Epidemiology of Acute Kidney Injury in England - 1998 to 2013
Epidemiology of Acute Kidney Injury in England - 1998 to 2013
The true population incidence of acute kidney injury (AKI), previously called acute renal failure, in England is not known. A better understanding of the epidemiology of AKI at a national level is essential to inform initiatives to prevent AKI as well as reduce the associated morbidity and mortality. The purpose of this study is to combine the national database of all hospital discharges with national census data to investigate trends in the incidence of both, AKI not requiring dialysis and AKI requiring dialysis, as well as its associated mortality and its determinants in England between 1998 and 2013
Subject Selection and Withdrawal We will identify all cases of AKI by using validated International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes in any of he diagnoses codes, in keeping with the objective of the study. Patients with any of the following codes will be included: N17.0 for acute renal failure with tubular necrosis, N17.1 for acute renal failure with acute cortical necrosis, N17.2 for acute renal failure with medullary necrosis, N17.8 for other acute renal failure and N17.9 for acute renal failure, unspecified. Acute renal failure has been replaced by new terminology, acute kidney injury, but due to lack of ICD-10-CM codes for AKI, we will use the ICD-10-CM codes for acute renal failure and henceforth, will be referred to as AKI in this protocol.
Subject Recruitment and Screening We will extract 1998 to 2013 data from the Hospital Episode Statistics (HES), a data warehouse containing details of all admissions, outpatient appointments and A&E attendances at National Health Service (NHS) hospitals in England. HES collects a detailed record for each 'episode' of admitted patient care delivered in England, either by NHS hospitals, primary care trusts, mental health trusts or delivered in the independent sector but commissioned by the NHS. These data are collected during a patient's time in hospital and stored as a large collection of separate records, one for each period of care, in a secure data warehouse. We will extract secondary diagnosis codes for up to 14 diagnoses available till 2006-07 and up to 20 additional diagnoses from 2007 till 2013. In addition to demographic data, we will also extract up to 15 procedure codes. To identify AKI patients requiring dialysis, we will include procedure code of X40.3 for hemodialysis or X40.4 for hemofiltration in any of the 15 procedures. To exclude patients with chronic kidney disease starting dialysis, we will exclude patients who had chronic kidney disease (CKD) stage 5 (N18.5) and end stage renal disease (N18.6) and procedure codes for arteriovenous fistula (L74.2) or arteriovenous shunt (L74.3) during the inpatient admission (figure 1). This algorithm has been shown to be sensitive and specific, with a high positive and negative predictive value (all >90%. Patients who were admitted, but were not discharged during the study period will not be included in the study.
We will obtain data on patient demographics, hospital characteristics, in-hospital mortality, disposition, length of stay (LOS), and up to 20 diagnosis and 15 procedure codes that are based on the ICD-10-CM and Office of Population Censuses and Surveys Classification of Surgical Operations and Procedures, 4th revision (OPCS-4) from the HES database. To obtain population incidence of AKI requiring dialysis, we will obtain mid-year populations of England in each year from 1998 to 2013 from the Office of National Statistics (ONS). We will also obtain length of stay and mortality for all patients who did not have acute kidney injury coded as N17 in any of the diagnoses codes, in England over the same time period.
Patients will not be contacted. This study will not obtain informed consent from the patients' as we will be working with anonymous patient data. No screening requirements (laboratory or diagnostic testing) are necessary to meet the inclusion or exclusion criteria. Unique Subject Identification numbers are already allocated to the patients and the HES database does not include any patient's name or hospital number.
Sample Size Determination The data from HES will be extracted for the period between April 1998 and March 2013. HES data is collected during a patients' time at hospital and is submitted to allow hospitals to be paid for the care they deliver. As this is a retrospective population cohort study using a national database, all cases of AKI as defined by the inclusion criteria, from the HES database will be used.
Statistical Methods All analyses will be performed using validated statistical software. Data during the 15-year period will be divided into three 5-yr periods (April 1988 to March 2003, April 2003 to March 2008 and April 2008 to March 2013) and also presented by individual year. Age will be categorized into the groups <65, 65-74, 75-84 and>85. Charlson scores greater than 5 will be grouped into one category. Method of admission will be defined as one of elective admission; emergency admission; maternity admission; other admission or unknown. Ethnicity will be grouped as White, Mixed, Asian, Black, other ethnic groups and ethnicity not stated/unknown.
Continuous variables will be expressed as mean with 95% confidence interval (CI). Categorical variables will be expressed as proportions and compared with the chi square test. Incidence rates per 1000 person-years and 95% confidence intervals will be calculated using Poisson regression analysis and compared using a chi-squared test for proportions. Age-adjusted incidence rates for AKI will be calculated based on the UK standard population using the direct standardization method. Further analysis will be done to identify regional variation in epidemiology of both AKI not requiring dialysis and AKI requiring dialysis in England.
The Poisson regression model will be used to analyse the effect of gender, age group, period of admission, method of admission, Charlson co-morbidity index and ethnicity on incidence of AKI. Cox Proportional Hazard Models will be applied to test the association between mortality and these variables. Multivariable logistic regression will be considered to assess the association between AKI and all these variables simultaneously. We will use the middle period of 2003-2008 as reference, which will enable us to evaluate improvement in mortality as compared to 1998-2003 and also to evaluate whether the improvement has been sustained in the later period of 2008-2013. The effect of AKI on mortality and length of stay will be tested using the Cox Proportional Hazard Models and the Linear Regression respectively. Long-term survival of patients will be plotted using Kaplan Meir curves after obtaining the date and cause of death for all patients obtained from Office of National Statistics.
We will randomly select a number of patients who had a recorded diagnosis of AKI in HES database since 1998. Controls will also be selected and matched to each patient by age, sex, date of diagnosis and region. The association between AKI and each exposure will be analyzed using conditional logistic regression.
Data Handling and Record Keeping The data obtained from HES will be anonymised and will not have any patient identifiable characteristics. The study protocol, documentation, data and all other information generated will be held in strict confidence. No information concerning the study or the data will be released to any unauthorized third party, without prior written approval of the sponsoring institution. Authorized representatives of the sponsoring institution may inspect all documents and records, required to be maintained, by the Investigator, including but not limited to, medical records (office, clinic or hospital) for the subjects in this study. The clinical study site will permit access to such records. Information about study subjects will be kept confidential and managed according to the requirements of the Data Protection Act, NHS Caldicott Guardian, The Research Governance Framework for Health and Social Care, Ethics Committee Approval and Trust IM&T Policy. The Chief/Principal Investigator will be the "Custodian" of the data
Source Documents The source document will be the HES database. HES data is held as financial year data-sets, although the source data is collected monthly from NHS - and originates from commissioning messages sent by healthcare providers to healthcare commissioners via the secondary User Services (SUS). At pre-arranged date during the year, SUS takes an extract from their database and sends it to HES. HES then validates and clean the extract, before deriving new items and making the information available in the data warehouse. Data quality reports and checks are completed at various stages in the cleaning and processing cycle.
Records Retention The data will be kept for 15 years after the completion of the study, in accordance with the requirements of the Therapeutic Goods Administration and Health Privacy Principals.
All research data and study related documents will be stored, following the end of the study. The data will be secured in password-protected computer with access to the chief investigator. The data will be deleted from the hard drive of the computer at the end of the archival period.
Study Monitoring, Auditing, and Inspecting The investigator will permit study-related monitoring, audits and inspections by the Ethics Committee, the Sponsor and the Research Governance Manager. This study will be monitored by the Research Governance Manager according to the Research & Development Office procedure for monitoring all non-commercial research carried out at the Derby Hospitals NHS Foundation Trust and the Southern Derbyshire Primary Care Trusts. In line with the responsibilities set out in the Research Governance Framework, the Investigator will ensure that the research governance manager or other regulatory monitoring authority is given access to all study-related documents and study related facilities.
Inclusion Criteria:
Exclusion Criteria: