Earlier Intervention for Acute Kidney Injury
Earlier Intervention for Acute Kidney Injury
The median [Interquartile range (IQR)] time interval between the previous and alert creatinines was 35 (7–170) and 51 (9–197) days for the before and after groups, respectively (P = 0.67, Mann–Whitney test). In the before study, from 264 alerts there were 157 first episodes of AKI identified in adult patients. In four patients in the before group, a further AKI episode was identified during that phase. A total of 103 further false alerts were excluded on review (see below).
In the after study, from 361 alerts there were 251 first episodes of AKI identified in adult patients. A total of 101 further false alerts were excluded on review. In six patients who were included in the before group, a further AKI episode was identified during the after phase. Three patients in the after phase had repeat episodes of AKI identified. These nine episodes were eligible for an outreach call, so 260 alerts in 259 patients were eligible for one or more outreach calls. In both phases, there were approximately five true alerts per day, 7 days a week (data not shown). In the before and after phases, respectively, 24 and 22% of the first alerts were for patients in the community at the time of the blood test (a small number were later admitted).
Across the two phases, the reasons for the 204 'false alerts' were similar to our previous work: identification of haemodialysis patients (n = 144); repeat alerts for the same AKI episode (20); patient already on haemodialysis at time of alert (15); suppressed previous creatinine value (13); alerts in children (4) and other reasons (8).
The characteristics of the before and after group are shown in Table 1. The groups were comparable in their clinical characteristics. As might be expected the patients were older, with considerable comorbidity. In the before and after groups, 35.7 and 35.9%, respectively, had an estimated GFR <60 mL/min/1.73 m, calculated from the previous creatinine. The ≥75% delta check produced similar proportions of patients with stages R, I and F AKI (equivalent to stages 1–3). The nature of the delta check excludes patients with smaller rises in creatinine (≥26 μmol/L or 0.3 mg/dL) now designated stage 1 AKI. Few patients had acute-on-chronic kidney disease in the before and after phases (0.6 and 2.0%, respectively), due to the requirement for a ≥75% rise in creatinine. The use of haemodialysis was comparable in both groups.
In the after study, 260 AKI episodes were eligible for AKI outreach. Of these, completed outreach calls were made in 229 (88.1%); 12 patients were deceased before the call could be made (4.6%); in 8 patients no contact could be made with the primary clinical team (3.1%); in 7 patients there was insufficient outreach time (2.7%) and in 4 cases (1.5%), the patient was referred for nephrology review before or at the time of the call.
The outreach call for the 229 successful contacts was made at a median (IQR) time after the alert of 14.6 (3.8–19.8) hours. A median (IQR) of three (two to four) specific recommendations were made and recorded on the outreach form. The details of the recommendations are shown in Table 2. Larger numbers of recommendations were made regarding investigations, volume or fluid management and drugs. General medical recommendations were required in a minority of patients.
There was no difference in peak creatinine between the groups. There was a non-significant reduction in length of stay of a little over half a day in the after group. The patients in the after phase showed a modest improvement in survival, compared with those in the before phase, initially about 6%. This difference appeared rapidly and was initially sustained during follow-up (Figure 2), with a later attenuation of any survival benefit (P = 0.38, log rank test).
(Enlarge Image)
Figure 2.
Survival of before and after groups.
Results
Alerts
The median [Interquartile range (IQR)] time interval between the previous and alert creatinines was 35 (7–170) and 51 (9–197) days for the before and after groups, respectively (P = 0.67, Mann–Whitney test). In the before study, from 264 alerts there were 157 first episodes of AKI identified in adult patients. In four patients in the before group, a further AKI episode was identified during that phase. A total of 103 further false alerts were excluded on review (see below).
In the after study, from 361 alerts there were 251 first episodes of AKI identified in adult patients. A total of 101 further false alerts were excluded on review. In six patients who were included in the before group, a further AKI episode was identified during the after phase. Three patients in the after phase had repeat episodes of AKI identified. These nine episodes were eligible for an outreach call, so 260 alerts in 259 patients were eligible for one or more outreach calls. In both phases, there were approximately five true alerts per day, 7 days a week (data not shown). In the before and after phases, respectively, 24 and 22% of the first alerts were for patients in the community at the time of the blood test (a small number were later admitted).
Across the two phases, the reasons for the 204 'false alerts' were similar to our previous work: identification of haemodialysis patients (n = 144); repeat alerts for the same AKI episode (20); patient already on haemodialysis at time of alert (15); suppressed previous creatinine value (13); alerts in children (4) and other reasons (8).
Comparison of Before and After Groups
The characteristics of the before and after group are shown in Table 1. The groups were comparable in their clinical characteristics. As might be expected the patients were older, with considerable comorbidity. In the before and after groups, 35.7 and 35.9%, respectively, had an estimated GFR <60 mL/min/1.73 m, calculated from the previous creatinine. The ≥75% delta check produced similar proportions of patients with stages R, I and F AKI (equivalent to stages 1–3). The nature of the delta check excludes patients with smaller rises in creatinine (≥26 μmol/L or 0.3 mg/dL) now designated stage 1 AKI. Few patients had acute-on-chronic kidney disease in the before and after phases (0.6 and 2.0%, respectively), due to the requirement for a ≥75% rise in creatinine. The use of haemodialysis was comparable in both groups.
Outreach Calls in After Phase
In the after study, 260 AKI episodes were eligible for AKI outreach. Of these, completed outreach calls were made in 229 (88.1%); 12 patients were deceased before the call could be made (4.6%); in 8 patients no contact could be made with the primary clinical team (3.1%); in 7 patients there was insufficient outreach time (2.7%) and in 4 cases (1.5%), the patient was referred for nephrology review before or at the time of the call.
Interventions Made During the Outreach Call
The outreach call for the 229 successful contacts was made at a median (IQR) time after the alert of 14.6 (3.8–19.8) hours. A median (IQR) of three (two to four) specific recommendations were made and recorded on the outreach form. The details of the recommendations are shown in Table 2. Larger numbers of recommendations were made regarding investigations, volume or fluid management and drugs. General medical recommendations were required in a minority of patients.
Outcomes
There was no difference in peak creatinine between the groups. There was a non-significant reduction in length of stay of a little over half a day in the after group. The patients in the after phase showed a modest improvement in survival, compared with those in the before phase, initially about 6%. This difference appeared rapidly and was initially sustained during follow-up (Figure 2), with a later attenuation of any survival benefit (P = 0.38, log rank test).
(Enlarge Image)
Figure 2.
Survival of before and after groups.
