Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “data quality”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources
PubMed1989

Assessing data quality: a computerized approach.

With the growing reliance on large health care data bases, the need to verify data quality increases as well. Because of the considerable costs involved in checks using primary data collection, a computerized methodology for performing such checks is suggested. The technique seems appropriate for any situation where two data collection systems (i.e. hospital discharge abstracts and physician claims for payment) relate to the same event, such as a patient's hospitalization. After reviewing other approaches, this paper suggests linking physician claims for performing particular surgical procedures with hospital discharge abstracts for the stay in which the surgery took place. Physician and hospital data for adults age 25 and over in Manitoba from 1 April, 1979 to 31 March, 1984 were used to address the questions: 1. How well can the two data sets be linked? 2. Given linkage of the two data sets, how much agreement is there as to procedure and diagnosis? Linkage between hospital and physician data was excellent (over 95%) for 5 out of 11 surgical procedures (hysterectomy, prostatectomy, total hip replacement, coronary artery bypass surgery, and heart valve replacement); there was over 90% perfect agreement for three other procedures (cholecystectomy, cataract surgery and total knee replacement). Problems with matching the Manitoba Health Services Commission tariffs (on physician claims) with ICD-9-CM operation codes (on hospital data) led to only 77% perfect agreement for vascular surgery and 84% for gallbladder and biliary tract operations other than cholecystectomy; over 10% of the cases linked on surgeon and date but not on the designated procedures.(ABSTRACT TRUNCATED AT 250 WORDS)

Data Collection↗
PubMed1993

Multi method approach to the assessment of data quality in the Finnish Medical Birth Registry.

OBJECTIVE: To assess comprehensively the validity of the data in the Finnish Medical Birth Registry (MBR) by the combined use of several controls and internal analysis of the data. DESIGN: The MBR data were individually linked to a medical record sample (n = 775) and to all perinatal death certificates in 1987. The data were also compared with annual hospital statistics. The distributions of birth weights and gestational ages were examined. SUBJECTS: All stillborn and liveborn babies registered in the MBR in 1987 (n = 59,370). SETTING: The nationwide MBR data were compared with medical records from one third of the Finnish hospitals, with statistics for all hospitals, and with nationwide cause of death registry data. MEASUREMENTS AND MAIN RESULTS: With regard to most variables, the data quality was good or satisfactory (agreement with medical records 95% or more). Allowing for minor deviations in variables with continuous scales improved the agreement rates further. Explanations could be deduced for items with poor agreement values. For most variables, the amount of missing data was less than 1%. With the exception of caesarean sections, medical procedures were registered in only 30 to 72% of the cases, and the proportion varied strongly between the hospitals. Common diagnoses (32 to 86%) and primary causes of death (59 to 78%) were also poorly recorded. CONCLUSIONS: Combined use of several control materials and internal analyses was successful in investigating the whole data content. The data in the MBR were generally valid but diagnoses and most data on medical procedures were not of sufficiently good quality.

Birth Certificates↗
PubMedDate not supplied

Data quality of a computerized medical birth registry.

Processed by a computerized medical birth registry system, the birth records of 20,103 deliveries, from February 1992 to February 1993, were digitized with medium registry. From 1 January to 28 February 1993, the original records (n = 2840 cases) of all 10 collaborative hospitals were requested for assessment of data quality. Thirty-six items were scored, data of poor quality was found in eight; acceptable quality in four; and good quality in 28. The feasibility of data transfer by floppy disc and per modem was evaluated. This registry system had effectively shortened data processing time and improved mutual feedback between the data centre and the delivery units. Errors resulting from technical faults originating in the preparation of data for computerization at hospital level could be effectively reduced. The validity of diagnosis remained as the major source of errors.

Birth Certificates↗
PubMed1979

An investigation into the distribution of radial immunodiffusion quality control data.

Quality control data from routine radial immunodiffusion assays for IgG, IgA, IgM, C3, C4 and alpha1-antitrypsin were tested by the Kolmogorov--Smirnov procedure for gaussian distribution. All but alpha1-antitrypsin were nongaussian in type. Further analysis of these date by plotting on log-normal probability paper showed them to have a log-normal distribution. Treatment of the date by either gaussian or nonparametric statistical methods produced little difference in confidence limits. It does not appear necessary to use nonparametric methods to calculate confidence limits from quality control data for the procedures studied.

Complement C3↗
PubMed1995

Data quality after restructuring a national medical registry.

The validity of the 1991 Finnish Medical Birth Registry data was assessed, with special emphasis on the effects of changes made to the data collection form in 1990. Data abstracted from medical records for all births occurring in 49 hospitals during a five-day sample period (n = 865) were compared to the register information. Good or satisfactory validity was found for 32 of 33 variables, when minor error was tolerated in variables with continuous scales. For diagnoses and procedures, recorded in check-box format, satisfactory validity was found for 10 of 45 variables. Validity could not be assessed for 18 variables because of insufficient number of cases (13 items) or definition problems (5 items). When the results were compared to a 1987 data quality study, many of the variables that had been changed to the check-box format showed improvement in validity. In addition, in some cases a small change in question alternatives or instructions caused a noticeable change in validity.

Abstracting and Indexing↗
PubMed1995

[The national patient registry. Evaluation of data quality].

The Danish National Patient Register, which includes information on all patients admitted to hospitals, has been evaluated as concerns the quality of the data included. The material examined consisted of a representative sample of 1094 patients from departments all over the country (gynaecology and obstetrics, medicine, surgery and paediatrics). Recoding of data, clinical as well as administrative, based on copies of the case records from the hospitals was carried out by two clinically working physicians (registrars). For the administrative data e.g. length of stay, satisfactory concordance was found. The validity of clinical information depended on clinical speciality and degree of diagnostic specificity. Based on the international classification the agreement on the three digit diagnostic level was better than on the five digit diagnostic level. For surgery the agreement was better than for medicine. The agreement between the diagnostic information (primary diagnosis) and the recoder in choosing primary diagnosis varied from 66-83 percent on the five digit level and between 73-89 percent on the three digit diagnostic level. If cases where the diagnosis in the registry could be regarded as an acceptable alternative were included, the agreement between the registry and recoding was 75-90%. In a subsample of the material double coding by the two coders was carried out and it was remarkable that, taken as a whole, the degree of agreement between the two coders was of the same size as between recoder and the registry. It is anticipated, however, that introduction of ICD-10 with more clear-cut rules for choice of primary diagnosis in morbidity coding will contribute to better validity and consequently improved hospital statistics.

Denmark↗
PubMed1989

Ensuring data quality in a multicenter clinical trial: remote site data entry, central coordination and feedback.

In an ongoing multicenter clinical trial, "Treatment Strategies in Schizophrenia," the five participating sites have the capacity to perform a variety of tasks or study functions independently. These tasks include (a) verification of diagnostic eligibility through the use of computerized decision algorithms; (b) assignment of patients to treatment based on prognostic indicators using a computerized randomization algorithm; (c) entry of data into a microcomputer using a clinical trial data management system that performs simple range and missing data item checks; and (d) regular transfer of all data to the central coordinating team. The clinical trial data management system employed allows for both independent site functioning and assurance of consistency across sites. The integration of a variety of software outside the main data management system provides the central coordinators with the tools to monitor critical data as it is collected, as well as the capacity to assess the flow, quality, and uniformity of the ongoing trial.

Clinical Trials as Topic↗
PubMed1996

Analyses of data quality in registries concerning diabetes mellitus--a comparison between a population based hospital discharge and an insulin prescription registry.

To evaluate the data quality in the Danish National Registry of Patients (DNRP) and the Prescription Registry in the country of Northern Jutland (487,000 inhabitants) concerning insulin dependent diabetes mellitus (IDDM) and insulin treated diabetes mellitus, a comparison between data in the two registries was made. From the Regional Hospital Registry in the County of Northern Jutland, containing discharge diagnoses from all admissions to hospitals in the county, we identified all patients with the IDDM diagnosis between 1987 and 1993. From the Regional Prescription Registry all insulin prescriptions taken up at pharmacies in the county in 1993 were identified. All persons were identified by their individual identification number (CPR-number), and a record linkage between the two data sources was made. The predictive value of an IDDM-registration in the DNRP was 96% and the corresponding completeness 91%. In the Prescription Registry the completeness was 96%. Both registries seem to be valuable study bases for epidemiological research in diabetes mellitus.

Denmark↗
PubMed1994

Data quality in a DRG-based information system.

The aim of this study initiated in May 1990 was to evaluate the quality of the medical data collected from the main hospital of the "Hospices Civils de Lyon", Edouard Herriot Hospital. We studied a random sample of 593 discharge abstracts from 12 wards of the hospital. Quality control was performed by checking multi-hospitalized patients' personal data, checking that each discharge abstract was exhaustive, examining the quality of abstracting, studying diagnoses and medical procedures coding, and checking data entry. Assessment of personal data showed a 4.4% error rate. It was mainly accounted for by spelling mistakes in surnames and first names, and mistakes in dates of birth. The quality of a discharge abstract was estimated according to the two purposes of the medical information system: description of hospital morbidity per patient and Diagnosis Related Group's case mix. Error rates in discharge abstracts were expressed in two ways: an overall rate for errors of concordance between Discharge Abstracts and Medical Records, and a specific rate for errors modifying classification in Diagnosis Related Groups (DRG). For abstracting medical information, these error rates were 11.5% (SE +/- 2.2) and 7.5% (SE +/- 1.9) respectively. For coding diagnoses and procedures, they were 11.4% (SE +/- 1.5) and 1.3% (SE +/- 0.5) respectively. For data entry on the computerized data base, the error rate was 2% (SE +/- 0.5) and 0.2% (SE +/- 0.05). Quality control must be performed regularly because it demonstrates the degree of participation from health care teams and the coherence of the database.(ABSTRACT TRUNCATED AT 250 WORDS)

Data Collection↗
PubMed1986

Design and data quality of a mixed longitudinal study to elucidate the role of dietary calcium and phosphorus on bone mineralization in pre-, peri-, and postmenopausal women.

The study design and data quality control of an ongoing study (10 yr duration) in a few hundred women are presented. Good variables with respect to their longitudinal usefulness are: body weight, body height, and span-width. Reasonable variables are the bone parameters of the radius (BMC, BW, and BMC/BW). Poor variables are: dietary calcium and phosphorus intake, dietary calcium-to-phosphorus ratio, urinary calcium-to-creatinine ratio, urinary sodium-to-creatinine ratio, hematocrit, serum alkaline phosphatase activity, serum gamma-GT activity, and serum parathyroid-hormone concentration. Bad variables are: urinary phosphorus-to-creatinine ratio, urinary hydroxyproline-to-creatinine ratio, creatinine clearance, hemoglobin, MCHC, serum calcium, serum ionized calcium, serum phosphorus, serum total protein, serum albumin, and serum creatinine. In conclusion, it is possible to relate bone loss to food intake and to changes in anthropometric variables on an individual basis. However, quantification of the metabolic process is not possible.

Anthropometry↗
PubMed1985

Data quality and DRGs: an assessment of the reliability of federal beneficiary discharge data in selected Manhattan hospitals.

New York County Health Services Review Organization (NYCHSRO), the physicians' professional standards review organization of Manhattan, examined whether diagnostic coding errors identified in Manhattan hospitals would affect reimbursement under a diagnostic-related group (DRG) method of financing inpatient services. A sampling of 1,027 Medicare and Medicaid cases representing discharges from 18 Manhattan hospitals during 1982 and 1983 revealed incorrect DRG assignment for 17.5% of patient record abstracts, but these appear to have been unsystematic rather than deliberate errors. The difference between estimated reimbursement based on original and reabstracted records was not statistically significant either in the aggregate or for specific hospitals. It is emphasized that while New York State's Prospective Hospital Reimbursement Methodology (in effect during the study period) is not solely dependent upon DRG's case-mix is one of several factors used to make adjustments to existing per diem rates. A key recommendation is that hospitals conduct internal monitorings with all involved departments to improve the quality of the data abstracting process.

Diagnosis-Related Groups↗
PubMed1990

Multicentric quality assurance in cardiac surgery. QUADRA study of the German Society for Thoracic and Cardiovascular Surgery (QUADRA: quality data retrospective analysis).

A method for initiating quality assurance in cardiac surgery was developed multicentrically by a commission of the German Society for Thoracic and Cardiovascular Surgery (QUADRA Study). To appraise the quality of cardiosurgical action, variables were compiled from the preoperative, intraoperative, and postoperative treatment course. The data collection was carried out at the same time as treatment. On the basis of unicentric data profiles, multicentric hospital profiles, and problem profiles, a quality comparison could be carried out and the variability of cardiosurgical action which may occasion interventions could be identified. A reduction of perioperative blood consumption during the study period could be observed in four out of five hospitals as the first result. The data collection also revealed epidemiological features. On average, women were older than men at the time of the heart-valve and coronary operations. To ensure data validity and the organization of quality assurance, a documentation assistant and a specially trained physician are necessary at every cardiovascular surgery center. The multicentric external comparison is indispensable and must be carried out by means of a central data collection, for which intrumental and staff capacities are also to be provided. With modern methods of data processing, an additional and new approach to the improvement of quality in cardiac surgery can thus be made.

Blood Transfusion↗
PubMed1988

Problems associated with collecting drinking water quality data for community studies: a case example, Fresno County, California.

This paper discusses methodology in developing exposure data for the water supply contaminant dibromochloropropane (DBCP) in Fresno County, California. There are 532 drinking water systems (49 large and 483 small) within Fresno County plus 14,000 private wells. We determined the number of wells per system, the output per well, and the population served by each system. The task of deriving water quality estimates for each census tract was complicated by the fact that a single census tract can be served by more than one system; each system usually has more than one well; and a single well can have several episodes of testing for various contaminants. We calculated a series of weighted averages for concentrations of DBCP, arsenic, and nitrates for each census tract, using water production figures for each well as the weighting factor. Water quality data were derived from a total of 14,861 laboratory reports, although the majority did not report on all contaminants. Mean DBCP levels ranged from 0.0041 ppb to 5.7543 ppb among the census tracts. We found no correlation between DBCP levels per census tract compared to either arsenic or nitrates. We believe that we made as complete an exposure assessment as practically feasible.

Arsenic↗
PubMed1993

Markers of data quality in computer audit: the Manchester Orthopaedic Database.

This study investigates the efficiency of the Manchester Orthopaedic Database (MOD), a computer software package for record collection and audit. Data is entered into the system in the form of diagnostic, operative and complication keywords. We have calculated the completeness, accuracy and quality (completeness x accuracy) of keyword data in the MOD in two departments of orthopaedics (Departments A and B). In each department, 100 sets of inpatient notes were reviewed. Department B obtained results which were significantly better than those in A at the 5% level. We attribute this to the presence of a systems coordinator to motivate and organise the team for audit. Senior and junior staff did not differ significantly with respect to completeness, accuracy and quality measures, but locum junior staff recorded data with a quality of 0%. Statistically, the biggest difference between the departments was the quality of operation keywords. Sample sizes were too small to permit effective statistical comparisons between the quality of complication keywords. In both departments, however, the poorest quality data was seen in complication keywords. The low complication keyword completeness contributed to this; on average, the true complication rate (39%) was twice the recorded complication rate (17%). In the recent Royal College of Surgeons of England Confidential Comparative Audit, the recorded complication rate was 4.7%. In the light of the above findings, we suggest that the true complication rate of the RCS CCA should approach 9%.

Data Collection↗
PubMed1991

Case finding, data quality aspects and comparability of myocardial infarction registers: results of a south German register study.

The population-based Augsburg Coronary Event Register (330,000 residents, age 25-74 years) has registered a total of 1012 cases of acute myocardial infarction (AMI) in 1985 and 1021 AMI in 1986 and categorized them on the basis of the current WHO diagnostic algorithm for AMI. The register is designed for longitudinal comparisons of annual AMI risk (incidence, attack rate, death rate), and the risk to the AMI patients themselves (28-day case fatality). The methodology and specific issues encountered during registration and data evaluation are described. With an estimated 95% completeness of case finding, the quality control data review which the register conducts annually shows a consistency of specific data structures which indicate stable case finding and validation procedures. However, local conditions which affect case finding and data completeness per case are responsible for the creation of subsets of AMI which are in turn distinguished by differences in diagnostic category structures. With regard to the study objectives, the differences among subsets appear to have the least effect on rate calculations if DEFINITE and POSSIBLE AMI are combined. The implications of methodological variations and subset differences within and across registers on annual rate calculations and result comparisons are discussed.

Adult↗