Bioscience and health technology sector statistics 2020 – user guide

1. Introduction

This User Guide accompanies the Bioscience and Health Technology Sector Statistics 2020 release. It is intended to be useful reference guide, defining key terms used in the publication, and explaining issues involved in producing the report.

We welcome feedback on the Bioscience and Health Technology Sector Statistics publications, its accompanying datasets and this User Guide. In addition, we continuously aim to improve the accessibility of our documents. If you experience any problems reading the report or using data, or have any feedback relating to accessibility or otherwise, please contact by email at analysis@officeforlifesciences.gov.uk, including “BAHTSS publication feedback” in the subject line.

2. Data partners acknowledgement statement

The Office for Life Sciences gratefully acknowledge the contribution of the following regional and national organisations in the compilation of the Bioscience and Health Technology Sector Statistics database over the past eleven years.

The content of the database has been derived from a variety of proprietary data sources which have been provided under licence. The Office for Life Sciences would like to acknowledge the assistance given by the owners of these data sources.

Business Information was accessed under licence by Dun & Bradstreet (D&B) Limited and the FAME database from Bureau van Dijk Electronic Publishing. More details on how this data is used can be found in the Methodology.

The database construction, data integration, data analysis and commentary preparation were completed by a consortium led by Kepier & Company Ltd. The consortium included Cels Business Services (CBSL) Ltd and Lindum Research.

2.1 Data partners

• Association of British Healthcare Industries (ABHI)
• Association of the British Pharmaceutical Industry (ABPI)
• AXREM
• BioIndustry Association (BIA)
• BioNow
• Bioindustry
• Biopartner
• Biosciences Knowledge Transfer Network (KTN)
• British Healthcare Trade Association (BHTA)
• British In Vitro Diagnostics Association (BIVDA)
• HealthTech and Medicines Knowledge Transfer Network (KTN)
• Innovate UK
• Invest Northern Ireland
• MedCity
• Medicines Discovery Catapult
• Medilink East Midlands
• Medilink North of England
• Medilink South West
• Medilink West Midlands
• MediWales
• MHRA
• OBN
• One Nucleus
• Scottish Enterprise
• South East Health Technologies Alliance (SEHTA)
• TechUK
• Welsh Government
• West of England LEP

3. Data quality principles

These statistics are released as an Official Statistics publication, and we aim to collect data and produce this report in line with principles of the Code of Practice for Statistics^{[footnote 1]}. This allows us to engender trust in our data and encourage the use of this report as a reliable source of information on the UK life sciences sector.

The following data quality statement covers the fourteen principles under the three pillars of the Code: trustworthiness, quality and value.

3.1 Trustworthiness

T1: Honesty and integrity – Data is collected, processed and quality assured by an independent contractor. The initial technical specification is set by professional analysts within the Office for Life Sciences (OLS) who also engage regularly with the contractor, review methodological aspects, and undertake further quality assurance checks before publication.

T2: Independent decision making and leadership – OLS analysts abide by the Code of Practice, keeping pre-publication access to the data strictly to those involved in the report’s creation and ensuring the statistical integrity of content. The Department’s Head of Profession for Statistics is engaged when necessary. A copy of the Pre-Release Access list for the report can be found on the BEIS statistics website.

T3: Orderly release – Pre-publication access to the report is restricted to those involved in the report’s creation and publication. The report meets Government Statistical Service (GSS) standards of statistical impartiality, separating statistical commentary from any political, press or ministerial statements. Subsequent statements by the government using data from this report quote this source and non-governmental users are encouraged to do the same. Unscheduled corrections are released as soon as is practicable, alongside an explanatory note on both the cause and impact of the error, in line with the Code of Practice.

T4: Transparent processes and management – Substantial financial and administrative resources are employed to enable this data collection and effective quality assurance, including a proportion for further development of the report each year in light of new user requirements or new methodology / collection possibilities. We are transparent about our methodology and approach to quality, providing a summary of the approach used in the Methodology section of this User Guide.

T5: Professional capability – Data is collected, processed and quality assured by a contractor. Each individual working for the contractor has appropriate analytical capabilities, data protection awareness and industry-specific expertise, and has significant experience producing the report and underlying database. Those involved in producing the report within OLS are badged government analysts.

T6: Data governance – All professionals involved in the creation, publication and storage of this dataset are well-versed in data protection and operate in compliance with data protection legislation. We publish the maximum amount of data available without contravening third-party licence agreements, utilising GSS best practice for statistical disclosure control (see Rounding and Disclosure control for more information).

3.2 Quality

Q1: Suitable data sources – A description of the methodology used, including a description of third-party administrative data sources and their suitability, can be found in the Methodology section of this User Guide. The report’s classification system is explained in the Segmentation section, together with a glossary defining the terms used.

Q2: Sound methods – A description of the methodology used can be found in the Methodology section of this User Guide, alongside assumptions made. Terminology is consistent through the report and accompanying files, with clear descriptions in the glossary. Year-on-year trends are for real growth only based on like-for-like data against the previous year. The method used for the real growth calculations is explained in the Methodology.

Q3: Assured quality – Rigorous quality assurance has been undertaken by the contractor and OLS analysts. Quality assurance is a significant part of the technical specification and contract tendering process and is reviewed each year. When an unscheduled revision was necessary following the post-publication identification of an error in a previous report, we immediately alerted users, engaged with the Department’s Head of Profession for Statistics, and published an explanation of the cause and impact of the error alongside the revised report, all in accordance with the Code of Practice.

3.3 Value

V1: Relevance to users – We review content each year based on user needs, allowing a proportion of resource for that year’s topic of interest. In previous years this has led to the inclusion of Digital Health and Genomics as chapters, with a new cross-cutting classification designed to identify businesses operating in the field of genomics.

V2: Accessibility – Data is free and equally available to all, published on gov.uk with no restrictions to access. Underlying data is published up to the extent our commercial licenses allow, with banded variables where we cannot provide exact figures. Commentary is objective and a range of graphical visualisations are used to aid comprehension.

V3: Clarity and insight – Commentary on the current size and shape of the life sciences sector is objective, focussing on impartial statistical messages. Charts and maps are used to illustrate these. Key statistical messages are highlighted at the beginning of each report section. The database itself is created through collaboration with a range of industry experts, including region-specific and sector-specific representation through trade bodies and other network organisations.

V4: Innovation and improvement – We review content, presentation and methodology each year based on user needs. Past development has primarily been around scope and how to identify new and emerging segments of the life sciences sector, such as digital health and genomics. Each new approach to scope is explored and tested with our data partners, and the statistical impact is fully considered before implementation. Other developments have included extending the scope of publicly available data fields.

V5: Efficiency and proportionality – Where possible, the database draws on existing information using third party sources, such as the Dun & Bradstreet and Bureau Van Dijk FAME datasets and company accounts. All data partners are voluntary contributors. The need for this report and underlying database arises from difficulties encountered when describing the life sciences sector from existing ONS sources. ONS data relies on SIC codes to classify companies, however SIC codes are considered to fail to capture the full extent of the life sciences sector. Furthermore, SIC codes were last revised in 2008, and do not allow easy identification of new and emerging segments within the medical technology sector, such as Digital Health. This publication provides a valuable and robust evidence base on the size and shape of the UK life science sector.

4. Methodology

The annual update of the database is carried out in four main phases: compiling information on new businesses and existing businesses; classification or segmentation of new businesses; matching of business details with economic data from external databases; and validation of the data set.

4.1 Compiling information

Information on new businesses is sourced from data partners and also by searches of publicly available and subscription databases. Data partners provide lists of businesses from their internal databases, which contain both new businesses (those businesses that have been formed in the period after the last annual update) and existing businesses (those businesses that are already in the database or are part of an existing business following merger and acquisition activity). For existing businesses, this includes information data partners have obtained on address changes, any information on employees at a location, or suggested segmentation changes.

4.2 Segmentation of new businesses

The information from data partners and other information sources is cleansed to remove duplicates and records already in the database and is then assigned a label within a hierarchical classification system, a process referred to as ‘segmentation’ in the context of this report.

Segmentation assigns each new business and site to a sector and segment. In some cases, allocation can be to more than one sector or segment, for example some large multi-national businesses produce both pharmaceutical and medical devices. If, based on the information available, a business cannot be assigned to a sector and segment, it is deemed to be not-in-scope (NIS). Such NIS business information is retained but is not included in the data set used to analysis the industry.

In order for a business to be classified as in-scope and its data to be included, it is assessed against the following criteria:

• has a legal entity in the UK;
• is a private limited company (this excludes universities, publicly owned institutions, NHS activities, and charities);
• and has 20% of its total UK turnover derived from one or more of the segments shown in the Segmentation section of this User Guide^{[footnote 2]}.

Businesses proposed for inclusion or identified through a search of new incorporations, are checked for “proof of life” i.e. signs of economic activity such as employees, turnover, award of funding, or an active website with contact details. Businesses which fail this test but appear to be in scope are reviewed again in the next project cycle.

4.3 Matching against external datasets

Once the cleansed data set is prepared, it is used to source data on turnover and employment from reliable external data sources, and from examination of published company reports or data. Matching between datasets is achieved using the company name, company registration number and the postcode for the site. The turnover figures include turnover on the sale of products wholly or partially manufactured outside the UK. The external databases used are Dun & Bradstreet’s database and Bureau van Dijk’s FAME database. For more information on this data matching, refer to the section on Data quality principles.

4.4 Validation

The data returns from D&B and FAME are carefully checked to ensure a correct match with the business location. Further detailed validation of the data is then carried out by examining changes in the employment and turnover data. These changes are investigated to detect any anomalies through verification against other sources.

For example, large changes in employment at a business site are scrutinised to see if information is available from press releases or other information in the public domain to verify the change. From 2018 onwards, Gender Pay Gap reporting was used both to detect potential anomalies by using the compulsory employment band data^{[footnote 3]}, and to verify or update using more detailed information provided by businesses within their own reports. The data for individual sites under one business is examined to ensure that there is no double-counting of employment or turnover data.

Once the validation analysis is completed, the data set is “locked” for the annual update cycle, ready for analysis for this publication.

4.5 Creating a time series

Each year, the new dataset is added to the datasets from previous years to form a time series dataset which allows trends to be observed and analysed. This edition of the report considers the ten-year period from 2011 to 2020.

The time series dataset is augmented by additional economic information (e.g. employment and turnover) from third-party sources including Dun & Bradstreet (D&B), FAME, and published company-filed accounts to supplement the data available from Companies House. These additional sources are used because smaller businesses are not required to publish full financial information in their annual accounts.

GDP deflators^{[footnote 4]} are applied to the turnover value for each business in previous years to ensure turnover is reported in 2020 prices.

5. Data quality considerations

Users of the data provided in this report should be aware of the following data limitations and caveats:

• Postcodes attached to records in the database allow geographical analysis of employment and turnover at site level. Where available, we have validated employment data for the large businesses by using information such as annual reports or websites to identify the number and types of employment.
• The primary allocation of turnover to location is based on the legal entity information sourced from third party databases, validated for large businesses from annual accounts. This method of turnover reporting is used throughout the document.
• Turnover is reported in 2020 prices, using GDP deflators^{[footnote 5]} to take account of inflation across the years.

6. Publication changes

The 2020 edition of this report incorporates several changes made to improve accessibility, data quality and the overall user experience.

6.1 Changes to methodology

Improvements in the annual snapshot dataset and the time series dataset have allowed the two datasets to become fully aligned in the 2020 publication. This means that the 2020 counts of businesses, value of turnover and number of employees are the same in both datasets. This is a change from previous editions of this report where counts and totals in the snapshot dataset did not match counts and totals in the time series dataset. This has simplified the presentation and reporting of UK Life Sciences in the 2020 publication and has removed the need to caveat comparisons between the snapshot and time series data.

To achieve this alignment, it has been necessary to exclude a small number of unincorporated businesses from the datasets. This is because unincorporated businesses do not have any accessible historic information that would allow them to be included in the time series dataset.

In 2020, 140 unincorporated businesses were excluded from the datasets, accounting for an estimated 1,000 employees and a turnover of £95m.

6.2 Changes to the report content

Two new sections relating to activity manufacturing sites have been added. The first new section reports 2020 data on manufacturing sites, and the second appears as part of the Industry and sector trends chapter.

In the report, manufacturing sites are defined as those sites where manufacturing activities take place, which is any site with the Business Activity Code ‘BAB’ in the database. Some of these sites will undertake types of work in addition to manufacturing, and therefore the number of employees and turnover from these sites includes contributions from non-manufacturing activities.

6.3 Changes to presentation of the report and accompanying datasets

The report is now presented in HTML format, and explanatory information accompanying previous reports has been collected into a single User Guide (this document). Accessible versions of the report’s infographics have been made available alongside the standard versions.

Data used to produce the charts and tables in the report are presented in the ‘Underlying Data’ spreadsheet. This spreadsheet also includes a supplementary table containing information that appeared in previous editions of the report, which has been added following consideration of user feedback.

7. Conventions used in the report

7.1 Rounding and Disclosure control

In this report and accompanying datasets, data have been rounded, as the figures are subject to the inaccuracies inherent in any large database. There are a range of different types of errors possible, such as those resulting from recording errors or misclassifications.

In addition, the licencing arrangement in place for the data used to compile this report restrict the disclosure of individual company details, such as the number of employees and annual turnover.

Rounding and cell suppression are forms of ‘statistical disclosure control’ and have been considered in line with the ‘Code of Practice for Official Statistics’.

The Code of Practice for Official Statistics states ‘Ensure that arrangements for confidentiality are sufficient to protect the privacy of individual information, but not so restrictive as to limit unduly the practical utility of official statistics.’ This emphasises the dual purposes of statistical disclosure control. Detail relating to individual statistical units is to be protected but the released data must still be of high practical utility for users^{[footnote 6]}.

In the report and the Underlying Data tables, employment figures for sectors and the whole industry are rounded to the nearest 100, turnover is rounded to the nearest £0.1bn (unless otherwise stated in a data table), and the number of sites to the nearest 10.

In the underlying data tables, data for segments where the number of employees is less than 500 have been suppressed for reasons of disclosure control, but count towards table totals. Suppressed data are indicated by a ‘*’.

In the public database, which contains company-level data, employment and turnover figures are provided within bands to avoid disclosure of individual company information.

8. Segmentation

The life sciences database contains information on businesses in the UK structured at the level of trading address corresponding to the 7,180 records in the database for 2020. Using this as the lowest level of information the data is aggregated to site and company level to give the estimate of total number of life sciences businesses in the UK (6,330). Each trading address or site is examined to allocate the activity carried out to one of the segments in a sector. As a small proportion of businesses in the database have more than one trading address or site and can operate in more than one sector (for example can have activity in medical technology and biopharmaceuticals).

Each business and their individual sites are segmented depending on the main type of final medicinal product or device produced. Businesses that produce products that are directly used in healthcare are designated “Core” businesses to distinguish them from businesses that are active only in the Service & Supply chain.

It should be noted that the Biopharma sector suppliers of over the counter (OTC) medicines are included along with generic suppliers and manufacturers.

Within the database, codes are used to allocate businesses and sites to one or more segments. Where a company has products that fall in more than one category, these are all coded, however only the code that represents the majority of the business activity is used in the analysis. Figure UG1 breaks down the count of records in the database from the total number of businesses in life sciences down to the allocation of sites to business activity.

The count of records in the database at each level of classification from site level through segment, sector and industry for 2020.

Figure UG1A: Counts of businesses at the industry level.

There are 6,330 businesses in the UK Life Sciences Industry.

Figure UG1B: Counts of businesses at the principal sector level.

There are 2,190 businesses in the Biopharma sector and 4,060 businesses in the Med Tech sector. 80 businesses operate in both sectors. There are 6,330 businesses in total.

Figure UG1C: Counts of businesses at the sector level.

There are 720 businesses in the Biopharma Core sector, 1,450 businesses in the Biopharma Service and Supply Chain sector, 2,820 businesses in the Med Tech Core sector and 1,210 businesses in the Med Tech Service and Supply Chain sector. 130 businesses operate in more than one sector. There are 6,330 businesses in total.

Figure UG1D: Counts of sites at the sector level.

There are 930 sites (business locations) in the Biopharma Core sector, 1,690 sites in the Biopharma Service and Supply Chain sector, 3,220 sites in the Med Tech Core sector and 1,340 sites in the Med Tech Service and Supply Chain sector. There are 7,180 sites in total.

Segmentation was reviewed for all businesses and sites in the 2014 update. During the 2015 update a number of the businesses that have large contributions to employment and turnover were reviewed for segmentation and their turnover in scope (TOS). The Pharmaceutical and Medical Biotechnology sectors were also combined into a new sector: Biopharma.

Additional segmentation codes are used to further classify company activities by both product type and business activity. For example, in vitro diagnostics is further segmented into in vitro diagnostic products that involve clinical chemistry, immunochemistry etc. The business activity codes are used to code businesses and sites dependent on whether they undertake R&D, manufacturing, Service & Supply (of their products), and sales/distribution (of their products).

The codes for each sector containing Core businesses are shown in the Segmentation section of this User Guide. The Service & Supply chain sectors that serve the Biopharma and Med Tech sectors are coded with the prefix BP and MT respectively followed by the appropriate number to define the type of service or supply.

Table UG1: Counts of Businesses by Sector, showing counts of businesses in single and multiple sectors

Note*: There are a total of 10 businesses that are active in multiple sectors where the total for that combination of sectors is 4 or less. These have been designated as “Other” in the above table.

9. Alignment with Standard Industry Classification (SIC) codes

Standard Industry Classification (SIC) codes are used to classify businesses by industry in administrative statistics. This was last updated in 2008^{[footnote 7]}. This classification system has categories for businesses whose primary activity is the manufacture of pharmaceuticals, manufacture of types of medical equipment, and those whose primary activity is biotechnology R&D.

The SIC system, however, does not allow identification of the full range of life sciences businesses. A bespoke industry segmentation based on this wider range, specifically to be used in the database, was defined with the assistance of the data partners and is summarised in the Segmentation section of this User Guide. This is the classification system used in this report.

We have analysed the SIC codes of the businesses within the database and only 25% of businesses fall into the standard SIC codes used to identify the life sciences industry. The remaining businesses fall into another 250 SIC codes, demonstrating the ongoing need for this report and for the life sciences database to describe and analyse the full breadth of this industry.

To allow comparison between the two classification systems, Table UG2 shows the number of sites, total employment and turnover for businesses in the database with SIC codes typically used to define the life sciences industry, and for all businesses included in the life sciences database.

Table UG2: Turnover, employment, and number of sites based on the SIC codes that cover the main sectors in the life sciences industry

The additional benefit of the segmentation approach used in the life sciences database is the ability to make a more granular assessment of the sector, including growth rates and trends. For example, this is the only source of definitive information that shows employment and growth rates in digital health or allows us to understand the growth of advanced therapy medicinal products.

10. Segmentation Coding and Classification Scheme

The following tables describe the identifiers (codes) and descriptions for each element of the classification schemes used in this report and accompanying dataset. Reference should be made to the methodology section in this User Guide.

Principal Sector and Sector codes and descriptions.

Biopharma Core (BP) segment codes and descriptions.

Medical Tech Core (MT) segment codes and descriptions.

Digital Health (MTT) sub-segment codes and descriptions.

Service & Supply Chain (MX/BX) segment codes and descriptions.

Note*: Healthcare service providers have been designated as “out of scope” for this study. No new records have been added to the dataset in 2020. Existing records will be removed in 2021 and adjustments made accordingly.

Business Activity codes and descriptions.

Genomics codes and descriptions.

Employment bands

Turnover bands

Alternative Turnover Bands

11. Public Dataset Field Descriptions

The public dataset which accompanies this report lists the site-level records and variables used to derive figures in 2020 report. The public dataset is presented in a Comma Separated Variables (CSV) data file and uses the data structure described in the following table. When using this data structure description table, reference should be made to the methodology and segmentation coding scheme sections of this User Guide.

Public Dataset Column Names and Descriptions