INTRODUCTION: The value of a data registry for research, benchmarking, and quality improvement activities depends on the underlying quality of the data. This pilot study was conducted to design and implement a validation process assessing data quality in the American Association of Birth Centers' online data registry, the Uniform Data Set (UDS). METHODS: Site visits were conducted in 5 midwifery practices attending births in freestanding birth centers and hospitals to compare data from health records with data in the registry. Practices meeting the inclusion criteria and representative of the overall population of midwifery practices contributing data to the UDS were audited. Between 2% and 5% of each practice's total UDS records were randomly selected and then compared with their matched health records for 29 key variables from the 189 variables in the UDS. RESULTS: A total of 3966 variables were reviewed for 152 records. There were 126 records for which complete maternity care was provided; thus, all 4 parts of the UDS were expected. There were 26 records for women who left care during pregnancy; thus, only parts 1 and 2 of the UDS were expected. Quality of the UDS data was evaluated, with the health records serving as the criterion standard, and the 2 sources were found to be consistent for 97.1% of the variables. DISCUSSION: Results from the study suggest that the data registry can provide valid data for use in research describing the process and outcomes of midwifery care and for benchmarking and quality improvement activities.
INTRODUCTION: The value of a data registry for research, benchmarking, and quality improvement activities depends on the underlying quality of the data. This pilot study was conducted to design and implement a validation process assessing data quality in the American Association of Birth Centers' online data registry, the Uniform Data Set (UDS). METHODS: Site visits were conducted in 5 midwifery practices attending births in freestanding birth centers and hospitals to compare data from health records with data in the registry. Practices meeting the inclusion criteria and representative of the overall population of midwifery practices contributing data to the UDS were audited. Between 2% and 5% of each practice's total UDS records were randomly selected and then compared with their matched health records for 29 key variables from the 189 variables in the UDS. RESULTS: A total of 3966 variables were reviewed for 152 records. There were 126 records for which complete maternity care was provided; thus, all 4 parts of the UDS were expected. There were 26 records for women who left care during pregnancy; thus, only parts 1 and 2 of the UDS were expected. Quality of the UDS data was evaluated, with the health records serving as the criterion standard, and the 2 sources were found to be consistent for 97.1% of the variables. DISCUSSION: Results from the study suggest that the data registry can provide valid data for use in research describing the process and outcomes of midwifery care and for benchmarking and quality improvement activities.
Authors: Sarah Toler; Susan Stapleton; Kim Kertsburg; Tiffany J Callahan; Marie Hastings-Tolsma Journal: Midwifery Date: 2018-04-03 Impact factor: 2.372