Developing and validating a diabetes database in a large health system.
ABSTRACT One component of clinical information systems is a registry of patients. Registries allow providers to identify gaps in care at the population level. Registries also allow for rapid cycle continuous quality improvement, targeted practice change and improved outcomes. Most registries are built based on membership with an insurer or other selection criteria. Little, if any data exist on registries representing demographically heterogeneous populations.
Administrative and clinical data for the period 1/1/2000-12/30/03 were examined. In total, 46,082,941 lab reports, 233,292,544 medical records, and 9,351,415 medical record abstracts, representing approximately 2 million unique patients were searched. The diabetes source population was identified by presence of any one of the following criteria: ICD-9 code 250 (diabetes) for inpatient, emergency room or outpatient visits; any hemoglobin A1c result; blood glucose >200mg/dl; or diabetes medication. A diagnosis of diabetes was verified by trained chart reviewers on a sample of patients. Single indicators and combinations were examined to determine optimal identification of these cases.
In two separate validation studies, using two or more indicators or outpatient diagnosis maximized positive predictive value (PPV) (96 and 97%) and sensitivity (99 and 100%) and identified 55,807 individuals. When all patients with a single indicator of outpatient diagnosis (which had the highest single PPV of 94 and 95%) were included together with those having >or=2 indicators, the final sample size was 65,725.
Two or more indicators or an out-patient-diagnosis identifies a sizeable and unselective diabetes database which can be used to track processes and outcomes.
- SourceAvailable from: Grace Joshy[Show abstract] [Hide abstract]
ABSTRACT: Background Lack of population based data is a critical problem in diabetes surveillance in New Zealand. This thesis looks at the feasibility, strengths and weaknesses of linking existing databases to create a regional diabetes register in the Waikato. Methods Completeness and validity of key databases and agreement between common data items have been studied using the following audits and studies linking multiple data sources: • A pilot study in a rural town (Taumarunui), linking multiple data sources including the secondary care based Waikato Regional Diabetes Service (WRDS) database and the Get Checked data from primary care. • A general practice based study in Hamilton, linking primary care data (diagnosis codes, prescriptions, laboratory tests, Get Checked) with the WRDS database. • Another general practice based study in Rotorua, a town with high Maori population, linking primary care data with deprivation scores. • Audits using WRDS data and Waikato DHB hospital systems to assess data agreement. • Retention of patients in the Get Checked programme was examined using Waikato Primary Health’s data. • Three retrospective studies linking the WRDS data with Waikato DHB hospital systems and national mortality data, which looked at hospital admissions, progression of renal disease and mortality. The studies used several methods of data validation including comparison of datasets, manual search of patient records, direct contact with patients and comparison of data from external sources. Linked datasets were used to identify disparities in prevalence of diabetes, access to diabetes care, diabetes complications and mortality. Results iii • The coverage of the WRDS database was high (86%-91%), but newly diagnosed patients and older patients not needing retinal screening are underrepresented. Case identification using primary care systems was high, but the coverage of the “Get Checked” programme (62%-80%) varied depending on practice IT systems, data handling procedures and patient characteristics. • The Rotorua study shows that diabetes prevalence rises with increasing deprivation among Europeans, but not among Maori. • Maori and Asian patients were less likely to access retinal screening in Hamilton. Patients aged<40 years, those of Maori or Asian origin, and those with Type 1 diabetes were less likely to be retained in the Get Checked programme with regular checks. Almost all patients had barriers to diabetes care in Taumarunui. Psychological barriers to diabetes care rank highly for all subgroups of ethnicity, age, gender, duration of diabetes and insulin treatment. • Outcomes analyses showed that compared with Europeans with diabetes, Maori diabetes patients had a significantly higher risk of end-stage renal disease (ESRD), renal admission and renal death (46-fold, seven-fold and fourfold increases, respectively). Maori patients progressed at a significantly faster rate from first hospital admission for chronic renal disease to ESRD. Maori were more likely than Europeans to have diabetes reported on mortality coding. They were also were more likely to die from cardiovascular disease, cancer and renal disease [Hazard-ratios 2.31(1.6-3.3), 1.83(1.1-3), and 11.74(4.8-29) respectively]. Discussion The advantages and the difficulties of linking primary care and secondary care databases to identifying diagnosed diabetes patients, the potential barriers to implementation of a diabetes register and the critical factors for a successful system are discussed. This research has demonstrated the potential of linking databases to monitor diabetes care and outcomes, but implementation would need substantial policy changes and financial backing.
- [Show abstract] [Hide abstract]
ABSTRACT: Among other risk factors, renal disease and ethnicity have been associated with diabetic lower extremity amputation (LEA) in Western populations. However, little is known about risk factors for LEA among Asian patients. The objective was to assess the proportion of hospitalized patients with diabetes who have a LEA among all hospital patients with diabetes mellitus (DM) and to investigate risk factors for diabetic LEA (especially renal disease and ethnicity) using hospital discharge database. A retrospective study of hospital discharge database (2004-2009) was performed to identify patients with DM, LEA and renal disease using the International Statistical Classification of Diseases and Related Health Problems, Ninth Revision, Australian Modification codes. Of 44 917 hospitalized patients with DM during the 6 years, 7312 (16.3%) patients had renal disease, and 1457 (3.2%) patients had LEA. DM patients with renal disease had significant higher rates of LEA (7.1%) compared to DM patients without renal disease (2.5%, P < .001). The differences were present for foot (2.7% vs. 1.2%), ankle or leg (2.8% vs. 0.9%), and knee or above amputation (1.6% vs. 0.4%, all P<.001). Malays had the highest rate of diabetic LEA (5.1%), followed by Indians (3.0%), Chinese (3.0%), and others (2.3%, P < .001). In logistic regression analyses, renal disease and ethnicity were significant predictors of diabetic LEA (renal disease: odds ratio 3.2, 95% confidence interval 2.8-3.6; ethnicity: odds ratio, 1.6, Malays vs. Chinese, P < .001; 1.0, Indians vs. Chinese, P = .784) after adjustment for age, gender, and year of discharge. DM patients with renal disease and Malay ethnicity had higher rates of LEA in this Asian patient population. Malay patients with DM and diabetic patients with renal disease should be considered as high-risk groups for LEA and therefore screened and monitored systematically.Journal of diabetes and its complications 11/2011; 25(6):382-6. · 2.11 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Large-scale epidemiological studies of primary biliary cirrhosis (PBC) have been hindered by difficulties in case ascertainment. To develop coding algorithms for identifying PBC patients using administrative data--a widely available data source. Population-based administrative databases were used to identify patients with a diagnosis code for PBC from 1994 to 2002. Coding algorithms for confirmed PBC (two or more of antimitochondrial antibody positivity, cholestatic liver biochemistry and/or compatible liver histology) were derived using chart abstraction data as the reference. Patients with a recorded PBC diagnosis but insufficient confirmatory data were classified as 'suspected PBC'. Of 189 potential PBC cases, 119 (60%) had confirmed PBC and 28 (14%) had suspected PBC. The optimal algorithm including two or more uses of a PBC code had a sensitivity of 94% (95% CI 71% to 100%) and positive predictive values of 73% (95% CI 61% to 75%) for confirmed PBC, and 89% (95% CI 82% to 94%) for confirmed or suspected PBC. Sensitivity analyses revealed greater accuracy among women, and with the use of multiple data sources and one or more years of data. Inclusion of diagnosis codes for conditions frequently misclassified as PBC did not improve algorithm performance. Administrative databases can reliably identify patients with PBC and may facilitate epidemiological investigations of this condition.Canadian journal of gastroenterology = Journal canadien de gastroenterologie 03/2010; 24(3):175-82. · 1.97 Impact Factor