Keywords
ICD-11 - German healthcare legislation - German hospital reform - sleepwake disorders - Alpha-ID-SE - electronic patient record
From ICD-10 to ICD-11
The International Classification of Diseases, the ICD, is the central reference
classification in the family of international classifications of the World Health
Organisation (WHO) [1]. It enables
diseases to be recorded independently of the respective national language. The ICD
is a central pillar of billing systems in many countries. It is used in the context
of pharmacovigilance and quality assurance. It forms the basis for morbidity and
mortality statistics at many different levels. It is used for care management. And
it enables comparisons of morbidity and mortality data across institutions, regions,
countries and historical periods.
The ICD-10, which was adopted in the mid-1990s, has been used worldwide for all these
purposes in recent decades. It had its strengths because it was more comprehensive
than the previous version, but its considerable weaknesses have also become
increasingly apparent [2]. It was not
developed in the digital era, which means that it does not utilise the full
potential of a modern coding system. Fundamental changes to the international
version are barely possible anymore, as the WHO has planned further development for
emergency codes only (e. g., for Covid-19). Internationally, the uncoordinated,
non-compliant expansion of the ICD-10 catalogue by individual nations is a further
problem, as this impairs the international comparability of ICD-coded data.
The ICD-11 [3] aims to eliminate these
deficits. It is nothing less than a “reset of the system”, according to the WHO in
2018 [2]. The ICD-11 is designed as a
relational database which, continues to enable hierarchical coding as before, but
can additionally map a variety of relationships between diseases, symptoms,
functions and body sites (see info box). Unlike in ICD-10, an element can be
assigned to several higher-level elements.
Info box: The ICD-11 as an agile, learning system with ontological infrastructure
[4]
Info box: The ICD-11 as an agile, learning system with ontological infrastructure
[4]
-
Few master codes and many extension codes enable precise coding via
digital recording
-
Foundation: Standardised basis that is to be further developed jointly
and can incorporate current medical and cultural developments
-
Hierarchical organisation into medical units (multidimensional, i. e.
diseases, disorders, injuries, external causes, signs and symptoms), which
can optionally be further defined with attributes (body site, body system
and causal mechanism)
-
An element can be assigned to several higher-level elements
(multi-parenting)
-
Completely electronic, real-time updates
-
Linearisation: Extracts as statistical classification with clear
hierarchical tree structure for coding (single-parenting)
-
Annual update
Objectives and innovations of ICD-11
Objectives and innovations of ICD-11
A distinction must be made between the formal/procedural and medical/content-related
objectives of the ICD-11. Formal and procedural are among the goals of the WHO [5]
[6],
-
To utilise the digital possibilities from the point of care to the
statistical analysis of data to achieve a more flexible and long-lasting
classification system, including annual updates, whereby the established
processes and experience with the annual updates of ICD-10 German
Modification (-GM) were expanded and supplemented for input from the general
public
-
To enable the use of digital tools to support coding in everyday life,
-
To improve data quality,
-
To improve international comparability and
-
To make the ICD-11 more robust and adaptable so that national extensions are
no longer necessary or can take place within the ICD-11 Foundation.
In terms of content and medicine, the previous Chapter 3 of ICD-10 “Diseases of the
blood and haematopoietic organs and certain disorders involving the immune system”
has been divided into Chapter 3 “Diseases of the blood or blood-forming organs” and
Chapter 4 “Diseases of the immune system” [7] in ICD-11. However, the addition of two completely new chapters is
particularly noteworthy. The first is Chapter 7 “Sleep-wake disorders”, which means
that a sleep-wake disorder is no longer classified as a symptom [7].
The second is Chapter 17 “Conditions related to sexual health”, including gender
incongruence.
With these chapters, health conditions that were previously categorised as
psychological so-called F diagnoses in the ICD-10 are given their own
classification. The complex post-traumatic stress disorder (CPTSD) included in the
ICD-11 should also be mentioned. This describes a more complex symptom picture and
is often associated with repeated or prolonged trauma from which it is difficult or
impossible to escape, such as torture, prolonged domestic violence or repeated
physical or sexual abuse in childhood, as well as chronic pain syndromes (see [Fig. 2]), which have been summarised in a
chapter (MG30) for the first time ([Fig.
2])[7]
[9]
[10].
Fig. 2 Chronic pain syndromes in the ICD-10 German Modification (-GM)
and ICD-11 (Fig. based on data from [7]
[8]).
Also new are Chapter 26 “Supplementary Chapter Traditional Medicine Conditions” and
Chapter V “Supplementary Section for Functioning Assessment” [7]. In order to ensure consistency with the
ICF, the International Classification of Functioning, Disability and Health, the ICF
has been incorporated into the ICD-11 Maintenance Platform (e. g., for Intentional
Communication, see [Fig. 3]) [11].
Fig. 3 Implementation of linking with other coding systems in ICD-11
(vs. ICD-10-GM and ICF) using the example of Intentional Communication (Fig.
based on data from [7]
[8]
[11]
[12]
[13]).
Chapter X “Extension codes” is also new. The additional codes can be used to describe
ICD-11 master codes in more detail – for example with regard to anatomical
localisation. It is also possible to add associated information – for example about
secondary diseases – to the master code. Due to its diagnostic breadth, the ICD-11
is ideal for specialist care. Its interoperability with standardised medical
terminologies used in the hospital setting such as SNOMED CT [14] (see example of Aortic Aneurysm, [Fig. 4]), as well as primary care (ICPC-3)
[15] is important.
Fig. 4 Aortic Aneurysm from SNOMED CT – ICD-11 harmonisation (Fig.
based on data from [7]
[16]). The SNOMED CT concept
“Aneurysm of the thoracic aorta” is matched with ICD-11 code BD50.3Z
(“Aneurysm of the thoracic aorta, unspecified”).
Status of implementation in Germany
Status of implementation in Germany
The ICD-11 has been effective and ready for adoption since 1 January 2022, including
in Germany. However, it will still be a few years before it arrives in healthcare.
An important deadline for the implementation of ICD-11, not only in Germany, is
2027: from then on, deaths should be reported to the WHO in ICD-11 coded form. In
Germany, the German translation of ICD-11 has been available since February 2022
[7]. The initial translation process
is largely complete ([Fig. 5]), but
quality assurance by the scientific and medical societies is still ongoing.
Fig. 5 Status of the translation of ICD-11 into German.
To ensure the smoothest possible transition to ICD-11, a transition analysis
sponsored by the Federal Ministry of Health (BMG) has taken place, the results of
which are currently being discussed in the ICD-11 working group of the Board of
Trustees for Classification Issues in Healthcare (KKG). Key recommendations for the
upcoming changeover planning are
-
To ensure the availability of the ICD-11 in German,
-
To create a bidirectional transition between ICD-10-GM and ICD-11,
-
To initiate further specific transition analyses and prospective studies
and
-
To draw up a roadmap for the introduction of ICD-11 in Germany.
Opportunities of ICD-11 from the perspective of research and industry
Opportunities of ICD-11 from the perspective of research and industry
The progress that ICD-11 can potentially bring for health service research is best
illustrated with examples. A typical health service research question in the field
of rheumatology is: “What exactly is the standard of care for rheumatoid arthritis
(RA) in the population of statutory health insurance (SHI)-insured patients?” To
answer this question, it is important to differentiate between seropositive and
seronegative RA and whether the serostatus is known at all. It would be interesting
to know how the medication is selected regarding the different sites of
manifestation of RA and whether and how it is related to secondary diseases of RA.
The possibilities offered by ICD-10 for these analyses are limited. Firstly,
although the ICD-10 recognises seropositive chronic polyarthritis, seronegative RA
only appears much later in the hierarchy and only as a subheading of 'other
chronic polyarthritis' [8]. It is
therefore rarely coded in everyday practice [17]. The ICD-11 offers a clearer structure in this example and makes it
possible to specify the location with the help of X codes and to link concomitant
diseases regarded as associated with RA with the help of a slash [7]. For example, the ICD-11 code
FA20.0&XA86T5 describes a seropositive RA in which the metacarpophalangeal
joints are affected. And FA20.0&XA86T5/CB05.1 would signal that there is also an
interstitial lung disease associated with the RA. In contrast, ICD-10 would only
allow the coding of interstitial lung disease as a parallel code, without the causal
link to RA becoming transparent [8].
The example shows that ICD-11 allows coding at a level of detail that would be a huge
step forward for health service research. However, it also shows how complex the
ICD-11 can be and how great the need is for solutions that support documentation
performed by the end users at the point of care – especially in a digital way – in
the treating institutions. The usage of ICD-11 in outpatient and inpatient routine
care can only be recommended by means of digital applications – but this is exactly
what the ICD-11 was developed for.
From the perspective of the healthcare industry, the advantages of ICD-11 compared to
ICD-10 are similar to those of health service research. Relevant questions for drug
development and drug sales can hardly be answered based on ICD-10 data or
conventional billing data, at least for some diseases. For example, in adults with
lung cancer, only 25% of cases could be clearly identified as small cell or
non-small cell lung cancer based on German health insurance datasets, and this rate
was only achieved when the medication was analysed in addition to the ICD-10 coding
[18]. In addition, ICD-11-based coding
would make the evidence base for calculating the total number of patients affected
in the AMNOG transparent and comprehensible and help to significantly improve
planning.
Opportunities from the perspective of medical care
Opportunities from the perspective of medical care
It is not only research that benefits from the ICD-11, but also the direct care of
those affected. The two newly included chapters “Sleep-wake disorders” and
“Conditions related to sexual health” are an important step forward here. Many
people do not realise it, but classifications can have a very direct influence on
how certain illnesses are perceived in a healthcare system. In the case of
sleep-wake disorders and conditions related to sexual health, this has meant that
conditions in these areas often do not receive the attention they deserve.
This can be illustrated by insomnia disorder, a disorder that affects around 6% of
the adult population [19]
[20]. The chronic form of this disorder,
chronic insomnia, is characterised by difficulty falling asleep and/or sleeping
through the night at least three times a week for at least three months, which is
accompanied by significant distress and/or impairment during the day. Insomnia was
first classified as an independent disorder in the DSM-5 (Diagnostic and Statistical
Manual of Mental Disorders) in 2013 [21]
[22] and was adopted in the
ICSD-3 (International Classification of Sleep Disorders) [23]. Prior to this, insomnia was not
considered a separate disorder, but was largely seen as a symptom of other illnesses
or dismissed as a mood disorder. In the ICD-10, short-term insomnia is shown in
Chapter VI “Diseases of the nervous system” as “Episodic and paroxysmal disorders”.
In contrast, chronic insomnia, like other sleep disorders, is classified in Chapter
V “Mental and behavioural disorders”. Accordingly, patients receive an F diagnosis
when coding, specifically F51.0 “Nonorganic sleep disorders” ([Fig. 1]) [8].
Fig. 1 Insomnia in the ICD-10 German Modification (-GM) and ICD-11
(Fig. based on data from [7]
[8]).
This classification as an F diagnosis is problematic because it is associated with
the stigmatisation of those affected as “mentally ill”. The consequence of the fear
of stigmatisation is that chronic insomnia is not even coded in cases of doubt. The
fact that this actually happens is illustrated by the 2019 health report from the
SHI Barmer [24]. The researchers found a
strong divergence between the frequency of reported sleep problems and the actual
documentation of these problems reported as a medical F51.0 diagnosis: only around
one in four people who reported relevant insomnia had also received a medical
diagnosis in this study [24]. However,
only time will tell whether the new classification system with a separate chapter on
“sleep-wake disorders” will help to catalyse a change in behaviour.
This is scientifically unsatisfactory because chronic insomnia disorder is
systematically underrepresented in statistics as a result. Furthermore, it is also
highly problematic from a healthcare perspective because a formal diagnosis is only
available if it can be adequately coded, based on which suitable and, above all,
reimbursable therapies can then be initiated. The ICD-11 will therefore contribute
to greater visibility in the case of chronic insomnia disorder and therefore
indirectly to better treatment. The same applies to gender medicine [25], the problem of abuse [9]
[10], and to pain medicine [26].
The patients’ perspective
The patients’ perspective
The example of chronic insomnia disorder shows that coding is not just an abstract
billing and healthcare research topic, but can have direct relevance for patients
and their environment, i. e., those affected. Another example that illustrates this
is the approximately 8000 known rare diseases, of which only around 500 are
categorised as separate codes in ICD-10 [8]. Here, too, the ICD-11 will lead to greater visibility [7]
[27]
[28], not least because –
also due to the planned regular updates – more diseases will become codable in the
future, thus appearing in the general statistics and also becoming more accessible
for health service research on the other.
In the case of rare diseases in particular it is not only about visibility, but also
about patient-relevant aspects such as prompt, accurate diagnosis and needs-based
therapeutic and nursing care at various levels. In the immediate care context, it is
often not so much the ICD diagnosis that is relevant, but rather the adequate
treatment of symptoms and/or functional limitations, depending on the diagnostic or
therapeutic context. For this reason, the National Action Plan for People with Rare
Diseases (NAMSE) published by the National Action Alliance for People with Rare
Diseases in 2013 envisages the coding of rare diseases using Orpha codes as measure
19 and a web-based diagnostic tool for primary care providers as measure 20. This
was followed by a national project “Coding of Rare Diseases” from 2013 to 2019, in
which the non-classifying diagnosis code Alpha-ID, which was introduced in Germany
in 2005 based on ICD-10 [29], was
supplemented by Orpha codes. In 2019, the German Federal Joint Committee (G-BA)
decided in its resolution on subsidies for centres for rare diseases [30] that coding with Alpha-ID and Orpha
code (Alpha-ID-SE) [36] is a quality
requirement. The Digital Care and Nursing Modernisation Act [32] then made Alpha-ID-SE coding mandatory
for inpatients from 2023. This is a great success for the field of rare diseases.
However, there is still no obligation to code in the outpatient sector, and medical
documentation and coding are not linked in hospitals either, resulting in a very
heterogeneous structure. To make matters worse, hospitals use different software for
individual components (e. g. patient administration system/ clinical workplace
system/ medical coding software), and the previous ICD-10-GM [8] continues to serve as the basis for
billing using German Diagnostic Related Groups (G-DRGs) [33] and for statistical purposes. The
Alpha-ID is also a national, parallel system to the ICD, which must be constantly
expanded. However, the findings from the Alpha-ID-SE [31] introduction in the inpatient sector
can be regarded as a test run for the changeover from ICD-10 to ICD-11 in Germany
for the field of rare diseases. A coding system such as ICD-11 is of great value,
since it maps rare diseases far more comprehensively than ICD-10, it enables a
better linking of symptoms, diagnoses, functionality and concomitant diseases, and
it can form the basis for digital applications for symptom-related diagnostic
support or software solutions that digitally implement therapeutic pathways. The
expectation for ICD-11 is that it will be easier to implement in different
healthcare contexts in the future. Of particular importance with regard to rare
diseases is the further development of coding in primary care. This is because it is
local care that has to pave the way for the 35 current centres for rare diseases in
Germany with their special outpatient clinics and that plays an important role in
the coordination of care. It therefore appears necessary to switch to a uniform,
standardised system across sector boundaries. The ICD-11 offers this
opportunity.
The reimbursement perspective
The reimbursement perspective
From a payer perspective, the migration from an ICD-10 world to an ICD-11 world is a
major challenge, not least because ICD codes are used extensively in the German
healthcare system for management and reimbursement purposes. This is most obvious
and best known in inpatient billing via G-DRGs, which relies almost entirely on ICD
coding [33]. Less present in the public
perception is the fact that ICD-coded diseases also form the basis for the risk
structure equalisation of health insurance funds in the German Federal Health Fund
[34]. Accordingly, changes in coding
may lead to relevant shifts in cash flows at different levels, which the system must
then deal with; this is not unsolvable but must be considered at an early stage with
regard to changeover planning and ICD-11 roadmaps.
The great opportunity of ICD-11 from the reimbursement perspective lies in the
improved differentiation options in many areas: More precise coding not only
facilitates health service research, as discussed above, but of course also
allocation of resources and expenditure control with the aim of improving care
across health care sectors. At the same time, however, this is also one of the
greatest risks: The potential of ICD-11 will remain untapped if it is not possible
to achieve coding of the necessary quality and depth. Incentives for users, and in
particular physicians, to familiarise themselves intensively with ICD-11 are
currently lacking. In the worst-case scenario, a poorly executed changeover could
even lead to a deterioration in coding quality with considerable consequences in
many areas of the system. This needs to be actively counteracted using digital
aids.
Some fields of action
Given the complexity of the transition from an old to a new classification system, it
seems clear that it is highly advised to address the challenges of such a migration
at an early stage. It is important to identify possible obstacles that could lead to
delays in the implementation of ICD-11. The potential of ICD-11 can only be
leveraged if the changeover leads to more detailed, higher quality documentation.
The challenges – also for health service research – include the latency of
implementation and uniform level of utilisation as well as the required changes in
coding practice. The question is how ICD-11 and thus more focused coding can
contribute to improving healthcare across health care sectors, because for the few
diseases for which coding of disease severity is already possible in ICD-10 (e. g.,
stage according to New York Heart Association (NYHA) in heart failure, or stage of
renal failure), such coding is used in less than half of the cases in the outpatient
setting [35]. Instead, ‘not further
classified’ is often coded. However, it needs to be mentioned that in the outpatient
setting – in contrast to inpatient care [33] – there is no link between ICD coding and reimbursement.
We will therefore conclude by outlining some areas of action that we highly advise to
work on or discuss more intensively. This should be started now, even if the
required implementation for mortality as per the WHO Assembly resolution by 2027 may
currently still seem far away. As in many other areas, forward-looking action will
also pay off here and will help to avoid time-consuming and costly loops in an
already dynamic regulatory environment.
-
Creating a common understanding among all stakeholders
Ideally, all stakeholders in the healthcare system will agree to use the
changeover to ICD-11 to sustainably advance documentation in the German
healthcare system. Outpatient care and outpatient specialist medical care
(Ambulante Spezialfachärztliche Versorgung, ASV, the ‘third sector’) should
also be included here. It is by no means guaranteed that such an agreement
will be reached. In this context, it is worth remembering the so-called
‘crocodile bite debate’ when ICD-10 was introduced [36] and the discussions in
connection with the introduction of coding guidelines in SHI-accredited
medical care [36]. It is clear
that a common understanding, even if it were to be achieved, would not be
enough. But it would be a good basis for all further endeavors.
-
Planning for digital embedding at an early stage
A key success factor for an ICD-11 implementation that goes beyond coding the
codes that are absolutely necessary for billing is the digital embedding of
the coding in the information systems of outpatient and inpatient care. This
must go beyond the provision of a pure thesaurus. Instead, coding
suggestions should be derived as automatically as possible from the standard
documentation, which then only need to be confirmed or discreetly
supplemented by the users. Ultimately, only with such technical
implementation will it be possible to achieve lasting acceptance for “deep”,
high-quality ICD-11 coding that leverages all the possibilities that ICD-11
offers with regard to research, quality assurance and patient management.
The implementation of corresponding tools will not be a definite success. In
addition, this technical implementation must also follow the “document only
once for all purposes” approach and thus embed ICD-11 in the ecosystem of
coding systems for various use cases.
-
Thinking about incentive systems
The incentive for the introduction of ICD-11 comes on the one hand from the
advantages of the new, current and modern classification itself, which
brings an advantage for users simply by using it, and on the other hand from
the successful embedding in IT systems in interaction with other coding
systems such as SNOMED CT ([Fig.
4]) [16], so that users
ideally do not perceive the changeover or coding as a burden or even do not
notice it at all. In this way, most coders could recognise the benefits of
higher-quality coding as significant, which would be preferable to
conventional incentive systems.
-
Involving medical societies more closely
There is a particular need for action when it comes to involving the medical
and scientific societies and their umbrella organisation, the Association of
the Scientific Medical Societies in Germany (AWMF). Medical societies must
be involved in the quality assurance of ICD-11 and its translation, a
process that is already underway and is being coordinated by the BfArM.
However, they are also important contacts for the implementation of ICD-11,
especially at a time when the definition of (ultimately ICD-10/11-based)
quality indicators and their automated analysis are becoming increasingly
important for healthcare policy. The role of the AWMF is also particularly
important with regard to the guidelines, and in particular the National
Health Care Guidelines and the Oncological Guidelines Programme of the
German Cancer Society, German Cancer Aid and AWMF [37], which can help to raise
awareness of the changeover in the medical profession. However, the G-BA is
also called upon here with regard to the disease management programmes [38] and ASV [39].
-
Check regulatory framework for ICD-11 compatibility
Keyword politics: In general, the ICD-11 should already be factored into upcoming
reforms in order to prevent facts being created that have to be revised again in the
course of the ICD-11 introduction. This concerns digitisation projects such as the
readjustment of the electronic patient record (ePA) from the beginning of 2025 [40]. Harmonisation with other coding
systems is necessary to smoothly integrate the ICD-11 into the ecosystem of coding
systems of the ePA. SNOMED CT [16], for
example, is the most comprehensive international health terminology. In order to
avoid burdening users with multiple coding, the principle of “document only once for
all purposes” is also an important goal for the ePA. This common use of coding
systems is also required internationally, as can be seen from a statement at the
last World Health Assembly in May 2023 [41]. However, it also concerns legislative projects that are more distant
from digitalisation, such as the hospital reform, where a future transition to
ICD-11 could at least be considered. Another example is the topic of
present-on-admission indicators [42],
which has been addressed but remained unsolved in healthcare policy since the
noughties. They are used to differentiate between diagnoses already existing at the
time of admission and those acquired in hospital and are relevant for the assessment
of complication rates and the risk assessment of the inpatient case mix, among other
things. Here, too, the question arises as to how to deal with this issue, which has
been postponed several times and is therefore now a priority from the point of view
of some stakeholders, in view of the foreseeable transition from ICD-10 to ICD-11.
On the one hand, it does not seem expedient to “postpone” this important element to
ICD-11, as it has already been postponed for far too long. On the other hand, it is
clear that the transition to ICD-11 will have an impact on these indicators, which
should at least be taken into account in planning.
Conclusion
The introduction of ICD-11 can be a great opportunity for the German healthcare
system on several levels. Firstly, it can lead to significantly more detailed
coding, which will enable a more comprehensive utilisation of the data. Secondly,
the ICD-11 will modernise the disease catalogue in several areas. Diseases that
previously did not appear in the ICD at all, including many rare diseases, will in
future be able to be mapped and coded in the regular system. In addition, sleep-wake
disorders and conditions related to sexual health will be given their own headings
and thus lose the stigma that was previously attached to them due to their
historical categorisation as F diagnoses.
In order to realise the medical, scientific and healthcare policy potential of
ICD-11, the topic should be given high priority at the level of healthcare policy,
specialist societies, supervisory authorities and the software industry. With a view
to the planned latest introduction date for mortality in 2027, it is necessary to
determine which measures should be taken in which order to ensure the simplest
possible coding at the point of care from the outset. In the best case this will be
perceived by the users themselves as progress and not as another additional
bureaucratic imposition, which will help to ensure that resources can be allocated
sensibly in the German healthcare system and thus that care in the interests of
those affected and their social environment will also be guaranteed in the
future.
