Data quality archive

This is an archive page of the monthly DQMI publications.

Links to previous DQMI publications, the associated methodology document and the interactive report can be found below.

Diagnostic codes for conditions that do not remit should not be lost from patient records over time - this "diagnostic persistence" is a measure of data quality.

The below factsheet highlights the proportion of patient episodes that include, or omit, a code for one of five clinical diagnoses where there is reason to expect that a code should continue to be recorded.

We reviewed diagnostic codes for all episodes of hospital care over a three month period, and then looked back to compare whether a relevant code was previously recorded.

February 2018

DQMI Comorbidity Outputs 17-18 Q2 Jul to Sep v1.0 Final [859kb](Opens in a new window)

February Diagnostic Persistence Infographic_FINAL v0.1 [518kb](Opens in a new window)

November 2017

DQMI Comorbidity Outputs Q1 Apr to Jun 17-18 v1.0 Final [209kb](Opens in a new window)

November Diagnostic Persistence Infographic_FINAL [519kb](Opens in a new window)

August 2017

Comorbidities analysis raw data [166kb](Opens in a new window) 

Comorbidities analysis Infographic [517kb](Opens in a new window)

Note: The analysis is to show diagnostic coding for conditions that do not remit should not be lost from patient records over time - this "diagnostic persistence" is a measure of data quality, it is not to show poor diagnosis or treatment of conditions.

The two pieces of code below are used by NHS Digital’s MHSDS team to generate the raw data provided to the DQ team that we then process to produce the DQMI scores and associated reports.

The code is run against the MHSDS data after the data has been submitted and processed through the DPS Portal. It will, therefore, require modification locally to be run against the pre-submission data. 

Whilst we cannot assist in this modification, we would be willing to work with a small number of Mental Health Services providers to understand the technical steps required to use the code locally and to then share this learning alongside the code.

Create validity measures SQL

Create integrity measures SQL