The taxonomy and classification of diseases is an area that is arguably as old as medicine itself. During the 17th and 18th centuries, emerging disease classifications were mainly symptom based (1, 2). Well-known examples are the classifications of Sauvage, Linnaeus, and Cullen. In Genera Morborum from 1763, Swedish botanist and physician Carl Linnaeus classified diseases in 11 categories, one of them being the painful diseases – Morbi Temperati Nervini Sensationis Dolorosi. The consequence of categorizing diseases based on symptoms was of course that diseases with different etiologies and pathophysiologies were often incorrectly lumped together. When French statistician Jacques Bertillon presented the precursor of what is now the International Classification of Diseases (ICD), he instead “adopted for main headings the anatomical site rather than the nature of disease” (2) thus overlooking or disqualifying pain as disease category. As of today, the structure of ICD remains overall strongly influenced by this anatomical and topographical structure. However, it is interesting to note that the eleventh version (ICD-11), adopted by the World Health Assembly in 2019, for the first time contains a structured section about chronic pain – thereby somehow reconnecting with Linnaeus' idea of pain as a separate diagnostic category.

The need for a pain classification system has long been recognized (3), and it has been argued since decades that defective and inconsistent pain taxonomies hamper the development of pain research (4). In an ideal classification system, as for instance the periodic table in chemistry, the different categories are mutually exclusive and exhaustive (5). Although this ideal will barely ever be achievable in pain medicine, the International Association for the Study of Pain (IASP) previously issued an extensive, expert-based multidimensional classification system with five axes: region of the body; organ system involved; temporal characteristics; intensity and time since onset; etiology (5). This resulted in a five-digit code, reflecting the complexity and heterogeneity of pain conditions. As of today, in countries such as Sweden where ICD-11 has not yet been implemented (it is presently being translated), diagnoses are still made according to ICD-10. Pain-relevant diagnoses are scattered throughout the ICD-10 manual and are often based on either anatomical location, duration, and/or etiology. The need for a more mechanism-based classification system has long been recognized (6, 7). A first step in that direction is the categorization of pain as nociceptive, neuropathic or nociplastic as part of an in-depth clinical assessment. This trichotomy is widely accepted by clinicians and has treatment implications, but whereas neuropathic pain is part of the ICD-11 classification scheme, the relatively new concept of nociplastic pain (8) is not.

Quality pain registries aiming to include information about diagnoses struggle with the fact that pain diagnoses are sometimes based on location, or on duration, or etiology. The aim of the present Methods paper is to describe how we envision that the new chronic pain category MG30 in ICD-11 can be used in the Swedish Quality Registry for Pain rehabilitation (SQRP).

The Swedish Quality Registry for Pain rehabilitation (SQRP) is a well-established clinical registry for adult patients with complex chronic pain conditions. SQRP registers patient-reported outcome measures (PROMs) data from a majority of specialist chronic pain units/departments in Sweden, the patients being mainly referred by primary care physicians. There are no strict inclusion criteria other than assessing that pain is chronic, with significant consequences, motivating a biopsychosocial assessment and if relevant an interdisciplinary intervention. Patients enrolled in SQRP can be characterized as complex as their health profiles often include psychiatric comorbidities such as depression and anxiety, dysfunctional coping behaviors as well as decreased working life and prolonged sick leave, low participation in social activities, and/or unresponsiveness to routine pharmacological or physiotherapeutic treatments delivered in a monodisciplinary fashion. General exclusion criteria are drug or alcohol abuse, severe psychiatric disease, pain due to a non-treated or under-treated cancer, medical conditions that do not allow physical exercise, and red flag pain conditions (i.e., other treatments are available). PROMs are completed by the patients on up to three occasions: before the first visit (baseline assessment) and for those who later participate in an interdisciplinary pain rehabilitation program (IPRP), immediately after completion of IPRP, and, finally, on follow-up one year after completion of IPRP. IPRP (also labelled as multimodal rehabilitation, multidisciplinary rehabilitation, biopsychosocial pain rehabilitation, pain management program) is an interdisciplinary intervention (with physical, occupational, psychological, social, and educational components) according to the International Association for the Study of Pain (IASP). This complex intervention is provided by a multidisciplinary team collaborating in assessment and treatment using a shared biopsychosocial approach and goals. For a detailed description of SQRP including research, see Gerdle et al. (9). There is also a primary care component in SQRP which only partly uses the same variables.

The PROMs capture a patient's background, pain intensity, pain-related cognitions, and psychological distress symptoms as well as activity/participation aspects and health-related quality of life variables. In Table 1 an overview of the variables in SQRP is presented. Diagnoses according to ICD-10 are registered in SQRP. It is possible to register up to four diagnoses in SQRP, and the first diagnosis must be a pain-related one. The most frequent ICD-10 diagnoses (first diagnosis) at baseline in SQRP are presented in Table 2. They constitute 84% of all first diagnoses registered.

The feasibility to register not just one but several diagnoses in SQRP makes it possible to combine the new MG30 category with other ICD-11 diagnoses. In SQRP, we envision the first diagnosis to be a MG30 diagnosis. Making a MG30 diagnosis mandatory as first diagnosis in the registry would ensure a broad and thorough implementation of ICD-11 in Swedish pain care, starting with the units that register in SQRP. At first glance, however, there seems to be drawbacks with the MG30 category. For example, both chronic low back pain and chronic cervical pain would be coded as MG30.02 – chronic primary musculoskeletal pain. Although this may (presumably) make a lot of sense from the point of view of pathophysiology, the heterogeneity of MG30.02 (in this case concerning pain location) would nonetheless create problems registry-wise when it comes to identifying a relatively homogeneous study population. For instance, to study patients with chronic low back pain, it would not be possible to easily identify them by just selecting the ICD-11 diagnosis. Hence, there is a need for more specificity.

Prima facie, one solution to this problem would be to combine MG30 diagnoses with other parts of the ICD-11 system. Despite the new MG30 chronic pain category, ICD-11 retains traditional pain diagnoses outside the MG30 group, such as “low back pain, unspecified” (ME84.2Z) or “cervical spine pain” (ME84.0). Hence, such traditional diagnoses could (if used judiciously) be used as a complement in quality registries to augment diagnostic specificity. By combining a first mandatory MG30 diagnosis with a more traditional pain diagnosis as second diagnosis, “diagnostic pairs” would thereby be created. For instance, the pair MG30.02+ME84.2Z would be readily distinguishable from MG30.02+ME84.0, i.e., chronic primary musculoskeletal pain (low back pain) would be searchable and distinguished from chronic primary musculoskeletal pain (cervical back pain). Is this the way forward, or are there are alternatives available in ICD-11 to enhance diagnostic specificity in registries such as SQRP?

We believe that the new MG30 category of ICD-11 has several advantages. It has the potential to increase the visibility of pain medicine as a medical specialty and academic discipline, it can help validate the experience and suffering of chronic pain patients, and it might lead to better research, e.g., when it comes to the use of registries such as SQRP. Concerning the diagnostic specificity issue described above and its prima facie solution, it is important to realize that ICD-11 offers more than meets the eye. By using the full potential of ICD-11, registries such as SQRP can indeed achieve higher diagnostic specificity in three ways. First, the entries at the level of the so-called foundational layer have unique resource identifiers (URI) that can be used to enhance specificity. Second, ICD-11 contains numerous extension codes that can be combined with the MG30 codes – for instance, concerning the anatomical location of pain. Third, a variant of the intuitive solution of “diagnostic pairs” already exists in ICD-11, namely the possibility to create clusters of stem codes. These three options will now be briefly delineated.

Granted the hierarchical structure imbedded in evidence-based medicine thinking, registry cohort studies such as the ones based on SQRP cannot provide the same level of evidence as randomized controlled trials (RCT) and systematic reviews (SR). However, we argue that register cohort studies are a necessary complement to ensure that the evidence reported in RCTs and SRs also holds for a consecutive non-selected flow of patients in real-world practice settings (15). In other words, registry cohort studies can help ascertain the external validity of RCTs and SRs (their generalizability). A crucial step for registries being able to do that is the accuracy of reported diagnoses. The diagnoses listed in Table 2 illustrate the inherent vagueness of ICD-10 concerning chronic pain, and the new MG30 categorization in ICD-11 is therefore an important clarifying step, even though only a preliminary one. As we have shown, the full potential of the MG30 category can be better exploited in registries if foundational codes, extension codes and/or clustering of stem codes are used to enhance diagnostic specificity. Moreover, from a precision medicine point of view (16), more specific diagnostic data could be used in conjunction with biomarker data to better understand the pathophysiology of different chronic pain conditions. By analyzing the multi-omic pattern of chronic pain patients with unsupervised cluster analysis (i.e., regardless of diagnosis) (17, 18), and by then comparing the frequency of diagnoses in the different clusters, it might also conceptually be possible to confirm the biological validity of traditional diagnoses and perhaps also define chronic pain conditions that are not captured well by traditional diagnoses – hence leading to future diagnostic categories that are difficult to envisage today. For instance, it is conceptually possible that what we call fibromyalgia might be a group of different diseases with only partly overlapping pathophysiology. Only the future will tell if this is so. For the time being, using ICD-11 foundational and/or extension codes and perhaps using the clustering possibility of stem codes, and combining this with the three well-known mechanistic descriptors defined by IASP (nociceptive, neuropathic, nociplastic), seems to be the most specific way to succinctly label the most common chronic pain conditions in pain quality registries such as SQRP.