NHS QUESTIONS

How Can You Check Your Mental Health Services Data Before Submitting It?

Before you submit your Mental Health Services Data Set (MHSDS) return in England, compare the extract with your service records for the same period. Follow up gaps, then check the NHS England validation reports after submission.

IntelliFront BI can give your reporting team a shared view of local counts and exceptions while you prepare the return.

  
Start with the gap

Compare the extract with the activity you expected

Imagine a reporting analyst preparing the monthly return for a community mental health service. One team’s activity looks low. Its local report shows 64 contacts the analyst expects to see in the extract. The extract contains 58.

Before asking anyone to re-enter records, the analyst checks that both reports cover the same team and reporting period. They also check which contacts each report includes. A count from a diary or activity report can mean something different from a count of records eligible for submission.

Write down what you counted

Record the team, period and inclusion rules alongside each total. The next person looking at the gap should be able to reproduce both figures.

  
Find the cause

Follow the missing records back to the source

Once both counts use the same scope, the analyst compares the eligible records with the extract in the team’s approved reporting environment. In this example, all six missing contacts were recorded after the service began using a new team code. The extract’s filter includes the old team code and leaves out the new one.

The analyst asks the owner of the extract to update the approved team-code filter, then reruns the comparison after the fix. A gap can also come from late entry or a record that falls outside the return’s rules. Find the cause before changing a filter or the underlying record. Check submission rules against the version of the NHS England guidance and technical specification that applies to your return.

Give each exception an owner

An extract filter issue goes to the person who maintains the extract. A missing source record goes to the team responsible for recording the activity.

  
Make the check repeatable

Give the reporting team a view they can use each month

If a workbook already makes the comparison clear and the team can review it together, it may be all you need. Keep the check and its extract reference together in your approved workspace so another analyst can retrace the decision.

When several people need the comparison each month, IntelliFront BI can show counts from an approved reporting dataset. One row per team could show the number expected, the number in the extract and the difference. Give each view a reporting period and an extract date or version. Use summary counts in the shared view and keep record-level investigation in your approved environment.

Checks to run on the prepared extract
  • Confirm the provider, reporting period and extract version.
  • Compare counts by service or team using the same inclusion rules.
  • Check for missing required values, duplicates and links between records where the specification calls for them.
  • Record who is following up each exception and rerun the checks when the extract changes.

Watch for a quiet team

A total for the whole service can look plausible even when one team is missing. Review the counts by team or service before you sign off the extract locally.

  
Read the response

Check the NHS England reports after the upload

The local comparison tells you whether the extract contains the activity you expected. Upload through your usual NHS England submission process and review the reports it returns. NHS England explains that an accepted submission can still contain rejected records. The Validation file contains rejection messages, and the Data Quality file contains warning messages. NHS England’s guide to validation and data quality reports.

Match each message to the file you actually uploaded. Check file-level rejections first. For an individual warning or rejection, look up its code in the Technical Output Specification for the return. If several rejected records depend on the same source record, investigate that record first. The earlier gap of six was a local extract problem, so the analyst checks that fix separately from the portal’s response.

Keep the upload and its reports together

Record which extract each response belongs to. That way the team can tell whether a message came from the file they corrected or from an earlier upload.

  
Your next step

Try the check with one service and one month

In the example, the corrected extract now includes all 64 contacts expected for that team. The analyst can explain the original gap and knows what to check when the NHS England reports come back.

Start with one service and one count you can reconcile. If you’re evaluating IntelliFront BI, use the guide to plan how you’ll show the comparison. The NHS pharmacy reporting evaluation gives an example of checking a dashboard figure against its source during a trial.

Download the IntelliFront BI Internal Evaluation Guide

Contributors:

Christian Ofori-Boateng

Christian Ofori-Boateng

CEO, ChristianSteven Software