Standardizing the Early Identification of Acute Kidney Injury: The NHS England National Patient Safety Alert

In June 2014, a national Patient Safety Alert was issued to all healthcare providers in England entitled ‘Standardising the Early Identification of Acute Kidney Injury'. Patient Safety Alerts are used by the National Health Service (NHS) in England to rapidly inform all healthcare organisations of risks to patients along with guidance on preventing harm. This alert was the first major output of the NHS England ‘Think Kidneys' Acute Kidney Injury (AKI) Programme and was designed to eliminate the considerable variability that existed in both uptake and methodology of electronic systems to detect AKI based on changes in serum creatinine concentration. In addition to the primary aim of embedding a tool in all hospitals to help improve recognition of AKI in clinical practice, it provides a framework for an ambitious plan to prospectively capture data at a national level on the incidence and outcomes of AKI. This has the potential to form a powerful tool for organisational quality improvement as well as clinically focused research. This review will summarise the clinical need that the patient safety alert addresses, the process of its development and the anticipated outcomes of this initiative.

AKI is common, harmful and potentially modifiable. AKI occurs in up to 22% of hospital admissions [1] and is associated with hugely elevated mortality rates [2,3]. Hospital stays for these patients are often prolonged and healthcare costs are considerable [4]. There is also a growing appreciation that renal function does not always recover following AKI, contributing to the development and progression of chronic kidney disease [5]. There are several reasons to explain the poor outcomes associated with AKI: its presence is associated with increased severity of acute illness and greater burden of comorbidity, whilst AKI itself exerts additional adverse effects on distant organ function, probably mediated via systemic inflammation [6]. However, the silent nature of AKI means that it is often poorly recognised by patients and clinicians alike, and an accumulating number of reports from different healthcare systems have demonstrated that deficiencies of basic medical care can often contribute to poor patient outcomes [7,8,9,10,11]. One of the most frequently observed deficiencies in care relates to delay in diagnosis. At the same time, emerging evidence has confirmed what ‘good medical practice' would decree; prompt intervention focusing on the basic elements of care (fluid balance, haemodynamic observations, medication review and appropriate investigation) can improve the outcomes of AKI [12]. Whilst recognition of these factors has been a major driver for change across the entire AKI pathway, one area that has attracted particular interest is improving recognition of AKI using electronic systems.

Since we reported the first hospital wide e-alert system based on current diagnostic criteria [13], there has been a burgeoning interest in the UK in the development of such systems for clinical use. Whilst the term ‘e-alert' has been used widely in the setting of AKI, it is worthwhile clarifying its 2 essential components. Firstly, there is the detection element, essentially a mathematical process that requires an algorithm to compare a serum creatinine test result to that individual's baseline or reference value, in line with current internationally accepted definitions. Secondly, there is the alerting process that refers to the method by which occurrence of any such changes are communicated to the relevant clinicians. Initial development of AKI detection and alerting systems in the UK occurred in individual hospitals in a sporadic way, usually led by local enthusiasts working in isolation; development was therefore ad hoc without any degree of oversight or coordination. This led to a situation in which there was variation in uptake of AKI detection as well as a wide variation in methodology used to apply the diagnostic criteria. Many UK hospitals (particularly those without onsite nephrology services) did not have electronic detection and alerting systems, whilst in those that did patients identified by one hospital's detection system may not have been identified by that of another.

It was therefore unsurprising that tackling this issue of variability was one of the 5 priority recommendations from the UK Acute Kidney Injury Consensus Conference in 2012 that stated ‘a national group should be established to develop agreed standards for e-alert systems … including an agreed definition of AKI based on the KDIGO classification and a standardised methodology for the derivation of baseline serum creatinine' [14]. From this, a series of initiatives arose under the umbrella of professional organisations (Association of Clinical Biochemists, Renal Association, British Association of Paediatric Nephrologists); their output was the design of an algorithm, based on the KDIGO diagnostic criteria but adapted for clinical use by consensus of expert opinion, that would identify changes in serum creatinine concentration consistent with AKI (AKI detection algorithm shown in figure 1). Importantly, these initiatives were in communication with the commercial suppliers of laboratory information management systems (LIMS) so that the proposed detection algorithm was within the technical bounds of the majority of IT systems in UK clinical chemistry laboratories. Despite the design and publication of this ‘standardised' algorithm, these groups did neither focus on implementation nor on the alerting arm of the process.

In 2013, in response to a growing concern about AKI, the NHS England formed a partnership with the UK Renal Registry to convene a national AKI programme (‘Think Kidneys' Programme [15]). Within a tenure of 3 years, its stated aim is to prevent avoidable harm caused by AKI. It will attempt to do so by developing new ways of working intended to reduce the number of preventable AKI episodes and ultimately to reduce mortality associated with AKI. This will include the development of clinical tools and resources to improve patient care. There is also a desire to raise public and professional awareness of AKI, with an ultimate ambition to position ‘kidney status' alongside other assessments of patients' physiological state.

This programme has given impetus to all aspects of AKI quality improvement and encompasses a wide variety of stakeholders including both expert professionals and patients. Underneath a steering committee, the programme is organised into a number of different work streams that each focuses on a different element of the AKI pathway (risk, detection, education, intervention, commissioning and national data collection). The focus on commissioning pathways is designed to utilise institutional financial incentives to leverage quality improvements. Although the standardised AKI detection algorithm had not been developed within the ‘Think Kidneys' programme, a major step to move from a design phase to implementation was taken by the programme endorsing the algorithm and then working with the NHS England (Patient Safety Domain) to issue a level 3 Patient Safety Alert in June 2014 entitled ‘Standardising the Early Identification of AKI'.

The patient safety domain of NHS England has a vital role in identifying, understanding and managing risks that pose a danger to patients. Patient safety alerts are a crucial part of this and are used to rapidly alert the healthcare system to risks, as well as providing guidance on preventing harm. Incidents are identified using a reporting system to discern emerging patterns at a national level, so that appropriate guidance can be developed and issued. Patient Safety Alerts are disseminated via the Central Alerting System, a web-based cascading system for issuing alerts, important public health messages and other safety critical information to the NHS and other organisations, including independent providers of health and social care. Since January 2014, patient safety alerts have been categorised into a 3-stage system: stage 1 (warning of emerging risk), stage 2 (resource and information sharing) and stage 3 (directive - requires organisations to implement specific solutions or actions within a defined timeframe, and then confirm this via Central Alerting System).

The ‘Standardising the Early Identification of AKI' safety alert was sent to all healthcare provider trusts across England in this manner, and its level 3 status requires them to install the national AKI detection algorithm in their biochemistry LIMS. When integrated in this way, the algorithm will identify changes in serum creatinine concentration compatible with potential AKI and produce a test result. The laboratory system will then send the test result, using existing IT connections, to patient management systems where it can be viewed by clinicians, acting as a prompt to assess the patient and instigate AKI treatment pathways. Of note, the Patient Safety Alert is limited to standardising the process of AKI detection; the ‘alerting' part of the process can be further configured by individual hospitals to suit requirements/capabilities. As the output of the LIMS based algorithm is a test result, this may be via transfer of the result to other more sophisticated alerting systems, aiming to increase impact. In addition, the Patient Safety Alert requires that AKI test results are also sent to a central data collection point at the UK Renal Registry. The timeframe for compliance with the Patient Safety Alert is March 2015 and a variety of supporting documents have been released alongside the Patient Safety Alert, including frequently asked questions and a best practice guideline [15].

The primary aim of the Patient Safety Alert is to improve clinical outcomes. It seeks to provide the ability to ensure that a timely and consistent approach to the identification of patients with potential AKI is taken across the NHS. The Patient Safety Alert is an extremely potent tool by which uptake and usage of real-time electronic detection can be driven forward in all centres across the country, whilst at the same time working towards a nationally standardised process. However, is there evidence to support this approach? A number of previous reports have suggested that electronic detection of AKI can be effective in changing physician behaviour, triggering earlier intervention and introduction or change of therapy [16]. These data would support the premise that the uptake of AKI detection can improve standards of delivered care. However, there are less data to show that e-alerts can improve patient outcomes. Despite the lack of outcome data, we would argue the Patient Safety Alert to be timely and justified for several reasons. Firstly, the degree of variation between different AKI detection systems embedded in clinical use within the NHS was a situation that needed to be addressed urgently. These differences in detection systems between hospitals were largely driven by the limited capabilities of LIMS, rather than outcome data suggesting superiority of any one particular approach. Secondly, the ‘Think Kidneys' programme has been designed with a strong measurement theme throughout. As AKI detection is rolled out, there will be opportunities for time-series and stepped wedge quality improvement evaluations that will generate an evidence base alongside the process of implementation. There will be a central process by which the performance of the detection algorithm can be reviewed after roll out and an initial period of clinical usage, and any modifications or updates that are required can happen in a stepwise fashion within a framework of regulated LIMS upgrades. This approach recognises the effect that differing applications of diagnostic criteria may have on the incidence of AKI, and that there are certain situations (e.g., patients with low baseline serum creatinine concentrations) in which AKI based on percentage increases from baseline may perform less well [17]. There are further secondary benefits of the Patient Safety Alert. Whilst the process of detection is mandated by the Patient Safety Alert, it deliberately does not specify how results consistent with AKI are communicated to clinicians. This will allow ongoing innovation in more sophisticated messaging or alerting, in addition to AKI stage appearing alongside other biochemical results. Important wider effects include encouraging healthcare organisations to recognise AKI as a patient safety issue, which has improved awareness amongst senior management and other specialities, has helped integrate AKI into clinical governance structures and will promote sustainability of AKI initiatives.

The second aspect of the Patient Safety Alert is the establishment of data links to transmit data on AKI episodes from every biochemistry laboratory in England to the UK Renal Registry. This will form the building blocks of an ambitious plan to allow prospective measurement of the incidence and outcomes of AKI on a national level from over 150 acute hospitals and is possible because of the relatively unfragmented nature of the NHS in England. Data will be compiled into a Master Patient Index of those who have had AKI as defined by the standardised detection algorithm; this can then be combined with other health databases to build an AKI registry, reporting incidence and outcomes of AKI at a national level (mortality, chronic renal replacement therapy in particular). Such a registry would be an extremely powerful tool for improvement. It would allow organisations to benchmark their performance against others, which in itself may be a lever to drive quality; it will also allow measurement of effect of organisational level interventions. This may be as part of quality improvement, but there may even be potential opportunities for registry-based clinical trials [18].

Contribution from the AKI & CRRT 2015 Symposium at the 20th International Conference on Advances in Critical Care Nephrology, Manchester Grand Hyatt, San Diego, Calif., USA, February 17-20, 2015.