Background

Injuries are a significant cause of death and disability worldwide, regardless of the development status of the country [1]. The World Health Organization (WHO) reported on the magnitude of the entire problem at the Global Forum on Trauma Care in 2009 [2], and also declared the targeted Decade of Action for Road Safety 2011–2020 [3]. Trauma systems have been developed to address this growing threat to public health, of which the Victorian State Trauma System (VSTS) is a world-class example. A key component system wide monitoring and governance, based on accurate data from the state trauma registry. Following implementation in 2000, risk adjusted mortality decreased by 50% [4].

The Kingdom of Saudi Arabia (KSA) is the largest Arab country in the Middle East and is highly developed. It has a rapidly growing population estimated at 32.4 million people in 2017 [5] but is suffering a well-documented rise in the incidence of serious injuries that occur on the roads, in the workplace and elsewhere [6, 7]. Health care is provided largely by the Saudi Arabian Ministry of Health (MOH) with 407 primary health care centres. Other health care sectors include university, military and private groups [8]. MOH hospitals provide emergency care at Emergency Departments and it is to these that most trauma patients are transported by a combination of Saudi Red Crescent Society ambulances, private ambulances, police or military vehicles and private transport.

In response to the increasing burden of traumatic injuries on the community, the MOH requested the King Saud Medical City (KSMC) to engage the Alfred Hospital and Monash University to assist with trauma service development, including a trauma registry program based on the Victorian model which could be scaled to the country as a whole. This study aimed to examine the process of establishing the registry at the hospital with a view to rolling the prototype out nationally. This was a challenge as the hospital had no previous involvement with a trauma registry and it required establishment without any prior experience from clinical staff.

KSMC is located in the capital city, Riyadh, which is home to approximately six million people. It is the MOH centre that receives the most emergency patients in the Kingdom [9] and is therefore the logical initial site for a trauma registry that ultimately will be rolled out nationally. Recent reports reveal that approximately 350 people a day present to the Emergency Department (ED) at the KSMC, a significant proportion of whom will be admitted to one of the facility’s 1400 beds. Over 300 people per month presented with the level of trauma that potentially fell within the population of interest for a registry. To capture these data accurately required considerable effort and a robust methodology.

Methods

Work plan

A plan was created by the project team, based on the operating principles in the 2008 Australian Commission on Safety and Quality in Healthcare (ACSQH) [10], that stipulated 12 steps (Table 1).

Table 1 Trauma registry implementation 12 step plan
Full size table

The initial steps were taken in sequence, with the latter steps taken either contemporaneously or out of sequence, depending on the progress at that time. Each step included a rationale and examples taken from other registries administered by the Alfred and Monash University, particularly the Australian Trauma Registry (ATR), the Victorian State Trauma Registry (VSTR) and the Alfred Health Trauma Registry (AHTR).

Selection of the inclusion and exclusion criteria and Minimum Dataset (MDS)

Initial steps required the selection of inclusion and exclusion criteria and the selection of a Minimum Dataset (MDS). These are critical as the former identifies the target population of interest [11] and the latter specifies the required information, arguably forming the framework of the entire registry.

Our criteria were based on those of the VSTR and designed to include cases that added the most value to the knowledge gained from the registry, and exclude cases that the experience of the VSTR showed added minimal value (Table 2).

Table 2 STAR inclusion and exclusion criteria
Full size table

The literature reveals that other trauma registries have selected their MDS using a Delphi technique [12, 13], however our project chose to work with the trauma MDS collated by the Global Alliance for Care of the Injured (GACI). This option was available via the Monash University Accident Research Centre (MUARC), which is a World Health Organisation (WHO) collaborating centre. The MDS was in draft form at that time and has been referenced as such in the resultant KSMC data dictionary.

The final dataset necessarily included some variables that were specific to the KSA, evolving from an initial selection of 67 elements to the final listing of 83 elements. A significant number of these are drawn from those proposed by the WHO GACI working group. Additional File 1 lists each variable, the dictionary reference and dictionary definition.

Site visits

Five site visits by the Alfred/Monash project team were necessary to implement the registry. Both visits included activities that were targeted at taking steps on the plan. The first visit in April 2017 achieved steps one to six. The second visit in August 2017 delivered a two-day training course to the on-site team as per step 12. A training program and materials had been created during the intervening months to complete steps 10 and 11. The course curriculum included presentations on data, registries and injury coding that were considered sufficient for the basic knowledge needed to implement the registry.

At the time of the second visit, all the steps of the 12-point plan had been taken except for step nine, construction of the database. However, this did not preclude the commencement of data collection on paper forms, which began on Wednesday 16th August.

The third, fourth and fifth visits incrementally progressed the implementation of the registry, culminating with User Acceptance Testing (UAT) of the database in March 2018 and the official Saudi TraumA Registry (STAR) launch in April 2018.

Construction of the database

There are guides available to organisations that provide details on technical expectations and standards of clinical registries [14]. However, the decision on which data collection system to adopt is subjective, and depends on the organisation’s needs and resources. The KSMC project team found that the primary differential in the choice of database was whether it was acquired externally, or was a bespoke system that was built internally.

Specifications for the database were prepared by the project team and provided to the KSMC IT Department. No programming code was provided, as this is particular to development of the software and would be unique to the construction.

Once the project team considered the arguments for and against building the database internally, and the IT Department assessed whether they had the technical expertise to do so, the decision was made to proceed. Construction commenced in August 2017.

Timeline

The time taken to implement the KSMC STAR was 16 months, from the commencement of project activities in January 2017 to the move into production of the database in April 2018 (Fig. 1).

Fig. 1
figure 1

STAR implementation timeline

Full size image

STAR operations

The day to day operations of registries differs between organisations according to the registry’s need and purpose. There are valuable recommendations in the literature [15] on the various strategies and methodologies, but as with the choice of data collection system there is no ‘one size fits all’. In the absence of any generic Standard Operating Procedures (SOPs) that would suffice, the STAR has written those necessary at this first site, that will also be applicable at ensuing sites.

An SOP was created for the methodology based on the experience of AHTR and the VSTR, which is that of retrospective data collection of cases that present to hospital and are found to meet inclusion criteria. The daily routine commenced on the 16th August. Cases were backdated to those with dates of injury beginning on the 1st August 2017 and the dataset was documented on paper forms initially. The database was ready for direct data entry in April 2018, after which time the registry was considered to be fully operational.

Results

During the first year of data collection an average of 216 records per month were entered into the STAR database, screened from the approximately 300 cases of potential interest that presented to the KSMC ED. From this monthly average, 175 were confirmed to meet inclusion criteria and progressed to completion of the case. A data quality and cleaning regime was initiated within one month of reporting being available from the database, to immediately ensure prompt monitoring of data integrity. Inbuilt database edits disallowed blank fields, therefore no null values were found. The cleaning regime assessed completeness and quality as the proportion of ‘unknown’ values; and anomalous, or mismatched, values. Each report interrogated the injury event and definitive care variables, and has consistently revealed the variables of concern to be those relevant to the injury event; injury coding and operations coding (Table 3).

Table 3 Error rates of concerning STAR variables at one month post go live of the database
Full size table

There was a marked difference between the demographics variables, and the variables of concern, which had higher error rates. The sample size of ‘Descriptions if Cause/Place/Activity is other’ variables was small, however the rate reflects the consistent error of coding the value as ‘Other’, when in fact the value could have been captured in the codeset of the primary variable.

It is not possible to determine at which point newly implemented registries should be achieving a particular standard of data quality and completeness. What is agreed is that data quality should be optimised and that missing data is a known problem [16, 17]. These quality reports have provided insight into the work still needing to be done at any given time.

In addition to ongoing quality reports, a comprehensive audit in October 2018 interrogated approximately 10% of records from the first 12 months of operations. The resulting recommendations addressed the data quality issues and will ensure continuous improvement as the STAR matures.

An inaugural epidemiological report of the first 12 months of data collection has been prepared and was presented at the Saudi Trauma Conference held in Riyadh in February 2019. Ongoing annual reports of calendar years are planned.

Discussion

The challenge facing the project team tasked with implementing a trauma registry at the KSMC was that trauma data had not formally been collected at the site previously, and there was no existing local example of how to do so. Despite this, the STAR has been successfully established at the KSMC and was fully operational eight months ahead of the project deliverable date of December 31st 2018. The 12 step implementation plan that was especially created for the project provided the framework for the process. Each step gave the project team confidence to proceed to the next step or schedule the required tasks accordingly. This process may be useful for other jurisdictions.

Through its partnership with the Monash University Collaborating Centre, the STAR was able to refer to the WHO draft trauma MDS for assistance in selecting its variables. KSMC has made a contribution to the WHO goal of promoting “wider and more standardised collection of data for injury surveillance and trauma care quality improvement.” [18] The education curriculum that was developed to meet the initial training needs of the STAR personnel; the specifications for the construction of the database and the myriad of other tasks that were required have showcased our plan as a project that the WHO can consider to be an example of a case study of good implementation practice. Collaboration will be ongoing as the project continues.

The timeline of the implementation showed the constant tasks required to implement the registry. Some were done by the Alfred-Monash team; some were done by the KSMC; some were done on site in Riyadh and some of them were done off site in Melbourne. The successful completion of each task contributed to the completion of the next and no single task could be achieved in isolation from another.

The project team were required to make a difficult decision on which data collection system would meet the KSMC’s needs. Choosing the option to construct the database on site has proven to be the correct one, as the system is purpose built and amenable to an upgrade that will enable it to accept data from multiple sites, thus meeting the needs of a national registry.

Of local significance is that for the first time the hospital has access to data that will inform process indicators such as time spent in the Emergency Department; rate of trauma team activation; time to operation and length of stay in the Intensive Care Unit. This information alone will assist the KSMC with its goal of achieving verification as a Level 1 Trauma Centre.

Of national significance to the KSA is that the STAR will mature to the point where it can be rolled out across the country, offering robust operating procedures and a functional database. The multi-site experience of the VSTR and the ATR will support the KSMC to be the locus for registry activity elsewhere in the greater Riyadh region and across the Kingdom. Last but not least, neighbouring countries will benefit from the knowledge gained of the implementation process from KSMC.

The literature [19] identifies several features of successful registries including but not limited to organisational support; existence of a steering committee; viable funding options; surety of purpose; committed registry management; training of registry staff; attention to data quality; organised data collection; secure data storage system and flexible reporting options. However, to the best of our knowledge the literature has not described any other step-by-step guide used in the establishment of other trauma registries and indeed O’Reilly et al [20] cited a perception among their study participants that no such guide was available. Our process may therefore be a first, and useful for other jurisdictions.

The STAR registry will enable robust assessment of outcomes and international benchmarking as it accrues increasingly high quality, accurate and accessible data that are available for interrogation and research.

Conclusion

By taking a systematic approach to registry development, the KSMC- The Alfred International trauma Program has implemented a scalable regional trauma registry with robust and rigorous data collection, collation and potential to drive trauma system improvements. The STAR experience demonstrates essential development steps for implementation which will be effective in other regions.