Appendices - MHSDS
MHSDS LDA provisional and final data
Published data which are labelled as ‘provisional’ are produced from primary monthly submissions, (i.e. the first submission of each month’s data). Provisional statistics are provided in some instances in order to give the timeliest information possible for some key measures. Provisional statistics are indicative and subject to change.
Providers are able to make an optional 'refresh' submission at the time of the next month's primary submission. Statistics which have been published under the label ‘final’ are a combination of both the primary and refresh submissions. Where a refresh submission is made, the primary submission is not used. If a refresh submission is not made then the primary submission is used. This ensures the most complete and latest picture for the MHSDS is made available. Once final statistics are published then any provisional statistics for the same period should no longer be used.
Interpreting inpatient statistics
A group of measures about hospital spells and ward stays open at the end of the month can be used to provide a detailed picture of inpatient caseload at National, Provider and CCG of registration or residence level. Each measure has a unique identifier (referenced in the metadata file), to ensure that users know what is represented by each measure.
This national picture is incomplete without full submissions from all providers of inpatient services and gaps in the provision of information from independent sector providers will also affect the figures at subnational level to varying degrees. See the data quality section of this report for more details. These are experimental statistics and should be used with caution, using all the other available information about data quality to aid interpretation.
Statistical disclosure control
To prevent the release of disclosive information, for the majority of statistics within this publication any numbers less than five (including zero) are replaced by a “*” symbol. Please note for tables 15-18 of the data tables, this also denotes where a submission has been made but data in the relevant table has not been provided. All other numbers are rounded to the nearest five. Calculated values in the data files are based on unrounded numbers but rounded to the nearest whole percent to prevent backward calculation. Where a value is suppressed, the corresponding percentage is also suppressed. This approach prevents identification of a person through cross referencing different publications.
In line with the NHS Anonymisation Standard, since the total population of inpatients identified as having a learning disability and/or autism is less than 10,000, this disclosure control has been applied to naional figures as well as sub-national figures.
Last edited: 14 July 2021 9:50 am