User guide for Feasibility Self-Service tool

The Feasibility Self-service tool contains 4 datasets. These are:

• Personal Demographics Service (updated daily)
• Hospital Episode Statistics (HES) - Admitted Patient Care (updated monthly)
• Hospital Episode Statistics (HES) - Outpatient (updated monthly)
• Medicines dispensed in primary care (updated monthly)

Only specific elements relevant to feasibility assessments, such as diagnoses and procedures, are displayed.

The tool does not contain GP codes and related data.

Currently, it’s not possible to distinguish which integrated care board (ICB) or hospital provided a specific diagnosis. For example, if a patient is seen at different hospitals for different diagnoses. The data is linked to the ICB associated with the patient's GP practice or postcode. Whether results are displayed by ICB, region, or partial postcode, the data is always based on the GP practice's location.

Once you are logged in to SDE you will see a list of your agreements. Select the agreement you want to access.

On the next page you need to select the ‘Launch the virtual desktop’ option.

The virtual desktop will load, note this can take a couple of minutes. SDE is allocating resources to set up a virtual computer in the cloud for your session.

The virtual desktop will load a Windows desktop, here you need to double click, or right click and select open on the icon for Feasibility Self Service.

A Google Chrome browser will then open asking that you sign in, however this step should be skipped by selecting the ‘Don’t sign in’ option.

Within the Chrome browser you will be asked to log in again to be able to access the Feasibility Self Service. The credentials to enter here are your SDE credentials that you used during the SDE log in process.

As before with the SDE log in, you will also be required to go through the two-factor authentication step.

Once log in is successful you will be redirected to the Feasibility Self-Service tool.

Note, you will automatically be logged out after a period of inactivity.

This video:

• provides an overview of the tool
• describes how to enter demographic criteria
• describes how to enter inclusion and exclusion criteria

Our welcome page

Our welcome page provides information about Feasibility Self-Service. You will see a diagram explaining where we get our data from, and more information about the data and codes used within the service.

Clicking on the ‘Start Search’ button will take you directly to the Search Page where you can run your query.

Our search page

Our search page is where you can run queries to evaluate the feasibility of your study.

Our search is broken down into a 5 step process:

1. Demographic Criteria (Age range, Gender, Ethnicity, Region and GP Practice) 
2. Clinical Inclusions – Diagnoses and Procedures 
3. Clinical Inclusions – Medication 
4. Clinical Exclusions – Diagnoses and Procedures 
5. Clinical Exclusions – Medication 

The guidance on this page covers some of the functionality on the Search Page in more detail.

Here is an example of the first page:

How to run a query

Demographics criteria

All filters are optional and you can use as many as you like.

Using the ‘More info’ icons

If you click the ‘More info’ links a pop-up box will appear.

The box contains metadata and offers further detail on the fields within each section of the search page.

Finding out more about the codes (Clinical Exclusions section)

Find out more about the codes by typing into your browser the links listed in the Clinical Exclusions section.

SDE restrictions mean you won’t be able to click directly on the links on the Search Page.

If you don’t want to type in the links, you can access them here:

ICD10 codes
OPCS codes
BNF codes

Inputting inclusions and exclusions

Inputting inclusions

You choose which information to filter in or out of your query.

Hospital diagnoses and procedures - If you choose to complete both of the diagnoses and procedure sections, your results will return:

• people who have had those diagnoses
• people who have had those procedure
• people who have had both those diagnoses and procedures

Medication - This section can be completed on its own, but it can also be used with a diagnoses and procedures search to locate people in this cohort taking particular medication, or it can be used on its own to include patients who are taking specific medication from your query.

Inputting exclusions

Hospital diagnoses and procedures - If you choose to complete both the diagnoses and procedure sections, your results will not return:

• people who have had those diagnoses
• people who have had those procedure
• people who have had both those diagnoses and procedures

Medication - This section can be completed alongside the diagnosis and procedures search to further refine your results or it can be used on its own to simply exclude patients who are taking specific medication from your query.

Using the ‘Reset’ buttons

The ‘Reset buttons are under the ‘More info’ links.

Clicking on ‘Reset’ will clear the information inputted within each section of the Search Page. It will also reset the ‘Yes/No’ button selections.

How to select multiple codes

If you want to select more than one code within a certain filter:

1. Select the first code you want (A02 in our example).
2. Hold down SHIFT on your keyboard and then click the last code you want to select (A18 in our example). This will then select all codes between the two.
3. Click on ‘Apply’ - this will add the codes to your query.

You can repeat this as many times as you like.

Checking your query

After entering all your criteria in the search page you will need to click the ‘Continue’ button:

This will allow you to review the criteria before running the query:

Getting your results

Once you are happy with the criteria selected, click the ‘Accept and run query’ button.

When the query has completed, you will automatically be taken to the results page.

Please note: it may take up to 5 minutes to return your results. 

How to use our results page

There are 2 ways of viewing the results of your query:

• table results
• map results - showing where in England your potential trial participants are located

Table results

This displays the data in columns and rows. The table is interactive, and you can choose what data is displayed in the table.

To change the data that is displayed in the columns and rows, use the drop-downs on the top-right hand side.

Rows can be changed from Integrated Care Board (ICB) to Region, GP Practice, Trust, Postcode Sector, or IMD- Deprivation Decile.

Columns can be changed from Age to Ethnicity, Gender, or IMD – Deprivation Decile.

Please note: To help protect the identity of patients, suppression has been applied to small numbers. This means that any count below 8 is set to * and any count 8 and above is rounded to the nearest increment of 5. Additionally, if the total count of the query results (of valid feasibility participants) is less than 100 the table and map view will not display. 

Small GP Practices with less than 2,500 valid feasibility participants have been grouped into an ‘Unknown’ category to help prevent identification of patients.

Postcode Sectors with less than 1,500 valid feasibility participants have been grouped into an ‘Unknown’ category to help prevent identification of patients.

Map results

The map shows where the greatest number of potential trial participants are located. The larger the dot, the higher the number of patients in that location.

The map can display dots to represent different geographical units; Integrated Care Board (ICB), Trust or Region.

Using the 'Search Parameters' page

The Search Parameters page can be reached from the results page.

It shows which criteria is currently filtering the results displayed on the Results Page.

The criteria are broken down into ‘Demographic filters’, ‘Clinical Inclusions’ and ‘Clinical Exclusions’ so that you can view each section individually.

Once you have run a query and viewed your results, this video explains how to export your result.

1. Click the 'Download' button at the top-right of the Feasibility Self-Service screen.
2. Select 'Crosstab' from the pop-up box.
3. Select either 'Map Results' or 'Table Results'.
4. Select the 'CSV' option to get your results in a .csv file.
5. Click the 'Download' button.
6. Open 'File explorer' from the Secure Data Environment (SDE) virtual desktop.
7. Navigate to the 'Downloads' folder.
8. Drag and drop your chosen file onto the data-out-storage icon on the SDE virtual desktop.
9. You will receive an email within 5 minutes which will contain a URL link for downloading your results.

For help with this service please contact the National Service Desk by email at [email protected].

Explaining the derivations and lineage for search fields in the Feasibility Self-Service tool.

PDS - Personal Demographics Service
HES - Hospital Episode Statistics
OP - Outpatient
APC - Admitted patient care

Age

The minimum/maximum age of inclusion. 

The age of patients is calculated from the date on which the patient was born or is officially deemed to have been born.

Data source = PDS (or HES OP if not available in PDS, then HES APC if not available in HES OP, then given an ‘unknown’ value if no date of birth (DoB) data available from any source).

Update method/frequency = Daily refresh

British National Formulary (BNF)

Patient records of dispensed medication, with the codes validated against Data Steward Service data to ensure that they exist, and for the retrieval of the latest description of the medication. The date used for BNF filtering is the first day of the month of the latest dispensing event during which data is submitted by pharmacists to NHSBSA for processing.

Data source = PCareMeds

Update method/frequency = Daily refresh, source is updated on approximately 20th day of every month.

Data available from = 1 April 2018

Ethnicity

The ethnicity of the patient, as specified by the patient. This is the classification used for the 2001 census. From April 2001 the codes were changed to represent the ethnic data categories as defined in the 2001 census. However, HES continued to accept the old codes as well as the new codes for the 2001-02 and 2002-03 data years.

Any patient who we don’t have HES data for will also be defaulted to having an ethnicity of ‘Unknown’. This means that if you were to run a query against the whole population, without including any clinical criteria, you will see a higher proportion of patients with an ‘Unknown’ ethnicity, as compared to when you run a query that includes any diagnosis or procedure criteria. This is due to patients who are registered with a GP who we don’t have any HES data for being included in the results.

Data source = HES OP if available or HES APC if HES OP not available. Otherwise, the ethnicity is logged as ‘Unknown'.

Update method/frequency = Daily refresh

International Classification of Diseases (ICD)

Patient records of condition diagnosis, with the codes validated against Data Steward Service data to ensure that they exist, and for the retrieval of the latest description of the diagnosis.

The date used for ICD filtering is when the patient was under the care of a particular consultant. If a patient has more than one episode in a spell, for each new episode there is a new value. However, the admission date which is copied to each new episode in a spell will remain unchanged and will be equal to the episode start date of the first episode in hospital.

Only the latest episode will be available from the database.

Data source = HES

Update method = Daily refresh, source data updated monthly by month end.

Data available from = 1 April 1989

Classification of Surgical Operations (OPCS)

Patient records of operating procedures, with the codes validated against Data Steward Service data to ensure that they exist, and for the retrieval of the latest description of the procedures. The date used for OPCS filtering is when the patient was under the care of a particular consultant, based on the scheduled appointment date. If a patient has more than one episode in a spell, for each new episode there is a new episode start date. Only the latest episode will be available from the database.

Data source = HES

Update method = Daily refresh, source data updated monthly by month end.

Data available from = 10 November 1992

Region

Region location is based on the patient's post code taken from PDS address data. We check the post code value against DSS Corporate post code reference data to ensure that it’s valid, then we retrieve the region value from same DSS Corporate reference table.

Data source = PDS

Update method = Daily refresh

Sex

Defines the sex of the patient. The classification is phenotypical rather than genotypical, meaning it does not provide codes for medical or scientific purposes.

Notes:

• National Code(Not Known) means that the sex of a person has not been recorded
• National Code(Not Specified) means indeterminate, meaning unable to be classified as either male or female.

Data source = PDS (or HES OP if not available in PDS, then HES APC if not available in HES OP, then given an ‘unknown’ value if no sex data available from any source).

Update method = Daily refresh

Patient location data (map view lat, lon)

Latitude and longitude values for patient locations based on different geographical units. This data is used by the map view to show patient counts from query results on the results page map view. 

Currently the results page map view handles the following geographical units:

• ICB
• Closest Trust
• Region

Data source = DSS Corporate Geometry

Update method = Daily refresh

Patient business rules

1. How do patients appear on the tool?

Patients who meet the DigiTrials valid feasibility patient criteria will have:

• patient is alive
• patient has a valid NHS number
• patient is registered with a GP in England
• patient's demographic record (PDS) is not flagged as 'Sensitive'
• patient's demographic record confidentiality code (PDS) is not flagged as 'I' for invalid

Furthermore, patients that are registered with a military or prison GP practice are also excluded from query results.

2. What demographic criteria is used in the tool?

All patient demographics must be the most recent values on record - most recent gender for example.

3. Which trusts are used to calculate the patient's nearest trust?

The calculation of the patient's nearest trust is based on factors such as:  

• active and open trusts only
• trusts that are not excluded as a high-level health geography exclusion (code Q99)
• trusts that are not excluded as a trust national grouping exclusion (code W00)
• benchmark group acute type trusts only - should return only secondary care trusts
• a mapping of all patient Lower Layer Super Output Area (LSOA) latitude/longitude values to all Acute Trust LSOA, latitude/longitude values
• individual patient's LSOA used to obtain the nearest trust value from the mapping
• trust LSOAs give the very central location of a trust

4. How is a patient associated to an ICB?

To identify a patient's integrated care board (ICB), we:

1. Identify their registered GP practice.
2. Identify the parent ICB for this GP practice.

5. How is suppression applied to GP practices?

To suppress low population groups (active GP practices with less than 2,500 registered patients that also meet the criteria in section How do patients appear on the tool?) and protect patient data, we mask GP practice names and codes with the value 'Unknown', so that those results are grouped together under the 'Unknown' value in the FSS Tool when viewing results by GP practice.

Additionally, we filter out any GP practices that are not classed as active, so practices that are; dormant, temporarily closed, or completely closed, migrating patients to another practice.

6. How is the age calculated for the patient?

To get the best possible age value for a patient, we attempt to pull date of birth values if present from PDS, HES APC and HES OP to then derive the age value. Once we have these three values, we then check for the first real non null value (as per the order above) and use that as the Feasibility age value for the patient.

7. How is the Index of Multiple Deprivation calculated?

We obtain the Index of Multiple Deprivation (IMD) score for a patient by cross referencing the patient's postcode (LSOA11) with NHS England corporate reference data (Indices of Deprivation corporate reference table).

If we cannot validate the patient's postcode then we set that patients IMD score to being 'Unknown'.

8. What age range can be used to run a query?

If the user selects to run a query with an age range of less than 3 years, the query will automatically be adjusted to increase the maximum age to give a 3 year age range and the following warning message will be displayed 'Min age range 3 years – ages will be adjusted in search', such as entering 19-20 will be auto-adjusted to 19-21.

9. How are small numbers suppressed in the tool?

To protect the identity of patients and ensure confidentiality is maintained, HES suppression has been applied to small numbers. For example:

• any count below 8 is set to *
• any count equal to or above 8 is rounded to the nearest increments of 5
• if the grant total is below 8, * will be displayed
• if there are no patients for a combination of Row/Column values result, the result will be blank
• if the sum of the patient count for the whole search is below 100, then the table will not be displayed
• if the Rows selection is GP Practice, practices with less than 2500 registered people that meet the DigiTrials valid feasibility patient criteria are grouped into an Unknown category
• if the Rows selection is Postcode sector, postcode sectors with less than 1500 people that meet the DigiTrials valid feasibility patient criteria are grouped into an Unknown category 

10. How are low population postcode sectors suppressed in the tool?

To protect the identity of patients and ensure confidentiality, postcode sectors with less than 1,500 valid feasibility patients (see section How do patients appear on the tool?) are suppressed and are grouped under the value ‘Unknown’ in the results page table.

Medical code business rules

1. How are diagnosis and procedure dates calculated?

The data that the tool uses only contains all instances of any diagnosis (when the appointment was scheduled for), procedure or dispensed prescribed medication code for a patient. This enables users to be able to search by their own date range values for each medical dataset inclusion and/or exclusion.

There are static start/from dates for each medical code dataset which are applied within the Feasibility Self Service Tool, see the meta data section of this page, or the tool itself for the specific dates. These dates are based on the earliest available data within each medical code dataset.

The end/to dates for each medical code dataset continue to move over time as they are based on the most recent processing of the source data into the database used by the Feasibility Self Service tool.

To see the latest available date this will be detailed underneath the end/to date field in the tool itself.

Note: Users can enter dates that exceed the date which the data runs up to for a medical code dataset as the tool does not restrict what is entered into this field. If a user does enter a date greater than the maximum available dated data for the dataset, then the query results will still be based on the maximum available dated data for the dataset.

2. How is diagnosis data calculated?

It is not only the patient’s initial or primary diagnosis that is used - any outpatient appointment or any inpatient hospital stay tied to the provided ICD code(s) determine a diagnosis.

It is important to note that a patient will only ever appear once within a single search. As an example, patient X had separate outpatient appointments or inpatient hospital stays that are associated with a a specific ICD-code on a variety of different dates across 2022 and 2023, as per the underlying HES data. If you ran the query with a date range of the calendar year 2022, and again for the calendar year 2023, you would have patient X included in both (once). If you ran the query with a date range of 1/1/22 to 31/12/23, again patient X would only appear once (not twice, as a patient can only appear once in a set of results).

3. Which prescriptions are included in the tool?

Medication codes only include those that were prescribed in England.

4. What other validation is done against the codes in the tool?

Using NHS England corporate  reference tables, we check that the medication codes, diagnosis codes and procedure codes we retrieve for a patient are valid. If any codes cannot be found in the NHS England corporate reference tables then we filter these records out of our database.