|
|
 |
How To... Review Data Quality - Periodic Data
Quality
Record Volume Update Section
Total Records Submitted
The Total records submitted is a description of the number and percent
of records on the transaction file that were submitted to CDC for analysis
by the month and year of the child’s initial visit date for PedNSS and the
mother’s expected date of delivery (EDD) or last menstrual period (LMP) if EDD is not available or
infant’s birth date for PNSS. Individual months are listed only if they
comprise > 5% of the total records on the file. The cumulative percent of
records should always be 100%.
Rejected Records
Duplicate Records
A duplicate record is a record submitted from the contributor that is
mostly or entirely identical to another record in the same transaction
file. When duplicate records are identified the first reported record is
retained. There are four types of duplicate errors:
- All fields match is an exact match of all fields on one
record to all fields on another record. This can be caused when the
contributor selects the same record more than once to be included in the
CDC transaction file. Duplicate records where all fields match are more
likely to occur when the contributor extracts three separate monthly
files and then combines them into a quarterly transaction file. The
extract program may not create duplicate records within a month’s file,
however it may extract the same record on the next month’s file.
- Key fields match is an exact match of key fields. Key fields
are defined differently for PedNSS and PNSS; however, they both include fields
such as state, substate, date of visit, and individual identifier. For
example, in PNSS, duplicate records with key field matches occur when
two prenatal only records are selected and extracted for the same woman.
- Select fields is a match of selected fields. The selected
fields are defined differently for PedNSS and PNSS.
For
PedNSS, the select fields include state, substate, individual ID, date
of birth, height (English/Metric), and weight (English/Metric) on more
than 1% of records. A match of PedNSS select fields indicates that both
records are for the same child, with the same height and weight data,
but different dates of visit.
For PNSS the select fields include: state, substate, woman’s ID and
certain combinations of infant’s date of birth, LMP and EDD, depending
on record type. A match of PNSS select fields indicates that both
records are for the same woman and the same pregnancy.
- Multiple births are PNSS records that have key fields that
match but the infant sequence number does not match. These duplicate
records, although listed in the report, are NOT a data quality problem.
This error occurs because in the previous PNSS system, a record
is submitted for each infant in a multiple birth. However, in the new
PNSS system, only one of these records is maintained on the master file.
Critical Errors
A critical error is a field that has missing or invalid data for the data
that are critical for data analysis. Critical fields are defined differently
for PedNSS and PNSS.
- For PedNSS, the critical fields include: state, substate, date of
visit, individual identifier, date of birth and the sex code. Records
are also rejected if the date of visit is not after the date of birth.
- For PNSS, the critical fields include: state, substate, initial
visit date, woman’s individual identifier, and completion code.
- For PNSS, records are also rejected if the infant’s date of birth and
date of LMP (last menstrual period) or EDD are all missing or miscoded.
To reduce the number of records rejected for critical errors in both
PedNSS and PNSS, ensure that all critical fields are required to be
collected by clinic staff. For PNSS, ensure that a woman’s postpartum
visit date is copied to the initial visit date field on postpartum only
records.
Review PedNSS edit criteria for Duplicate
Records and Critical Errors.
Review PNSS edit criteria for Duplicate
Records and Critical Errors.
Records Replaced on Master File
Records Replaced on Master File indicates records on the transaction
file that will replace records already on the master file. This occurs
when all key fields match on both transaction and master file records.
If the data quality report indicates that records were replaced on your
master file:
- verify that this transaction file is a replacement file to correct
records on the master file
- verify that the number and % of records replaced is as expected, and
- if you are unable to verify the records that were replaced, then
correct the extract and resubmit the file.
Records Added to the Master File
Records added to the master file includes the number and percent of
records from the transaction file that were added to the master file. For
PNSS, this includes information on the proportion of complete, prenatal
only, and postpartum only records on the file. A lower or higher expected
proportion of specific record types can indicate problems with assignment
of completion codes. Generally the proportion of complete records should
not be less than 50% of records and the proportion of either prenatal only
or postpartum only records should not be less than 5% of records. These
criterion are based on national PNSS that show about 65% of records are
complete records 12% of records are prenatal only records, and 23% of
records are postpartum only records. The Completion Code and Record
Linkage Errors section of this report provides more information on the
types of errors that result in incorrect assignment of a completion code
for a record.
Total % of Transaction Records Added to the Master File
Total % of transaction records added to the master file is on the last
line of the report summarizes the percent of records from the transaction
file that were added to the master file.
Updated Number of PNSS Master File Records
Updated Number of PNSS Master File Records lists the prior years and
current year-to-date monthly master file record volume including records
from this transaction file that were added to the master file. Some
problems that this portion of the PNSS data quality report can help
identify are monthly record volume that is not similar to other months in
the
same year, and current year monthly record volumes that are not similar to
prior year monthly record volumes.
If record volume is inconsistent:
- lower than other months in the same year, confirm that all quarterly
files related to these months were submitted
- higher than other months in the same year, confirm if mass changes
were made to clinic or individual identifications creating duplicate
records for participants that cannot be identified as duplicates.
- lower or higher record volume, confirm if a change in the program
that extracts the data from the computer information system into the
transaction file can explain the difference
- current and prior year monthly record volume are not similar, confirm
if shifts in program caseload can explain the difference
back to top
|
|
