findings

Referral patterns

Patients were referred to PACU by GPs (49%), A&E (20%), self-referral (14%), community nurses (9%) and other medical staff (8%).

Access issues

Most parents (88%) reported that referral to PACU was “easy”. In terms of access to the building, 76% reported this was “easy”, with 19% saying it was “average” and the remainder (5%) saying it was “difficult”. In contrast, 95% of parents felt access to A&E was “easy”, with the remainder (5%) saying “average” (p = 0.058).

Waiting time

The median time for a child to be assessed (“dealt with”) in PACU, as perceived by parents, was 35 min (interquartile range (IQR): 10 to 180). Over half (57%) of parents felt the time taken was “quick” or “very quick” (24%), while over a quarter (29%) said “average” and 14% said “slow”. The perceived time matches audit data. It also matched the perceived time found in A&E (30 min; IQR: 10 to 165), although A&E parents were more likely to state the time taken was “slow” (25%) (p = 0.381). On average, families referred to PACU by A&E would have waited a further 30 min in A&E.

Self-reported parental anxiety

There was a statistically significant reduction in perceived anxiety after use of the PACU service (55% of parents were “anxious” before and 13% were “anxious” after; p<0.001, McNemar test). Similar results were found in A&E (before: 63%, after: 5%). There was no significant difference between the PACU and A&E groups regarding anxiety (before: p = 0.356, after: p = 0.198).

Information provision

Information given to parents was mainly verbal (74%), with some also receiving written information (14%). Most said they were given the “right amount” (93%).

Confidence and understanding

Information delivery appeared to enhance parents’ understanding of the child’s problem (89% “fully understood”) and “confidence” in managing their sick child (87%).

Appropriateness of clinical decision

The majority (85%) of parents felt the final care decision was “appropriate”.

Expectations and overall satisfaction

Although most A&E parents were “satisfied” (referred to as “happy” in the questionnaire to allow for cultural diversity), overall perception towards the two services elicited key differences. Figure 1 displays the results from two separate questions, where parents were asked to rate (i) their level of satisfaction and (ii) how far the service met expectations. This illustrates that PACU parents were significantly more likely to feel “very satisfied” and that their expectations had been “fully met”, than their A&E counterparts (satisfaction: p = 0.03, expectations: p = 0.049). In addition, fig 2 shows that parents reported a better experience in PACU than previous family experiences in A&E.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Overall parental perceptions of the service: PACU vs A&E. The first three rows are derived from the “satisfaction” question: PACU, n = 96; A&E, n = 35. The last two rows are derived from the “meeting expectations” question: PACU, n = 98; A&E, n = 39. Satisfaction: p = 0.03; expectations: p = 0.049.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 How did parental experience in PACU compare with previous experiences in A&E? (n = 64).

Influence of demographic characteristics

We analysed whether demographics may have acted as confounders by influencing the “satisfaction” variable. No statistically significant effects were found regarding ethnicity (p = 0.472), first language (p = 0.995), relationship to child (p = 0.562), parent’s age (p = 0.923), child’s age (p = 0.332), child’s gender (p = 0.754), presentation (p = 0.683) or care decision (p = 0.384). Parents who spoke English as their first language were more likely than their non-English speaking counterparts to feel that referral to PACU was “very easy” (65% vs 37%, respectively; p = 0.002) and that their expectations had been “fully met” (82% vs 68%; p = 0.02).

Patient journey

We observed the journey of 10 families with children of different gender (six males, four females), ethnicity, ages (0–6 years) and presentation. The purpose was to gain an in-depth, real-time view of care processes. Figure 3 shows the typical pathway in PACU and fig 4 the typical pathway in an A&E service we have studied (traditional A&E pathway). In PACU the journey is reduced from six (traditional) to four stages, mainly due to the close liaison between staff. There is a reduction in repeated clinical assessments and the need for users to repeat their story. The unit reflected the flow of the patient journey with a nursery area at reception, including toys and apparatus, leading into a brightly decorated long corridor with dedicated assessment and treatment rooms (including four bed and four cot spaces), with a nurse station directly facing these rooms, followed by a major treatment room and observation area for any ongoing assessment. There was no similar family-friendly area or structure in A&E.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Typical A&E pathway (n = 25 patient maps).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Typical PACU pathway (n = 10 patient maps). *Multi-disciplinary specialist paediatric team including paediatric nurses, junior doctors and consultants.

Staff perceptions

Interviews were conducted with 10 PACU staff (four paediatric nurses, three junior doctors, one consultant and two clerks) to explore perceived impact on working practices. Key themes were that: (i) PACU was perceived as a “superior” model to standard A&E (in which staff had experience), and where the family’s needs could be addressed more holistically; (ii) specialist nurses were “facilitators” for teamworking and patient flow; and (iii) there were concerns over the “clarity of the role” of PACU amongst other departments and external agencies, and the need for “better induction training” so that the philosophy of the ambulatory approach was understood by all. While nurses and doctors felt communication between themselves was good, clerical staff felt communication could be improved (eg, doctors informing them of children due to attend).

Staff reported that the lack of understanding of the role of PACU amongst external agencies often led to inappropriate referrals. According to a paediatric nurse:

PACU works well, but it’s used and abused, as we see kids (with minor illness) who could be seen in outpatients… doctors want quick answers.

Referrer views

A systematic random sample of 70 GPs (sampling frame: 280) was selected for a 10 min structured telephone interview. The response rate was low (23%, n = 16), as most GPs felt they were “too busy”. All GPs were satisfied with PACU and its attempt to provide rapid diagnosis but reported they were told of its role mainly in an informal fashion (eg, via colleagues).

DISCUSSION

Paediatric ambulatory care is becoming an acceptable form of service delivery, but there is limited knowledge of the impact of this approach.7 This multi-method evaluation provides evidence that parents and staff prefer the PACU model to traditional A&E services, and that it enhances parental satisfaction and the user’s journey. While some areas around multi-disciplinary working require improvement, the model appeared able to meet its key objective of providing a user-centred service within a modernised environment. Given that current emergency services for children in the UK fall short of national recommendations16 and a recent review found 40% of A&E departments do not provide specialist paediatric services,17 these findings contribute to evidence suggesting that such an approach should be adopted.7 11 12 18 At the same time, future planning of services should still consider inpatient admission where appropriate.19

We were able to conduct a thorough evaluation which included triangulation (ie, providing a view of PACU from different angles and sources).14 15 However, our study has some possible limitations. Firstly, we were only able to focus on one site. Secondly, the study began after PACU had opened and therefore we could not assess before and after changes (although comparison with A&E helps redress the balance). Thirdly, due to resource limitations, the final A&E survey sample was relatively small, and there may have been some selection bias in the survey overall, as in previous surveys.7 18 We were reliant on staff to administer the forms and on parents who spoke English.

However, staff received clear direction via emails and a workshop to ensure that the questionnaires were administered whenever possible, the recruitment process was random and response rate was high. The survey included parents of those admitted and discharged, unlike a previous key survey that only included those discharged.18 Moreover, the sample closely resembled the population of parents using the service.20 The survey groups were generally well matched (table 1), although there were significantly more fathers and non-English speaking respondents in the A&E group, and some differences in ethnicity. We attempted to include equal numbers in the two survey groups, but the clinician managing administration of the forms in A&E was only available for 1 month. For the qualitative methods, purposive sampling was used to select an appropriate range and number of respondents, in line with published guidance.14

In our analysis of whether parental satisfaction differed between the groups, we attempted to account for a number of key demographics that may have confounded the results (table 1). However, as A&E parents were sampled at night when PACU was closed, there may be some group differences that were not accounted for. This, together with the small A&E sample size, limits our ability to address all possible confounding variables. Those demographics that were measured did not appear to influence satisfaction, so the differences found between the survey groups were unlikely to confound the results. There is no evidence to suggest that disease severity differed, as measured by presentation, care decision and baseline parental anxiety levels. Waiting times also were similar. This assessment, in addition to the fact that the same PACU doctors attended to families in A&E, increases our confidence that satisfaction is likely to be related to core differences in environment, processes and practices between the two settings. PACU benefited from the presence of a paediatric multi-disciplinary team, including highly regarded specialist nurses, a family-friendly environment and reduced stages in the patient journey, in agreement with national recommendations.2 21 22 An enhanced journey may explain the more favourable perception towards waiting times in PACU. A key difference in favour of the A&E service was that access to the building was seen as easier. Since the study was conducted, the approach to PACU has been integrated with a paediatric A&E department at the front of the hospital.

Our research indicates that a number of areas need further evaluation when a PACU model is being considered. These include the expectations of and access for non-English speaking families, and improvements in joint induction for staff. Further, paediatric ambulatory care is but one element of a complex acute system, which includes primary care, and the role of the PACU model needs to be more clearly defined within this context and parents questioned about the whole experience. There is also scope for managing demand by referrers, as found in other settings,23 24 through educational outreach and guidelines on appropriate referrals.25 Such considerations would ensure that a seamless service is available to families, and that ambulatory care is provided at an appropriate stage in the patient journey.