Validation of The Health Improvement Network (THIN) database for epidemiologic studies of chronic kidney disease

Abstract

Purpose Chronic kidney disease (CKD) is a prevalent and important outcome and covariate in pharmacoepidemiology. The Health Improvement Network (THIN) in the U.K. represents a unique resource for population-based studies of CKD. We compiled a valid list of Read codes to identify subjects with moderate to advanced CKD. Methods A cross-sectional validation study was performed to identify codes that best define CKD stages 3–5. All subjects with at least one non-zero measure of serum creatinine after 1-Jan-2002 were included. Estimated glomerular filtration rate (eGFR) was calculated according to the Schwartz formula for subjects <18 years and the Modification of Diet in Renal Disease formula for subjects ≥18 years of age. CKD was defined as an eGFR <60 ml/min/1.73m2 on at least two occasions, more than 90 days apart. Results The laboratory definition identified 230,426 subjects with CKD, for a period prevalence in 2008 of 4.56% (95% CI: 4.54, 4.58). A list of 45 Read codes was compiled which yielded a positive predictive value of 88.9% (95% CI: 88.7, 89.1), sensitivity of 48.8%, negative predictive value of 86.5%, and specificity of 98.2%. Of the 11.1% of subjects with a code who did not meet the laboratory definition, 83.6% had at least one eGFR <60. The most commonly used code was for CKD stage 3. Conclusions The proposed list of codes can be used to accurately identify CKD when serum creatinine data are limited. The most sensitive approach for the detection of CKD is to use this list to supplement creatinine measures.

Use of proton pump inhibitors is associated with fractures in young adults: a population-based study

“…We used a large, population-based dataset that has been extensively validated for health outcomes [12,13]. This dataset, THIN, is fundamentally different from an insurance claims database; it is an electronic medical record maintained for the purpose of providing care.…”

“…For practitioners who use THIN, it represents the entire medical record, and THIN data has previously been validated for health outcomes research [12,13]. Because patients in the UK must register with a general practitioner to receive non-emergent medical care through the National Health Service, THIN broadly represents the population.…”

Is routine hospital episode data sufficient for identifying individuals with chronic kidney disease? A comparison study with laboratory data

“…It may also reflect that those with CKD at younger ages are likely to have fewer comorbidites when admitted to hospital and therefore have this recognised when discharge coding is carried out. 31 Denburg et al 17 looked at the recording of biochemistry results at a general practice level compared to the recognition of CKD on general practice coding, which again found low sensitivity but excellent specificity and high PPV. It is unclear, however, how many of the biochemistry results had been entered into GP systems manually.…”

“…Although this would be a useful additional source of data, it was not available to us, would require assessment in its own right, and has been explored at least at a GP biochemistry recording level before. 17 Our biochemistry definition of no-CKD was all-inclusive, including impaired eGFR (at least one eGFR <60ml/min/1.73m 2 but not sustained) and eGFR not measured. However, we performed sensitivity analyses, defining "no CKD" as those with normal eGFR only and found that this only improved PPV and worsened NPV.…”

“…Two recent systematic reviews 14,15 , and recent studies [16][17][18][19] , have evaluated the degree to which administrative coding accurately identified individuals with kidney diseases, reporting a large variation in sensitivity (3%-88%). Only a few studies have compared hospital administrative data to laboratory data employing the 2002 KDOQI definition of stages [3][4][5] CKD, of at least two eGFR <60 ml/min/1.73m 2 at least 90 days apart.…”