Referrals to a pediatric emergency room of a tertiary care teaching hospital before and after introduction of a referral education module- a qualitative study

Background: Provision of timely care to critically ill children is essential for good outcome. With development of specialized pediatric critical care units, referral from community and smaller peripheral hospitals has received greater impetus. Our tertiary care hospital caters to referrals from a wide geographical area. Since there is no standard referral and feedback system in India, we decided to evaluate the quality of referrals coming to our Emergency Room (ER) with respect to their demography, association with severity of illness and mortality. We plan to use this data to establish a standard referral and feedback process to streamline our ER referrals. Methods: Our study was completed in three phases in Pediatric ER;Pre-intervention, Intervention and Post intervention phases. Quality of referrals was graded by granting one score to presence of each item when matched with a quality checklist performa. A referral was graded ‘good’, ‘fair’ and ‘poor’ if it scored >7, 5 –7 and < 5 points respectively. Case studies, expert opinions, and lacunae observed in first phase were taken into consideration while preparing the referral education module, that was administered to health care providers of referring hospitals. Quality of referrals was compared between pre and post intervention phases. Results: Majority of inpatient admissions were ‘referred’ patients (99.3%). Most referrals belonged to the neighboring states of Punjab (48.2%) and Haryana (22.4%). Major referrals were from public sector hospitals (80.9%) of which the teaching hospitals topped the list (53.6%). Most common mode of transport was government run ambulance services(85.5%) and the common reason for referral was need for PICU bed and/or mechanical ventilation (50.4%). The post intervention phase saw a significant decline in the proportion of poor referrals (93.2 vs.78.2%;p=0.001) and a significant increase in the proportion of fair(6.1 vs 18%; p=0.001)and good referrals (0.7 vs 18%;p=0.001) Proportion of children presenting triage decompensation to poor referrals in post


Background
Provision of optimal and timely care to critically ill children is crucial for best outcome. With development of specialized pediatric critical care units, referral from community, district or other peripheral heath care facilities has gained greater momentum (1). Referral is a process in continuum; beginning from the referring and ending at the referred facility (2). Maintenance of optimal physiology during referral is vital (3,4).India lacks an organized referral network. This is further compounded by multiple barriers which include poor pre-referral communication and transport facilities (5). The Emergency Room (ER) at the Advanced Pediatrics Centre is a standalone unit that caters to patients referred from 5 neighboring states viz Punjab, Haryana, Himachal Pradesh, Rajasthan and Uttar Pradesh. It sees about 18,000 to 20,000 admissions per year, of which more than three fourths are referred patients both from public (community health centers, district hospitals, and medical colleges) and corporate sector hospitals of the above states. Since there is no standard referral and feedback system in India, we decided to conduct this study to ascertain the quality of referrals coming to our Emergency Room and their association with severity of illness and mortality before and after introduction of a referral education module. Using this data, we plan to establish a standard referral and feedback process that will help streamline our emergency referrals.

Methods
We conducted a prospective before and after study which was divided into three phases; Preintervention (Phase-I), Intervention (Phase II) and Post intervention (Phase III). All consecutive children aged > 1 month to 12 years of age referred to our Emergency Room were included in the study.
The pre-intervention phase lasted for 4 months from January 2016 to April 2016. During this phase a total 1403 children referred from different hospitals were enrolled. The referral letters accompanying these patients were collected and studied for quality and completeness by comparing them with a quality checklist performa. This comparator checklist was made a priori based on WHO recommendations (6).The quality of each referral was estimated by granting one score to the presence of each item in the performa. Some items were subdivided into their integral components and each component was granted equal scores between 0.2 or 0.5 in accordance with the number of components. All the components included were given equal importance. Each referral letter was given an objective score and graded as 'good' if it scored > 7 points, 'fair' if it scored between 5-7 and 'poor' if the score was less than 5.
In Phase-II (May to August 2016) we developed a referral education module with inputs from emergency and social pediatric experts (JM,AB,BB).The lacunae in referrals observed during the first phase were analyzed and discussed and steps to prevent them were included in the module.
Anonymized case studies in which identity of patient, referring hospital and physician were kept confidential, were prepared to highlight both good referrals and common errors during referrals. The content of the module emphasized on 1) How to identify a sick child who needs referral 2) importance of timely referral,3) key components to be addressed during the referral process like stabilization, and maintenance of airway, breathing and circulation, fluids and medications 4) Importance of good prereferral communication, 5) Importance of good written documentation, and feedback. This module was peer reviewed by all the co-investigators before finalizing it. The final module (duration 2 hours) was then administered onsite to all the doctors, and nurses working in the emergency area of various hospitals from where the maximum referrals had taken place in Phase I. The list of hospitals to be visited was prepared and the concerned hospitals were informed at least 1 week prior to the scheduled visit to ensure maximum participation from concerned health care providers of that site.
On the proposed date and time, the team of investigators visited the site and administered the module to the target group. Written handouts emphasizing key aspects of a good referral, were also given. Small group discussion addressing their barriers and problems were also conducted.
In the post intervention phase, which lasted for 5 months from September 2016 to January 2017, a total 957 referred cases were enrolled. The referral letters and their quality were assessed once again with the same quality checklist performa and objectively scored as described for Phase I.
The pre and post intervention phase were then compared with respect to demography of referral, association with severity of illness and mortality using Chi-square test or Student's t test as applicable. Quality of referral is expressed as proportion. The association between quality of referral and the physiological status at triage was also determined using Chi-square test. P < 0.05 was taken as significant.

Outcome parameters
Change in proportion of good, fair and poor-quality referrals before and after the referral education module was taken as primary outcome. Reasons for referral, proportion of patients transported in ambulance with a trained emergency personnel, association between the quality of referral and triage physiological status were secondary outcome parameters.

Discussion
This study was designed firstly to observe the quality of referrals coming to a pediatric Emergency Room of a tertiary care multispecialty teaching hospital and secondly to develop a referral education module for health care providers at different referring healthcare facilities; and thirdly to assess the difference in the quality of referrals between both the phases.
A total of 12,490 patients visited our ER during the pre and post intervention phases combined.
Majority were 'referred' patients, primarily from public sector teaching hospitals. Furthermore, most referred patients required admission. A study from another teaching and tertiary hospital has reported findings to the contrary; only 22.5% of referred patients required admission (7 Since most patients in our set up are also ignorant about the treatment given, the healthcare providers at the receiving facility are highly dependent on the referring doctors documentation or verbal information. We found that almost all referral letters lacked the most essential information required for maintaining continuity of care. The referral letters in the pre-intervention phase had missing information related to illness, clinical examination, treatment given, investigations, procedures and pre-referral stabilization. Studies have shown that nearly one third of specialist referrals from general practitioner (GP) lack essential clinical information and are often inadequate (8)(9)(10).Similarly another study evaluating the quality of GP referrals to a South African tertiary care hospital, reported that certain important components related to pre referral treatment (6.3%), laboratory tests (8.3%) and special tests (4%) were mentioned in very few referral letters only (11).
On the other hand, nearly half of the referral letters in both phases contained reasons for referral.
Though much lower than that reported by Langalibaele et al (11) and Leonard et al (12) which was 88% and 100% respectively. Although indications were mentioned, pre-referral communication with the referred center was absent. In a resource limited setting like ours, absence of prior information results in poor preparedness at the tertiary level, the case in example being need of a ventilator.
We observed that all the referral letters were hand written and unstructured. Majority of the referral letters were in the 'letter format' similar to that reported by Xiang et al (13). Lack of structured format could have been one of the reasons for the poor quality and deficiency of several domains in these referral letters similar to that reported by Manis et al (14,15). This study paves the way for development of an organized referral network which connects tertiary care facilities with the other public and corporate hospitals in the periphery. The initial sensitization achieved by our study should be further consolidated by a more sustained multifaceted outreach and continuous feedback process. These interventions will also help reduce the unnecessary referrals and consequent overcrowding of a tertiary care facility. The major limitation of our study was that we could not cover all the health care facilities that were referring to us as they were distributed over a widely geographical area. We had targeted our interventions on hospitals from where maximum referrals were received during the pre-intervention phase. Secondly this being a before and after 9 study design, there could have been Hawthorne effect in the post intervention phase.

Conclusion
Referral education of healthcare providers had improved the quality of referral by improving some domains of referral content. Proportion of children presenting to triage with physiological decompensation had significantly decreased after introduction of this module. This change suggests sensitization of the peripheral hospitals towards a better referral process. Continued multifaceted approach will be required for sustained and increased benefits. This will pave the way for building an Comparison of quality of referrals ( pre vs postintervention phase )