Official Statistics

UK service personnel medical discharges: background quality report 2022

Updated 13 July 2023

1. Contact

Defence Statistics Health Deputy Head

Defence Statistics Health

Analysis-Health-PQ-FOI@mod.gov.uk

2. Introduction & Statistical Presentation

Defence Statistics produce a publication detailing medical discharges within the UK regular armed forces which can be found at the gov.uk website. The Medical Discharges in the UK Regular Armed Forces is an annual publication that provides information on medical discharges among UK regular armed forces personnel for the previous five financial years. Each of the three services are presented separately; Royal Navy/Royal Marines, army and Royal Air Force (RAF). The statistics display:

• The number of medical discharges within the UK regular armed forces by key demographic factors: service, gender, age, officer or other ranks, trained or untrained status.

• The number of medical discharges within the UK regular armed forces by principal and contributory causes of medical discharge.

• Crude rates of medical discharges within the UK regular armed forces.

The first Medical Discharge Official statistic was published in 2011.

Prior to the 2013/14 report, published 16 July 2015, this publication included only the principal condition leading to medical discharge. Since two in five UK regular armed forces personnel medically discharged have more than one condition affecting their ability to continue their role in the UK armed forces (both principal and contributory), the addition of contributory causes provided a more comprehensive picture. Contributory cause information is intended to provide information on the total number of personnel medically discharged from service that had their ability to perform their duties affected by specific medical conditions. It cannot be used to identify the complexity of conditions for which personnel have been medically discharged.

The report covers regular service personnel (including Gurkhas and Military Provost Guard (MPGS)) only and excludes all reservist personnel. This is because the medical discharge process and medical record information for reservist personnel is not comparable to that of regular personnel. Most reserve personnel do not receive their primary medical care from MOD, but instead receive their primary medical care from the NHS. Therefore, Defence Statistics Health are unable to verify the quality of information relating to the discharge of reservist personnel and it has not been deemed appropriate to include information on this population until further understanding is gained.

Personnel described in this report as “untrained” are personnel who are in Phase 1 and Phase 2 training. Untrained personnel are sometimes discharged under administrative categories, albeit on medical grounds. These discharges usually concern individuals who have failed their initial training for medical reasons, or who at their initial medical, failed to disclose medical reasons which may later affect their application and training. As these cases are not defined as medical discharges, they are not included in this report.

Medical discharges are the result of a number of specialists (medical, occupational, psychological, personnel, etc.) coming to the conclusion that an individual is suffering from a medical condition that pre-empts their continued service in the armed forces. Statistics based on these discharges do not represent measures of true morbidity or pathology. At best they indicate a minimum burden of ill-health in the armed forces. Furthermore, the number and diversity of processes involved with administering a medical discharge introduce a series of time lags, as well as impact on the quality of data recorded.

Although medical boards recommend medical discharges, they do not attribute the principal condition leading to the board to the individual’s time in service. A medical board could take place many months or even years after an event or injury and it is not clinically possible in some cases to link an earlier injury to a later problem which may lead to a discharge.

Medical discharges in the UK regular armed forces involve a series of processes, at times complex, which differ in each service to meet their specific employment requirements^{[footnote 1]}. Due to these differences, comparisons between the single service statistics are judged to be invalid.

3. Statistical Processing

3.1 Source Data

Joint Personnel Administration System (JPA)

JPA is the system used by the UK armed forces to deal with matters of pay, leave and other personnel administrative tasks. JPA replaced a number of single-Service IT systems and was implemented in April 2006 for the RAF, November 2006 for the Royal Navy and April 2007 for the army.

FMed 23

FMed 23s are official medical documents used to record all medical board proceedings, including recording the outcome of a medical board held for members of the UK armed forces leading to medical discharge.

Defence Medical Information Capability Programme (DMICP)

The Defence Medical Information Capability Programme (DMICP) commenced during 2007 and comprises an integrated primary Health Record (iHR) for clinical use, and a pseudo-anonymised central data warehouse. It is the source of electronic, integrated healthcare records for primary healthcare and some MOD specialist care providers. Prior to this data warehouse, medical records were kept locally, at each individual medical centre. By 2010, DMICP was in place for the UK and the majority of Germany. Rollout to other overseas locations commenced in November 2011 and is ongoing. Please note DMICP data prior to 2010 is considered incomplete due to rollout of the programme. DMICP is a live data source and is subject to change.

3.2 Frequency of data collection

Data collected from JPA, FMed 23s and DMICP are done so on a monthly basis.

3.3 Data collection

The following data was collected from JPA:

• The service personnel who were discharged from service having outflowed to civil life, including demographic fields such as; service, rank, gender and date of birth

• The numbers of personnel serving in the UK regular armed forces in each financial year, which was used to calculate the population at risk.

Defence Statistic Health receive paper and electronic copies of FMed 23s from the armed forces medical boards (unless the individual concerned has withheld consent). A record of the FMed 23s received are stored in an electronic receipt table. The primary purpose of these medical documents is to ensure the appropriate administration of each individual patient’s discharge. Statistical analysis and reporting is a secondary function.

Data from JPA is matched to FMed 23s to collect the principal and contributory cause of discharge. Any records present on JPA for which Defence Statistics Health have not received an FMed 23 are queried in DMICP. This was done using the “Principal Condition” and “Contributory Condition” fields on each of the single service medical board templates. This information was extracted as Read codes. Any further records are confirmed with single service representatives.

Due to COVID-19 limiting access to place of work, Defence Statistics Health were not able to access the FMed 23s paperwork for confirming cause of medical discharge for part of 2019/20 and 2020/21 – 202/22. During this time period DMICP was used as the primary source, therefore, cause information for services in these years should be considered provisional and subject to change. Defence Statistics Health will assess the methodology with the potential to revise the cause information using FMed 23 as the primary source.

At the point of medical board, service personnel have the opportunity to withhold or give consent to their medical information relating to the medical board being forwarded to Defence Statistics Health. If consent was withheld, personnel were still counted as a medical discharge as indicated on JPA, however their reason for medical discharge is not held by Defence Statistics Health. Therefore, their principal or contributory conditions leading to medical discharge were not presented in this report. This was indicated as “Withheld Consent” in report tables

3.4 Data Validation

Monthly validation procedures are used to collate and check the data.

The main sources of potential error in the Medical Discharge statistics are as follows:

• Incomplete or inaccurate data from the DMICP.

• Data processing errors resulting in incorrect data outputs.

• Manual error during production of report tables and commentary.

To ensure that potential errors are identified and resolved, Defence Statistics Health implement a series of data quality checks throughout the report production. For further information, please see section 6.1 for more details.

After compilation data is compared to the previous statistical output and is signed off by both a SEO and the Deputy Head of Defence Statistics Health.

3.5 Data Compilation

Coding of Medical Conditions

Cause information extracted from FMed 23 forms, and as Read codes from DMICP were mapped, where possible, to the International Classification of Diseases and Related Health Problems version 10 (ICD-10). This data was then validated and tabulated to form the principal and contributory cause of discharge figures presented in the report. As a result of public interest, some ICD-10 groups were provided in more detail allowing the presentation of specific conditions.

Pseudo-Anonymisation

Prior to analysis data sources were linked using a pseudo-anonymisation process. The individual identifiers were stripped from datasets and replaced by a pseudo-anonymiser, generated by an automated sequential numbering system. The key to the system is that it recognises previous occurrences of a given service number and allocates the same pseudo-anonymiser on each occasion. The pseudo-anonymisation process can only be reversed in exceptional circumstances controlled by the Caldicott Guardian under strict protocols.

Crude Rates

Crude rates enable comparison between groups by removing the issue of different populations at risk (group sizes). The rates in this bulletin present the number of personnel per 1,000 on strength that are medically discharged each year. As the size of the armed forces varies over time, this is a more accurate means of comparing the proportion of personnel medically discharged from service in different years than utilising counts of the personnel medically discharged.

Crude rates were calculated by dividing the number of events (in this case medical discharges for each year) by the population at risk (in this case the average number of service personnel on strength in each year ^{[footnote 2]}). They are presented as an overall summary and for demographics including age group, gender, rank, training status for each service by year. The crude rates for the overall summary do not take the changing demographic profile (e.g. gender and age structure) into account. This is because the structure of the UK armed forces has seen limited change over past years. Whereas, the crude rates presented for the demographic breakdowns, do account for changing profiles.

Z Test for Independent Proportions

The Z test for independent proportions evaluates if two rates are different to a statistically significant degree. In order to identify age groups with a significantly higher or lower rate of medical discharge, Z tests for a single proportion were performed comparing each age group to the remaining age groups rate of discharge (the mean for data fitting a normal distribution or the median for non-parametric data). In some cases, Defence Statistics Health also performed Z tests for two proportions between specific age groups to provide greater clarity on the relationship between age and rate of discharge. Where appropriate, a Fisher’s exact test was used. A Fisher’s exact test is the more accurate statistical test when one of the numerators in the rates is less than 5.

The confidence level to which these tests were run in this report was 95%: this means that if the test determined two populations to have different medical discharge rates, this was true in greater than 95% of cases.

4. Quality Management

4.1 Quality Assurance

The MOD’s quality management process for Official Statistics consists of three elements: (1) Regularly monitoring and assessing quality risk via an annual assessment; (2) Providing a mechanism for reporting and reviewing revisions/corrections to Official Statistics; (3) Ensuring BQRs are publishing alongside reports and are updated regularly.

A large proportion of data processing has now been automated. This will reduce human error which has been the cause of quality incidents in the past. Defence Statistics Health are planning to improve data processing further to fully automate data collection, and collation.

4.2 Quality Assessment

The Medical Discharges in the UK Regular Armed Forces Official Statistic was last assessed in 2021 and was deemed to be of low risk.

5. Relevance

The report was initially created in response to a number of Freedom of Information requests for medical discharge figures for the UK armed forces. Key internal and external stakeholders were consulted in its creation.

Defence Statistics Health invite feedback from customers within the publication and seek feedback from a wider range of internal and external customers.

5.1 User needs

The report is used to inform internal and external stakeholders, including to inform policy and decision making within the Department. The statistics are also used to inform general debate in government, parliament and the wider public.

It is thought that this information may be of use in understanding the number and rate of UK regular armed forces, the demographic populations at risk with this and the types of injuries and conditions that personnel are medically discharged for.

The report was provided in response to the increasing number of requests for information about UK armed forces personnel medical discharges. The nature of the requests varied from more detail on the injuries and illnesses causing medical discharges, to information on the long-term outcome for personnel that were medically discharged.

Data from the report is utilised in the annual health of the armed forces report, as well being used to answer parliamentary questions and Freedom of Information requests. In addition, this information is used to prioritise resources used for the rehabilitation and reintegration of personnel leaving the armed forces for medical reasons and to help inform discussions on injury prevention in the armed forces.

The principal customers for the medical discharge publication include:

• Single service medical boards

• Defence Business Services (DBS)

• Chief of Defence Personnel (CDP), People Secretariat

• Defence Medical Services

• Armed Forces Occupational Health

• Department of Health

• External Organisations concerned with former service personnel welfare (including charitable organisations)

• Academic Researchers

• Journalists

The report has an accompanying ministerial submission.

6. Accuracy and Reliability

6.1 Overall Accuracy

This report uses a variety of different data sources, each of which has different properties which affect the difference between published and true values.

JPA

Extracts are taken from JPA each month taken six calendar days after the end of the month and the situation as at the first of the month is calculated. This ensures most late-reporting is captured. As a result of improvements in the quality of data sourced from JPA and the monthly data validation processes, all JPA data is considered to be fit for purpose.

There is a known small issue where an exit date can be altered on JPA, after the extract has been taken by Defence Statistics. This results in a person appearing in the discharge data twice. Defence Statistics Health are aware of the small issue, and as part of routine validation check and review any duplicate medical discharges within the dataset. This may result in small revisions in future versions of the report where the duplicate discharges have fallen outside of the reporting year.

FMed 23

Whilst FMed 23 forms received by Defence Statistics Health do include some reservists, the number and coverage of reservists captured is currently unknown and reliable denominator data is not available for the entire reporting period. Therefore, numbers and rates were calculated using only strengths for regular personnel and for this report all known reservists have been removed.

From 2013/14 onwards, Defence Statistics Health have not received FMed 23 forms for a portion of the regular army personnel listed as medical discharges on JPA. In previous years, the Army Personnel Centre (APC) retrieved any FMed 23 forms not received by Defence Statistics Health. This service is no longer provided by APC and Defence Statistics Health are therefore unable to determine the principal or contributory causes for their discharge. Whilst some of this information has been sourced by utilising DMICP, there remain a number of trained army personnel for which principal and contributory cause information has not been obtained.

DMICP

The DMICP system is a large clinical and administrative database and is subject to the data quality issues of any large administrative system with data collated by many medical and administrative staff for clinical delivery purposes.

In 2019/20 – 2021/22, due to the COVID-19 pandemic, a temporary method was used whereby cause information was acquired from DMICP before an FMed 23 form, please see section 3.3 for more details.

In some cases Defence Statistics Health were not able to utilise DMICP for this missing cause information:

• Some personnel did not have a medical board template entered on their record.

• Some personnel had a medical board template entered on their record, but no cause information.

• Some personnel had “Principal Condition” and “Contributory Condition” codes entered on their records that were not possible to convert to ICD-10.

To ensure that potential errors are identified and resolved, Defence Statistics implement a series of data quality checks throughout the report production.

6.2 Data Revisions

Data revisions are handled in accordance with the MOD’s Official Statistics Revisions and Corrections Policy.

Some of the data sources used in this report are live systems that are constantly updated. This means figures can occasionally change. Any amendments made since the last release will be indicated by an ‘r’. Revisions can be addressed in two ways. For this report, the first of these two methods has been applied:

• Where the number of figures updated in a table is small, figures will be updated and those which have been revised will be identified with the symbol “r”. An explanation for the revisions will be provided in the section below.

• Where the number of figures updated in a table is substantial, the revisions to the table, together with the reason for the revisions will be identified in the commentary at the beginning of the relevant chapter / section, and in the commentary above the affected tables. Revisions will not be identified by the symbol “r” since where there are a large number of revisions in a table this could make them more difficult to read.

In the 2021/22 release of this bulletin, the following revisions were made:

• The rates of army medical discharges for all years presented have been updated. This was due to an undercount in the army strength numbers from which the rates were calculated. This had no significant impact on findings presented in previous reports.

• In 2019/20 and 2020/21, a small number of personnel (n = 7, 0.3%) were included twice within the reported figures in error. The duplicate record has been removed from the reported figures, and the numbers and rates of medical discharges have been updated. This had no significant impact on findings presented in previous reports.

Revised figures have been represented with an ‘r’. Where figures previously marked as provisional have been revised and updated, and are still provisional, they have been marked as ‘p’ rather than as revised ‘r’.

Occasionally updated figures will be provided to the editor during the course of the year. Since this Bulletin is published electronically, it is possible to revise figures during the course of the year. However to ensure continuity and consistency, figures will only be adjusted during the year where it is likely to substantially affect interpretation and use of the figures.

7. Timeliness and Punctuality

7.1 Timeliness

The report for the previous financial year is released annual in early to mid-July. The publication dates ensure data is available and at a suitable level of accuracy and allow sufficient time for processing and producing the reports.

7.2 Punctuality

All official statistics reports have been published on time to meet preannounced release dates. A one-year release schedule outlining the following financial year’s publication date is published on the gov.uk Official Statistics Release Calendar. Future publication dates will also be announced on the UK Statistics Authority hub at least one month in advance.

8. Coherence and Comparability

The Defence Statistics Health figures on the medical discharge of personnel from the UK regular armed forces are the definitive statistics in the MOD. There are no other publicly available regular publications on the medical discharge of UK Armed Forces personnel with which to ensure coherence. Within the MOD, direct queries of DMICP for medical board data will produce differing results due to quality issues.

Each annual edition of UK Defence Statistics (UKDS) since 1992 and historic Tri-Service publications back to 2002 are available on the MOD National and Official Statistics website page. Medical discharges have been included as a discrete outflow type in UKDS since 1994. The official statistics reports covering the period 2005/06 onward are also available on Gov.uk.

The numbers of personnel medically discharged from each service are comparable over time. However, medical discharges in the UK armed forces involve a series of processes, at times complex, which differ in each service to meet their specific employment requirements and policy changes may influence rates. Due to these differences between the three services, comparisons between the single service statistics are judged to be invalid.

9. Accessibility and Clarity

All reports are published on the Defence Gov.uk website. Publications are available from 0930 hours on the day of release. 24-hour pre-release access to the report is available to a limited distribution list within the MOD. The full list can be found in the pre-release access list.

The report is published in an accessible PDF file. The report includes the following sections;

• Key points and trends

• Introduction,

• Main points,

• Royal Navy/Royal Marines – Trends, Demographics, Cause

• Army - Trends, Demographics, Cause

• RAF - Trends, Demographics, Cause

• Glossary

• Methodology

• Further Information

• Contact Us

• Annex A

• References

The report commentary identifies and analyses the key changes in the data and summary statistics. It discusses the quality of the underlying data and identifies specific issues and estimates their impact. Within each service section, graphs are used to visually compare the rates of medical discharges between different demographic groups, and the principal cause of medical discharge over time.

Data tables from each report are separately available in both accessible Excel and Open Data Source (ODS) formats to download. This allows for use in individual research and reports. Defence Statistics are currently ensuring all published information is equally accessible by everyone.

Each table/figure has a number of notes clarifying what is included and excluded and provides appropriate caveats. Within the report, these are found in table footnotes, and within the data tables these are found in the Notes tab.

10. Trade-offs between Output Quality Components

Timeliness versus quality of the data and depth of analysis are the most notable trade-offs for this report. If additional time was allowed after the reporting period for acquisition of FMed 23s, the proportion of personnel without a known principal condition of discharge may fall. This would however reduce the timeliness of the report.

The time allowed to process the raw data and compile the completed report is around three weeks. Counts, rates and standardised rates are included within the report, as well as some statistical testing - as is some commentary on trends and patterns. However, further statistical analysis is not included within the report.

The trade-off between timeliness and accuracy/depth of analysis has ensured that the information is made available as soon as possible after the end of the reporting period.

11. Cost and Respondent Burden

Defence Statistics Health has a coder whose role is to code the cause information on FMed 23s and DMICP to ICD-10. This work is overseen by an analyst that collates the data, conducts monthly validation and produces the annual report. In 2019/20 – 2021/22, due to the COVID-19 pandemic, the temporary method to obtain cause information from DMICP before an FMed 23 form (see 3.3. for more details) has resulted in a lesser burden for the coder.

There is some respondent burden as FMed 23 forms completed by medical boards should be copied and sent to Defence Statistics Health unless consent is withheld. A move away from sending paper forms in the post, to sending forms electronically has reduced this burden.

JPA data is automatically obtained from administrative systems, however it is validated and supplemented with small amounts of data and input from other areas by the Defence Statistics workforce branches prior to its use by Defence Statistics Health.

12. Confidentiality and Security

12.1 Confidentiality – Policy

All staff involved in the production process adhere to the MOD, Civil Service and data protection regulations. Defence Statistics adhere to the principles and protocols laid out in the Code of Practice for Official Statistics and comply with pre-release access arrangements. The Defence Statistics Pre-Release Access lists are available on the gov.uk website.

12.2 Confidentiality – Data Treatment

Publications do not contain any identifiable personal data. The information presented in each publication has been structured in such a way to release sensitive medical information into the public domain that contributes to the MOD accountability to the British public, but which doesn’t compromise the operational security of UK Armed Forces personnel nor that risk breaching the rights of UK Armed Forces personnel.

In line with the directives of the JSP 200, disclosure control is conducted on all statistical information provided by the MOD to safeguard the confidentiality of individuals. Within these statistics a risk of disclosure has been considered to be high where numbers presented are fewer than five. In cases where a risk of disclosure exists the following appropriate disclosure control method has been applied:

• Figures have been suppressed: In most cases where there may be a risk of disclosure, numbers fewer than five have been suppressed and marked as ‘~’ or [c]. Where there is only one cell in a row or column that is fewer than five, secondary suppression has been applied where the next smallest number has also been suppressed so that numbers cannot simply be derived from totals.

In order to protect personnel confidentiality, pseudo-anonymisation was employed. For further information, see section 3.5 of this Background Quality Report.

12.3 Security

The data is stored, accessed and analysed using the MOD’s restricted network and IT systems, and the access to raw data is password protected.

1. As set out in JSP 346 and/or the single services retention standards for their career group. ↩

2. The average number of service personnel on strength in each year is calculated using the 13 month average. I.e. For RAF 13/14 the number of RAF personnel at the first of each month from April 2013 to, and including, April 2014 are summed and divided by 13. ↩