Acute kidney injury electronic alerts in primary care - findings from a large population cohort

Exploring the role of Artificial Intelligence in Acute Kidney Injury management: a comprehensive review and future research agenda

This study reviews the studies utilizing Artificial Intelligence (AI) and AI-driven tools and methods in managing Acute Kidney Injury (AKI). It categorizes the studies according to medical specialties, analyses the gaps in the existing research, and identifies opportunities for future research directions. PRISMA guidelines were adopted using the three most common databases (PubMed, Scopus, and EBSCO), which resulted in 27 eligible studies, published between 2012 and 2023. The study showed significant heterogeneity in the design of the models, with variations in clinical settings, patient characteristics, cohort regions, and statistical methods. Most models were developed for AKI in hospitalized patients, particularly those undergoing surgery or in intensive care units. Compact models with a subset of significant predictors were deemed more clinically applicable than full models with all predictors. The findings suggest that AI tools, such as machine learning (ML) algorithms, have high prediction capabilities despite the dynamic and complex association among the influencing factors and AKI. Based on these findings and the recognized need for broader inclusivity, future studies should consider adopting a more inclusive approach by incorporating diverse healthcare settings, including resource-limited or developing countries. This inclusivity will lead to a more holistic understanding of AKI management challenges and facilitate the development of adaptable and universally applicable AI-driven solutions. Additionally, further investigations should focus on refining AI models to enhance their accuracy and interpretability, promoting seamless integration and implementation of AI-based tools in real-world clinical practice. Addressing these key aspects will elevate the effectiveness and impact of AI-driven approaches in managing AKI.

Artificial intelligence and predictive models for early detection of acute kidney injury: transforming clinical practice

Acute kidney injury (AKI) presents a significant clinical challenge due to its rapid progression to kidney failure, resulting in serious complications such as electrolyte imbalances, fluid overload, and the potential need for renal replacement therapy. Early detection and prediction of AKI can improve patient outcomes through timely interventions. This review was conducted as a narrative literature review, aiming to explore state-of-the-art models for early detection and prediction of AKI. We conducted a comprehensive review of findings from various studies, highlighting their strengths, limitations, and practical considerations for implementation in healthcare settings. We highlight the potential benefits and challenges of their integration into routine clinical care and emphasize the importance of establishing robust early-detection systems before the introduction of artificial intelligence (AI)-assisted prediction models. Advances in AI for AKI detection and prediction are examined, addressing their clinical applicability, challenges, and opportunities for routine implementation.

A Review of Electronic Early Warning Systems for Acute Kidney Injury

Acute kidney injury (AKI) is characterized by impaired renal function that can result in irreversible severe renal impairment or lifelong dependence on renal replacement therapy in some cases. Early intervention can significantly slow down the progression of AKI and reduce mortality. In recent years, electronic early warning systems for patients with AKI have been gaining attention as a potential clinical decision-support option. This paper presents a review of the application of electronic early warning systems for AKI from four aspects: development process, types of output, influencing factors, and system evaluation.

Digital health and acute kidney injury: consensus report of the 27th Acute Disease Quality Initiative workgroup

Acute kidney injury (AKI), which is a common complication of acute illnesses, affects the health of individuals in community, acute care and post-acute care settings. Although the recognition, prevention and management of AKI has advanced over the past decades, its incidence and related morbidity, mortality and health care burden remain overwhelming. The rapid growth of digital technologies has provided a new platform to improve patient care, and reports show demonstrable benefits in care processes and, in some instances, in patient outcomes. However, despite great progress, the potential benefits of using digital technology to manage AKI has not yet been fully explored or implemented in clinical practice. Digital health studies in AKI have shown variable evidence of benefits, and the digital divide means that access to digital technologies is not equitable. Upstream research and development costs, limited stakeholder participation and acceptance, and poor scalability of digital health solutions have hindered their widespread implementation and use. Here, we provide recommendations from the Acute Disease Quality Initiative consensus meeting, which involved experts in adult and paediatric nephrology, critical care, pharmacy and data science, at which the use of digital health for risk prediction, prevention, identification and management of AKI and its consequences was discussed.

Utilising routine clinical laboratory data to support quality improvement in health care: Application of a national acute kidney injury alert system as a proof of concept

BACKGROUND

Acute kidney injury (AKI) is a global health issue known to cause avoidable harm and death. Improvement in its prevention and management is therefore considered an important goal for the health-care sector. The work here aimed to develop a tool which could be used to robustly and reliably measure, monitor, and compare the effectiveness of health-care interventions related to AKI across the Welsh NHS, a mechanism which did not exist previously.

METHODS

Using serum creatinine (SCr) as a biomarker for AKI and a validated national data-set collected from the all Wales Laboratory Information Management System, work involved applying Donabedian's framework to develop indicators with which to measure outcomes related to AKI, and exploring the potential of statistical process control (SPC) techniques for analysing data on these indicators.

RESULTS

Rate of AKI incidence and 30-day AKI-associated mortality are proposed as valid, feasible indicators with which to measure the effectiveness of health-care interventions related to AKI. The control chart, funnel plot, and Pareto chart are proposed as appropriate, robust SPC techniques to analyse and visualise variation in AKI-related outcomes.

CONCLUSIONS

This work demonstrates that routinely collected large SCr data offer a significant opportunity to monitor and therefore inform improvement in patient outcomes related to AKI. Moreover, while this work concerns utilisation of SCr data for improvement in AKI strategies, it is a proof of concept which could be replicated for other routinely collected clinical laboratory data, to improve the prevention and/or management of the conditions to which they relate.

Effect of clinical decision support systems on clinical outcome for acute kidney injury: a systematic review and meta-analysis

Background

Clinical decision support systems including both electronic alerts and care bundles have been developed for hospitalized patients with acute kidney injury.

Methods

Electronic databases were searched for randomized, before-after and cohort studies that implemented a clinical decision support system for hospitalized patients with acute kidney injury between 1990 and 2019. The studies must describe their impact on care processes, patient-related outcomes, or hospital length of stay. The clinical decision support system included both electronic alerts and care bundles.

Results

We identified seven studies involving 32,846 participants. Clinical decision support system implementation significantly reduced mortality (OR 0.86; 95 % CI, 0.75–0.99; p = 0.040, I² = 65.3 %; n = 5 studies; N = 30,791 participants) and increased the proportion of acute kidney injury recognition (OR 3.12; 95 % CI, 2.37–4.10; p < 0.001, I² = 77.1 %; n = 2 studies; N = 25,121 participants), and investigations (OR 3.07; 95 % CI, 2.91–3.24; p < 0.001, I² = 0.0 %; n = 2 studies; N = 25,121 participants).

Conclusions

Nonrandomized controlled trials of clinical decision support systems for acute kidney injury have yielded evidence of improved patient-centered outcomes and care processes. This review is limited by the low number of randomized trials and the relatively short follow-up period.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12882-021-02459-y.

Recurrent acute kidney injury: predictors and impact in a large population-based cohort

BACKGROUND

This study examined the impact of recurrent episodes of acute kidney injury (AKI) on patient outcomes.

METHODS

The Welsh National electronic AKI reporting system was used to identify all cases of AKI in patients ≥18 years of age between April 2015 and September 2018. Patients were grouped according to the number of AKI episodes they experienced with each patient's first episode described as their index episode. We compared the demography and patient outcomes of those patients with a single AKI episode with those patients with multiple AKI episodes. Analysis included 153 776 AKI episodes in 111 528 patients.

RESULTS

Of those who experienced AKI and survived their index episode, 29.3% experienced a second episode, 9.9% a third episode and 4.0% experienced fourth or more episodes. Thirty-day mortality for those patients with multiple episodes of AKI was significantly higher than for those patients with a single episode (31.3% versus 24.9%, P < 0.001). Following a single episode, recovery to baseline renal function at 30 days was achieved in 83.6% of patients and was significantly higher than for patients who had repeated episodes (77.8%, P < 0.001). For surviving patients, non-recovery of renal function following any AKI episode was significantly associated with a higher probability of a further AKI episode (33.4% versus 41.0%, P < 0.001). Furthermore, with each episode of AKI the likelihood of a subsequent episode also increased (31.0% versus 43.2% versus 51.2% versus 51.7% following a first, second, third and fourth episode, P < 0.001 for all comparisons).

CONCLUSIONS

The results of this study provide an important contribution to the debate regarding the need for risk stratification for recurrent AKI. The data suggest that such a tool would be useful given the poor patient and renal outcomes associated with recurrent AKI episodes as highlighted by this study.

Using electronic AKI alerts to define the epidemiology of acute kidney injury in renal transplants

Background

Little is known regarding the impact of acute kidney injury (AKI) on renal transplant outcome. Our aim was to define the incidence and outcome of AKI in renal transplant patients using data collected from a national AKI electronic alert system

Methods

The study represents a prospective national cohort study collecting data on 1224 renal transplants recipients with a functioning renal transplant, between April 2015 and March 2019.

Results

Four hundred forty patients experienced at least one episode of AKI giving an incidence rate of 35.4%. Sixty-four point seven% of episodes were AKI stage 1, 7.3% AKI stage 2 and 28% AKI stage 3. Only 6.2% of episodes occurred in the context of rejection. Forty-three point five% of AKI episodes were associated with sepsis. AKI was associated with pre-existing renal dysfunction, and a primary renal diagnosis of diabetic nephropathy. AKI was more prevalent in recipients from a donor after cardiac death (26.4% vs. 21.4%, p < 0.05) compared to the non-AKI cohort. Following AKI, 30-day mortality was 19.8% and overall mortality was 34.8%, compared to 8.4% in the non AKI cohort (RR 4.06, 95% CI 3.1–5.3, p < 0.001). Graft survival (GS), and death censored graft survival (DCGS) censored at 4 years, in the AKI cohort were significantly lower than in the non AKI group (p < 0.0001 for GS and DCGS).

Conclusion

The study provides a detailed characterisation of AKI in renal transplant recipients highlighting its significant negative impact on patient and graft survival.

Acute kidney injury demographics and outcomes: changes following introduction of electronic acute kidney injury alerts-an analysis of a national dataset

BACKGROUND

Electronic alerts for acute kidney injury (AKI) have been widely advocated. Our aim was to describe the changes in AKI demographics and outcomes following implementation of a national electronic AKI alert programme.

METHODS

A prospective national cohort study was undertaken to collect data on all cases of AKI in adult patients (≥18 years of age) between 1 April 2015 and 31 March 2019.

RESULTS

Over the period of data collection, there were 193 838 AKI episodes in a total of 132 599 patients. The lowest incidence of AKI was seen in the first year after implementation of electronic alerts. A 30-day mortality was highest in Year 1 and significantly lower in all subsequent years. A direct comparison of mortality in Years 1 and 4 demonstrated a significantly increased relative risk (RR) of death in Year 1: RR = 1.08 [95% confidence interval (CI) 1.054-1.114 P < 0.001]. This translates into a number needed to treat in Year 4 for one additional patient to survive of 69.5 (95% CI 51.7-106.2) when directly comparing the outcomes across the 2 years. The increase in the number of cases and improved outcomes was more pronounced in community-acquired AKI, and was associated with a significant increase in patient hospitalization.

CONCLUSIONS

This study represents the first large-scale dataset to clearly demonstrate that a national AKI alerting system which highlights AKI is associated with a change in both AKI demographics and patient outcomes.

Acute Kidney Injury in Children Based on Electronic Alerts

OBJECTIVE

To define the incidence and outcome of acute kidney injury (AKI) in pediatrics using data collected from a national electronic alert system.

STUDY DESIGN

A prospective national cohort study was undertaken to collect data on all cases of pediatric AKI, excluding neonates, identified by an e-alert, from April 2015 to March 2019.

RESULTS

There were 2472 alerts in a total of 1719 patients, giving an incidence of 77.3 per 100 000 person-years. Of the patients, 84.2% of all AKI were stage 1 and 58.3% occurred with a triggering creatinine within the reference range. The incidence of AKI was associated with measures of social deprivation. Thirty-day mortality was 1.7% but was significantly higher in hospital-acquired AKI (2.1%), compared with community-acquired AKI (0.8%, P < .001) and was associated with the severity of AKI at presentation. A significant proportion of patients had no repeat measure of creatinine (39.8%). This was higher in community-acquired AKI (69.7%) compared with hospital-acquired AKI (43.0%, P < .001), and higher in patients alerting with patients triggering with a creatinine within the reference range (48.4% vs 24.5%, P < .001). The majority of patients (84.7%) experienced only 1 AKI episode. Repeated episodes of AKI were associated with increased 30-mortaltiy (11.6% vs 4.6%, P < .001) and higher residual renal impairment (13.3% vs 5.4%, P < .001).

CONCLUSIONS

The results suggest that the significance of the alert is missed in many cases reflecting that a large proportion of cases represent modest elevations in serum creatinine (SCr), triggered by a SCr level that may be interpreted as being normal despite a significant increase from the baseline for the patient.

Using chronic kidney disease trigger tools for safety and learning: a qualitative evaluation in East London primary care

BACKGROUND

An innovative programme to improve identification and management of chronic kidney disease (CKD) in primary care was implemented across three clinical commissioning groups (CCGs) in 2016. This included a falling estimated glomerular filtration rate (eGFR) trigger tool built from data in the electronic health record (EHR). This tool notifies GP practices of falling eGFR values. By alerting clinicians to patients with possible CKD progression the tool invites clinical review, a referral option, and written reflection on management.

AIM

To identify practitioner perceptions of trigger tool use from interviews, and compare these with reflections on clinical management recorded within the tools.

DESIGN AND SETTING

A qualitative analysis set in 136 practices across East London during 2016-2018.

METHOD

Eight semi-structured interviews with GPs and practice staff were recorded, and thematic analysis was undertaken using framework analysis. The reflective comments recorded in the trigger tools of 1921 cases were categorised by age group, referral status, and by the drop in eGFR (>15 or >25 ml/min).

RESULTS

Three themes emerged from the interviews: getting started, patient safety, and trigger tools for learning. Well-organised practices found the tool was readily embedded into workflow and expressed greater motivation for using it. The tool was seen to support patient safety, and was used for learning about CKD management, both individually and as a practice. Reflective comments from 1921 trigger tools were reviewed. These supported the theme of patient safety. The free-text data, stratified by age, challenged the expectation that younger cases, at higher risk of progressive CKD, would have higher referral rates.

CONCLUSION

Building electronic trigger tools from the EHR can identify patients with a falling eGFR, prompting review of the eGFR trajectory and management plan. Interview and reflective data illustrated that practice use of the tool supports the patient safety agenda and encourages learning about CKD management.

Acute Kidney Injury, Age, and Socioeconomic Deprivation: Evaluation of a National Data Set

Introduction

This study examined the relationship among age, measures of social deprivation, and incidence and outcome of acute kidney injury (AKI).

Methods

The Welsh National electronic AKI reporting system was used to identify all cases of AKI in patients 18 years or older between March 2015 and January 2017. Socioeconomic classification of patients was derived from the Welsh Index of Multiple Deprivation (WIMD). Patients were grouped according to their WIMD score, and Multivariate Cox proportional hazard modeling was used to adjust the data for age. The ranked data were categorized into percentiles and correlated with incidence, and measures of AKI severity and outcome.

Results

Analysis included 57,654 patients. For the whole cohort, the highest 90-day survival was associated with the most socially deprived cohorts. There was a significant negative relationship between age-adjusted incidence of AKI and the WIMD score. In patients 60 years or older, there was an inverse correlation between WIMD score and survival that was not evident in those younger than 60. AKI severity at presentation was worse in patients from areas of social deprivation. Social deprivation was associated with a significantly higher proportion of preexisting chronic kidney disease (CKD) in patients with AKI older than 60, but not in those younger than 60.

Conclusion

Overall mortality following AKI was higher in least-deprived areas, reflecting an older patient cohort. In contrast, social deprivation was associated with higher age-adjusted AKI incidence and age-adjusted mortality following AKI. The excess mortality observed in more deprived areas was associated with more severe AKI and a higher proportion of preexisting CKD.

Improved management of acute kidney injury in primary care using e-alerts and an educational outreach programme

PURPOSE

Acute kidney injury (AKI) detected in primary care is associated with increased morbidity and mortality. AKI electronic alerts (e-alerts) and educational programmes have recently been implemented but their contribution to improve AKI care is unknown. This project aimed to improve response to AKI detected in primary care and used a factorial design to evaluate the impact of the UK National Health Service (NHS) AKI e-alert and AKI educational outreach sessions on time to response to primary care AKI stages 2 and 3 between April and August 2016.

METHODS

A total of 46 primary care practices were randomized into four groups. A 2 × 2 factorial design exposed each group to different combinations of two interventions. The primary outcome was 'time to repeat test' or hospitalization following AKI e-alert for stages 2 and 3. Yates algorithm was used to evaluate the impact of each intervention. Time to response and mortality pre- and post-intervention were analysed using Mann-Whitney U test and chi-square test respectively. The factorial design included two interventions: an AKI educational outreach programme and the NHS AKI e-alerts.

RESULTS

1807 (0.8%) primary care blood tests demonstrated AKI 1-3 (78.3% stage 1, 14.8% stage 2, 6.9% stage 3). There were 391 stage 2 and 3 events from 251 patients. E-alerts demonstrated a reduction in mean response time (-29 hours). Educational outreach had a smaller effect (-3 hours). Median response time to AKI 2 and 3 pre- and post-interventions was 27 hours versus 16 hours respectively (P = 0.037). Stage 2 and 3 event-related 30-day all-cause mortality decreased following the interventions (15.6% versus 3.9% P = 0.036).

CONCLUSION

AKI e-alerts in primary care hasten response to AKI 2 and 3 and reduce all-cause mortality. Educational outreach sessions further improve response time.

The influence of socioeconomic status on presentation and outcome of acute kidney injury

AIM

Although socioeconomic background is known to impact on the incidence and progression of chronic kidney disease, its influence of on the presentation and outcome for acute kidney injury is not known and is the subject of this study.

DESIGN

The Welsh National electronic AKI reporting system was used to identify all cases of AKI in patients >18 years of age between March 2015 and November 2017.

METHODS

Socioeconomic classification of patients was derived from the Welsh Index Multiple Deprivation score (WIMD). Patients were grouped according to the WIMD score by their postcode, and the ranked data were categorized into percentiles and correlated with incidence and measures of AKI severity and outcome.

RESULTS

Date was collected on a total of 57 654 patients. Increased deprivation was associated with higher AKI incidence rates, more episodes of AKI per patient and more severe AKI at presentation. In contrast 90-day mortality was highest in the most affluent areas. Mortality in affluent areas was driven by increased patient age. Corrected for age 90-day mortality was higher in areas of increased deprivation.

CONCLUSION

This study highlights that AKI incidence presentation and outcomes are adversely affected by social deprivation. Further studies are required to understand the extent to which these differences reflect patient related factors or regional differences in provision and access to care.

Impact of introducing electronic acute kidney injury alerts in primary care

Background

Acute kidney injury (AKI) is associated with decreased survival, future risk of chronic kidney disease and longer hospital stays. Electronic alerts (e-alerts) for AKI have been introduced in the UK in order to facilitate earlier detection and improve management. The aim of this study was to establish if e-alerts in primary care were acted on by examining timing of repeat creatinine testing.

Methods

The National Health Service England Acute Kidney Injury electronic alert algorithm was introduced in April 2015 across both primary and secondary care in NHS Tayside accompanied by a programme of education. Data from a 12-month period (2012) predating introduction of the e-alerts were compared with a 12-month period following implementation of e-alerts for AKI. Biochemistry testing following the AKI episode, timing of repeat tests and numbers of patients hospitalized within 7 days of episode were compared between the two time periods.

Results

During the 12 months after e-alert introduction, 9781 AKI e-alerts were generated. Of these, 1460 (14.9%) alerts were generated in primary care. Median duration to repeat blood testing for these primary care alerts was 5 days for AKI Stage 1 [interquartile range (IQR) 2–10], 2 days for Stage 2 (IQR 1–5) and 1 day (IQR 0–2) for Stage 3. During 2012 (prior to e-alert implementation) 8812 AKI episodes were identified. Of these, 2650 tests (30.1%) were requested by primary care staff. Median duration to repeat creatinine testing was longer: 55 days (IQR 20–142) for Stage 1, 38 days (IQR 15–128) for Stage 2 was and 53 days (IQR 20–137) for Stage 3. More patients had biochemistry tests repeated within 7 days of AKI onset, pre-alert implementation; 252 (9.5%) versus 857 (58.7%) (P < 0.001). Rates of hospitalization within 7 days of AKI increased from 342 (12.9%) pre-implementation to 372 (25.5%) post-implementation (P < 0.001).

Conclusions

Within primary care, e-alert implementation was associated with higher rates of creatinine monitoring, but also higher rates of hospitalization.

Adding a new dimension to the weekend effect: an analysis of a national data set of electronic AKI alerts

BACKGROUND

Increased mortality related to differences in delivery of weekend clinical care is the subject of much debate.

AIM

We compared mortality following detection of acute kidney injury (AKI) on week and weekend days across community and hospital settings.

DESIGN

A prospective national cohort study, with AKI identified using the Welsh National electronic AKI reporting system.

METHODS

Data were collected on outcome for all cases of adult AKI in Wales between 1 November 2013 and 31 January 2017.

RESULTS

There were a total of 107 298 episodes. Weekday detection of AKI was associated with 28.8% (26 439); 90-day mortality compared to 90-day mortality of 31.9% (4551) for AKI detected on weekdays (RR: 1.11, 95% CI: 1.08-1.14, P < 0.001, HR: 1.16 95% CI: 1.12-1.20, P < 0.001). There was no 'weekend effect' for mortality associated with hospital-acquired AKI. Weekday detection of community-acquired AKI (CA-AKI) was associated with a 22.6% (10 356) mortality compared with weekend detection of CA-AKI, which was associated with a 28.6% (1619) mortality (RR: 1.26, 95% CI: 1.21-1.32, P < 0.001, HR: 1.34, 95%CI: 1.28-1.42, P < 0.001). The excess mortality in weekend CA-AKI was driven by CA-AKI detected at the weekend that was not admitted to hospital compared with CA-AKI detected on weekdays which was admitted to hospital (34.5% vs. 19.1%, RR: 1.8, 95% CI: 1.69-1.91, P < 0.001, HR: 2.03, 95% CI: 1.88-2.19, P < 0.001).

CONCLUSION

'Weekend effect' in AKI relates to access to in-patient care for patients presenting predominantly to hospital emergency departments with AKI at the weekend.

Seasonal pattern of incidence and outcome of Acute Kidney Injury: A national study of Welsh AKI electronic alerts

OBJECTIVES

To identify any seasonal variation in the occurrence of, and outcome following Acute Kidney Injury.

METHODS

The study utilised the biochemistry based AKI electronic (e)-alert system established across the Welsh National Health Service to collect data on all AKI episodes to identify changes in incidence and outcome over one calendar year (1st October 2015 and the 30th September 2016).

RESULTS

There were total of 48 457 incident AKI alerts. The highest proportion of AKI episodes was seen in the quarter of January to March (26.2%), and the lowest in the quarter of October to December (23.3%, P < .001). The same trend was seen for both community-acquired and hospital-acquired AKI sub-sets. Overall 90 day mortality for all AKI was 27.3%. In contrast with the seasonal trend in AKI occurrence, 90 day mortality after the incident AKI alert was significantly higher in the quarters of January to March and October to December compared with the quarters of April to June and July to September (P < .001) consistent with excess winter mortality reported for likely underlying diseases which precipitate AKI.

CONCLUSIONS

In summary we report for the first time in a large national cohort, a seasonal variation in the incidence and outcomes of AKI. The results demonstrate distinct trends in the incidence and outcome of AKI.