Ey Value of Health Care Data v20 Final

Realising the
value of health
care data: a
framework for
the future
Realising the value of heath care data: a framework for the future
How can we place

a value on health
care data?
A framework proposal focused on the UK’s health care
ecosystem and patient records held across the National
Health Service (NHS)
Medicine is no longer a clinical science just supported by data, it’s moving to a field
defined as clinical science in collaboration with data science. Patient data is one
of the most important drivers of this change. Unlocking the insights contained in
patient genomic and phenotypic data is of high value to all the key stakeholders in
the health care ecosystem: patients, providers, payers and the life sciences sector. In
this paper, we show the methodological tools that can be used to estimate the value
of patient data, specifically the data held by the UK’s NHS. We show how realising
the value of this data can help the UK Government achieve its health priorities, and
place the UK at the forefront of global health care innovation.
Contents
1. Patient data as a desirable intangible asset 2

2. Drivers of health care data value 4
3. Valuation approaches to consider 6
4. Application of the market-based approach 8
5. Application of the income approach 16
6. Concluding comments 20
7. Appendix 22
Executive summary
Unlocking the power of health care data to fuel innovation in medical research and improve
patient care is at the heart of today’s health care revolution. When curated or consolidated into
a single longitudinal data set, patient-level records will trace a complete story of a patient’s
health, wellness, diagnosis, treatments, medical procedures and outcomes. Health care
providers need to recognise patient data for what it is: a valuable intangible asset desired by
multiple stakeholders, a treasure trove of information.
Among the universe of providers holding significant data assets, the UK’s NHS is the single
largest integrated health care provider in the world, its patient records covering the entire UK
population from birth through to death. There are two primary approaches to quantifying the
value of data:
A market-based approach, calculating the implied “per record”

1 valuation multiples of comparable data assets or valuation
multiples of companies with significant patient data assets
An income-based approach, which quantifies value based

2 on the economic benefit to be generated from the curated
data set
Applying these approaches, we estimate that the 55 million patient records held by the NHS today
may have an indicative market value of several billion pounds to a commercial organisation. We
estimate also that the value of the curated NHS data set could be as much as £5bn per annum and
deliver around £4.6bn of benefit to patients per annum — generated through potential operational
savings for the NHS, enhanced patient outcomes and creation of wider economic benefits to the UK,
generated through ‘big data’, artificial intelligence and personalised medicine.
The curated NHS data set is an intangible asset with a current valuation of several billion pounds
and a realisation of £9.6bn per annum in benefits that could be unlocked following the generation
of insights.
There will be a significant process and technology costs associated with aggregation, cleaning,
curating, hosting, analysing and protecting the transformation of these raw data records into a
consolidated longitudinal patient-level data set. The costs associated with this data transformation
requires further research and clearly could impact data set valuation. To ensure success the NHS
and the UK government will need to partner with companies that can help unlock these valuable
patient insights.
It is critical that the analyses and innovations adhere with medical ethics and research regulations.
Patients have to be informed and need to be confident that their data is being used for their own
and public good, and that their privacy and rights are safeguarded.
Ultimately, analysis and insights generated from this unique NHS data set can help the UK
Government achieve its health priorities on prevention, care and costs, place the NHS and the UK
at the forefront of health care innovation, and make the NHS the envy of the world.
1
1
Patient data
as a desirable
intangible asset
Unlocking the power of health care data to fuel innovation Given these considerations, health care providers need
in medical research and improve patient care is at the heart to recognise the patient data they accumulate for what it
of today’s health care revolution. It is made possible by truly is — a valuable intangible asset desired by multiple
advances in health care technologies and data digitalisation, stakeholders, and a treasure trove of information detailing
and achieved through the analysis of real-world evidence health, wellness, illness and the associated care pathways.
contained within a given patient’s medical records. When
curated or consolidated into a single longitudinal data set,
Realising the value of NHS data
these patient-level records will trace a complete story of a
patient’s health, wellness, diagnosis, treatments, medical Amongst the providers holding significant data assets,
procedures and outcomes. the UK’s NHS is the single largest integrated health care
provider in the world. Its patient records covering the
Unlocking insights from such medical records would be entire UK population from birth to death is an estimated
of high value and benefit to all key stakeholders in the 55 million lives.
health care ecosystem — patients, health care providers,
payers, pharmaceutical companies and medical device Ultimately, the analytical insights derived from this unique
manufacturers. Consider a few examples: data set can help the UK Government achieve its health
priorities on prevention, care and costs. In the near term,
• Deeper disease understanding monetising access to the data set can yield a compelling
• Treatment effectiveness and safety or pharmacovigilance revenue stream for the NHS — especially until other
• Increased in the quality of care, such as faster and early countries amass and put on the market their digitised
diagnosis patient records of comparable depth and scale.
• Observation of real-world patient outcomes and clinical We estimate that the 55 million patient records held by
*
pathway efficiency the NHS today may have an indicative market value of
• Improved patient access to therapies several billion pounds to a commercial organisation.
• Evidence of cost-effectiveness and outcomes to inform This figure is based on benchmarks derived from the
value-based payments following:
• Efficient target identification for new treatments and
• The trading values of health and life sciences companies
medicines
with significant patient data assets
• Clinical trial design for target populations and reduced
• Recent M&A transactions involving such companies and
time to market for new therapies
data
• Operational and cost-effectiveness of health care delivery,
and workforce planning
• Enablement of personalised medicine — right treatments A patient record is defined as quantitative or qualitative information about a
*
patient of a single patient’s medical history and care held by the NHS in any
for the right patients format. The term ‘patient record’ includes electronic patient records, electronic
• Socioeconomic impacts of health, wellness and health care medical records, records of episodic care and supplementary registries such as
genomic or other ‘omic’ profiling with or without associated medical histories.
2
This estimate will likely increase as the proportion of The worth of data — three questions to
genomic records in the NHS data set increases, with consider
sequencing costs falling over time. Increases will occur
as the volume records available to curate and link grows. Such considerations raise the pressing tactical question of
We estimate that in the next 5 years the total number how one places a value on patient health data — NHS-owned
of patient records associated whole genome sequences or any other. Quantifying this value is a complex endeavour
will increase from the current 100,000 to over 5 million. dependent on multiple variables:
In addition to this increase in genomic profiles, the UK’s • Each major stakeholder from across the health care
Department of Health and Social Care has announced that ecosystem will have a different view of what value is — its
patients will be asked to give consent for their genome data perceived value would not necessarily be commercial in
to be securely analysed by approved researchers. nature.
We expect additional value to be unlocked from patient • There are significant costs that need to be taken into
records by emerging scientific and medical innovations. account associated with curating, processing and
**
These include panomic profiling where advances in analysing electronic patient records.
molecular biological technologies are leading to detailed
mapping of a patient’s panomic profile. • Challenges around interoperability and diverse data
contents need to be addressed.
To unlock further value of its patient records, the NHS
would need to create longitudinal patient-level data sets by As a starting point, one needs to explore the following
combining all care settings (e.g., primary, secondary and three stage-setting questions:
social care) and available genomic profile data for each 1. What is each stakeholder group’s key objective or the
patient. It should be noted, however, that there would be a issue it needs to resolve?
significant cost associated with this data transformation.
2. How does each stakeholder intend to use the data set?
We estimate that the value of the curated NHS data
set could be as much as £5bn per annum, delivering 3. Is the proposed use ethical and in line with the
around £4.6bn worth of benefit to patients per annum, Government’s regulatory guidance?
if this effort is undertaken. The value to patients would Once these perspective-setting questions are considered,
come from potential operational savings for the NHS, two approaches can be used to quantify the data’s value.
enhanced patient outcomes and the generation of wider Both approaches focus on the fundamental drivers of
economic benefits to the UK. health care data value.
**
Panomics refers to a range of molecular biology technologies including:
1. Proteomics — the analysis and quantification of proteins
2. Metabolomics — the quantification of multiple small molecule types, such as amino acids
3. Microbiomics — the examination of all the microorganisms of a community simultaneously
4. Epigenomics — the characterisation of reversible modifications of DNA
5. Transcriptomics — the examination of RNA levels genome-wide
6. WGS — Whole Genome Sequencing
3
2
Drivers of health
care data value
The health care industry and health care providers have In fact, the understanding of the associations and trends
historically generated vast amounts of data as a result within health data is already being used to study health at
of clinical and regulatory requirements. However, the both the population level, and the personalised, individual
collection and storage of this data was generally in paper level. Companies are utilising health and data analytics to
form. Recently, the accessibility and amount of digitalised build predictive models that identify and predict high-risk
data has exploded, growing at an exponential rate. patients, and are developing early intervention plans to
improve health and care proactively.
Though all data can have value, this will vary depending
on its characteristics. To measure the monetary value Analytics and predictive models are helping address one
of data, one must first understand how these different of the biggest costs to health care — the treatment of
characteristics increase or decrease its value. chronic diseases. The main factors that influence the risk
of developing chronic diseases include medical history,
On the basis of our experience and the review of published
demographic or socioeconomic profiles and comorbidities.
literature on the subject, the characteristics that affect the
Thus, by aggregating data related to these factors,
value of health care data can be broadly categorised into
predictive analytics can help identify those individuals who
four main pillars as presented in figure 1. Please refer to
are at risk of chronic conditions.
the appendices for more details.
Advanced data mining, such as artificial intelligence (AI)
and natural language processing, will help reveal patterns
and trends that will allow health care providers and other
stakeholders to improve diagnoses, treatments and
operational effectiveness.
“
Having genetic information and longitudinal data
allows us to paint the clearest picture on patient
epidemiology, progression, and overall experience.
Director, Field Health Outcomes, Pharmaceutical company
4
Figure 1. Framework for analysing characteristics that impact the value of a data set
Nature • Data type (patient, payer, product, provider and scientific research)
1 • Data availability or time frame (contemporaneous vs. historical with time lag)
• Exclusivity or scarcity (available from a single source vs. multiple)
• Granularity or detail (aggregated vs. transaction level)
• Source or seller (original source/generator of the data vs. reseller)
Data quality, maturity and • Raw (unorganised with potential data gaps and inconsistencies)
2 embedded analytic insight • Curated (i.e., organised and easy to work with)
• Aggregated longitudinally for the same patient or record
• Analysed with descriptive statistics, insights and predictions or forecasts provided
Complexity of data capture • Source or the party generating the data

3 • Accessibility of data (open source vs. paid)
• Data capture (auto-captured vs. collected with human intervention)
Use/application • Use and potential impact

4 • Exclusivity (exclusive licence vs. data being offered to multiple buyers)
• Limitations on use
• Usage by other businesses or competitors
5
3
Valuation
approaches
to consider
The value of patient data is proportional to its content, and comes from the value of insights and innovations that can be
unlocked from the curated data.
We see two primary approaches to quantifying this value.
A ‘top-down’ or market-based approach A ‘bottom-up’ or income-based approach

1 This approach estimates the value of a data
2 This approach quantifies value based on the
set based on its profile — size, content, nature, economic benefit to be generated from the
availability, maturity, uniqueness and the quality curated data set under its proposed use. Value
of the data — and on the observed market can be expressed in terms of economic uplift and
pricing of functionally comparable data assets. commercial opportunity:
Market pricing is analysed by calculating the
Economic uplift: The benefits to patients include
implied ‘per-record’ valuation multiples of
faster and more accurate diagnosis, improved
comparable data assets or, in the absence of
outcomes and new products, improved care
such information for ‘pure’ data assets, valuation
pathways, and operational efficiency. Benefits to
multiples of companies with significant patient
the UK economy include socioeconomic benefits or
data assets. By applying the benchmark per-
increased gross value added.
record valuation multiples to the NHS data set,
one can estimate the data’s value based on Commercial opportunity: This refers to the
recent ‘market transactions’. incremental profit or licensing income generated
from a successful business deployment of insights
gained from the data. For example, if the insights are
deployed by a life sciences company, such business
deployment may translate into reduced drug
development time and the launch of a novel therapy.
6
“
We estimate that the 55 million patient records
held by the NHS today may have an indicative
market value of several billion pounds to a
commercial organisation, based on recent M&A
transactions and public company valuations.
7
4
Application of
the market-based
approach
Simply put, the application of the market-based approach To quantify the rise in value, one should then compare
calls for a thoughtful analysis of two matters: the characteristics of the longitudinal data set with
those that have been the target of recent transactions
1. Strengths and weaknesses of the subject data set
or public company trading activity within the relevant
relative to comparable data sets that are being traded in
data subcategory — e.g., primary care, secondary care or
the market
genomic profile. This can be achieved by reviewing publicly
2. Per-record valuation multiples inferred from comparable available data, private databases, company reports and
market transactions or company valuations press releases.
The application of this framework is illustrated in figure 2,
with a focus on the NHS data set.
Understanding the NHS data and its ecosystem
We assume that there are 55 million longitudinal primary
Analysing the NHS data set’s strengths and care records or electronic health records (EHRs) in the UK.
These are held by the patients’ general practitioner (GP)
weaknesses — the power of longitudinal data surgery where the patients are registered and seen regularly.
Each fragmented and isolated data set will have a certain
In the ideal world, every patient would have a single EHR,
value, albeit low. True value is generated when these data
all records would be digital and centrally housed, and all
sets are aggregated, curated, processed and linked to
health care platforms and data sets would be standardised
create a longitudinal data set.
and interoperable. But in reality, the NHS health care
The transformed data set will contain a patient’s unique data ecosystem is made up of multiple fragmented and
phenotypical and genomic information that has been geographically dispersed data silos. Each data set contains
captured over a period of time. Panomic data generated multiple data types and formats.
from patient’s tissue and fluid samples will become
These isolated data sets are often organised at the patient,
increasingly available over the next few years as molecular
population and organisation levels rather than solely at the
biology technologies advance. Greater value can be
patient level. The data contents and complexity vary, as
unlocked when the data set undergoes analysis using
does the data quality. Many records contain both structured
basic or advanced analytical methodologies to identify
and unstructured data detailing demographics, diagnosis,
commercially valuable insights. A range of diagnostic,
procedures, medical images, treatments, outcomes and
descriptive, predictive and prescriptive analytical
genomic data.
methodologies should be applied to maximise insight
generation and extract as much value as possible. It is The extent to which these patient records are fully
important to bear in mind that there will be a significant digitalised has to be established, as many of the earlier
process and technology costs associated with aggregation, patient records will be non-digital and maintained within
cleaning, curating, hosting, analysing and protecting these paper files, with attached test results, medical images and
transformed data sets. x-rays, or stapled electrocardiogram (ECG) cardiac traces.
8
Figure 2. Framework for estimating the value of the NHS data set using the market-based approach
1. Understand the NHS

EHRs or electronic medical records (EMRs)
data set linked to episodic care records and genomic
profile; all at patient level
2. Characterise and evaluate

EHR/EMR records linked at patient level to
individual and linked episodic care records Genomic profile and
associated patient history
data set using qualitative
framework with the Episodic care records
EHRs and EMRs covering secondary
following categories covering primary care and tertiary or
specialty care
3. Analyse comparable market

transaction and select
valuation multiples
Central NHS data sets (local, regional and national level) include
prescriptions, medicine expenditure, supply chain, procurement, financial,
resourcing and workflow or process details, and for local authorities, data
4. Apply valuation multiples sets will include the details of social care
to NHS data set
5. Estimate NHS data

set value and identify
necessary limitations
9
We must assume that in 2019, most patients will have an There are separate records that detail social care that
EHR, although how far back the data is digitalised is a key patients are receiving and are collated by local authorities
question. Higher-level and less detailed extracts are held by from the social care providers.
the local clinical commissioning groups (CCGs) and these
Data sets (both digital and paper) are held by NHS data
are used primarily for health and financial reporting.
controllers — e.g., NHS Trusts for secondary or tertiary care
In addition to the 55 million primary care records, there and by CCGs or GPs for primary care.
are approximately 23 million patient-episodic care records
Other health care-related data sets are created for
or electronic medical records (EMRs); these are captured
specific needs — e.g., Genomics England has sequenced
when patients receive secondary or tertiary (or specialist)
the DNA of 100,000 patients with cancer, rare diseases
care, and are held by the hospital trusts. These EMRs will
and infectious diseases, and collate these sequences
include details on admissions and medical interventions in
for the purpose of understanding the disease, earlier
accident and emergency departments.
Figure 3. Transformation of population-level data to patient-level data
The health ecosystem is made up of multiple types of real-world data (patient-, population- and organisation-level
data). Moving from siloed raw data to structured patient-level data generates unique data sets curated over time and
captures longitudinal patient data combining long-term health or medical records with episodic periods of treatment.
These can be analyses to generate insights and provide value to patients, providers, payers, pharmaceutical companies
and the wider economy as a whole.
Population-level data (CCGs and trust-level data) Patient-level data (anonymised and linked
longitudinal, grouped by disease, phenotype,
demographics, care setting, etc.)
Primary care Secondary care
Genomic data Patient A Primary Secondary Genomic
Patient medical Patient medical Data

and health records records transformation Patient B Primary Secondary
Operational and Operational and Moving from siloed raw Patient C Primary Secondary Genomic
process data process data data to structured patient-

level data generates unique
data sets curated over time Patient D Primary Secondary
Financial data Financial data and captures longitudinal
patient data combining
NHS centralised data, including regional
long-term health or medical
records with episodic Operational and
periods of treatment Financial data
process data
Curation of Cleansed, ordered, Repeatable data Data linkages and Feature engineered Data usage to solve
raw data mined and extraction and scale data for analysis business problem
normalised data cleansing
Descriptive Diagnostic Predictive Prescriptive

(Explore the data) (Understand the data) (Use the data) (Provide value from the data)
Increasing data value
10
“There is little commercial use of genomic data on its own,

but it becomes powerful when wedded to clinical outcomes.
For example, the combination of EMR and genomic data
is essential for biomarker identification, enhancing the
predictive power of models.
Director, Roche, Sequencing Solutions
diagnosis and, ultimately, personalised care. In addition One should also be aware of the increase in complexity (and
to data controllers, there are also data processors — e.g., cost) of the longitudinal data when records and episodes of
commissioning support units (CSUs), the Office for National care are linked; and that the operational risks of updating
Statistics and Public Health England. Their role is to provide and refreshing the vast data set increases accordingly.
population-level analyses of health, wellness, disease and
In the case of NHS records, resolving the operational
cost or operation effectiveness. There are some examples
challenge of linking the raw data is aided by the availability
where the data processors may be both controllers and
of unique patient identifier numbers. The entire population
processors — e.g., NHS Digital, CCGs, local authorities and
of the UK has a unique 10-digit patient identifier, the
social care providers.
NHS number, that stays with individuals from birth until
death. This NHS number will be associated with any
Curating and linking fragmented patient data sets
care the patient receives in their lifetime, and as such
into a single longitudinal patient record
will cover primary, secondary, and tertiary health and
For each patient, there is both health and medical medical records; any social care received; and any data
information contained in EHRs (detailing primary care) and collected through health-related registries that patients
EMRs (detailing episodes of secondary or tertiary care). have volunteered information to — e.g., by participating in
This is illustrated in figure 3. Recently, genomic profiling Genomics England’s 100,000 Genomes project.
(genotype or whole genome sequencing) is becoming more
Per-record valuation multiples — analysing comparable
frequent as sequencing costs are reduced. This has the
market transactions or company valuations
potential to vastly increase complexity of analyses and
the data set size. Additional information from registries A number of public and private companies with assets that
and social care may also be available. Linking data sets, include patient data can be identified by a detailed review
both geographically and longitudinally, along the patient of publicly available and transactional databases, and from
pathway offers the potential of a significant increase in these the implied or observed pricing for patient data can
the number and value of insights that can be derived to be analysed. The implied business valuation multiples on
improve patient care and accelerate medical research. a per-record basis can be calculated for data they own or
can access. These valuations can be aligned to the data
Interoperability between health care solutions and data categories we described and illustrated in the framework
platforms is pivotal to success. Part of the challenge (and above in figure 2.
cost) of unlocking value is developing the most suitable
processes and data workflows to aggregate, clean and Caution must be taken in analysing transactions — the
convert these fragmented and isolated data records into a EHR or EMR data is often acquired as part of a broader
single high-quality, analysable data set. Protection of these set of capabilities, such as advanced analytics, including
data files is vital and an appropriate cybersecurity solution AI. Hence, isolating the value of data becomes a challenge
is critical. and there will be differences in the nature, quality,
complexity and application of data in each data set and
with each transaction.
11
Electronic patient records (EHR and EMR) approximately two million oncology practices. Based purely
From our review, it is clear that the EHR or EMR data, on the acquisition cost and number of records, one can
in isolation, has lower value than genomic data. Recent estimate a value of US$950 per patient record; however,
transactional activity in these patient data categories has the data content is of the premium quality, and the value
been motivated by acquirers seeking to broaden their includes access to Flatiron’s health care technology, which
capabilities. The range of estimated values per patient connects oncologists on a shared platform.
record and company-level analyses are listed in the
In some cases, the investment information is not in the
appendix — supplement.
public domain, such as the Great Point Partners’ acquisition
Examples where we are unable to quantify the implied of Corrona, which held an observational registry with data
data value vs. other capabilities, include the IMS Health- on approximately 65,000 patients.
Quintiles merger in 2016, which included 15 petabytes
We can make estimates from other deals where EHR data
of prescription, promotional, claims and approximately
sets were acquired for testing and refining AI solutions.
530 million patients’ records or EHR data, leading to
IBM spent approximately US$4bn on acquisitions of Phytel,
the combined entity (IQVIA) having a market capital of
Explorys, Merge and Truven in order to enhance its Watson
approximately US$20bn and 50,000 employees.
Health (AI) offerings. The individual acquisitions allow us to
Again, Roche’s acquisition of Flatiron was a combination of estimate that IBM paid Merge, e.g., approximately US$30
patient data, an EHR and an oncology platform that brings per patient record for an anonymised pool of approximately
all available data together for review by the oncology board. 30 billion medical images, including x-rays, computerised
In this case, Roche acquired real-world data or evidence, axial tomography (CAT) and magnetic resonance imaging
collated from Flatiron’s network of approximately 280 (MRI) scans.
oncology communities, which provided patients’ records of
Figure 4. Observed values (£) per patient record based on recent data transactions (January 2019)
Typical estimated values (£) per patient record based on recent data transactions
EHR or EMR data Genomic data Deals combining Partnerships

has an estimated aggregators have genomic and combining genomic
value of greater raised capital from phenotypic data and phenotypic
than £100 per private equity and from patient data from patient
patient record. pharmaceutical records have been records have been
companies valued between valued between
at estimated £1,000 and £1,000 and
valuations of over £5,000 per patient £5,000 per patient
£1,500 per DNA record. record.
sample.
12
Genomic profiles and medical histories across the pharmaceutical industry to expedite drug
Genomic data is recognised by experts as a key component discovery and focus on personalised medicines. Evotec
for future drug discovery and also personalised medicine. and Celgene have entered into a third long-term strategic
However, this data is of limited value in isolation without drug discovery and development partnership in the field of
longitudinal health and medical data. targeted protein degradation. In this collaboration, Evotec
and Celgene will leverage Evotec’s panomics platform
There is currently a scarcity of large data sets due to
in order to identify drug targets. There is currently no
the high historical costs of sequencing and this scarcity
available data on panomics record transactions and our
is driving the valuation of per deoxyribonucleic acid
value estimation for the NHS data set does not include
(DNA) sample to be less than US$1500 (i.e., each
value that could be unlocked by curation of patient records
individual patient profile). Some examples are national-
with panomic profiles.
level initiatives, such as the UK Biobank and Genomics
England. These are often open-access for academic Consolidated genomic profile and electronic patient
researchers and clinicians to analyse as well as contribute records (EHR and EMR)
to, but require a subscription fee when pharmaceutical There are relatively few historical examples of genotypic
companies request access. An example of this is the nine and phenotypic data being linked, and this provides the
months’ exclusive data access granted by UK Biobank NHS with a unique opportunity. The NHS holds EHR and
to Regeneron and GSK: the first data set has just been EMR records from all care settings and has the ability to
released to health researchers from this collaboration link these to the Genomics England 100,000 Genomes
offering an unprecedented ‘big data’ resource to enhance Project.
understanding of human biology and aid in therapeutic
Transactions or partnerships for these linked data
discovery. Details of the estimated record value are
sets command high valuations per patient record
described in the paragraph below. The range of estimated
(approximately US$1,000–US$5,000 per record) as they
values per patient record and company-level analyses are
provide an holistic perspective on patient cohorts. These
listed in the appendix — supplement.
longitudinal phenotypic and genotypic patient-level data
Observed valuations of consumer-facing genomic profiling sets are inherently valuable for drug discovery and wider
providers, such as 23andMe and Ancestry.com provide population health analysis due to the scale and depth of
useful valuation benchmarks for genomic data. It should data provided for analysis. An example of this is Roche’s
be noted that these companies mainly analyse genome acquisitions in 2018, which gave it analytic capabilities
samples rather than whole genome sequences and the and ownership of phenotypic and genotype data for its
genetic analysis offered combines genetic profile and the oncology franchise.
persons’ associated health and family history (note, this
Examples where we can provide an estimated value
does not include EHR or EMR data).
for any transaction or partnership investment, include
With this caveat, Silverlake’s and GIC’s investment Amgen‘s success in the public bidding held by the Icelandic
in Ancestry.com indicates an estimated value of Parliament to create a genealogical database linking
approximately US$1700 per genotype record, based genotypic to phenotypic data. Data was sourced from the
on 1.5 million DNA samples sourced from consumers Icelandic population and contained 600 genetic samples,
and linked to more than 20 billion digitised historical from a population of 320,000, and we can estimate a value
records and over 80 million family trees. The research of approximately US$1,300 per record for this partnership.
partnership between 23andMe and Genentech indicates
A second example is Genomics Medicine Ireland (GMI)
an approximate value of US$5,000 per patient
which has a project objective to undertake one of the
record, highlighting that a premium is placed on more
world’s largest whole genome sequencing programme,
‘medicalised’ consumer genotype profiles with associated
creating 600 new jobs in five years; here, we estimate a
medical history (in this case, focusing on Parkinson’s
value of US$1,000 per record. The partnership between
disease). The 23andMe data was sourced directly from
Regeneron and GSK to sequence 500,000 participants
over one million consumers, with 23andMe recruiting and
in the UK biobank project (with nine months of exclusive
profiling the genotypes of 12,000 Parkinson’s patients
access to the sequenced data) yields an estimate of
(and 1,300 parents and siblings), while a subset of 3,000
US$3,000 per record.
Parkinson’s patients samples underwent whole genome
profiling. One final example is the investments that lead to Roche’s
acquisition of Flatiron as it demonstrated how value
Panomic profile and medical histories
increases as phenotypic and genotype data sets are
Panomics, which analyses patient tissue and body samples
combined, with a further value-add when additional
in more detailed data than genomics, is gaining traction
13
Figure 5. Summary of the estimate ranges of market services are added on top (such as health care platforms
value for NHS data sets to collect patient data and advanced analytics capabilities).
Foundation Medicine and Flatiron jointly launched a
Benchmark valuation multiples are estimated clinico-genomic database following the integration of
primarily using the observed valuations of companies Foundation Medicine’s genomic profiling assays with
with assets that include patient data. The data is Flatiron’s EHR outcomes data, building an initial database
understood to be curated (clean) and actively used. of approximately 20,000 patients as a subset of the
The range of estimated values per patient record and 400,000 cancer patients in the EHR database. The
company-level analyses are listed in the appendix — objective was to help researchers and biopharmaceutical
supplement. (Analyses as of January 2019). partners accelerate drug development in oncology.
Foundation Medicine and Flatiron were both acquired by
Roche at a cost which an estimate of US$6,000 per record
for Foundation Medicine genomic data and US$950 per
record for Flatiron EHR data. This is based on the price
Roche paid for data, oncology EHRs, health care platform
Typical estimated values (in £) per patient record and analytical services, which was approximately US$2.4bn
based on recent data transactions for Foundation Medicine and US$2.6bn for Flatiron, which
would equate to US$250, per clinico-genomic record.
Curated or analysed data in commercial use
Considerations to take into account whilst
estimating the value of patient records held by
Market capitalisation Post-money valuations the NHS
of selected publicly of private companies and For the UK’s NHS data set, we have made the assumption
traded companies pharma collaborations that the patient records fall into the following categories
and in each, we know the approximate number:
EHR Genomics
• Around 55 million longitudinal primary care records
£21–£192 £208–£2,024 • Around 23 million episodic specialty care patient records
Episodic records Oncology • Around 100,000 DNA codes of patients with cancer, rare
diseases and infectious diseases collated by Genomics
£30–£619 £40–£5,549
England
Genomics General
For a simple estimate, we would take the number of patient
£2,354–£4,773 £8–£486 records in each category and multiply by the relevant
estimated value for each patient record (on the basis of
Drug development recent data transactions described in figure 5 and the text
above). However, we need to take into account that the
£0.20–£35
value of these data sets increases when they are linked.
Population data There should be :
£3–£21 • Approximately 23 million linked records where the

patient’s primary and specialty care records are curated
We derived the following benchmark valuation • Approximately 100,000 linked patient records that
multiples using primarily the observed valuations detail primary care, episodic specialty care and genomic
of companies with assets that include patient data. profiles based on overlaps and our ability to link using
The data is understood to be curated (clean) and the patient’s NHS number (for this, a premium value per
actively used. record should be applied)
As mentioned above, we have listed down two factors
to consider for a simple estimate by stating ‘should’
as completed digital records for some patients may be
unavailable, or there could be missing data. It is essential
that when estimating the total value of the patient records
held by the NHS, one should consider the following factors
14
affecting the value of patient data (these are described in 6. Variability in benchmark valuation multiples (analysis
more detail in the appendix — supplement): in figure 5 above was done by relying on business
valuations as opposed to raw data asset valuations; i.e.,
1. Available digitised patient records — not all patient
they rely on valuations of public and private companies
historical data will be digitalised, and an estimate of the
that possess significant patient data assets)
period that corresponds to the EHR or EMR must be
made 7. Understanding the minimal statistically significant data
sample size needed for research and unlocking value
2. Nature of the data, including demographics, health,
from the patient records
wellness, diagnosis, treatments, medical procedures
and outcomes 8. Diminishing returns on increasing data set size
(additional improvements in data algorithms and
3. Data access availability, availability of substitutes or
corresponding extraction of insights will start becoming
alternative data sources, granularity and levels of detail
marginal in nature once the user begins exceeding
4. Data quality, maturity and embedded analytical the minimal statistically significant data sample size
insights (factors contributing to this include whether milestone)
the data is raw, unstructured or structured, curated,
Additionally, a key consideration when estimating value is
quality assured, aggregated across individual records,
the uniqueness of the data: are similar data sets available
aggregated longitudinally at the patient-level, or
from other sources, or countries? The more unique
analysed by descriptive statistics or predictive analytics)
the data, the greater will be the value, especially if a
5. Overlap across three data categories and synergy value commercial option is also desirable.
• Amongst the 55 million patient records, most Finally, caution must be taken when directly applying fair
patients would have received multiple episodes of market values for each data category. As noted previously,
secondary or tertiary care. often when patient data is acquired, especially for data
contained within EHR or EMR, the transactional deal will
• But 23 million patients or fewer would have patient-
often include acquisition of additional capabilities such
level data sets with both primary and secondary or
advanced analytical or health care platforms in addition to
tertiary care since there are only 23 million episodic
the curated patient data.
care records.
• Lastly, fewer than 100,000 records will contain
genomic and phenotypic data, because there
currently only 100,000 patients who has undergone
genome sequencing; this number is expected to rise
to five million in five years.
“
We have estimated that the value to
the NHS could be as much as £5bn
per annum and deliver around £4.6bn
p.a. of benefit to patients. The value to
NHS would be in terms of operational
savings for the NHS, enhanced patient
outcomes and generation of wider
economic benefits to the UK.
15
5
Application
of the income
approach
The NHS, through its analysis of the patient-level data it quantitative outputs. It also takes into account the costs
holds, can unlock significant operational savings, enhanced of transformation of the data sets, as the services and
patient outcomes, and wider economic benefits through technology costs will all contribute to the gross value added
the application of ‘big data’, AI and personalised medicine. (GVA) for the local economy. The longitudinal patient-level
If the NHS were to curate these data sets and generate data set and full use of use of ‘big data’ will enable better
longitudinal patient-level records, it would be able to unlock decision-making and an ability to explore AI and further
significant value, although there would be a significant cost advancements in personalised medicine.
associated with this data transformation and generation of
The economic uplift approach is an evidence-based review
insights.
methodology that draws on a combination of expert
We estimate that the value to the NHS could be as consultation, literature review and assessment of working
much as £5bn per annum and deliver around £4.6bn papers which enable the development of an impact
per annum of benefit to patients. The value to the NHS framework to underpin economic modelling and analysis
can come from operational savings for the NHS, enhanced (figure 6). These analyses are split into three categories
patient outcomes and generation of a wider economic of impact — ‘big data’, AI and personalised medicine —
benefits to the UK. The economic analyses reveals a and takes into account the ‘customers’ of the insights
number of short-, medium- and long-term impacts which generated: in this case, the NHS and three industry sectors
yield financial and economic benefits as a result of more (i.e., pharmaceutical, medical technology and insurance
cost-effective and informed spending. companies). The outputs are estimated for each of the
‘beneficiaries’ of the insights: patients, the NHS and the
wider economy. It is important to consider the realisation
Framework for valuing patient data based on
of the economic benefits as each of the three categories
economic benefits to patient, the NHS and has differential impact across short-term (less than three
the UK economy years) and medium- or long-term (greater than three years)
Estimating the economic value (i.e., the benefits) that timeframes for each of these categories (see figure 7).
would be generated from the curation, processing and ‘Big data’ are large, static data sets that can be analysed to
analyses of a large data set provides an alternative provide real-world insights on trends and patterns. Uniting
approach to assessing the value of data. It also allows all NHS data sets into a single analysable patient- level data
us to include the wider benefits to the economy in set would create opportunities to improve NHS productivity
addition to patient benefits and economic uplift for and delivery of care. Our analyses estimate that productivity
the NHS. This approach generates both qualitative and savings could be equivalent to 2% of the annual NHS budget
16
Figure 6. Schematic of impact framework for estimating the economic benefits and uplift
This impact framework demonstrates a number of short-, medium- and long-term impacts which yield financial and
economic benefits as a result of more cost-effective and informed spending (QALY — quality-adjusted life year; GDP —
gross domestic product).
NHS longitudinal Customers of Short-term outcomes Medium/long term Final long-term

data set the data set (0-3 years) outcome (over three years) impacts
NHS • Improves • Improved NHS First-order effect Second-order effect

• NHS curates
• Real-world operational productivity and
patient records evidence (RWE) efficiency lower costs Reduce disease Reduce economic
to generate • Makes better • Prevention of burden for burden for patients
longitudinal clinical decisions disease and patients (QALYs) (more productive,
patient-level (Dx) episodes of ill health increased
data sets, • Makes better use of • NHS resource consumption)
existing medicines allocation changes Impact on health
combining all
or treatments — time spent by care costs (NHS
care settings spend)
staff on activities
and available • Patients receive Third-order effect
genomic the right treatment
profiles more quickly Positive GDP impact
(employment,
• NHS applies output)
‘big data’ Pharmaceutical • Supports clinical NHS medicines and
analytics, AI companies trials and post- treatment mix
and advances • RWE market surveillance changes Public finance
personalised • Clinical advisory • Better clinical trial
impact
medicines service design
• Medicine discovery • New or more
efficacious
medicines
discovered
Medical technology New or refined

companies medical equipment Drive direct
• New and improved discovered investment
medical devices
Insurance Insurance develops Companies in health,

companies new products or adjust insurance, life
• Identify patient existing products sciences gain extra
cohorts and revenue or reduce
population costs
trends to develop
propositions
17
and equate to £2.7bn per annum in productivity savings. quickly and accurately. Our economic modelling reveals
Moreover, these savings could start to be realised as soon that the economic benefits to the NHS of applying AI to the
as one year after NHS data set curation (figure 6). The longitudinal patient-level NHS data set could be realised in
NHS could use a big NHS data set to identify best practice the medium term (from three years onwards) with savings
care pathways to be adopted. This would improve NHS to the NHS of £1.7bn per annum and benefits to patients of
productivity and make its limited resources go further. £2.5bn per annum (figure 6).
A unified NHS data set would be necessary to achieve
These benefits will continue long into the future as the data
integrated care, currently a policy goal, according to
set grows, and the breadth and depth of analyses continue
several influential voices in the health sector. The existence
to reveal insights and suggest solutions to help the NHS
of this unique data set could attract investment from life
address clinical and financial problems. Early adoptions of
sciences sector in terms of research and development
AI within the health care ecosystems have been promising.
activities. Patients and the wider economy are also likely
Applying AI to the NHS data set will yield insights that can
to benefit from better health and better care through
be used to improve patient outcomes through improved
improved pathways, and more R&D spending in the UK,
monitoring and better use of medicines. As a follow-on from
but there is, to date, limited evidence on the scale of these
this, AI can help reduce errors, provide better diagnostics
potential benefits.
and improve demand planning. Ultimately, improved patient
The full potential of a single longitudinal patient-level health will also have implications for the UK economy and
NHS data set can be realised following development and consequent positive impacts on public finances. AI could
curation with the application of AI. AI here means the be applied to the NHS data set to make its operations more
application of computational processes that mimic human efficient and more accurate, including reducing adverse
intelligence (e.g., reasoning, learning and adaptation) to reactions to medicines, improving the speed and accuracy
big, real-time data sets. AI has been used in health care of diagnostics, and, from an operational perspective,
for early detection, diagnosis, decision-making, treatment leading to more effective and accurate theatre scheduling
and research. AI could support the NHS to work more and capacity planning.
Figure 7. Summary of economic value to the NHS benefit to patients
The value would be in terms of operational savings for the NHS, enhanced patient outcomes and generation of a wider
economic benefits to the UK.
The NHS data set could unlock significant operational savings for the NHS, enhanced patient outcomes, and
wider economic benefits through the application of big data, AI and personalised medicine.
AI Personalised medicine (PM)

‘Big data’
NHS — £1.7bn p.a. NHS — £0.6bn p.a.
NHS — £2.7bn p.a.
Patients — £2.5bn p.a. Patients — £2.1bn p.a.
Total NHS saving = £5.0bn p.a. and total patients saving = £4.6bn p.a.
Short-term benefits Medium-term benefits Long-term benefits realisation

realisation from year 1 realisation from year 3 from year 10
18
Personalised medicine tailors treatments to the revenue gains from introducing new medicines. There
characteristics of a particular patient. Personalised may also be benefits from clinical trials that are becoming
medicine will be enabled if the NHS creates a patient- faster and more cost-effective by better targeting of
level longitudinal data set containing both phenotypic and participants through genomics. By helping pharmaceutical
genotype-linked data. Genomic data can be used for drug companies to target the right participants for clinical trials,
discovery, to give more precise diagnoses, make more the NHS could reduce the length and cost of trials in the
efficient use of medicines, and increase the quality and UK. Evidence indicates that better targeting could reduce
length of life. Through its promise of more targeted and the time it takes to get a medicine to market by 1.7 years.
effective treatment, personalised medicine could have Moreover, a healthier and more productive workforce would
significant impacts on the NHS, patients, the life sciences have economy-wide benefits and positive public finance
industry and the economy. Our analyses estimate that implications. The GVA benefits to the UK from a healthier
the economic benefits of personalised medicine could be and larger workforce could be as high as £1bn per annum.
£0.6bn per annum in savings for the NHS, with a patient
Realisation of the economic benefits. The application
benefit of £2.1bn per annum (figure 6). The benefits of
of ‘big data’, AI and PM will require a level of set up and
personalised medicine would be realised only in the longer
initial investment for the benefits to realise. The economic
term, over a period of 10 years or more.
analyses highlight that ‘big data’ will materialise first,
The NHS would be able to dedicate fewer resources to followed by AI, and then PM last but with potentially the
ineffective treatments for non-responders, and patients most significant long-term impact (figure 7).
should be treated more efficiently with medicines or
technologies identified as likely to be effective for them.
Early evidence indicates that personalised medicine could
both save the NHS from wasting money on ineffective
treatments, and lead to improved morbidity and reduced
mortality. Life sciences companies could benefit through
Figure 8. Realisation of the economic benefits for ‘big data’, AI and personalised medicine resulting from the creation
of a NHS longitudinal patient-level data set and the analytical identification of insights
Big data AI PM
Short term NHS bodies procure technology Data set created and leads to Pharma begins process of
(less than 3 they need useful insights on best practice developing new medicines
years)
Medium term Best practices and learnings Best practices in short run New medicines go through
(3-9 years) spread through NHS implemented across NHS clinical trials
Long term All potential benefits realised Continued discovery of useful New medicines come to market
(greater than 10 insights
years)
19
6
Concluding
comments
The NHS data set holds immense value to all stakeholders Care will need to partner to ensure success, to access the
across the health care ecosystem — patients, the NHS, the expertise in technology and clinical analytics needed to
life sciences sector, the UK and society as a whole. unlock valuable insights and drive this transformation.
The curated NHS data set is an intangible asset with Analysis of the consolidated NHS data set has the potential
a current valuation of several billion pounds and a to unlock ground breaking data-driven innovations, allow
realisation of £9.6bn per annum in benefits (i.e., the the development of apps for patients and doctors, provide
NHS benefits worth £5bn per annum and the patient clinical decision support tools supported by intelligent
benefits worth £4.6bn per annum) that could be algorithms and advanced analytical methodologies. These
unlocked following the generation of insights. innovations have the potential to improve diagnosis,
treatment, experience of care, efficiency of the system and
There will be significant process and technology costs
overall outcomes for patient, the NHS, public health and
associated with aggregating, cleaning, curating, hosting,
the wider health and care system.
analysing and protecting the transformation of these
raw data records into a consolidated longitudinal patient- It is critical that the analyses and innovations adhere with
level data set. Estimating these costs should be a next medical ethics and research regulations. Patients have to
step in light of the multifaceted requirements needed to be informed and need to be confident that their data is
convert isolated NHS patient records into a longitudinal being used for their own and public good, and that their
curated patient-level data set. The need and associated privacy and rights are safeguarded. It is the duty of the
costs to address digitalisation and data transformation, NHS to consolidated patient records into the highest quality
cybersecurity, infrastructure as well as solving the longitudinal patient level data set, and then to use these
interoperability issues must not be under estimated. data-driven technologies, in a safe, ethical, evidenced and
transparent way. This will ensure that the insights and
A significant and appropriately sized team of highly skilled
real-world evidence generated can be used confidently
data scientists, who have access to cutting edge IT and
to improve the quality and safety of care. Analysis and
analytical technologies, will be required to generate the
insights generated from this unique NHS data set can
insights needed to produce the value as described. The
help the UK Government achieve its health priorities on
up-front and maintenance costs will come in the form
prevention, care and costs.
of people, process and technology, and one should not
underestimate the scale of the task or its requirements. This health care data transformation can place the NHS and
The NHS and the UK’s Department of Health and Social the UK at the forefront of health care innovation.
20
“
Having genomic data alone doesn’t allow you
to answer interesting and valuable questions.
Director, Bioinformatics & Biostatistics, pharmaceutical company
21
7
Appendix
Factors affecting value of data Another example of value evident in granularity is
genome sequencing data. By using predictive analytics
Nature of data on patient-level genome sequencing data, one can
identify individual patients that are at high risk for certain
Data access and availability conditions. Providers can then take preventative actions
Data access and availability refer to how quickly data is with the patient to decrease the likelihood of developing
available to the buyer after it is created. Contemporaneous or worsening a medical condition. According to the 2016
data is one that is immediately available for analysis. McKinsey report: The Age of Analytics — Competing in a
Data-Driven World, the impact from personalised medicine
The age of the data can increase or decrease its value, but
resulting from such data analytics could result in US$2tn to
in general, rapid access to fresh data is thought to be of
US$10tn in health care savings on a global basis.
more value, as it provides real-time insights and enables
timely informed decisions. Source or seller
Exclusivity or scarcity Data that is purchased directly from the party that
generated it will generally be more valuable as compared
When data is only available from a single source, the party
with the data purchased from a reseller. This is due to data
holding the data acquires monopoly pricing power over the
accuracy.
data. The larger the business disruption caused by losing
access to such data, the more significant will be the seller’s Altered or inaccurate data may result in misleading
ability to charge a premium price. conclusions and faulty decisions. It is presumed that data
purchased from a reseller is at a higher risk of having
Access to data can be limited via physical barriers, expense
been tampered with or altered, when compared with data
of the collection process, business strategy or contractual
purchased from the original source.
restrictions.
It should be noted, however, that many third-party data
Granularity of data
aggregators and resellers may clean and organise data to
In general, the increased granularity or specificity of a make it easier to analyse. The incremental increase of data
data set increases its value by allowing the user to extract value resulting from this analytical process is addressed in
additional insights absent from aggregated data sets. 5.1.2 of the framework, and should be weighted against
In health care context, granularity translates into patient- the increased risk of data alteration.
and transaction-level data (e.g., medical product dispense,
patient encounters, change in status, etc.).
By analysing dispense data at a transaction level,
pharmaceutical companies can hone in on dispensing
patterns in specific zip codes — which would not be possible
using aggregated data.
22
Data quality, maturity and embedded analytics In the case of government or public data sets, a monetary
insights payment set for accessing may not reflect full market
value of the data across multiple user groups. When using
Data maturity
such payments as price benchmarks in an fair market
The concept of data maturity addresses the progression of value (FMV) assessment, careful consideration should be
a data set along the data analytics process. given to understanding the price formation of such access
More often than not, raw data sets are disorganised and restrictions and their consistency with true market value of
complex, or lack structure. To extract insights from it, one the data.
must clean, organise and analyse it. For example, government data, such as census results,
allows for ready resource allocation, capital investment
Complexity of data capture planning, policy-making and monitoring numerous other
Data capture benefits. In the G20 countries, open-sourced data is
estimated to be valued between US$700bn and US$950bn
The value of data is also affected by the effort and risk
per annum according to a 2014 study by Lateral
associated with capturing it. In general, data that is auto
Economics.
captured as a by-product of everyday business processes
would generally be perceived as less valuable — on a
Use or application
cost-approach basis — than the data that is collected in a
separate process with human intervention and by including Use and potential impact
or not including other incremental resources involved in the The ultimate determinant of data value is its use and
process. application — its power to inform decisions. These
Accessibility of data may range from ongoing daily operating decisions of a
pharmaceutical company, to reducing time to market
Generally speaking, any restrictions on access to data that
for new medicines and improving patient adherence to
is perceived by multiple parties as useful to their decision-
therapy.
making processes will increase its value. Hence, the value
of data that is freely available to any interested party would The more uses the data set has to multiple buyers, and
be lower than the value of the data that calls for certain the more risky and impactful are the resulting decisions
qualifications from the potential buyer, ranging from a on the user’s long-term strategy and operations, the more
monetary payment or through qualifications, such as valuable it is to the buyers.
citizenship or a need to submit a use case request.
23
Factors affecting value of data
Observed or implied value in £ per data record

Q1 Median Q3 n=
Multiples implied by market capitalisation of selected benchmarks of publicly traded companies
EHR 21 42 4,990 9
Episodic care records 30 54 39,379 7
Genomics 2,354 2,719 6,826 3
Multiples implied by post-money valuations of private companies and by pharma collaborations
Genomics 208 1,153 2,338 6
Oncology 40 379 20,905 6
General 8 27 10,255 21
Drug development 0.2 17 51,577 6
Population data 3 9 45 4
Subset of above multiples for companies with combined data sets 4,318 6,297 10,255 2
Appendix — supplement
Company-level indicative estimates of market value — observed or implied price (in £) per patient record
Public companies and EHR (also known as primary care)

Indicative estimates of market value for NHS data sets
Market-based approach: guideline public companies benchmarking
Analysis as of 01 January 2019
# Company Ticker EV # records EV/record Domain
GPCs w/assets that include access to EHR data [1] [2]
1 Computer Programs and Systems, Inc. NasdaqGS:CPSI £373,769,930 18,000,000 £21 EHR
2 OptumInsight (UnitedHealth Subsidiary) NYSE:UNH £9,118,903,409 216,000,000 £42 EHR
3 Cerner Corporation NasdaqGS:CERN £13,336,402,955 100,000,000 £133 EHR
4 AllScripts Healthcare Solutions, Inc. NasdaqGS:MDRX £3,074,770,356 16,000,000 £192 EHR
5 athenahealth, Inc. NasdaqGS:ATHN £4,249,839,591 106,000,000 £40 EHR
6 NextGen Healthcare, Inc. NasdaqGS:NXGN £781,472,941 240,000,000 £3 EHR
7 Winning Health Technology Group SZSE:300253 £2,321,668,843 n/a n/a EHR
8 CompuGroup Medical Societas Europaea DB:COP £2,085,836,362 n/a n/a EHR
9 B-SOFT Co.,Ltd. SZSE:300451 £1,076,947,286 n/a n/a EHR
10 EMIS Group plc AIM:EMIS £548,698,594 40,000,000 £14 EHR
11 Medasys S.A. ENXTPA:MED £30,774,801 n/a n/a EHR
12 Pharmagest Interactive SA ENXTPA:PHA £673,709,001 135,000 £4,990 EHR
13 Alibaba Health Information Technology SEHK:241 £7,335,594,444 28,000,000 £262 EHR

Limited
14 Rayseach Laboratories OM:RAY B £300,776,801 n/a n/a EHR
First quartile £579,951,196 18,000,000 £21
Median £1,581,391,824 40,000,000 £42
Third quartile £3,956,072,282 106,000,000 £192
Notes:
EV=Estimated value
[1] Source: SP Capital IQ. Amount shown for Optum represents a portion of UnitedHealthcare’s EV attributed to Optum Insight by prorata revenue contribution.
[2] Source: Latest 10K filings or equivalent annual report, and/or company websites. For more detail on data source please refer to supporting schedules.
24
Appendix — supplement (continued)
Public companies and episodic records or EMR (also known as secondary, tertiary or specialty care)
GPCs w/assets that include access to episodes of care or [1] [2]

transaction data
1 IQVIA Holdings Inc. NYSE:IQV £26,030,825,543 530,000,000 £49 Episodic
2 Inovalon Holdings Inc. NasdaqGS:INOV £2,378,296,346 240,000,000 £10 Episodic
3 Medidata Solutions Inc. NasdaqGS:MDSO £3,067,675,815 3,800,000 £807 Episodic
4 Tabula Rasa HealthCare Inc. NasdaqGM:TRHC £1,040,371,508 n/a n/a Episodic
5 Veeva Systems Inc. NYSE:VEEV £9,408,263,746 n/a n/a Episodic
6 Guardant Health Inc. NasdaqGS:GH £2,756,527,429 70,000 £39,379 Episodic
7 Precipio Inc. NasdaqCM:PRPO £6,437,121 n/a n/a Episodic
8 Syneos Health Inc. NasdaqGS:SYNH £5,368,017,071 100,000,000 £54 Episodic
9 WuXi AppTec Co. Ltd. SHSE:603259 £8,988,144,252 n/a n/a Episodic
10 Evolent Health Inc. NYSE:EVH £1,162,121,376 2,700,000 £430 Episodic
11 Inovalon Holdings Inc. NasdaqGS:INOV £2,378,296,346 240,000,000 £10 Episodic
12 Craneware plc AIM:CRW £614,529,127 n/a n/a Episodic
, First quartile £1,131,683,909 3,250,000 £30
Median £2,567,411,888 100,000,000 £54
Third quartile £6,273,048,866 240,000,000 £619
Notes:
EV=Estimated value
Public companies and genomic profiles

GPCs w/assets that include access to genomic data [1] [2]
1 Invitae Corporation NYSE:NVTA £598,237,180 220,000 £2,719 Genomics
2 Myriad Genetics Inc. NasdaqGS:MYGN £1,785,846,015 n/a n/a Genomics
3 Natera Inc. NasdaqGS:NTRA £647,694,600 n/a n/a Genomics
4 Genomic Health Inc. NasdaqGS:GHDX £1,690,872,942 850,000 £1,989 Genomics
5 Exact Sciences Corporation NasdaqCM:EXAS £5,720,119,415 n/a n/a Genomics
6 Berry Genomics Co. Ltd SZSE:000710 £1,254,923,166 n/a n/a Genomics
7 Illumina Inc. NasdaqGS:ILMN £33,902,607,451 n/a n/a Genomics
8 OPKO Health Inc. NasdaqGS:OPK £1,459,479,999 n/a n/a Genomics
9 Twist Bioscience Corporation NasdaqGS:TWST £682,625,186 100,000 £6,826 Genomics
First quartile £682,625,186 160,000 £2,354
Median £1,459,479,999 220,000 £2,719
Third quartile £1,785,846,015 535,000 £4,773
Notes:
EV=Estimated value
25

Private companies w/assets that include access to genomic data
Market-based approach: private companies benchmarking
# Target Buyer Type of deal Year Domain Deal value Approx % Post-money Number of Implied
transacted valuation or records EV or
acquired price record
Private companies w/assets that include access to genomic data [1] [2]
1 23 and Me GSK Collaboration 2018 Genomics £236,658,000 12% £1,972,150,000 5,000,000 £394
2 NextCODE Health WuXi Acquisition 2015 Genomics £51,275,900 100% £51,275,900 350,000 £147
PharmaTech
3 Helix DFJ Growth Investment 2018 Genomics £19,721,500 8% £257,215,692 n/a n/a
Mayo Clinic
4 Fabric Genomics Artis Investment 2016 Genomics £18,143,780 19% £95,530,946 50,000 £1,911
Ventures
5 Color Genomics General Investment 2017 Genomics £70,555,638 29% £239,868,660 N/A N/A
Catalyst
6 DeCODE Genetics Amgen Acquisition 2012 Genomics £327,376,900 100% £327,376,900 140,000 £2,338
7 DeepGenomics Khosla Investment 2017 Genomics £10,255,180 78% £13,173,962 n/a n/a
Ventures
8 Genomics plc Invesco Investment 2018 Genomics £33,896,328 30% £113,449,983 800,000 £142
Perpetual
9 Nebula Genomics ARCH Investment 2018 Genomics £3,392,098 16% £20,612,912 10,000 £2,061
Inc Venture
Partners LP
10 Verge Genomics SUNU Investment 2018 Genomics £25,243,520 23% £108,878,457 n/a n/a
Inc. Ventures BV
11 Pathway Genomics International Investment 2015 Genomics £31,554,400 29% £107,545,284 n/a n/a
Corp. Business
Machines
Corp.
First quartile £73,403,423 72,500 £208
Median £108,878,457 245,000 £1,153
Third quartile £248,542,176 687,500 £2,024
Notes:
EV=Estimated value
[1] Source: SP Capital IQ, DowJones VentureSource.
[2] Source: EY research and/or company websites.
26

Private companies w/assets that include access to oncology data
# Target Buyer Type of deal Year Domain Deal value Approx % Post-money Number of Implied
transacted valuation or records EV or
acquired price record
Private companies w/assets that include access to oncology data [1] [2]
1 Flatiron Roche Acquisition 2018 Oncology £1,498,834,000 100% £1,498,834,000 2,200,000 £681
2 Cota Health IQVIA Investment 2018 Oncology £31,554,400 18% £172,831,337 n/a n/a
3 Tempus Baillie Gifford Investment 2018 Oncology £87,563,460 6% £1,577,720,000 220,000 £7,171
T. Rowe Price
Revolution
Growth New
Enterprise
Associates
(NEA)
4 Precision Health SymphonyAI Investment 2017 Oncology £15,777,200 21% £75,841,000 £76
AI Group
5 Paige.AI Breyer Capital Investment 2018 Oncology £19,721,500 21% £93,251,141 25,000,000 £4
6 Freenome Andreessen Investment 2017 Oncology £56,797,920 32% £175,868,448 n/a n/a
Horowitz
7 Foundation Merck Acquisition 2015 Oncology £1,893,264,000 45% £4,180,958,000 200,000 £20,905
Medicine
8 Cambridge Cancer AME Cloud Investment 2018 Oncology £3,549,870 15% £23,665,800 n/a n/a
Genomics Inc. Ventures
9 Mendel Health Inc. Boostrap Labs Investment 2017 Oncology £1,577,720 11% £13,954,933 n/a n/a
10 Oncora Medical Inc. Undisclosed Investment 2017 Oncology £1,490,945 9% £16,700,166 n/a n/a
investor
11 PathAI Inc 8VC Investment 2017 Oncology £11,785,568 17% £70,445,198 n/a n/a
12 SkinVision BV PHS Capital B.V. Investment 2018 Oncology £5,995,336 19% £31,775,281 1,156,359 £27
First quartile £29,747,911 415,000 £40
Median £84,546,071 1,078,180 £379
Third quartile £506,609,836 1,939,090 £5,549
Notes:
EV=Estimated value
27

Private companies w/assets that include access to drug development data
v Target Buyer Type of Year Domain Deal Approx % Post-money Number of Implied
deal value trans valuation or records EV or
acted acquired price record
Private companies w/assets that include access to drug development data [1] [2]
1 twoXAR SoftBank Ventures Investment 2018 Drug £7,888,600 13% £60,103,243 n/a n/a
Andreessen development
Horowitz OS Fund
2 Nuritas European Grant 2018 Drug £23,665,800 28% £84,076,699 n/a n/a
Investment Bank development
3 Insilico WuXi AppTec Investment 2018 Drug £4,733,160 7% £70,279,537 2,000,000 £35
Medicine Deep Knowledge development
Ventures
4 Recursion Square 1 Bank Bill Investment 2018 Drug £47,331,600 26% £178,905,559 5,200,000 £34
Pharmaceu & Melinda Gates development
ticals Foundation
5 Berg n/a n/a n/a Drug n/a n/a n/a n/a n/a
development
6 BenevolentAI Woodford Investment 2018 Drug £90,718,900 5% £1,668,438,900 50,000,000,000 £0

Ltd. Investment development
Management
7 Covance LabCorp Acquisition 2014 Drug £5,157,692,898 100% £5,157,692,898 100,000 £51577
development
8 Covance LabCorp Acquisition 2014 Drug £5,157,692,898 100% £5,157,692,898 30,000,000,000 £0

development
9 Evidera Pharmaceutical n/a n/a Drug n/a n/a n/a n/a n/a
Product development
Development Inc.
10 Atomwise Monsanto Investment 2018 Drug £35,498,700 20% £180,609,497 600,000,000 £0
development
11 Genetech Roche Investment 2009 Drug £36,918,648,000 100% £36,918,648,000 N/A n/a
development
12 Antidote Merck Global Investment 2017 Drug £8,677,460 16% £55,220,200 N/A n/a
Technologies Health Innovation development
Inc. LLC
13 Innoplexus AG HCS Beteiligung Investment 2016 Drug n/a n/a n/a n/a n/a
sgesellschaft mbH development
14 NuMedii Inc. Claremont Creek Investment 2015 Drug £1,577,720 11% £14,538,690 n/a n/a
Ventures development
15 Pharnext SA Truffle Capital Investment 2016 Drug £24,375,774 31% £79,201,544 n/a n/a
SAS repurposing
16 Qrativ Inc. Mayo Medical Investment 2017 Drug £6,547,538 17% £39,663,881 n/a n/a
Ventures development
17 Zephyr Health Anju Software Inc. Acquisition 2018 Drug n/a n/a n/a n/a n/a
Inc. development
18 e-Therapeutics Octopus Ventures Investment 2007 Drug £1,049,184 4% £29,448,144 n/a n/a
development
19 ExScientia Ltd. Evotec AG Investment 2017 Drug £11,832,900 31% £38,307,042 n/a n/a
development
20 ExScientia Ltd. GT Healthcare Investment 2019 Drug £20,510,360 n/a n/a n/a n/a
Capital Partners development
LP
28

Private companies w/assets that include access to drug development data
# Target Buyer Type of deal Year Domain Deal Approx Post-money Number of Implied
value % valuation or records EV or
trans acquired price record
acted
Private companies w/assets that include access to drug development data [1] [2]
21 Healx Ltd. Amadeus Capital Investment 2018 Drug £7,888,600 21% £37,825,837 n/a n/a
Partners Ltd. development
22 Numerate Inc. Foundation Investment 2014 Drug £6,310,880 15% £42,661,549 n/a n/a
Capital development
23 Owkin Inc. Cathay Investment 2018 Drug £14,199,480 16% £87,547,683 n/a n/a
Innovation development
24 ReviveMed Inc. Rivas Capital LLC Investment 2018 Drug £1,183,290 13% £8,843,121 n/a n/a
development
25 Alector Inc. OrbiMed Advisors Investment 2019 Drug £138,642,145 13% £1,105,776,616 n/a n/a
LLC development
26 BioAge Labs Felicis Ventures Investment 2017 Drug £8,598,574 20% £43,276,860 n/a n/a
Inc. development
27 Causaly Inc. Marathon Investment 2018 Drug £788,860 25% £3,131,774 n/a n/a
Venture Capital development
28 GTN Ltd. Octopus Ventures Investment 2018 Drug £2,208,808 30% £7,446,838 n/a n/a
development
29 Prellis Biologics SOSV LLC Investment 2017 Drug £1,459,391 29% £5,064,481 n/a n/a
Inc. development
30 Spring First Round Investment 2018 Drug £14,199,480 17% £84,108,253 n/a n/a
Discovery Inc. Capital development
31 XtalPi Inc. China Life Investment 2018 Drug £36,287,560 22% £165,810,483 n/a n/a
Investment development
Hodling Co. Ltd
32 Lantern Bios Partners Investment 2017 Drug £2,918,782 35% £8,440,802 n/a n/a
Pharma Inc. development
33 Resonant Mercury Fund Investment 2017 Drug £1,609,274 17% £9,229,662 n/a n/a
Therapeutics Ventures III LP development
Inc.
34 Envisagenics Notion Capital Investment 2017 Drug £2,761,010 13% £21,291,331 n/a n/a
Inc. development
35 Notable Labs Builders VC LLC Investment 2017 Drug £7,888,600 14% £57,886,547 n/a n/a
Inc. development
36 Celsius Third Rock Investment 2018 Drug £51,275,900 30% £168,326,947 n/a n/a
Therapeutics Ventures LLC development
Inc.
First quartile £25,369,738 2,800,000 £0
Median £57,886,547 302,600,000 £17
Third quartile £167,068,715 22,650,000,000 £35
Notes:
EV=Estimated value
29
Appendix-Supplement (continued)
Private companies w/assets that include access to population health data
Market approach: private companies benchmarking
# Target Buyer Type of deal Year Domain Deal Approx % Post-money Number of Implied
value trans valuation or records EV or
acted acquired price record
Private companies w/assets that include access to population health data [1] [2]
1 Health Fidelity UPMC Investment 2015 Pop Health £15,193,444 22% £68,859,589 200,000,000 $0
2 Lumiata Sandbox Investment 2018 Pop Health £8,677,460 10% £85,827,968 20,000,000 $4
Industries,
Khosla Ventures,
Intel Capital
3 Innovaccer WestBridge Investment 2018 Pop Health £27,610,100 20% £136,275,565 10,000,000 $14
Capital
4 Ginger.io Inc Kaiser Investment 2014 Pop Health £15,777,200 18% £89,330,506 2,000,000 $45
Permanente
Ventures
5 HealthCrowd Healthy Ventures Investment 2018 Pop Health £5,679,792 12% £48,270,343 n/a n/a
First quartile £68,859,589 8,000,000 £3
Median £85,827,968 15,000,000 £9
Third quartile £89,330,506 65,000,000 £21
Notes:
EV=Estimated value
30
Appendix-Supplement (continued)
Private companies w/assets that include synergies between multiple data types

Market approach: private companies benchmarking
# Target Buyer Type of deal Year Domain Deal Approx % Post-money Number of Implied
value trans valuation or acquired records EV or
acted price record
Private companies w/assets that include synergies between multiple data types [1] [2]
1 DeCODE Genetics Amgen Acquisition 2012 Genomics £327,376,900 100% £327,376,900 140,000 £2,338
2 The Advisory Board Optum Acquisition 2017 General £1,025,518,000 100% £1,025,518,000 100,000 £10,255
3 Ayasdi Kleiner Investment 2015 General £43,387,300 20% £214,656,695 n/a n/a
Perkins
Khosla
Ventures
First quartile 110,000 £4,318
Median 120,000 £6,297
Third quartile 130,000 £8,276
Notes:
EV=Estimated value
31
Acknowledgments Authors
Sincere thanks from EY teams to the following Chris Wayman, PhD

individuals who assisted in the creation of this report. EY UK&I Life Sciences & Health Care — Valuation,
Economics & Modelling
Pamela Spence, EY Global Health Sciences and cwayman@uk.ey.com
Wellness Industry Leader, and Rob Moody, TAS Chief +44 20 7951 6469
Innovation Officer and EY UK&I TAS Analytics Leader,
provided strategic guidance and feedback on the
narrative throughout its development. Natasha Hunerlach, CFA, CFE, MRICS
EY Americas Life Sciences & Health Care — Valuation,
Peter Arnold, EY UK&I TAS Economics Lead, David Economics & Modelling
Gudmunsen, EY UK&I TAS Valuations Lead, Natasha natasha.hunerlach@ey.com
Hunerlach, EY US TAS Valuations Lead, and Chris +1 404 817 5942
Wayman, EY UK&I TAS Business Modelling and Decision
Analytics, all provided key input on concepts of the
publication.
Rajiv Mehrotra, Himanshu Chaudhry, Hannah Kalozdi
and Wade Robbins, under the guidance of Natasha
Hunerlach, performed industry research, detailed fair
market value pricing for health care data sets, and
conducted analysis associated with estimating the value
of NHS data set value that underpinned the market-
based approach.
George Agathangelou, Emily Hutchison, Nida
Mahmud and Sabina Mamedova, under the guidance
of Peter Arnold, conducted health care research
and analysis to underpin the income-based approach
quantifying the economic benefits associated with
unlocking insights from NHS patient records.
James Evans reviewed the publication for quality and
consistency. Dennis Ryan created the memorable
accompanying visuals.
32
Contacts
EY GLOBAL HEALTH SCIENCES AND WELLNESS EY LIFE SCIENCE & HEALTH CARE VALUATIONS,
INDUSTRY LEADER ECOMONICS & MODELLING
Pamela Spence Peter Arnold

pspence2@uk.ey.com parnold@uk.ey.com
+44 207 951 3523 +44 20 7951 5587
David Gudmunsen
dgudmunsen@uk.ey.com
EY GLOBAL TRANSACTIONS ADVISORY SERVICES +44 20 7951 1519
Peter Behner Natasha Hunerlach

peter.behner@de.ey.com natasha.hunerlach@ey.com
+49 30 25471 12467 +14048175942
Rob Moody
rmoody@uk.ey.com
+44 20 7951 2515
Chris Wayman
cwayman@uk.ey.com
+44 20 7951 6469
33
EY | Assurance | Tax | Transactions | Advisory

About EY
EY is a global leader in assurance, tax, transaction and advisory
services. The insights and quality services we deliver help build trust
and confidence in the capital markets and in economies the world
over. We develop outstanding leaders who team to deliver on our
promises to all of our stakeholders. In so doing, we play a critical role
in building a better working world for our people, for our clients and
for our communities.
EY refers to the global organization, and may refer to one or more, of

the member firms of Ernst & Young Global Limited, each of which is
a separate legal entity. Ernst & Young Global Limited, a UK company
limited by guarantee, does not provide services to clients. Information
about how EY collects and uses personal data and a description of the
rights individuals have under data protection legislation are available
via ey.com/privacy. For more information about our organization,
please visit ey.com.
About EY Health Sciences and Wellness

The rise of the empowered consumer, coupled with technology
advancements and the emergence of digitally focused entrants, is
changing every aspect of health and care delivery. To retain relevancy
in today’s digitally focused, data-infused ecosystem, all participants
in health care today must rethink their business practices, including
capital strategy, partnering and the creation of patient-centric
operating models.
The EY Health Sciences and Wellness architecture brings together a

worldwide network of more than 20,000 professionals to build data-
centric approaches to customer engagement and improved outcomes.
We help our clients deliver on their strategic goals; design optimized
operating models; and form the right partnerships so they may thrive
today and succeed in the health systems of tomorrow. We work across
the ecosystem to understand the implications of today’s trends,
proactively finding solutions to business issues and to seize the upside
of disruption in this transformative age.
© 2019 EYGM Limited.

All Rights Reserved.
EYG no. 003378-19Gbl

1906-3182693 US CSG
ED none
This material has been prepared for general informational purposes only and is
not intended to be relied upon as accounting, tax or other professional advice.
Please refer to your advisors for specific advice.
The views of third parties set out in this publication are not necessarily the views
of the global EY organization or its member firms. Moreover, they should be
seen in the context of the time they were made.
ey.com/lifesciences
34

Ey Value of Health Care Data v20 Final

Uploaded by

Copyright:

Available Formats

Ey Value of Health Care Data v20 Final

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ey Value of Health Care Data v20 Final

Uploaded by

Copyright:

Available Formats

What is the main focus of the paper?

What is the main focus of the paper?

What are the two primary approaches for quantifying the value of data discussed in the paper?

What are the two primary approaches for quantifying the value of data discussed in the paper?

Realising the

How can we place

1. Patient data as a desirable intangible asset 2

A market-based approach, calculating the implied “per record”

An income-based approach, which quantifies value based

Director, Field Health Outcomes, Pharmaceutical company

Complexity of data capture • Source or the party generating the data

Use/application • Use and potential impact

A ‘top-down’ or market-based approach A ‘bottom-up’ or income-based approach

1. Understand the NHS

2. Characterise and evaluate

3. Analyse comparable market

to NHS data set

5. Estimate NHS data

Figure 3. Transformation of population-level data to patient-level data

Patient medical Patient medical Data

process data process data data to structured patient-

Descriptive Diagnostic Predictive Prescriptive

Increasing data value

“There is little commercial use of genomic data on its own,

EHR or EMR data Genomic data Deals combining Partnerships

£3–£21 • Approximately 23 million linked records where the

NHS longitudinal Customers of Short-term outcomes Medium/long term Final long-term

NHS • Improves • Improved NHS First-order effect Second-order effect

Medical technology New or refined

Insurance Insurance develops Companies in health,

Figure 7. Summary of economic value to the NHS benefit to patients

AI Personalised medicine (PM)

Short-term benefits Medium-term benefits Long-term benefits realisation

Director, Bioinformatics & Biostatistics, pharmaceutical company

Factors affecting value of data

Observed or implied value in £ per data record

Public companies and EHR (also known as primary care)

# Company Ticker EV # records EV/record Domain

GPCs w/assets that include access to EHR data [1] [2]

2 OptumInsight (UnitedHealth Subsidiary) NYSE:UNH £9,118,903,409 216,000,000 £42 EHR

3 Cerner Corporation NasdaqGS:CERN £13,336,402,955 100,000,000 £133 EHR

4 AllScripts Healthcare Solutions, Inc. NasdaqGS:MDRX £3,074,770,356 16,000,000 £192 EHR

5 athenahealth, Inc. NasdaqGS:ATHN £4,249,839,591 106,000,000 £40 EHR

6 NextGen Healthcare, Inc. NasdaqGS:NXGN £781,472,941 240,000,000 £3 EHR

7 Winning Health Technology Group SZSE:300253 £2,321,668,843 n/a n/a EHR

8 CompuGroup Medical Societas Europaea DB:COP £2,085,836,362 n/a n/a EHR

9 B-SOFT Co.,Ltd. SZSE:300451 £1,076,947,286 n/a n/a EHR

10 EMIS Group plc AIM:EMIS £548,698,594 40,000,000 £14 EHR

11 Medasys S.A. ENXTPA:MED £30,774,801 n/a n/a EHR

12 Pharmagest Interactive SA ENXTPA:PHA £673,709,001 135,000 £4,990 EHR

13 Alibaba Health Information Technology SEHK:241 £7,335,594,444 28,000,000 £262 EHR

14 Rayseach Laboratories OM:RAY B £300,776,801 n/a n/a EHR

First quartile £579,951,196 18,000,000 £21

Median £1,581,391,824 40,000,000 £42

Third quartile £3,956,072,282 106,000,000 £192

Appendix — supplement (continued)

# Company Ticker EV # records EV/record Domain

GPCs w/assets that include access to episodes of care or [1] [2]

1 IQVIA Holdings Inc. NYSE:IQV £26,030,825,543 530,000,000 £49 Episodic

2 Inovalon Holdings Inc. NasdaqGS:INOV £2,378,296,346 240,000,000 £10 Episodic